Show More
@@ -1,1943 +1,1946 b'' | |||||
1 | # changegroup.py - Mercurial changegroup manipulation functions |
|
1 | # changegroup.py - Mercurial changegroup manipulation functions | |
2 | # |
|
2 | # | |
3 | # Copyright 2006 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2006 Olivia Mackall <olivia@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import os |
|
10 | import os | |
11 | import struct |
|
11 | import struct | |
12 | import weakref |
|
12 | import weakref | |
13 |
|
13 | |||
14 | from .i18n import _ |
|
14 | from .i18n import _ | |
15 | from .node import ( |
|
15 | from .node import ( | |
16 | hex, |
|
16 | hex, | |
17 | nullrev, |
|
17 | nullrev, | |
18 | short, |
|
18 | short, | |
19 | ) |
|
19 | ) | |
20 | from .pycompat import open |
|
20 | from .pycompat import open | |
21 |
|
21 | |||
22 | from . import ( |
|
22 | from . import ( | |
23 | error, |
|
23 | error, | |
24 | match as matchmod, |
|
24 | match as matchmod, | |
25 | mdiff, |
|
25 | mdiff, | |
26 | phases, |
|
26 | phases, | |
27 | pycompat, |
|
27 | pycompat, | |
28 | requirements, |
|
28 | requirements, | |
29 | scmutil, |
|
29 | scmutil, | |
30 | util, |
|
30 | util, | |
31 | ) |
|
31 | ) | |
32 |
|
32 | |||
33 | from .interfaces import repository |
|
33 | from .interfaces import repository | |
34 | from .revlogutils import sidedata as sidedatamod |
|
34 | from .revlogutils import sidedata as sidedatamod | |
35 | from .revlogutils import constants as revlog_constants |
|
35 | from .revlogutils import constants as revlog_constants | |
36 | from .utils import storageutil |
|
36 | from .utils import storageutil | |
37 |
|
37 | |||
38 | _CHANGEGROUPV1_DELTA_HEADER = struct.Struct(b"20s20s20s20s") |
|
38 | _CHANGEGROUPV1_DELTA_HEADER = struct.Struct(b"20s20s20s20s") | |
39 | _CHANGEGROUPV2_DELTA_HEADER = struct.Struct(b"20s20s20s20s20s") |
|
39 | _CHANGEGROUPV2_DELTA_HEADER = struct.Struct(b"20s20s20s20s20s") | |
40 | _CHANGEGROUPV3_DELTA_HEADER = struct.Struct(b">20s20s20s20s20sH") |
|
40 | _CHANGEGROUPV3_DELTA_HEADER = struct.Struct(b">20s20s20s20s20sH") | |
41 | _CHANGEGROUPV4_DELTA_HEADER = struct.Struct(b">B20s20s20s20s20sH") |
|
41 | _CHANGEGROUPV4_DELTA_HEADER = struct.Struct(b">B20s20s20s20s20sH") | |
42 |
|
42 | |||
43 | LFS_REQUIREMENT = b'lfs' |
|
43 | LFS_REQUIREMENT = b'lfs' | |
44 |
|
44 | |||
45 | readexactly = util.readexactly |
|
45 | readexactly = util.readexactly | |
46 |
|
46 | |||
47 |
|
47 | |||
48 | def getchunk(stream): |
|
48 | def getchunk(stream): | |
49 | """return the next chunk from stream as a string""" |
|
49 | """return the next chunk from stream as a string""" | |
50 | d = readexactly(stream, 4) |
|
50 | d = readexactly(stream, 4) | |
51 | l = struct.unpack(b">l", d)[0] |
|
51 | l = struct.unpack(b">l", d)[0] | |
52 | if l <= 4: |
|
52 | if l <= 4: | |
53 | if l: |
|
53 | if l: | |
54 | raise error.Abort(_(b"invalid chunk length %d") % l) |
|
54 | raise error.Abort(_(b"invalid chunk length %d") % l) | |
55 | return b"" |
|
55 | return b"" | |
56 | return readexactly(stream, l - 4) |
|
56 | return readexactly(stream, l - 4) | |
57 |
|
57 | |||
58 |
|
58 | |||
59 | def chunkheader(length): |
|
59 | def chunkheader(length): | |
60 | """return a changegroup chunk header (string)""" |
|
60 | """return a changegroup chunk header (string)""" | |
61 | return struct.pack(b">l", length + 4) |
|
61 | return struct.pack(b">l", length + 4) | |
62 |
|
62 | |||
63 |
|
63 | |||
64 | def closechunk(): |
|
64 | def closechunk(): | |
65 | """return a changegroup chunk header (string) for a zero-length chunk""" |
|
65 | """return a changegroup chunk header (string) for a zero-length chunk""" | |
66 | return struct.pack(b">l", 0) |
|
66 | return struct.pack(b">l", 0) | |
67 |
|
67 | |||
68 |
|
68 | |||
69 | def _fileheader(path): |
|
69 | def _fileheader(path): | |
70 | """Obtain a changegroup chunk header for a named path.""" |
|
70 | """Obtain a changegroup chunk header for a named path.""" | |
71 | return chunkheader(len(path)) + path |
|
71 | return chunkheader(len(path)) + path | |
72 |
|
72 | |||
73 |
|
73 | |||
74 | def writechunks(ui, chunks, filename, vfs=None): |
|
74 | def writechunks(ui, chunks, filename, vfs=None): | |
75 | """Write chunks to a file and return its filename. |
|
75 | """Write chunks to a file and return its filename. | |
76 |
|
76 | |||
77 | The stream is assumed to be a bundle file. |
|
77 | The stream is assumed to be a bundle file. | |
78 | Existing files will not be overwritten. |
|
78 | Existing files will not be overwritten. | |
79 | If no filename is specified, a temporary file is created. |
|
79 | If no filename is specified, a temporary file is created. | |
80 | """ |
|
80 | """ | |
81 | fh = None |
|
81 | fh = None | |
82 | cleanup = None |
|
82 | cleanup = None | |
83 | try: |
|
83 | try: | |
84 | if filename: |
|
84 | if filename: | |
85 | if vfs: |
|
85 | if vfs: | |
86 | fh = vfs.open(filename, b"wb") |
|
86 | fh = vfs.open(filename, b"wb") | |
87 | else: |
|
87 | else: | |
88 | # Increase default buffer size because default is usually |
|
88 | # Increase default buffer size because default is usually | |
89 | # small (4k is common on Linux). |
|
89 | # small (4k is common on Linux). | |
90 | fh = open(filename, b"wb", 131072) |
|
90 | fh = open(filename, b"wb", 131072) | |
91 | else: |
|
91 | else: | |
92 | fd, filename = pycompat.mkstemp(prefix=b"hg-bundle-", suffix=b".hg") |
|
92 | fd, filename = pycompat.mkstemp(prefix=b"hg-bundle-", suffix=b".hg") | |
93 | fh = os.fdopen(fd, "wb") |
|
93 | fh = os.fdopen(fd, "wb") | |
94 | cleanup = filename |
|
94 | cleanup = filename | |
95 | for c in chunks: |
|
95 | for c in chunks: | |
96 | fh.write(c) |
|
96 | fh.write(c) | |
97 | cleanup = None |
|
97 | cleanup = None | |
98 | return filename |
|
98 | return filename | |
99 | finally: |
|
99 | finally: | |
100 | if fh is not None: |
|
100 | if fh is not None: | |
101 | fh.close() |
|
101 | fh.close() | |
102 | if cleanup is not None: |
|
102 | if cleanup is not None: | |
103 | if filename and vfs: |
|
103 | if filename and vfs: | |
104 | vfs.unlink(cleanup) |
|
104 | vfs.unlink(cleanup) | |
105 | else: |
|
105 | else: | |
106 | os.unlink(cleanup) |
|
106 | os.unlink(cleanup) | |
107 |
|
107 | |||
108 |
|
108 | |||
109 | class cg1unpacker(object): |
|
109 | class cg1unpacker(object): | |
110 | """Unpacker for cg1 changegroup streams. |
|
110 | """Unpacker for cg1 changegroup streams. | |
111 |
|
111 | |||
112 | A changegroup unpacker handles the framing of the revision data in |
|
112 | A changegroup unpacker handles the framing of the revision data in | |
113 | the wire format. Most consumers will want to use the apply() |
|
113 | the wire format. Most consumers will want to use the apply() | |
114 | method to add the changes from the changegroup to a repository. |
|
114 | method to add the changes from the changegroup to a repository. | |
115 |
|
115 | |||
116 | If you're forwarding a changegroup unmodified to another consumer, |
|
116 | If you're forwarding a changegroup unmodified to another consumer, | |
117 | use getchunks(), which returns an iterator of changegroup |
|
117 | use getchunks(), which returns an iterator of changegroup | |
118 | chunks. This is mostly useful for cases where you need to know the |
|
118 | chunks. This is mostly useful for cases where you need to know the | |
119 | data stream has ended by observing the end of the changegroup. |
|
119 | data stream has ended by observing the end of the changegroup. | |
120 |
|
120 | |||
121 | deltachunk() is useful only if you're applying delta data. Most |
|
121 | deltachunk() is useful only if you're applying delta data. Most | |
122 | consumers should prefer apply() instead. |
|
122 | consumers should prefer apply() instead. | |
123 |
|
123 | |||
124 | A few other public methods exist. Those are used only for |
|
124 | A few other public methods exist. Those are used only for | |
125 | bundlerepo and some debug commands - their use is discouraged. |
|
125 | bundlerepo and some debug commands - their use is discouraged. | |
126 | """ |
|
126 | """ | |
127 |
|
127 | |||
128 | deltaheader = _CHANGEGROUPV1_DELTA_HEADER |
|
128 | deltaheader = _CHANGEGROUPV1_DELTA_HEADER | |
129 | deltaheadersize = deltaheader.size |
|
129 | deltaheadersize = deltaheader.size | |
130 | version = b'01' |
|
130 | version = b'01' | |
131 | _grouplistcount = 1 # One list of files after the manifests |
|
131 | _grouplistcount = 1 # One list of files after the manifests | |
132 |
|
132 | |||
133 | def __init__(self, fh, alg, extras=None): |
|
133 | def __init__(self, fh, alg, extras=None): | |
134 | if alg is None: |
|
134 | if alg is None: | |
135 | alg = b'UN' |
|
135 | alg = b'UN' | |
136 | if alg not in util.compengines.supportedbundletypes: |
|
136 | if alg not in util.compengines.supportedbundletypes: | |
137 | raise error.Abort(_(b'unknown stream compression type: %s') % alg) |
|
137 | raise error.Abort(_(b'unknown stream compression type: %s') % alg) | |
138 | if alg == b'BZ': |
|
138 | if alg == b'BZ': | |
139 | alg = b'_truncatedBZ' |
|
139 | alg = b'_truncatedBZ' | |
140 |
|
140 | |||
141 | compengine = util.compengines.forbundletype(alg) |
|
141 | compengine = util.compengines.forbundletype(alg) | |
142 | self._stream = compengine.decompressorreader(fh) |
|
142 | self._stream = compengine.decompressorreader(fh) | |
143 | self._type = alg |
|
143 | self._type = alg | |
144 | self.extras = extras or {} |
|
144 | self.extras = extras or {} | |
145 | self.callback = None |
|
145 | self.callback = None | |
146 |
|
146 | |||
147 | # These methods (compressed, read, seek, tell) all appear to only |
|
147 | # These methods (compressed, read, seek, tell) all appear to only | |
148 | # be used by bundlerepo, but it's a little hard to tell. |
|
148 | # be used by bundlerepo, but it's a little hard to tell. | |
149 | def compressed(self): |
|
149 | def compressed(self): | |
150 | return self._type is not None and self._type != b'UN' |
|
150 | return self._type is not None and self._type != b'UN' | |
151 |
|
151 | |||
152 | def read(self, l): |
|
152 | def read(self, l): | |
153 | return self._stream.read(l) |
|
153 | return self._stream.read(l) | |
154 |
|
154 | |||
155 | def seek(self, pos): |
|
155 | def seek(self, pos): | |
156 | return self._stream.seek(pos) |
|
156 | return self._stream.seek(pos) | |
157 |
|
157 | |||
158 | def tell(self): |
|
158 | def tell(self): | |
159 | return self._stream.tell() |
|
159 | return self._stream.tell() | |
160 |
|
160 | |||
161 | def close(self): |
|
161 | def close(self): | |
162 | return self._stream.close() |
|
162 | return self._stream.close() | |
163 |
|
163 | |||
164 | def _chunklength(self): |
|
164 | def _chunklength(self): | |
165 | d = readexactly(self._stream, 4) |
|
165 | d = readexactly(self._stream, 4) | |
166 | l = struct.unpack(b">l", d)[0] |
|
166 | l = struct.unpack(b">l", d)[0] | |
167 | if l <= 4: |
|
167 | if l <= 4: | |
168 | if l: |
|
168 | if l: | |
169 | raise error.Abort(_(b"invalid chunk length %d") % l) |
|
169 | raise error.Abort(_(b"invalid chunk length %d") % l) | |
170 | return 0 |
|
170 | return 0 | |
171 | if self.callback: |
|
171 | if self.callback: | |
172 | self.callback() |
|
172 | self.callback() | |
173 | return l - 4 |
|
173 | return l - 4 | |
174 |
|
174 | |||
175 | def changelogheader(self): |
|
175 | def changelogheader(self): | |
176 | """v10 does not have a changelog header chunk""" |
|
176 | """v10 does not have a changelog header chunk""" | |
177 | return {} |
|
177 | return {} | |
178 |
|
178 | |||
179 | def manifestheader(self): |
|
179 | def manifestheader(self): | |
180 | """v10 does not have a manifest header chunk""" |
|
180 | """v10 does not have a manifest header chunk""" | |
181 | return {} |
|
181 | return {} | |
182 |
|
182 | |||
183 | def filelogheader(self): |
|
183 | def filelogheader(self): | |
184 | """return the header of the filelogs chunk, v10 only has the filename""" |
|
184 | """return the header of the filelogs chunk, v10 only has the filename""" | |
185 | l = self._chunklength() |
|
185 | l = self._chunklength() | |
186 | if not l: |
|
186 | if not l: | |
187 | return {} |
|
187 | return {} | |
188 | fname = readexactly(self._stream, l) |
|
188 | fname = readexactly(self._stream, l) | |
189 | return {b'filename': fname} |
|
189 | return {b'filename': fname} | |
190 |
|
190 | |||
191 | def _deltaheader(self, headertuple, prevnode): |
|
191 | def _deltaheader(self, headertuple, prevnode): | |
192 | node, p1, p2, cs = headertuple |
|
192 | node, p1, p2, cs = headertuple | |
193 | if prevnode is None: |
|
193 | if prevnode is None: | |
194 | deltabase = p1 |
|
194 | deltabase = p1 | |
195 | else: |
|
195 | else: | |
196 | deltabase = prevnode |
|
196 | deltabase = prevnode | |
197 | flags = 0 |
|
197 | flags = 0 | |
198 | protocol_flags = 0 |
|
198 | protocol_flags = 0 | |
199 | return node, p1, p2, deltabase, cs, flags, protocol_flags |
|
199 | return node, p1, p2, deltabase, cs, flags, protocol_flags | |
200 |
|
200 | |||
201 | def deltachunk(self, prevnode): |
|
201 | def deltachunk(self, prevnode): | |
202 | l = self._chunklength() |
|
202 | l = self._chunklength() | |
203 | if not l: |
|
203 | if not l: | |
204 | return {} |
|
204 | return {} | |
205 | headerdata = readexactly(self._stream, self.deltaheadersize) |
|
205 | headerdata = readexactly(self._stream, self.deltaheadersize) | |
206 | header = self.deltaheader.unpack(headerdata) |
|
206 | header = self.deltaheader.unpack(headerdata) | |
207 | delta = readexactly(self._stream, l - self.deltaheadersize) |
|
207 | delta = readexactly(self._stream, l - self.deltaheadersize) | |
208 | header = self._deltaheader(header, prevnode) |
|
208 | header = self._deltaheader(header, prevnode) | |
209 | node, p1, p2, deltabase, cs, flags, protocol_flags = header |
|
209 | node, p1, p2, deltabase, cs, flags, protocol_flags = header | |
210 | return node, p1, p2, cs, deltabase, delta, flags, protocol_flags |
|
210 | return node, p1, p2, cs, deltabase, delta, flags, protocol_flags | |
211 |
|
211 | |||
212 | def getchunks(self): |
|
212 | def getchunks(self): | |
213 | """returns all the chunks contains in the bundle |
|
213 | """returns all the chunks contains in the bundle | |
214 |
|
214 | |||
215 | Used when you need to forward the binary stream to a file or another |
|
215 | Used when you need to forward the binary stream to a file or another | |
216 | network API. To do so, it parse the changegroup data, otherwise it will |
|
216 | network API. To do so, it parse the changegroup data, otherwise it will | |
217 | block in case of sshrepo because it don't know the end of the stream. |
|
217 | block in case of sshrepo because it don't know the end of the stream. | |
218 | """ |
|
218 | """ | |
219 | # For changegroup 1 and 2, we expect 3 parts: changelog, manifestlog, |
|
219 | # For changegroup 1 and 2, we expect 3 parts: changelog, manifestlog, | |
220 | # and a list of filelogs. For changegroup 3, we expect 4 parts: |
|
220 | # and a list of filelogs. For changegroup 3, we expect 4 parts: | |
221 | # changelog, manifestlog, a list of tree manifestlogs, and a list of |
|
221 | # changelog, manifestlog, a list of tree manifestlogs, and a list of | |
222 | # filelogs. |
|
222 | # filelogs. | |
223 | # |
|
223 | # | |
224 | # Changelog and manifestlog parts are terminated with empty chunks. The |
|
224 | # Changelog and manifestlog parts are terminated with empty chunks. The | |
225 | # tree and file parts are a list of entry sections. Each entry section |
|
225 | # tree and file parts are a list of entry sections. Each entry section | |
226 | # is a series of chunks terminating in an empty chunk. The list of these |
|
226 | # is a series of chunks terminating in an empty chunk. The list of these | |
227 | # entry sections is terminated in yet another empty chunk, so we know |
|
227 | # entry sections is terminated in yet another empty chunk, so we know | |
228 | # we've reached the end of the tree/file list when we reach an empty |
|
228 | # we've reached the end of the tree/file list when we reach an empty | |
229 | # chunk that was proceeded by no non-empty chunks. |
|
229 | # chunk that was proceeded by no non-empty chunks. | |
230 |
|
230 | |||
231 | parts = 0 |
|
231 | parts = 0 | |
232 | while parts < 2 + self._grouplistcount: |
|
232 | while parts < 2 + self._grouplistcount: | |
233 | noentries = True |
|
233 | noentries = True | |
234 | while True: |
|
234 | while True: | |
235 | chunk = getchunk(self) |
|
235 | chunk = getchunk(self) | |
236 | if not chunk: |
|
236 | if not chunk: | |
237 | # The first two empty chunks represent the end of the |
|
237 | # The first two empty chunks represent the end of the | |
238 | # changelog and the manifestlog portions. The remaining |
|
238 | # changelog and the manifestlog portions. The remaining | |
239 | # empty chunks represent either A) the end of individual |
|
239 | # empty chunks represent either A) the end of individual | |
240 | # tree or file entries in the file list, or B) the end of |
|
240 | # tree or file entries in the file list, or B) the end of | |
241 | # the entire list. It's the end of the entire list if there |
|
241 | # the entire list. It's the end of the entire list if there | |
242 | # were no entries (i.e. noentries is True). |
|
242 | # were no entries (i.e. noentries is True). | |
243 | if parts < 2: |
|
243 | if parts < 2: | |
244 | parts += 1 |
|
244 | parts += 1 | |
245 | elif noentries: |
|
245 | elif noentries: | |
246 | parts += 1 |
|
246 | parts += 1 | |
247 | break |
|
247 | break | |
248 | noentries = False |
|
248 | noentries = False | |
249 | yield chunkheader(len(chunk)) |
|
249 | yield chunkheader(len(chunk)) | |
250 | pos = 0 |
|
250 | pos = 0 | |
251 | while pos < len(chunk): |
|
251 | while pos < len(chunk): | |
252 | next = pos + 2 ** 20 |
|
252 | next = pos + 2 ** 20 | |
253 | yield chunk[pos:next] |
|
253 | yield chunk[pos:next] | |
254 | pos = next |
|
254 | pos = next | |
255 | yield closechunk() |
|
255 | yield closechunk() | |
256 |
|
256 | |||
257 | def _unpackmanifests(self, repo, revmap, trp, prog, addrevisioncb=None): |
|
257 | def _unpackmanifests(self, repo, revmap, trp, prog, addrevisioncb=None): | |
258 | self.callback = prog.increment |
|
258 | self.callback = prog.increment | |
259 | # no need to check for empty manifest group here: |
|
259 | # no need to check for empty manifest group here: | |
260 | # if the result of the merge of 1 and 2 is the same in 3 and 4, |
|
260 | # if the result of the merge of 1 and 2 is the same in 3 and 4, | |
261 | # no new manifest will be created and the manifest group will |
|
261 | # no new manifest will be created and the manifest group will | |
262 | # be empty during the pull |
|
262 | # be empty during the pull | |
263 | self.manifestheader() |
|
263 | self.manifestheader() | |
264 | deltas = self.deltaiter() |
|
264 | deltas = self.deltaiter() | |
265 | storage = repo.manifestlog.getstorage(b'') |
|
265 | storage = repo.manifestlog.getstorage(b'') | |
266 | storage.addgroup(deltas, revmap, trp, addrevisioncb=addrevisioncb) |
|
266 | storage.addgroup(deltas, revmap, trp, addrevisioncb=addrevisioncb) | |
267 | prog.complete() |
|
267 | prog.complete() | |
268 | self.callback = None |
|
268 | self.callback = None | |
269 |
|
269 | |||
270 | def apply( |
|
270 | def apply( | |
271 | self, |
|
271 | self, | |
272 | repo, |
|
272 | repo, | |
273 | tr, |
|
273 | tr, | |
274 | srctype, |
|
274 | srctype, | |
275 | url, |
|
275 | url, | |
276 | targetphase=phases.draft, |
|
276 | targetphase=phases.draft, | |
277 | expectedtotal=None, |
|
277 | expectedtotal=None, | |
278 | sidedata_categories=None, |
|
278 | sidedata_categories=None, | |
279 | ): |
|
279 | ): | |
280 | """Add the changegroup returned by source.read() to this repo. |
|
280 | """Add the changegroup returned by source.read() to this repo. | |
281 | srctype is a string like 'push', 'pull', or 'unbundle'. url is |
|
281 | srctype is a string like 'push', 'pull', or 'unbundle'. url is | |
282 | the URL of the repo where this changegroup is coming from. |
|
282 | the URL of the repo where this changegroup is coming from. | |
283 |
|
283 | |||
284 | Return an integer summarizing the change to this repo: |
|
284 | Return an integer summarizing the change to this repo: | |
285 | - nothing changed or no source: 0 |
|
285 | - nothing changed or no source: 0 | |
286 | - more heads than before: 1+added heads (2..n) |
|
286 | - more heads than before: 1+added heads (2..n) | |
287 | - fewer heads than before: -1-removed heads (-2..-n) |
|
287 | - fewer heads than before: -1-removed heads (-2..-n) | |
288 | - number of heads stays the same: 1 |
|
288 | - number of heads stays the same: 1 | |
289 |
|
289 | |||
290 | `sidedata_categories` is an optional set of the remote's sidedata wanted |
|
290 | `sidedata_categories` is an optional set of the remote's sidedata wanted | |
291 | categories. |
|
291 | categories. | |
292 | """ |
|
292 | """ | |
293 | repo = repo.unfiltered() |
|
293 | repo = repo.unfiltered() | |
294 |
|
294 | |||
295 | # Only useful if we're adding sidedata categories. If both peers have |
|
295 | # Only useful if we're adding sidedata categories. If both peers have | |
296 | # the same categories, then we simply don't do anything. |
|
296 | # the same categories, then we simply don't do anything. | |
297 | adding_sidedata = ( |
|
297 | adding_sidedata = ( | |
298 | requirements.REVLOGV2_REQUIREMENT in repo.requirements |
|
298 | requirements.REVLOGV2_REQUIREMENT in repo.requirements | |
299 | and self.version == b'04' |
|
299 | and self.version == b'04' | |
300 | and srctype == b'pull' |
|
300 | and srctype == b'pull' | |
301 | ) |
|
301 | ) | |
302 | if adding_sidedata: |
|
302 | if adding_sidedata: | |
303 | sidedata_helpers = sidedatamod.get_sidedata_helpers( |
|
303 | sidedata_helpers = sidedatamod.get_sidedata_helpers( | |
304 | repo, |
|
304 | repo, | |
305 | sidedata_categories or set(), |
|
305 | sidedata_categories or set(), | |
306 | pull=True, |
|
306 | pull=True, | |
307 | ) |
|
307 | ) | |
308 | else: |
|
308 | else: | |
309 | sidedata_helpers = None |
|
309 | sidedata_helpers = None | |
310 |
|
310 | |||
311 | def csmap(x): |
|
311 | def csmap(x): | |
312 | repo.ui.debug(b"add changeset %s\n" % short(x)) |
|
312 | repo.ui.debug(b"add changeset %s\n" % short(x)) | |
313 | return len(cl) |
|
313 | return len(cl) | |
314 |
|
314 | |||
315 | def revmap(x): |
|
315 | def revmap(x): | |
316 | return cl.rev(x) |
|
316 | return cl.rev(x) | |
317 |
|
317 | |||
318 | try: |
|
318 | try: | |
319 | # The transaction may already carry source information. In this |
|
319 | # The transaction may already carry source information. In this | |
320 | # case we use the top level data. We overwrite the argument |
|
320 | # case we use the top level data. We overwrite the argument | |
321 | # because we need to use the top level value (if they exist) |
|
321 | # because we need to use the top level value (if they exist) | |
322 | # in this function. |
|
322 | # in this function. | |
323 | srctype = tr.hookargs.setdefault(b'source', srctype) |
|
323 | srctype = tr.hookargs.setdefault(b'source', srctype) | |
324 | tr.hookargs.setdefault(b'url', url) |
|
324 | tr.hookargs.setdefault(b'url', url) | |
325 | repo.hook( |
|
325 | repo.hook( | |
326 | b'prechangegroup', throw=True, **pycompat.strkwargs(tr.hookargs) |
|
326 | b'prechangegroup', throw=True, **pycompat.strkwargs(tr.hookargs) | |
327 | ) |
|
327 | ) | |
328 |
|
328 | |||
329 | # write changelog data to temp files so concurrent readers |
|
329 | # write changelog data to temp files so concurrent readers | |
330 | # will not see an inconsistent view |
|
330 | # will not see an inconsistent view | |
331 | cl = repo.changelog |
|
331 | cl = repo.changelog | |
332 | cl.delayupdate(tr) |
|
332 | cl.delayupdate(tr) | |
333 | oldheads = set(cl.heads()) |
|
333 | oldheads = set(cl.heads()) | |
334 |
|
334 | |||
335 | trp = weakref.proxy(tr) |
|
335 | trp = weakref.proxy(tr) | |
336 | # pull off the changeset group |
|
336 | # pull off the changeset group | |
337 | repo.ui.status(_(b"adding changesets\n")) |
|
337 | repo.ui.status(_(b"adding changesets\n")) | |
338 | clstart = len(cl) |
|
338 | clstart = len(cl) | |
339 | progress = repo.ui.makeprogress( |
|
339 | progress = repo.ui.makeprogress( | |
340 | _(b'changesets'), unit=_(b'chunks'), total=expectedtotal |
|
340 | _(b'changesets'), unit=_(b'chunks'), total=expectedtotal | |
341 | ) |
|
341 | ) | |
342 | self.callback = progress.increment |
|
342 | self.callback = progress.increment | |
343 |
|
343 | |||
344 | efilesset = set() |
|
344 | efilesset = set() | |
345 | duprevs = [] |
|
345 | duprevs = [] | |
346 |
|
346 | |||
347 | def ondupchangelog(cl, rev): |
|
347 | def ondupchangelog(cl, rev): | |
348 | if rev < clstart: |
|
348 | if rev < clstart: | |
349 | duprevs.append(rev) |
|
349 | duprevs.append(rev) | |
350 |
|
350 | |||
351 | def onchangelog(cl, rev): |
|
351 | def onchangelog(cl, rev): | |
352 | ctx = cl.changelogrevision(rev) |
|
352 | ctx = cl.changelogrevision(rev) | |
353 | efilesset.update(ctx.files) |
|
353 | efilesset.update(ctx.files) | |
354 | repo.register_changeset(rev, ctx) |
|
354 | repo.register_changeset(rev, ctx) | |
355 |
|
355 | |||
356 | self.changelogheader() |
|
356 | self.changelogheader() | |
357 | deltas = self.deltaiter() |
|
357 | deltas = self.deltaiter() | |
358 | if not cl.addgroup( |
|
358 | if not cl.addgroup( | |
359 | deltas, |
|
359 | deltas, | |
360 | csmap, |
|
360 | csmap, | |
361 | trp, |
|
361 | trp, | |
362 | alwayscache=True, |
|
362 | alwayscache=True, | |
363 | addrevisioncb=onchangelog, |
|
363 | addrevisioncb=onchangelog, | |
364 | duplicaterevisioncb=ondupchangelog, |
|
364 | duplicaterevisioncb=ondupchangelog, | |
365 | ): |
|
365 | ): | |
366 | repo.ui.develwarn( |
|
366 | repo.ui.develwarn( | |
367 | b'applied empty changelog from changegroup', |
|
367 | b'applied empty changelog from changegroup', | |
368 | config=b'warn-empty-changegroup', |
|
368 | config=b'warn-empty-changegroup', | |
369 | ) |
|
369 | ) | |
370 | efiles = len(efilesset) |
|
370 | efiles = len(efilesset) | |
371 | clend = len(cl) |
|
371 | clend = len(cl) | |
372 | changesets = clend - clstart |
|
372 | changesets = clend - clstart | |
373 | progress.complete() |
|
373 | progress.complete() | |
374 | del deltas |
|
374 | del deltas | |
375 | # TODO Python 2.7 removal |
|
375 | # TODO Python 2.7 removal | |
376 | # del efilesset |
|
376 | # del efilesset | |
377 | efilesset = None |
|
377 | efilesset = None | |
378 | self.callback = None |
|
378 | self.callback = None | |
379 |
|
379 | |||
380 | # Keep track of the (non-changelog) revlogs we've updated and their |
|
380 | # Keep track of the (non-changelog) revlogs we've updated and their | |
381 | # range of new revisions for sidedata rewrite. |
|
381 | # range of new revisions for sidedata rewrite. | |
382 | # TODO do something more efficient than keeping the reference to |
|
382 | # TODO do something more efficient than keeping the reference to | |
383 | # the revlogs, especially memory-wise. |
|
383 | # the revlogs, especially memory-wise. | |
384 | touched_manifests = {} |
|
384 | touched_manifests = {} | |
385 | touched_filelogs = {} |
|
385 | touched_filelogs = {} | |
386 |
|
386 | |||
387 | # pull off the manifest group |
|
387 | # pull off the manifest group | |
388 | repo.ui.status(_(b"adding manifests\n")) |
|
388 | repo.ui.status(_(b"adding manifests\n")) | |
389 | # We know that we'll never have more manifests than we had |
|
389 | # We know that we'll never have more manifests than we had | |
390 | # changesets. |
|
390 | # changesets. | |
391 | progress = repo.ui.makeprogress( |
|
391 | progress = repo.ui.makeprogress( | |
392 | _(b'manifests'), unit=_(b'chunks'), total=changesets |
|
392 | _(b'manifests'), unit=_(b'chunks'), total=changesets | |
393 | ) |
|
393 | ) | |
394 | on_manifest_rev = None |
|
394 | on_manifest_rev = None | |
395 | if sidedata_helpers: |
|
395 | if sidedata_helpers: | |
396 | if revlog_constants.KIND_MANIFESTLOG in sidedata_helpers[1]: |
|
396 | if revlog_constants.KIND_MANIFESTLOG in sidedata_helpers[1]: | |
397 |
|
397 | |||
398 | def on_manifest_rev(manifest, rev): |
|
398 | def on_manifest_rev(manifest, rev): | |
399 | range = touched_manifests.get(manifest) |
|
399 | range = touched_manifests.get(manifest) | |
400 | if not range: |
|
400 | if not range: | |
401 | touched_manifests[manifest] = (rev, rev) |
|
401 | touched_manifests[manifest] = (rev, rev) | |
402 | else: |
|
402 | else: | |
403 | assert rev == range[1] + 1 |
|
403 | assert rev == range[1] + 1 | |
404 | touched_manifests[manifest] = (range[0], rev) |
|
404 | touched_manifests[manifest] = (range[0], rev) | |
405 |
|
405 | |||
406 | self._unpackmanifests( |
|
406 | self._unpackmanifests( | |
407 | repo, |
|
407 | repo, | |
408 | revmap, |
|
408 | revmap, | |
409 | trp, |
|
409 | trp, | |
410 | progress, |
|
410 | progress, | |
411 | addrevisioncb=on_manifest_rev, |
|
411 | addrevisioncb=on_manifest_rev, | |
412 | ) |
|
412 | ) | |
413 |
|
413 | |||
414 | needfiles = {} |
|
414 | needfiles = {} | |
415 | if repo.ui.configbool(b'server', b'validate'): |
|
415 | if repo.ui.configbool(b'server', b'validate'): | |
416 | cl = repo.changelog |
|
416 | cl = repo.changelog | |
417 | ml = repo.manifestlog |
|
417 | ml = repo.manifestlog | |
418 | # validate incoming csets have their manifests |
|
418 | # validate incoming csets have their manifests | |
419 | for cset in pycompat.xrange(clstart, clend): |
|
419 | for cset in pycompat.xrange(clstart, clend): | |
420 | mfnode = cl.changelogrevision(cset).manifest |
|
420 | mfnode = cl.changelogrevision(cset).manifest | |
421 | mfest = ml[mfnode].readdelta() |
|
421 | mfest = ml[mfnode].readdelta() | |
422 | # store file nodes we must see |
|
422 | # store file nodes we must see | |
423 | for f, n in pycompat.iteritems(mfest): |
|
423 | for f, n in pycompat.iteritems(mfest): | |
424 | needfiles.setdefault(f, set()).add(n) |
|
424 | needfiles.setdefault(f, set()).add(n) | |
425 |
|
425 | |||
426 | on_filelog_rev = None |
|
426 | on_filelog_rev = None | |
427 | if sidedata_helpers: |
|
427 | if sidedata_helpers: | |
428 | if revlog_constants.KIND_FILELOG in sidedata_helpers[1]: |
|
428 | if revlog_constants.KIND_FILELOG in sidedata_helpers[1]: | |
429 |
|
429 | |||
430 | def on_filelog_rev(filelog, rev): |
|
430 | def on_filelog_rev(filelog, rev): | |
431 | range = touched_filelogs.get(filelog) |
|
431 | range = touched_filelogs.get(filelog) | |
432 | if not range: |
|
432 | if not range: | |
433 | touched_filelogs[filelog] = (rev, rev) |
|
433 | touched_filelogs[filelog] = (rev, rev) | |
434 | else: |
|
434 | else: | |
435 | assert rev == range[1] + 1 |
|
435 | assert rev == range[1] + 1 | |
436 | touched_filelogs[filelog] = (range[0], rev) |
|
436 | touched_filelogs[filelog] = (range[0], rev) | |
437 |
|
437 | |||
438 | # process the files |
|
438 | # process the files | |
439 | repo.ui.status(_(b"adding file changes\n")) |
|
439 | repo.ui.status(_(b"adding file changes\n")) | |
440 | newrevs, newfiles = _addchangegroupfiles( |
|
440 | newrevs, newfiles = _addchangegroupfiles( | |
441 | repo, |
|
441 | repo, | |
442 | self, |
|
442 | self, | |
443 | revmap, |
|
443 | revmap, | |
444 | trp, |
|
444 | trp, | |
445 | efiles, |
|
445 | efiles, | |
446 | needfiles, |
|
446 | needfiles, | |
447 | addrevisioncb=on_filelog_rev, |
|
447 | addrevisioncb=on_filelog_rev, | |
448 | ) |
|
448 | ) | |
449 |
|
449 | |||
450 | if sidedata_helpers: |
|
450 | if sidedata_helpers: | |
451 | if revlog_constants.KIND_CHANGELOG in sidedata_helpers[1]: |
|
451 | if revlog_constants.KIND_CHANGELOG in sidedata_helpers[1]: | |
452 | cl.rewrite_sidedata(sidedata_helpers, clstart, clend - 1) |
|
452 | cl.rewrite_sidedata(sidedata_helpers, clstart, clend - 1) | |
453 | for mf, (startrev, endrev) in touched_manifests.items(): |
|
453 | for mf, (startrev, endrev) in touched_manifests.items(): | |
454 | mf.rewrite_sidedata(sidedata_helpers, startrev, endrev) |
|
454 | mf.rewrite_sidedata(sidedata_helpers, startrev, endrev) | |
455 | for fl, (startrev, endrev) in touched_filelogs.items(): |
|
455 | for fl, (startrev, endrev) in touched_filelogs.items(): | |
456 | fl.rewrite_sidedata(sidedata_helpers, startrev, endrev) |
|
456 | fl.rewrite_sidedata(sidedata_helpers, startrev, endrev) | |
457 |
|
457 | |||
458 | # making sure the value exists |
|
458 | # making sure the value exists | |
459 | tr.changes.setdefault(b'changegroup-count-changesets', 0) |
|
459 | tr.changes.setdefault(b'changegroup-count-changesets', 0) | |
460 | tr.changes.setdefault(b'changegroup-count-revisions', 0) |
|
460 | tr.changes.setdefault(b'changegroup-count-revisions', 0) | |
461 | tr.changes.setdefault(b'changegroup-count-files', 0) |
|
461 | tr.changes.setdefault(b'changegroup-count-files', 0) | |
462 | tr.changes.setdefault(b'changegroup-count-heads', 0) |
|
462 | tr.changes.setdefault(b'changegroup-count-heads', 0) | |
463 |
|
463 | |||
464 | # some code use bundle operation for internal purpose. They usually |
|
464 | # some code use bundle operation for internal purpose. They usually | |
465 | # set `ui.quiet` to do this outside of user sight. Size the report |
|
465 | # set `ui.quiet` to do this outside of user sight. Size the report | |
466 | # of such operation now happens at the end of the transaction, that |
|
466 | # of such operation now happens at the end of the transaction, that | |
467 | # ui.quiet has not direct effect on the output. |
|
467 | # ui.quiet has not direct effect on the output. | |
468 | # |
|
468 | # | |
469 | # To preserve this intend use an inelegant hack, we fail to report |
|
469 | # To preserve this intend use an inelegant hack, we fail to report | |
470 | # the change if `quiet` is set. We should probably move to |
|
470 | # the change if `quiet` is set. We should probably move to | |
471 | # something better, but this is a good first step to allow the "end |
|
471 | # something better, but this is a good first step to allow the "end | |
472 | # of transaction report" to pass tests. |
|
472 | # of transaction report" to pass tests. | |
473 | if not repo.ui.quiet: |
|
473 | if not repo.ui.quiet: | |
474 | tr.changes[b'changegroup-count-changesets'] += changesets |
|
474 | tr.changes[b'changegroup-count-changesets'] += changesets | |
475 | tr.changes[b'changegroup-count-revisions'] += newrevs |
|
475 | tr.changes[b'changegroup-count-revisions'] += newrevs | |
476 | tr.changes[b'changegroup-count-files'] += newfiles |
|
476 | tr.changes[b'changegroup-count-files'] += newfiles | |
477 |
|
477 | |||
478 | deltaheads = 0 |
|
478 | deltaheads = 0 | |
479 | if oldheads: |
|
479 | if oldheads: | |
480 | heads = cl.heads() |
|
480 | heads = cl.heads() | |
481 | deltaheads += len(heads) - len(oldheads) |
|
481 | deltaheads += len(heads) - len(oldheads) | |
482 | for h in heads: |
|
482 | for h in heads: | |
483 | if h not in oldheads and repo[h].closesbranch(): |
|
483 | if h not in oldheads and repo[h].closesbranch(): | |
484 | deltaheads -= 1 |
|
484 | deltaheads -= 1 | |
485 |
|
485 | |||
486 | # see previous comment about checking ui.quiet |
|
486 | # see previous comment about checking ui.quiet | |
487 | if not repo.ui.quiet: |
|
487 | if not repo.ui.quiet: | |
488 | tr.changes[b'changegroup-count-heads'] += deltaheads |
|
488 | tr.changes[b'changegroup-count-heads'] += deltaheads | |
489 | repo.invalidatevolatilesets() |
|
489 | repo.invalidatevolatilesets() | |
490 |
|
490 | |||
491 | if changesets > 0: |
|
491 | if changesets > 0: | |
492 | if b'node' not in tr.hookargs: |
|
492 | if b'node' not in tr.hookargs: | |
493 | tr.hookargs[b'node'] = hex(cl.node(clstart)) |
|
493 | tr.hookargs[b'node'] = hex(cl.node(clstart)) | |
494 | tr.hookargs[b'node_last'] = hex(cl.node(clend - 1)) |
|
494 | tr.hookargs[b'node_last'] = hex(cl.node(clend - 1)) | |
495 | hookargs = dict(tr.hookargs) |
|
495 | hookargs = dict(tr.hookargs) | |
496 | else: |
|
496 | else: | |
497 | hookargs = dict(tr.hookargs) |
|
497 | hookargs = dict(tr.hookargs) | |
498 | hookargs[b'node'] = hex(cl.node(clstart)) |
|
498 | hookargs[b'node'] = hex(cl.node(clstart)) | |
499 | hookargs[b'node_last'] = hex(cl.node(clend - 1)) |
|
499 | hookargs[b'node_last'] = hex(cl.node(clend - 1)) | |
500 | repo.hook( |
|
500 | repo.hook( | |
501 | b'pretxnchangegroup', |
|
501 | b'pretxnchangegroup', | |
502 | throw=True, |
|
502 | throw=True, | |
503 | **pycompat.strkwargs(hookargs) |
|
503 | **pycompat.strkwargs(hookargs) | |
504 | ) |
|
504 | ) | |
505 |
|
505 | |||
506 | added = pycompat.xrange(clstart, clend) |
|
506 | added = pycompat.xrange(clstart, clend) | |
507 | phaseall = None |
|
507 | phaseall = None | |
508 | if srctype in (b'push', b'serve'): |
|
508 | if srctype in (b'push', b'serve'): | |
509 | # Old servers can not push the boundary themselves. |
|
509 | # Old servers can not push the boundary themselves. | |
510 | # New servers won't push the boundary if changeset already |
|
510 | # New servers won't push the boundary if changeset already | |
511 | # exists locally as secret |
|
511 | # exists locally as secret | |
512 | # |
|
512 | # | |
513 | # We should not use added here but the list of all change in |
|
513 | # We should not use added here but the list of all change in | |
514 | # the bundle |
|
514 | # the bundle | |
515 | if repo.publishing(): |
|
515 | if repo.publishing(): | |
516 | targetphase = phaseall = phases.public |
|
516 | targetphase = phaseall = phases.public | |
517 | else: |
|
517 | else: | |
518 | # closer target phase computation |
|
518 | # closer target phase computation | |
519 |
|
519 | |||
520 | # Those changesets have been pushed from the |
|
520 | # Those changesets have been pushed from the | |
521 | # outside, their phases are going to be pushed |
|
521 | # outside, their phases are going to be pushed | |
522 | # alongside. Therefor `targetphase` is |
|
522 | # alongside. Therefor `targetphase` is | |
523 | # ignored. |
|
523 | # ignored. | |
524 | targetphase = phaseall = phases.draft |
|
524 | targetphase = phaseall = phases.draft | |
525 | if added: |
|
525 | if added: | |
526 | phases.registernew(repo, tr, targetphase, added) |
|
526 | phases.registernew(repo, tr, targetphase, added) | |
527 | if phaseall is not None: |
|
527 | if phaseall is not None: | |
528 | if duprevs: |
|
528 | if duprevs: | |
529 | duprevs.extend(added) |
|
529 | duprevs.extend(added) | |
530 | else: |
|
530 | else: | |
531 | duprevs = added |
|
531 | duprevs = added | |
532 | phases.advanceboundary(repo, tr, phaseall, [], revs=duprevs) |
|
532 | phases.advanceboundary(repo, tr, phaseall, [], revs=duprevs) | |
533 | duprevs = [] |
|
533 | duprevs = [] | |
534 |
|
534 | |||
535 | if changesets > 0: |
|
535 | if changesets > 0: | |
536 |
|
536 | |||
537 | def runhooks(unused_success): |
|
537 | def runhooks(unused_success): | |
538 | # These hooks run when the lock releases, not when the |
|
538 | # These hooks run when the lock releases, not when the | |
539 | # transaction closes. So it's possible for the changelog |
|
539 | # transaction closes. So it's possible for the changelog | |
540 | # to have changed since we last saw it. |
|
540 | # to have changed since we last saw it. | |
541 | if clstart >= len(repo): |
|
541 | if clstart >= len(repo): | |
542 | return |
|
542 | return | |
543 |
|
543 | |||
544 | repo.hook(b"changegroup", **pycompat.strkwargs(hookargs)) |
|
544 | repo.hook(b"changegroup", **pycompat.strkwargs(hookargs)) | |
545 |
|
545 | |||
546 | for rev in added: |
|
546 | for rev in added: | |
547 | args = hookargs.copy() |
|
547 | args = hookargs.copy() | |
548 | args[b'node'] = hex(cl.node(rev)) |
|
548 | args[b'node'] = hex(cl.node(rev)) | |
549 | del args[b'node_last'] |
|
549 | del args[b'node_last'] | |
550 | repo.hook(b"incoming", **pycompat.strkwargs(args)) |
|
550 | repo.hook(b"incoming", **pycompat.strkwargs(args)) | |
551 |
|
551 | |||
552 | newheads = [h for h in repo.heads() if h not in oldheads] |
|
552 | newheads = [h for h in repo.heads() if h not in oldheads] | |
553 | repo.ui.log( |
|
553 | repo.ui.log( | |
554 | b"incoming", |
|
554 | b"incoming", | |
555 | b"%d incoming changes - new heads: %s\n", |
|
555 | b"%d incoming changes - new heads: %s\n", | |
556 | len(added), |
|
556 | len(added), | |
557 | b', '.join([hex(c[:6]) for c in newheads]), |
|
557 | b', '.join([hex(c[:6]) for c in newheads]), | |
558 | ) |
|
558 | ) | |
559 |
|
559 | |||
560 | tr.addpostclose( |
|
560 | tr.addpostclose( | |
561 | b'changegroup-runhooks-%020i' % clstart, |
|
561 | b'changegroup-runhooks-%020i' % clstart, | |
562 | lambda tr: repo._afterlock(runhooks), |
|
562 | lambda tr: repo._afterlock(runhooks), | |
563 | ) |
|
563 | ) | |
564 | finally: |
|
564 | finally: | |
565 | repo.ui.flush() |
|
565 | repo.ui.flush() | |
566 | # never return 0 here: |
|
566 | # never return 0 here: | |
567 | if deltaheads < 0: |
|
567 | if deltaheads < 0: | |
568 | ret = deltaheads - 1 |
|
568 | ret = deltaheads - 1 | |
569 | else: |
|
569 | else: | |
570 | ret = deltaheads + 1 |
|
570 | ret = deltaheads + 1 | |
571 | return ret |
|
571 | return ret | |
572 |
|
572 | |||
573 | def deltaiter(self): |
|
573 | def deltaiter(self): | |
574 | """ |
|
574 | """ | |
575 | returns an iterator of the deltas in this changegroup |
|
575 | returns an iterator of the deltas in this changegroup | |
576 |
|
576 | |||
577 | Useful for passing to the underlying storage system to be stored. |
|
577 | Useful for passing to the underlying storage system to be stored. | |
578 | """ |
|
578 | """ | |
579 | chain = None |
|
579 | chain = None | |
580 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): |
|
580 | for chunkdata in iter(lambda: self.deltachunk(chain), {}): | |
581 | # Chunkdata: (node, p1, p2, cs, deltabase, delta, flags, sidedata) |
|
581 | # Chunkdata: (node, p1, p2, cs, deltabase, delta, flags, sidedata) | |
582 | yield chunkdata |
|
582 | yield chunkdata | |
583 | chain = chunkdata[0] |
|
583 | chain = chunkdata[0] | |
584 |
|
584 | |||
585 |
|
585 | |||
586 | class cg2unpacker(cg1unpacker): |
|
586 | class cg2unpacker(cg1unpacker): | |
587 | """Unpacker for cg2 streams. |
|
587 | """Unpacker for cg2 streams. | |
588 |
|
588 | |||
589 | cg2 streams add support for generaldelta, so the delta header |
|
589 | cg2 streams add support for generaldelta, so the delta header | |
590 | format is slightly different. All other features about the data |
|
590 | format is slightly different. All other features about the data | |
591 | remain the same. |
|
591 | remain the same. | |
592 | """ |
|
592 | """ | |
593 |
|
593 | |||
594 | deltaheader = _CHANGEGROUPV2_DELTA_HEADER |
|
594 | deltaheader = _CHANGEGROUPV2_DELTA_HEADER | |
595 | deltaheadersize = deltaheader.size |
|
595 | deltaheadersize = deltaheader.size | |
596 | version = b'02' |
|
596 | version = b'02' | |
597 |
|
597 | |||
598 | def _deltaheader(self, headertuple, prevnode): |
|
598 | def _deltaheader(self, headertuple, prevnode): | |
599 | node, p1, p2, deltabase, cs = headertuple |
|
599 | node, p1, p2, deltabase, cs = headertuple | |
600 | flags = 0 |
|
600 | flags = 0 | |
601 | protocol_flags = 0 |
|
601 | protocol_flags = 0 | |
602 | return node, p1, p2, deltabase, cs, flags, protocol_flags |
|
602 | return node, p1, p2, deltabase, cs, flags, protocol_flags | |
603 |
|
603 | |||
604 |
|
604 | |||
605 | class cg3unpacker(cg2unpacker): |
|
605 | class cg3unpacker(cg2unpacker): | |
606 | """Unpacker for cg3 streams. |
|
606 | """Unpacker for cg3 streams. | |
607 |
|
607 | |||
608 | cg3 streams add support for exchanging treemanifests and revlog |
|
608 | cg3 streams add support for exchanging treemanifests and revlog | |
609 | flags. It adds the revlog flags to the delta header and an empty chunk |
|
609 | flags. It adds the revlog flags to the delta header and an empty chunk | |
610 | separating manifests and files. |
|
610 | separating manifests and files. | |
611 | """ |
|
611 | """ | |
612 |
|
612 | |||
613 | deltaheader = _CHANGEGROUPV3_DELTA_HEADER |
|
613 | deltaheader = _CHANGEGROUPV3_DELTA_HEADER | |
614 | deltaheadersize = deltaheader.size |
|
614 | deltaheadersize = deltaheader.size | |
615 | version = b'03' |
|
615 | version = b'03' | |
616 | _grouplistcount = 2 # One list of manifests and one list of files |
|
616 | _grouplistcount = 2 # One list of manifests and one list of files | |
617 |
|
617 | |||
618 | def _deltaheader(self, headertuple, prevnode): |
|
618 | def _deltaheader(self, headertuple, prevnode): | |
619 | node, p1, p2, deltabase, cs, flags = headertuple |
|
619 | node, p1, p2, deltabase, cs, flags = headertuple | |
620 | protocol_flags = 0 |
|
620 | protocol_flags = 0 | |
621 | return node, p1, p2, deltabase, cs, flags, protocol_flags |
|
621 | return node, p1, p2, deltabase, cs, flags, protocol_flags | |
622 |
|
622 | |||
623 | def _unpackmanifests(self, repo, revmap, trp, prog, addrevisioncb=None): |
|
623 | def _unpackmanifests(self, repo, revmap, trp, prog, addrevisioncb=None): | |
624 | super(cg3unpacker, self)._unpackmanifests( |
|
624 | super(cg3unpacker, self)._unpackmanifests( | |
625 | repo, revmap, trp, prog, addrevisioncb=addrevisioncb |
|
625 | repo, revmap, trp, prog, addrevisioncb=addrevisioncb | |
626 | ) |
|
626 | ) | |
627 | for chunkdata in iter(self.filelogheader, {}): |
|
627 | for chunkdata in iter(self.filelogheader, {}): | |
628 | # If we get here, there are directory manifests in the changegroup |
|
628 | # If we get here, there are directory manifests in the changegroup | |
629 | d = chunkdata[b"filename"] |
|
629 | d = chunkdata[b"filename"] | |
630 | repo.ui.debug(b"adding %s revisions\n" % d) |
|
630 | repo.ui.debug(b"adding %s revisions\n" % d) | |
631 | deltas = self.deltaiter() |
|
631 | deltas = self.deltaiter() | |
632 | if not repo.manifestlog.getstorage(d).addgroup( |
|
632 | if not repo.manifestlog.getstorage(d).addgroup( | |
633 | deltas, revmap, trp, addrevisioncb=addrevisioncb |
|
633 | deltas, revmap, trp, addrevisioncb=addrevisioncb | |
634 | ): |
|
634 | ): | |
635 | raise error.Abort(_(b"received dir revlog group is empty")) |
|
635 | raise error.Abort(_(b"received dir revlog group is empty")) | |
636 |
|
636 | |||
637 |
|
637 | |||
638 | class cg4unpacker(cg3unpacker): |
|
638 | class cg4unpacker(cg3unpacker): | |
639 | """Unpacker for cg4 streams. |
|
639 | """Unpacker for cg4 streams. | |
640 |
|
640 | |||
641 | cg4 streams add support for exchanging sidedata. |
|
641 | cg4 streams add support for exchanging sidedata. | |
642 | """ |
|
642 | """ | |
643 |
|
643 | |||
644 | deltaheader = _CHANGEGROUPV4_DELTA_HEADER |
|
644 | deltaheader = _CHANGEGROUPV4_DELTA_HEADER | |
645 | deltaheadersize = deltaheader.size |
|
645 | deltaheadersize = deltaheader.size | |
646 | version = b'04' |
|
646 | version = b'04' | |
647 |
|
647 | |||
648 | def _deltaheader(self, headertuple, prevnode): |
|
648 | def _deltaheader(self, headertuple, prevnode): | |
649 | protocol_flags, node, p1, p2, deltabase, cs, flags = headertuple |
|
649 | protocol_flags, node, p1, p2, deltabase, cs, flags = headertuple | |
650 | return node, p1, p2, deltabase, cs, flags, protocol_flags |
|
650 | return node, p1, p2, deltabase, cs, flags, protocol_flags | |
651 |
|
651 | |||
652 | def deltachunk(self, prevnode): |
|
652 | def deltachunk(self, prevnode): | |
653 | res = super(cg4unpacker, self).deltachunk(prevnode) |
|
653 | res = super(cg4unpacker, self).deltachunk(prevnode) | |
654 | if not res: |
|
654 | if not res: | |
655 | return res |
|
655 | return res | |
656 |
|
656 | |||
657 | (node, p1, p2, cs, deltabase, delta, flags, protocol_flags) = res |
|
657 | (node, p1, p2, cs, deltabase, delta, flags, protocol_flags) = res | |
658 |
|
658 | |||
659 | sidedata = {} |
|
659 | sidedata = {} | |
660 | if protocol_flags & storageutil.CG_FLAG_SIDEDATA: |
|
660 | if protocol_flags & storageutil.CG_FLAG_SIDEDATA: | |
661 | sidedata_raw = getchunk(self._stream) |
|
661 | sidedata_raw = getchunk(self._stream) | |
662 | sidedata = sidedatamod.deserialize_sidedata(sidedata_raw) |
|
662 | sidedata = sidedatamod.deserialize_sidedata(sidedata_raw) | |
663 |
|
663 | |||
664 | return node, p1, p2, cs, deltabase, delta, flags, sidedata |
|
664 | return node, p1, p2, cs, deltabase, delta, flags, sidedata | |
665 |
|
665 | |||
666 |
|
666 | |||
667 | class headerlessfixup(object): |
|
667 | class headerlessfixup(object): | |
668 | def __init__(self, fh, h): |
|
668 | def __init__(self, fh, h): | |
669 | self._h = h |
|
669 | self._h = h | |
670 | self._fh = fh |
|
670 | self._fh = fh | |
671 |
|
671 | |||
672 | def read(self, n): |
|
672 | def read(self, n): | |
673 | if self._h: |
|
673 | if self._h: | |
674 | d, self._h = self._h[:n], self._h[n:] |
|
674 | d, self._h = self._h[:n], self._h[n:] | |
675 | if len(d) < n: |
|
675 | if len(d) < n: | |
676 | d += readexactly(self._fh, n - len(d)) |
|
676 | d += readexactly(self._fh, n - len(d)) | |
677 | return d |
|
677 | return d | |
678 | return readexactly(self._fh, n) |
|
678 | return readexactly(self._fh, n) | |
679 |
|
679 | |||
680 |
|
680 | |||
681 | def _revisiondeltatochunks(repo, delta, headerfn): |
|
681 | def _revisiondeltatochunks(repo, delta, headerfn): | |
682 | """Serialize a revisiondelta to changegroup chunks.""" |
|
682 | """Serialize a revisiondelta to changegroup chunks.""" | |
683 |
|
683 | |||
684 | # The captured revision delta may be encoded as a delta against |
|
684 | # The captured revision delta may be encoded as a delta against | |
685 | # a base revision or as a full revision. The changegroup format |
|
685 | # a base revision or as a full revision. The changegroup format | |
686 | # requires that everything on the wire be deltas. So for full |
|
686 | # requires that everything on the wire be deltas. So for full | |
687 | # revisions, we need to invent a header that says to rewrite |
|
687 | # revisions, we need to invent a header that says to rewrite | |
688 | # data. |
|
688 | # data. | |
689 |
|
689 | |||
690 | if delta.delta is not None: |
|
690 | if delta.delta is not None: | |
691 | prefix, data = b'', delta.delta |
|
691 | prefix, data = b'', delta.delta | |
692 | elif delta.basenode == repo.nullid: |
|
692 | elif delta.basenode == repo.nullid: | |
693 | data = delta.revision |
|
693 | data = delta.revision | |
694 | prefix = mdiff.trivialdiffheader(len(data)) |
|
694 | prefix = mdiff.trivialdiffheader(len(data)) | |
695 | else: |
|
695 | else: | |
696 | data = delta.revision |
|
696 | data = delta.revision | |
697 | prefix = mdiff.replacediffheader(delta.baserevisionsize, len(data)) |
|
697 | prefix = mdiff.replacediffheader(delta.baserevisionsize, len(data)) | |
698 |
|
698 | |||
699 | meta = headerfn(delta) |
|
699 | meta = headerfn(delta) | |
700 |
|
700 | |||
701 | yield chunkheader(len(meta) + len(prefix) + len(data)) |
|
701 | yield chunkheader(len(meta) + len(prefix) + len(data)) | |
702 | yield meta |
|
702 | yield meta | |
703 | if prefix: |
|
703 | if prefix: | |
704 | yield prefix |
|
704 | yield prefix | |
705 | yield data |
|
705 | yield data | |
706 |
|
706 | |||
707 | if delta.protocol_flags & storageutil.CG_FLAG_SIDEDATA: |
|
707 | if delta.protocol_flags & storageutil.CG_FLAG_SIDEDATA: | |
708 | # Need a separate chunk for sidedata to be able to differentiate |
|
708 | # Need a separate chunk for sidedata to be able to differentiate | |
709 | # "raw delta" length and sidedata length |
|
709 | # "raw delta" length and sidedata length | |
710 | sidedata = delta.sidedata |
|
710 | sidedata = delta.sidedata | |
711 | yield chunkheader(len(sidedata)) |
|
711 | yield chunkheader(len(sidedata)) | |
712 | yield sidedata |
|
712 | yield sidedata | |
713 |
|
713 | |||
714 |
|
714 | |||
715 | def _sortnodesellipsis(store, nodes, cl, lookup): |
|
715 | def _sortnodesellipsis(store, nodes, cl, lookup): | |
716 | """Sort nodes for changegroup generation.""" |
|
716 | """Sort nodes for changegroup generation.""" | |
717 | # Ellipses serving mode. |
|
717 | # Ellipses serving mode. | |
718 | # |
|
718 | # | |
719 | # In a perfect world, we'd generate better ellipsis-ified graphs |
|
719 | # In a perfect world, we'd generate better ellipsis-ified graphs | |
720 | # for non-changelog revlogs. In practice, we haven't started doing |
|
720 | # for non-changelog revlogs. In practice, we haven't started doing | |
721 | # that yet, so the resulting DAGs for the manifestlog and filelogs |
|
721 | # that yet, so the resulting DAGs for the manifestlog and filelogs | |
722 | # are actually full of bogus parentage on all the ellipsis |
|
722 | # are actually full of bogus parentage on all the ellipsis | |
723 | # nodes. This has the side effect that, while the contents are |
|
723 | # nodes. This has the side effect that, while the contents are | |
724 | # correct, the individual DAGs might be completely out of whack in |
|
724 | # correct, the individual DAGs might be completely out of whack in | |
725 | # a case like 882681bc3166 and its ancestors (back about 10 |
|
725 | # a case like 882681bc3166 and its ancestors (back about 10 | |
726 | # revisions or so) in the main hg repo. |
|
726 | # revisions or so) in the main hg repo. | |
727 | # |
|
727 | # | |
728 | # The one invariant we *know* holds is that the new (potentially |
|
728 | # The one invariant we *know* holds is that the new (potentially | |
729 | # bogus) DAG shape will be valid if we order the nodes in the |
|
729 | # bogus) DAG shape will be valid if we order the nodes in the | |
730 | # order that they're introduced in dramatis personae by the |
|
730 | # order that they're introduced in dramatis personae by the | |
731 | # changelog, so what we do is we sort the non-changelog histories |
|
731 | # changelog, so what we do is we sort the non-changelog histories | |
732 | # by the order in which they are used by the changelog. |
|
732 | # by the order in which they are used by the changelog. | |
733 | key = lambda n: cl.rev(lookup(n)) |
|
733 | key = lambda n: cl.rev(lookup(n)) | |
734 | return sorted(nodes, key=key) |
|
734 | return sorted(nodes, key=key) | |
735 |
|
735 | |||
736 |
|
736 | |||
737 | def _resolvenarrowrevisioninfo( |
|
737 | def _resolvenarrowrevisioninfo( | |
738 | cl, |
|
738 | cl, | |
739 | store, |
|
739 | store, | |
740 | ischangelog, |
|
740 | ischangelog, | |
741 | rev, |
|
741 | rev, | |
742 | linkrev, |
|
742 | linkrev, | |
743 | linknode, |
|
743 | linknode, | |
744 | clrevtolocalrev, |
|
744 | clrevtolocalrev, | |
745 | fullclnodes, |
|
745 | fullclnodes, | |
746 | precomputedellipsis, |
|
746 | precomputedellipsis, | |
747 | ): |
|
747 | ): | |
748 | linkparents = precomputedellipsis[linkrev] |
|
748 | linkparents = precomputedellipsis[linkrev] | |
749 |
|
749 | |||
750 | def local(clrev): |
|
750 | def local(clrev): | |
751 | """Turn a changelog revnum into a local revnum. |
|
751 | """Turn a changelog revnum into a local revnum. | |
752 |
|
752 | |||
753 | The ellipsis dag is stored as revnums on the changelog, |
|
753 | The ellipsis dag is stored as revnums on the changelog, | |
754 | but when we're producing ellipsis entries for |
|
754 | but when we're producing ellipsis entries for | |
755 | non-changelog revlogs, we need to turn those numbers into |
|
755 | non-changelog revlogs, we need to turn those numbers into | |
756 | something local. This does that for us, and during the |
|
756 | something local. This does that for us, and during the | |
757 | changelog sending phase will also expand the stored |
|
757 | changelog sending phase will also expand the stored | |
758 | mappings as needed. |
|
758 | mappings as needed. | |
759 | """ |
|
759 | """ | |
760 | if clrev == nullrev: |
|
760 | if clrev == nullrev: | |
761 | return nullrev |
|
761 | return nullrev | |
762 |
|
762 | |||
763 | if ischangelog: |
|
763 | if ischangelog: | |
764 | return clrev |
|
764 | return clrev | |
765 |
|
765 | |||
766 | # Walk the ellipsis-ized changelog breadth-first looking for a |
|
766 | # Walk the ellipsis-ized changelog breadth-first looking for a | |
767 | # change that has been linked from the current revlog. |
|
767 | # change that has been linked from the current revlog. | |
768 | # |
|
768 | # | |
769 | # For a flat manifest revlog only a single step should be necessary |
|
769 | # For a flat manifest revlog only a single step should be necessary | |
770 | # as all relevant changelog entries are relevant to the flat |
|
770 | # as all relevant changelog entries are relevant to the flat | |
771 | # manifest. |
|
771 | # manifest. | |
772 | # |
|
772 | # | |
773 | # For a filelog or tree manifest dirlog however not every changelog |
|
773 | # For a filelog or tree manifest dirlog however not every changelog | |
774 | # entry will have been relevant, so we need to skip some changelog |
|
774 | # entry will have been relevant, so we need to skip some changelog | |
775 | # nodes even after ellipsis-izing. |
|
775 | # nodes even after ellipsis-izing. | |
776 | walk = [clrev] |
|
776 | walk = [clrev] | |
777 | while walk: |
|
777 | while walk: | |
778 | p = walk[0] |
|
778 | p = walk[0] | |
779 | walk = walk[1:] |
|
779 | walk = walk[1:] | |
780 | if p in clrevtolocalrev: |
|
780 | if p in clrevtolocalrev: | |
781 | return clrevtolocalrev[p] |
|
781 | return clrevtolocalrev[p] | |
782 | elif p in fullclnodes: |
|
782 | elif p in fullclnodes: | |
783 | walk.extend([pp for pp in cl.parentrevs(p) if pp != nullrev]) |
|
783 | walk.extend([pp for pp in cl.parentrevs(p) if pp != nullrev]) | |
784 | elif p in precomputedellipsis: |
|
784 | elif p in precomputedellipsis: | |
785 | walk.extend( |
|
785 | walk.extend( | |
786 | [pp for pp in precomputedellipsis[p] if pp != nullrev] |
|
786 | [pp for pp in precomputedellipsis[p] if pp != nullrev] | |
787 | ) |
|
787 | ) | |
788 | else: |
|
788 | else: | |
789 | # In this case, we've got an ellipsis with parents |
|
789 | # In this case, we've got an ellipsis with parents | |
790 | # outside the current bundle (likely an |
|
790 | # outside the current bundle (likely an | |
791 | # incremental pull). We "know" that we can use the |
|
791 | # incremental pull). We "know" that we can use the | |
792 | # value of this same revlog at whatever revision |
|
792 | # value of this same revlog at whatever revision | |
793 | # is pointed to by linknode. "Know" is in scare |
|
793 | # is pointed to by linknode. "Know" is in scare | |
794 | # quotes because I haven't done enough examination |
|
794 | # quotes because I haven't done enough examination | |
795 | # of edge cases to convince myself this is really |
|
795 | # of edge cases to convince myself this is really | |
796 | # a fact - it works for all the (admittedly |
|
796 | # a fact - it works for all the (admittedly | |
797 | # thorough) cases in our testsuite, but I would be |
|
797 | # thorough) cases in our testsuite, but I would be | |
798 | # somewhat unsurprised to find a case in the wild |
|
798 | # somewhat unsurprised to find a case in the wild | |
799 | # where this breaks down a bit. That said, I don't |
|
799 | # where this breaks down a bit. That said, I don't | |
800 | # know if it would hurt anything. |
|
800 | # know if it would hurt anything. | |
801 | for i in pycompat.xrange(rev, 0, -1): |
|
801 | for i in pycompat.xrange(rev, 0, -1): | |
802 | if store.linkrev(i) == clrev: |
|
802 | if store.linkrev(i) == clrev: | |
803 | return i |
|
803 | return i | |
804 | # We failed to resolve a parent for this node, so |
|
804 | # We failed to resolve a parent for this node, so | |
805 | # we crash the changegroup construction. |
|
805 | # we crash the changegroup construction. | |
806 | raise error.Abort( |
|
806 | raise error.Abort( | |
807 | b"unable to resolve parent while packing '%s' %r" |
|
807 | b"unable to resolve parent while packing '%s' %r" | |
808 | b' for changeset %r' % (store.indexfile, rev, clrev) |
|
808 | b' for changeset %r' % (store.indexfile, rev, clrev) | |
809 | ) |
|
809 | ) | |
810 |
|
810 | |||
811 | return nullrev |
|
811 | return nullrev | |
812 |
|
812 | |||
813 | if not linkparents or (store.parentrevs(rev) == (nullrev, nullrev)): |
|
813 | if not linkparents or (store.parentrevs(rev) == (nullrev, nullrev)): | |
814 | p1, p2 = nullrev, nullrev |
|
814 | p1, p2 = nullrev, nullrev | |
815 | elif len(linkparents) == 1: |
|
815 | elif len(linkparents) == 1: | |
816 | (p1,) = sorted(local(p) for p in linkparents) |
|
816 | (p1,) = sorted(local(p) for p in linkparents) | |
817 | p2 = nullrev |
|
817 | p2 = nullrev | |
818 | else: |
|
818 | else: | |
819 | p1, p2 = sorted(local(p) for p in linkparents) |
|
819 | p1, p2 = sorted(local(p) for p in linkparents) | |
820 |
|
820 | |||
821 | p1node, p2node = store.node(p1), store.node(p2) |
|
821 | p1node, p2node = store.node(p1), store.node(p2) | |
822 |
|
822 | |||
823 | return p1node, p2node, linknode |
|
823 | return p1node, p2node, linknode | |
824 |
|
824 | |||
825 |
|
825 | |||
826 | def deltagroup( |
|
826 | def deltagroup( | |
827 | repo, |
|
827 | repo, | |
828 | store, |
|
828 | store, | |
829 | nodes, |
|
829 | nodes, | |
830 | ischangelog, |
|
830 | ischangelog, | |
831 | lookup, |
|
831 | lookup, | |
832 | forcedeltaparentprev, |
|
832 | forcedeltaparentprev, | |
833 | topic=None, |
|
833 | topic=None, | |
834 | ellipses=False, |
|
834 | ellipses=False, | |
835 | clrevtolocalrev=None, |
|
835 | clrevtolocalrev=None, | |
836 | fullclnodes=None, |
|
836 | fullclnodes=None, | |
837 | precomputedellipsis=None, |
|
837 | precomputedellipsis=None, | |
838 | sidedata_helpers=None, |
|
838 | sidedata_helpers=None, | |
839 | ): |
|
839 | ): | |
840 | """Calculate deltas for a set of revisions. |
|
840 | """Calculate deltas for a set of revisions. | |
841 |
|
841 | |||
842 | Is a generator of ``revisiondelta`` instances. |
|
842 | Is a generator of ``revisiondelta`` instances. | |
843 |
|
843 | |||
844 | If topic is not None, progress detail will be generated using this |
|
844 | If topic is not None, progress detail will be generated using this | |
845 | topic name (e.g. changesets, manifests, etc). |
|
845 | topic name (e.g. changesets, manifests, etc). | |
846 |
|
846 | |||
847 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. |
|
847 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
|
848 | `sidedata_helpers`. | |||
848 | """ |
|
849 | """ | |
849 | if not nodes: |
|
850 | if not nodes: | |
850 | return |
|
851 | return | |
851 |
|
852 | |||
852 | cl = repo.changelog |
|
853 | cl = repo.changelog | |
853 |
|
854 | |||
854 | if ischangelog: |
|
855 | if ischangelog: | |
855 | # `hg log` shows changesets in storage order. To preserve order |
|
856 | # `hg log` shows changesets in storage order. To preserve order | |
856 | # across clones, send out changesets in storage order. |
|
857 | # across clones, send out changesets in storage order. | |
857 | nodesorder = b'storage' |
|
858 | nodesorder = b'storage' | |
858 | elif ellipses: |
|
859 | elif ellipses: | |
859 | nodes = _sortnodesellipsis(store, nodes, cl, lookup) |
|
860 | nodes = _sortnodesellipsis(store, nodes, cl, lookup) | |
860 | nodesorder = b'nodes' |
|
861 | nodesorder = b'nodes' | |
861 | else: |
|
862 | else: | |
862 | nodesorder = None |
|
863 | nodesorder = None | |
863 |
|
864 | |||
864 | # Perform ellipses filtering and revision massaging. We do this before |
|
865 | # Perform ellipses filtering and revision massaging. We do this before | |
865 | # emitrevisions() because a) filtering out revisions creates less work |
|
866 | # emitrevisions() because a) filtering out revisions creates less work | |
866 | # for emitrevisions() b) dropping revisions would break emitrevisions()'s |
|
867 | # for emitrevisions() b) dropping revisions would break emitrevisions()'s | |
867 | # assumptions about delta choices and we would possibly send a delta |
|
868 | # assumptions about delta choices and we would possibly send a delta | |
868 | # referencing a missing base revision. |
|
869 | # referencing a missing base revision. | |
869 | # |
|
870 | # | |
870 | # Also, calling lookup() has side-effects with regards to populating |
|
871 | # Also, calling lookup() has side-effects with regards to populating | |
871 | # data structures. If we don't call lookup() for each node or if we call |
|
872 | # data structures. If we don't call lookup() for each node or if we call | |
872 | # lookup() after the first pass through each node, things can break - |
|
873 | # lookup() after the first pass through each node, things can break - | |
873 | # possibly intermittently depending on the python hash seed! For that |
|
874 | # possibly intermittently depending on the python hash seed! For that | |
874 | # reason, we store a mapping of all linknodes during the initial node |
|
875 | # reason, we store a mapping of all linknodes during the initial node | |
875 | # pass rather than use lookup() on the output side. |
|
876 | # pass rather than use lookup() on the output side. | |
876 | if ellipses: |
|
877 | if ellipses: | |
877 | filtered = [] |
|
878 | filtered = [] | |
878 | adjustedparents = {} |
|
879 | adjustedparents = {} | |
879 | linknodes = {} |
|
880 | linknodes = {} | |
880 |
|
881 | |||
881 | for node in nodes: |
|
882 | for node in nodes: | |
882 | rev = store.rev(node) |
|
883 | rev = store.rev(node) | |
883 | linknode = lookup(node) |
|
884 | linknode = lookup(node) | |
884 | linkrev = cl.rev(linknode) |
|
885 | linkrev = cl.rev(linknode) | |
885 | clrevtolocalrev[linkrev] = rev |
|
886 | clrevtolocalrev[linkrev] = rev | |
886 |
|
887 | |||
887 | # If linknode is in fullclnodes, it means the corresponding |
|
888 | # If linknode is in fullclnodes, it means the corresponding | |
888 | # changeset was a full changeset and is being sent unaltered. |
|
889 | # changeset was a full changeset and is being sent unaltered. | |
889 | if linknode in fullclnodes: |
|
890 | if linknode in fullclnodes: | |
890 | linknodes[node] = linknode |
|
891 | linknodes[node] = linknode | |
891 |
|
892 | |||
892 | # If the corresponding changeset wasn't in the set computed |
|
893 | # If the corresponding changeset wasn't in the set computed | |
893 | # as relevant to us, it should be dropped outright. |
|
894 | # as relevant to us, it should be dropped outright. | |
894 | elif linkrev not in precomputedellipsis: |
|
895 | elif linkrev not in precomputedellipsis: | |
895 | continue |
|
896 | continue | |
896 |
|
897 | |||
897 | else: |
|
898 | else: | |
898 | # We could probably do this later and avoid the dict |
|
899 | # We could probably do this later and avoid the dict | |
899 | # holding state. But it likely doesn't matter. |
|
900 | # holding state. But it likely doesn't matter. | |
900 | p1node, p2node, linknode = _resolvenarrowrevisioninfo( |
|
901 | p1node, p2node, linknode = _resolvenarrowrevisioninfo( | |
901 | cl, |
|
902 | cl, | |
902 | store, |
|
903 | store, | |
903 | ischangelog, |
|
904 | ischangelog, | |
904 | rev, |
|
905 | rev, | |
905 | linkrev, |
|
906 | linkrev, | |
906 | linknode, |
|
907 | linknode, | |
907 | clrevtolocalrev, |
|
908 | clrevtolocalrev, | |
908 | fullclnodes, |
|
909 | fullclnodes, | |
909 | precomputedellipsis, |
|
910 | precomputedellipsis, | |
910 | ) |
|
911 | ) | |
911 |
|
912 | |||
912 | adjustedparents[node] = (p1node, p2node) |
|
913 | adjustedparents[node] = (p1node, p2node) | |
913 | linknodes[node] = linknode |
|
914 | linknodes[node] = linknode | |
914 |
|
915 | |||
915 | filtered.append(node) |
|
916 | filtered.append(node) | |
916 |
|
917 | |||
917 | nodes = filtered |
|
918 | nodes = filtered | |
918 |
|
919 | |||
919 | # We expect the first pass to be fast, so we only engage the progress |
|
920 | # We expect the first pass to be fast, so we only engage the progress | |
920 | # meter for constructing the revision deltas. |
|
921 | # meter for constructing the revision deltas. | |
921 | progress = None |
|
922 | progress = None | |
922 | if topic is not None: |
|
923 | if topic is not None: | |
923 | progress = repo.ui.makeprogress( |
|
924 | progress = repo.ui.makeprogress( | |
924 | topic, unit=_(b'chunks'), total=len(nodes) |
|
925 | topic, unit=_(b'chunks'), total=len(nodes) | |
925 | ) |
|
926 | ) | |
926 |
|
927 | |||
927 | configtarget = repo.ui.config(b'devel', b'bundle.delta') |
|
928 | configtarget = repo.ui.config(b'devel', b'bundle.delta') | |
928 | if configtarget not in (b'', b'p1', b'full'): |
|
929 | if configtarget not in (b'', b'p1', b'full'): | |
929 | msg = _(b"""config "devel.bundle.delta" as unknown value: %s""") |
|
930 | msg = _(b"""config "devel.bundle.delta" as unknown value: %s""") | |
930 | repo.ui.warn(msg % configtarget) |
|
931 | repo.ui.warn(msg % configtarget) | |
931 |
|
932 | |||
932 | deltamode = repository.CG_DELTAMODE_STD |
|
933 | deltamode = repository.CG_DELTAMODE_STD | |
933 | if forcedeltaparentprev: |
|
934 | if forcedeltaparentprev: | |
934 | deltamode = repository.CG_DELTAMODE_PREV |
|
935 | deltamode = repository.CG_DELTAMODE_PREV | |
935 | elif configtarget == b'p1': |
|
936 | elif configtarget == b'p1': | |
936 | deltamode = repository.CG_DELTAMODE_P1 |
|
937 | deltamode = repository.CG_DELTAMODE_P1 | |
937 | elif configtarget == b'full': |
|
938 | elif configtarget == b'full': | |
938 | deltamode = repository.CG_DELTAMODE_FULL |
|
939 | deltamode = repository.CG_DELTAMODE_FULL | |
939 |
|
940 | |||
940 | revisions = store.emitrevisions( |
|
941 | revisions = store.emitrevisions( | |
941 | nodes, |
|
942 | nodes, | |
942 | nodesorder=nodesorder, |
|
943 | nodesorder=nodesorder, | |
943 | revisiondata=True, |
|
944 | revisiondata=True, | |
944 | assumehaveparentrevisions=not ellipses, |
|
945 | assumehaveparentrevisions=not ellipses, | |
945 | deltamode=deltamode, |
|
946 | deltamode=deltamode, | |
946 | sidedata_helpers=sidedata_helpers, |
|
947 | sidedata_helpers=sidedata_helpers, | |
947 | ) |
|
948 | ) | |
948 |
|
949 | |||
949 | for i, revision in enumerate(revisions): |
|
950 | for i, revision in enumerate(revisions): | |
950 | if progress: |
|
951 | if progress: | |
951 | progress.update(i + 1) |
|
952 | progress.update(i + 1) | |
952 |
|
953 | |||
953 | if ellipses: |
|
954 | if ellipses: | |
954 | linknode = linknodes[revision.node] |
|
955 | linknode = linknodes[revision.node] | |
955 |
|
956 | |||
956 | if revision.node in adjustedparents: |
|
957 | if revision.node in adjustedparents: | |
957 | p1node, p2node = adjustedparents[revision.node] |
|
958 | p1node, p2node = adjustedparents[revision.node] | |
958 | revision.p1node = p1node |
|
959 | revision.p1node = p1node | |
959 | revision.p2node = p2node |
|
960 | revision.p2node = p2node | |
960 | revision.flags |= repository.REVISION_FLAG_ELLIPSIS |
|
961 | revision.flags |= repository.REVISION_FLAG_ELLIPSIS | |
961 |
|
962 | |||
962 | else: |
|
963 | else: | |
963 | linknode = lookup(revision.node) |
|
964 | linknode = lookup(revision.node) | |
964 |
|
965 | |||
965 | revision.linknode = linknode |
|
966 | revision.linknode = linknode | |
966 | yield revision |
|
967 | yield revision | |
967 |
|
968 | |||
968 | if progress: |
|
969 | if progress: | |
969 | progress.complete() |
|
970 | progress.complete() | |
970 |
|
971 | |||
971 |
|
972 | |||
972 | class cgpacker(object): |
|
973 | class cgpacker(object): | |
973 | def __init__( |
|
974 | def __init__( | |
974 | self, |
|
975 | self, | |
975 | repo, |
|
976 | repo, | |
976 | oldmatcher, |
|
977 | oldmatcher, | |
977 | matcher, |
|
978 | matcher, | |
978 | version, |
|
979 | version, | |
979 | builddeltaheader, |
|
980 | builddeltaheader, | |
980 | manifestsend, |
|
981 | manifestsend, | |
981 | forcedeltaparentprev=False, |
|
982 | forcedeltaparentprev=False, | |
982 | bundlecaps=None, |
|
983 | bundlecaps=None, | |
983 | ellipses=False, |
|
984 | ellipses=False, | |
984 | shallow=False, |
|
985 | shallow=False, | |
985 | ellipsisroots=None, |
|
986 | ellipsisroots=None, | |
986 | fullnodes=None, |
|
987 | fullnodes=None, | |
987 | remote_sidedata=None, |
|
988 | remote_sidedata=None, | |
988 | ): |
|
989 | ): | |
989 | """Given a source repo, construct a bundler. |
|
990 | """Given a source repo, construct a bundler. | |
990 |
|
991 | |||
991 | oldmatcher is a matcher that matches on files the client already has. |
|
992 | oldmatcher is a matcher that matches on files the client already has. | |
992 | These will not be included in the changegroup. |
|
993 | These will not be included in the changegroup. | |
993 |
|
994 | |||
994 | matcher is a matcher that matches on files to include in the |
|
995 | matcher is a matcher that matches on files to include in the | |
995 | changegroup. Used to facilitate sparse changegroups. |
|
996 | changegroup. Used to facilitate sparse changegroups. | |
996 |
|
997 | |||
997 | forcedeltaparentprev indicates whether delta parents must be against |
|
998 | forcedeltaparentprev indicates whether delta parents must be against | |
998 | the previous revision in a delta group. This should only be used for |
|
999 | the previous revision in a delta group. This should only be used for | |
999 | compatibility with changegroup version 1. |
|
1000 | compatibility with changegroup version 1. | |
1000 |
|
1001 | |||
1001 | builddeltaheader is a callable that constructs the header for a group |
|
1002 | builddeltaheader is a callable that constructs the header for a group | |
1002 | delta. |
|
1003 | delta. | |
1003 |
|
1004 | |||
1004 | manifestsend is a chunk to send after manifests have been fully emitted. |
|
1005 | manifestsend is a chunk to send after manifests have been fully emitted. | |
1005 |
|
1006 | |||
1006 | ellipses indicates whether ellipsis serving mode is enabled. |
|
1007 | ellipses indicates whether ellipsis serving mode is enabled. | |
1007 |
|
1008 | |||
1008 | bundlecaps is optional and can be used to specify the set of |
|
1009 | bundlecaps is optional and can be used to specify the set of | |
1009 | capabilities which can be used to build the bundle. While bundlecaps is |
|
1010 | capabilities which can be used to build the bundle. While bundlecaps is | |
1010 | unused in core Mercurial, extensions rely on this feature to communicate |
|
1011 | unused in core Mercurial, extensions rely on this feature to communicate | |
1011 | capabilities to customize the changegroup packer. |
|
1012 | capabilities to customize the changegroup packer. | |
1012 |
|
1013 | |||
1013 | shallow indicates whether shallow data might be sent. The packer may |
|
1014 | shallow indicates whether shallow data might be sent. The packer may | |
1014 | need to pack file contents not introduced by the changes being packed. |
|
1015 | need to pack file contents not introduced by the changes being packed. | |
1015 |
|
1016 | |||
1016 | fullnodes is the set of changelog nodes which should not be ellipsis |
|
1017 | fullnodes is the set of changelog nodes which should not be ellipsis | |
1017 | nodes. We store this rather than the set of nodes that should be |
|
1018 | nodes. We store this rather than the set of nodes that should be | |
1018 | ellipsis because for very large histories we expect this to be |
|
1019 | ellipsis because for very large histories we expect this to be | |
1019 | significantly smaller. |
|
1020 | significantly smaller. | |
1020 |
|
1021 | |||
1021 | remote_sidedata is the set of sidedata categories wanted by the remote. |
|
1022 | remote_sidedata is the set of sidedata categories wanted by the remote. | |
1022 | """ |
|
1023 | """ | |
1023 | assert oldmatcher |
|
1024 | assert oldmatcher | |
1024 | assert matcher |
|
1025 | assert matcher | |
1025 | self._oldmatcher = oldmatcher |
|
1026 | self._oldmatcher = oldmatcher | |
1026 | self._matcher = matcher |
|
1027 | self._matcher = matcher | |
1027 |
|
1028 | |||
1028 | self.version = version |
|
1029 | self.version = version | |
1029 | self._forcedeltaparentprev = forcedeltaparentprev |
|
1030 | self._forcedeltaparentprev = forcedeltaparentprev | |
1030 | self._builddeltaheader = builddeltaheader |
|
1031 | self._builddeltaheader = builddeltaheader | |
1031 | self._manifestsend = manifestsend |
|
1032 | self._manifestsend = manifestsend | |
1032 | self._ellipses = ellipses |
|
1033 | self._ellipses = ellipses | |
1033 |
|
1034 | |||
1034 | # Set of capabilities we can use to build the bundle. |
|
1035 | # Set of capabilities we can use to build the bundle. | |
1035 | if bundlecaps is None: |
|
1036 | if bundlecaps is None: | |
1036 | bundlecaps = set() |
|
1037 | bundlecaps = set() | |
1037 | self._bundlecaps = bundlecaps |
|
1038 | self._bundlecaps = bundlecaps | |
1038 | if remote_sidedata is None: |
|
1039 | if remote_sidedata is None: | |
1039 | remote_sidedata = set() |
|
1040 | remote_sidedata = set() | |
1040 | self._remote_sidedata = remote_sidedata |
|
1041 | self._remote_sidedata = remote_sidedata | |
1041 | self._isshallow = shallow |
|
1042 | self._isshallow = shallow | |
1042 | self._fullclnodes = fullnodes |
|
1043 | self._fullclnodes = fullnodes | |
1043 |
|
1044 | |||
1044 | # Maps ellipsis revs to their roots at the changelog level. |
|
1045 | # Maps ellipsis revs to their roots at the changelog level. | |
1045 | self._precomputedellipsis = ellipsisroots |
|
1046 | self._precomputedellipsis = ellipsisroots | |
1046 |
|
1047 | |||
1047 | self._repo = repo |
|
1048 | self._repo = repo | |
1048 |
|
1049 | |||
1049 | if self._repo.ui.verbose and not self._repo.ui.debugflag: |
|
1050 | if self._repo.ui.verbose and not self._repo.ui.debugflag: | |
1050 | self._verbosenote = self._repo.ui.note |
|
1051 | self._verbosenote = self._repo.ui.note | |
1051 | else: |
|
1052 | else: | |
1052 | self._verbosenote = lambda s: None |
|
1053 | self._verbosenote = lambda s: None | |
1053 |
|
1054 | |||
1054 | def generate( |
|
1055 | def generate( | |
1055 | self, commonrevs, clnodes, fastpathlinkrev, source, changelog=True |
|
1056 | self, commonrevs, clnodes, fastpathlinkrev, source, changelog=True | |
1056 | ): |
|
1057 | ): | |
1057 | """Yield a sequence of changegroup byte chunks. |
|
1058 | """Yield a sequence of changegroup byte chunks. | |
1058 | If changelog is False, changelog data won't be added to changegroup |
|
1059 | If changelog is False, changelog data won't be added to changegroup | |
1059 | """ |
|
1060 | """ | |
1060 |
|
1061 | |||
1061 | repo = self._repo |
|
1062 | repo = self._repo | |
1062 | cl = repo.changelog |
|
1063 | cl = repo.changelog | |
1063 |
|
1064 | |||
1064 | self._verbosenote(_(b'uncompressed size of bundle content:\n')) |
|
1065 | self._verbosenote(_(b'uncompressed size of bundle content:\n')) | |
1065 | size = 0 |
|
1066 | size = 0 | |
1066 |
|
1067 | |||
1067 | sidedata_helpers = None |
|
1068 | sidedata_helpers = None | |
1068 | if self.version == b'04': |
|
1069 | if self.version == b'04': | |
1069 | remote_sidedata = self._remote_sidedata |
|
1070 | remote_sidedata = self._remote_sidedata | |
1070 | if source == b'strip': |
|
1071 | if source == b'strip': | |
1071 | # We're our own remote when stripping, get the no-op helpers |
|
1072 | # We're our own remote when stripping, get the no-op helpers | |
1072 | # TODO a better approach would be for the strip bundle to |
|
1073 | # TODO a better approach would be for the strip bundle to | |
1073 | # correctly advertise its sidedata categories directly. |
|
1074 | # correctly advertise its sidedata categories directly. | |
1074 | remote_sidedata = repo._wanted_sidedata |
|
1075 | remote_sidedata = repo._wanted_sidedata | |
1075 | sidedata_helpers = sidedatamod.get_sidedata_helpers( |
|
1076 | sidedata_helpers = sidedatamod.get_sidedata_helpers( | |
1076 | repo, remote_sidedata |
|
1077 | repo, remote_sidedata | |
1077 | ) |
|
1078 | ) | |
1078 |
|
1079 | |||
1079 | clstate, deltas = self._generatechangelog( |
|
1080 | clstate, deltas = self._generatechangelog( | |
1080 | cl, |
|
1081 | cl, | |
1081 | clnodes, |
|
1082 | clnodes, | |
1082 | generate=changelog, |
|
1083 | generate=changelog, | |
1083 | sidedata_helpers=sidedata_helpers, |
|
1084 | sidedata_helpers=sidedata_helpers, | |
1084 | ) |
|
1085 | ) | |
1085 | for delta in deltas: |
|
1086 | for delta in deltas: | |
1086 | for chunk in _revisiondeltatochunks( |
|
1087 | for chunk in _revisiondeltatochunks( | |
1087 | self._repo, delta, self._builddeltaheader |
|
1088 | self._repo, delta, self._builddeltaheader | |
1088 | ): |
|
1089 | ): | |
1089 | size += len(chunk) |
|
1090 | size += len(chunk) | |
1090 | yield chunk |
|
1091 | yield chunk | |
1091 |
|
1092 | |||
1092 | close = closechunk() |
|
1093 | close = closechunk() | |
1093 | size += len(close) |
|
1094 | size += len(close) | |
1094 | yield closechunk() |
|
1095 | yield closechunk() | |
1095 |
|
1096 | |||
1096 | self._verbosenote(_(b'%8.i (changelog)\n') % size) |
|
1097 | self._verbosenote(_(b'%8.i (changelog)\n') % size) | |
1097 |
|
1098 | |||
1098 | clrevorder = clstate[b'clrevorder'] |
|
1099 | clrevorder = clstate[b'clrevorder'] | |
1099 | manifests = clstate[b'manifests'] |
|
1100 | manifests = clstate[b'manifests'] | |
1100 | changedfiles = clstate[b'changedfiles'] |
|
1101 | changedfiles = clstate[b'changedfiles'] | |
1101 |
|
1102 | |||
1102 | # We need to make sure that the linkrev in the changegroup refers to |
|
1103 | # We need to make sure that the linkrev in the changegroup refers to | |
1103 | # the first changeset that introduced the manifest or file revision. |
|
1104 | # the first changeset that introduced the manifest or file revision. | |
1104 | # The fastpath is usually safer than the slowpath, because the filelogs |
|
1105 | # The fastpath is usually safer than the slowpath, because the filelogs | |
1105 | # are walked in revlog order. |
|
1106 | # are walked in revlog order. | |
1106 | # |
|
1107 | # | |
1107 | # When taking the slowpath when the manifest revlog uses generaldelta, |
|
1108 | # When taking the slowpath when the manifest revlog uses generaldelta, | |
1108 | # the manifest may be walked in the "wrong" order. Without 'clrevorder', |
|
1109 | # the manifest may be walked in the "wrong" order. Without 'clrevorder', | |
1109 | # we would get an incorrect linkrev (see fix in cc0ff93d0c0c). |
|
1110 | # we would get an incorrect linkrev (see fix in cc0ff93d0c0c). | |
1110 | # |
|
1111 | # | |
1111 | # When taking the fastpath, we are only vulnerable to reordering |
|
1112 | # When taking the fastpath, we are only vulnerable to reordering | |
1112 | # of the changelog itself. The changelog never uses generaldelta and is |
|
1113 | # of the changelog itself. The changelog never uses generaldelta and is | |
1113 | # never reordered. To handle this case, we simply take the slowpath, |
|
1114 | # never reordered. To handle this case, we simply take the slowpath, | |
1114 | # which already has the 'clrevorder' logic. This was also fixed in |
|
1115 | # which already has the 'clrevorder' logic. This was also fixed in | |
1115 | # cc0ff93d0c0c. |
|
1116 | # cc0ff93d0c0c. | |
1116 |
|
1117 | |||
1117 | # Treemanifests don't work correctly with fastpathlinkrev |
|
1118 | # Treemanifests don't work correctly with fastpathlinkrev | |
1118 | # either, because we don't discover which directory nodes to |
|
1119 | # either, because we don't discover which directory nodes to | |
1119 | # send along with files. This could probably be fixed. |
|
1120 | # send along with files. This could probably be fixed. | |
1120 | fastpathlinkrev = fastpathlinkrev and not scmutil.istreemanifest(repo) |
|
1121 | fastpathlinkrev = fastpathlinkrev and not scmutil.istreemanifest(repo) | |
1121 |
|
1122 | |||
1122 | fnodes = {} # needed file nodes |
|
1123 | fnodes = {} # needed file nodes | |
1123 |
|
1124 | |||
1124 | size = 0 |
|
1125 | size = 0 | |
1125 | it = self.generatemanifests( |
|
1126 | it = self.generatemanifests( | |
1126 | commonrevs, |
|
1127 | commonrevs, | |
1127 | clrevorder, |
|
1128 | clrevorder, | |
1128 | fastpathlinkrev, |
|
1129 | fastpathlinkrev, | |
1129 | manifests, |
|
1130 | manifests, | |
1130 | fnodes, |
|
1131 | fnodes, | |
1131 | source, |
|
1132 | source, | |
1132 | clstate[b'clrevtomanifestrev'], |
|
1133 | clstate[b'clrevtomanifestrev'], | |
1133 | sidedata_helpers=sidedata_helpers, |
|
1134 | sidedata_helpers=sidedata_helpers, | |
1134 | ) |
|
1135 | ) | |
1135 |
|
1136 | |||
1136 | for tree, deltas in it: |
|
1137 | for tree, deltas in it: | |
1137 | if tree: |
|
1138 | if tree: | |
1138 | assert self.version in (b'03', b'04') |
|
1139 | assert self.version in (b'03', b'04') | |
1139 | chunk = _fileheader(tree) |
|
1140 | chunk = _fileheader(tree) | |
1140 | size += len(chunk) |
|
1141 | size += len(chunk) | |
1141 | yield chunk |
|
1142 | yield chunk | |
1142 |
|
1143 | |||
1143 | for delta in deltas: |
|
1144 | for delta in deltas: | |
1144 | chunks = _revisiondeltatochunks( |
|
1145 | chunks = _revisiondeltatochunks( | |
1145 | self._repo, delta, self._builddeltaheader |
|
1146 | self._repo, delta, self._builddeltaheader | |
1146 | ) |
|
1147 | ) | |
1147 | for chunk in chunks: |
|
1148 | for chunk in chunks: | |
1148 | size += len(chunk) |
|
1149 | size += len(chunk) | |
1149 | yield chunk |
|
1150 | yield chunk | |
1150 |
|
1151 | |||
1151 | close = closechunk() |
|
1152 | close = closechunk() | |
1152 | size += len(close) |
|
1153 | size += len(close) | |
1153 | yield close |
|
1154 | yield close | |
1154 |
|
1155 | |||
1155 | self._verbosenote(_(b'%8.i (manifests)\n') % size) |
|
1156 | self._verbosenote(_(b'%8.i (manifests)\n') % size) | |
1156 | yield self._manifestsend |
|
1157 | yield self._manifestsend | |
1157 |
|
1158 | |||
1158 | mfdicts = None |
|
1159 | mfdicts = None | |
1159 | if self._ellipses and self._isshallow: |
|
1160 | if self._ellipses and self._isshallow: | |
1160 | mfdicts = [ |
|
1161 | mfdicts = [ | |
1161 | (repo.manifestlog[n].read(), lr) |
|
1162 | (repo.manifestlog[n].read(), lr) | |
1162 | for (n, lr) in pycompat.iteritems(manifests) |
|
1163 | for (n, lr) in pycompat.iteritems(manifests) | |
1163 | ] |
|
1164 | ] | |
1164 |
|
1165 | |||
1165 | manifests.clear() |
|
1166 | manifests.clear() | |
1166 | clrevs = {cl.rev(x) for x in clnodes} |
|
1167 | clrevs = {cl.rev(x) for x in clnodes} | |
1167 |
|
1168 | |||
1168 | it = self.generatefiles( |
|
1169 | it = self.generatefiles( | |
1169 | changedfiles, |
|
1170 | changedfiles, | |
1170 | commonrevs, |
|
1171 | commonrevs, | |
1171 | source, |
|
1172 | source, | |
1172 | mfdicts, |
|
1173 | mfdicts, | |
1173 | fastpathlinkrev, |
|
1174 | fastpathlinkrev, | |
1174 | fnodes, |
|
1175 | fnodes, | |
1175 | clrevs, |
|
1176 | clrevs, | |
1176 | sidedata_helpers=sidedata_helpers, |
|
1177 | sidedata_helpers=sidedata_helpers, | |
1177 | ) |
|
1178 | ) | |
1178 |
|
1179 | |||
1179 | for path, deltas in it: |
|
1180 | for path, deltas in it: | |
1180 | h = _fileheader(path) |
|
1181 | h = _fileheader(path) | |
1181 | size = len(h) |
|
1182 | size = len(h) | |
1182 | yield h |
|
1183 | yield h | |
1183 |
|
1184 | |||
1184 | for delta in deltas: |
|
1185 | for delta in deltas: | |
1185 | chunks = _revisiondeltatochunks( |
|
1186 | chunks = _revisiondeltatochunks( | |
1186 | self._repo, delta, self._builddeltaheader |
|
1187 | self._repo, delta, self._builddeltaheader | |
1187 | ) |
|
1188 | ) | |
1188 | for chunk in chunks: |
|
1189 | for chunk in chunks: | |
1189 | size += len(chunk) |
|
1190 | size += len(chunk) | |
1190 | yield chunk |
|
1191 | yield chunk | |
1191 |
|
1192 | |||
1192 | close = closechunk() |
|
1193 | close = closechunk() | |
1193 | size += len(close) |
|
1194 | size += len(close) | |
1194 | yield close |
|
1195 | yield close | |
1195 |
|
1196 | |||
1196 | self._verbosenote(_(b'%8.i %s\n') % (size, path)) |
|
1197 | self._verbosenote(_(b'%8.i %s\n') % (size, path)) | |
1197 |
|
1198 | |||
1198 | yield closechunk() |
|
1199 | yield closechunk() | |
1199 |
|
1200 | |||
1200 | if clnodes: |
|
1201 | if clnodes: | |
1201 | repo.hook(b'outgoing', node=hex(clnodes[0]), source=source) |
|
1202 | repo.hook(b'outgoing', node=hex(clnodes[0]), source=source) | |
1202 |
|
1203 | |||
1203 | def _generatechangelog( |
|
1204 | def _generatechangelog( | |
1204 | self, cl, nodes, generate=True, sidedata_helpers=None |
|
1205 | self, cl, nodes, generate=True, sidedata_helpers=None | |
1205 | ): |
|
1206 | ): | |
1206 | """Generate data for changelog chunks. |
|
1207 | """Generate data for changelog chunks. | |
1207 |
|
1208 | |||
1208 | Returns a 2-tuple of a dict containing state and an iterable of |
|
1209 | Returns a 2-tuple of a dict containing state and an iterable of | |
1209 | byte chunks. The state will not be fully populated until the |
|
1210 | byte chunks. The state will not be fully populated until the | |
1210 | chunk stream has been fully consumed. |
|
1211 | chunk stream has been fully consumed. | |
1211 |
|
1212 | |||
1212 | if generate is False, the state will be fully populated and no chunk |
|
1213 | if generate is False, the state will be fully populated and no chunk | |
1213 | stream will be yielded |
|
1214 | stream will be yielded | |
1214 |
|
1215 | |||
1215 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. |
|
1216 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
|
1217 | `sidedata_helpers`. | |||
1216 | """ |
|
1218 | """ | |
1217 | clrevorder = {} |
|
1219 | clrevorder = {} | |
1218 | manifests = {} |
|
1220 | manifests = {} | |
1219 | mfl = self._repo.manifestlog |
|
1221 | mfl = self._repo.manifestlog | |
1220 | changedfiles = set() |
|
1222 | changedfiles = set() | |
1221 | clrevtomanifestrev = {} |
|
1223 | clrevtomanifestrev = {} | |
1222 |
|
1224 | |||
1223 | state = { |
|
1225 | state = { | |
1224 | b'clrevorder': clrevorder, |
|
1226 | b'clrevorder': clrevorder, | |
1225 | b'manifests': manifests, |
|
1227 | b'manifests': manifests, | |
1226 | b'changedfiles': changedfiles, |
|
1228 | b'changedfiles': changedfiles, | |
1227 | b'clrevtomanifestrev': clrevtomanifestrev, |
|
1229 | b'clrevtomanifestrev': clrevtomanifestrev, | |
1228 | } |
|
1230 | } | |
1229 |
|
1231 | |||
1230 | if not (generate or self._ellipses): |
|
1232 | if not (generate or self._ellipses): | |
1231 | # sort the nodes in storage order |
|
1233 | # sort the nodes in storage order | |
1232 | nodes = sorted(nodes, key=cl.rev) |
|
1234 | nodes = sorted(nodes, key=cl.rev) | |
1233 | for node in nodes: |
|
1235 | for node in nodes: | |
1234 | c = cl.changelogrevision(node) |
|
1236 | c = cl.changelogrevision(node) | |
1235 | clrevorder[node] = len(clrevorder) |
|
1237 | clrevorder[node] = len(clrevorder) | |
1236 | # record the first changeset introducing this manifest version |
|
1238 | # record the first changeset introducing this manifest version | |
1237 | manifests.setdefault(c.manifest, node) |
|
1239 | manifests.setdefault(c.manifest, node) | |
1238 | # Record a complete list of potentially-changed files in |
|
1240 | # Record a complete list of potentially-changed files in | |
1239 | # this manifest. |
|
1241 | # this manifest. | |
1240 | changedfiles.update(c.files) |
|
1242 | changedfiles.update(c.files) | |
1241 |
|
1243 | |||
1242 | return state, () |
|
1244 | return state, () | |
1243 |
|
1245 | |||
1244 | # Callback for the changelog, used to collect changed files and |
|
1246 | # Callback for the changelog, used to collect changed files and | |
1245 | # manifest nodes. |
|
1247 | # manifest nodes. | |
1246 | # Returns the linkrev node (identity in the changelog case). |
|
1248 | # Returns the linkrev node (identity in the changelog case). | |
1247 | def lookupcl(x): |
|
1249 | def lookupcl(x): | |
1248 | c = cl.changelogrevision(x) |
|
1250 | c = cl.changelogrevision(x) | |
1249 | clrevorder[x] = len(clrevorder) |
|
1251 | clrevorder[x] = len(clrevorder) | |
1250 |
|
1252 | |||
1251 | if self._ellipses: |
|
1253 | if self._ellipses: | |
1252 | # Only update manifests if x is going to be sent. Otherwise we |
|
1254 | # Only update manifests if x is going to be sent. Otherwise we | |
1253 | # end up with bogus linkrevs specified for manifests and |
|
1255 | # end up with bogus linkrevs specified for manifests and | |
1254 | # we skip some manifest nodes that we should otherwise |
|
1256 | # we skip some manifest nodes that we should otherwise | |
1255 | # have sent. |
|
1257 | # have sent. | |
1256 | if ( |
|
1258 | if ( | |
1257 | x in self._fullclnodes |
|
1259 | x in self._fullclnodes | |
1258 | or cl.rev(x) in self._precomputedellipsis |
|
1260 | or cl.rev(x) in self._precomputedellipsis | |
1259 | ): |
|
1261 | ): | |
1260 |
|
1262 | |||
1261 | manifestnode = c.manifest |
|
1263 | manifestnode = c.manifest | |
1262 | # Record the first changeset introducing this manifest |
|
1264 | # Record the first changeset introducing this manifest | |
1263 | # version. |
|
1265 | # version. | |
1264 | manifests.setdefault(manifestnode, x) |
|
1266 | manifests.setdefault(manifestnode, x) | |
1265 | # Set this narrow-specific dict so we have the lowest |
|
1267 | # Set this narrow-specific dict so we have the lowest | |
1266 | # manifest revnum to look up for this cl revnum. (Part of |
|
1268 | # manifest revnum to look up for this cl revnum. (Part of | |
1267 | # mapping changelog ellipsis parents to manifest ellipsis |
|
1269 | # mapping changelog ellipsis parents to manifest ellipsis | |
1268 | # parents) |
|
1270 | # parents) | |
1269 | clrevtomanifestrev.setdefault( |
|
1271 | clrevtomanifestrev.setdefault( | |
1270 | cl.rev(x), mfl.rev(manifestnode) |
|
1272 | cl.rev(x), mfl.rev(manifestnode) | |
1271 | ) |
|
1273 | ) | |
1272 | # We can't trust the changed files list in the changeset if the |
|
1274 | # We can't trust the changed files list in the changeset if the | |
1273 | # client requested a shallow clone. |
|
1275 | # client requested a shallow clone. | |
1274 | if self._isshallow: |
|
1276 | if self._isshallow: | |
1275 | changedfiles.update(mfl[c.manifest].read().keys()) |
|
1277 | changedfiles.update(mfl[c.manifest].read().keys()) | |
1276 | else: |
|
1278 | else: | |
1277 | changedfiles.update(c.files) |
|
1279 | changedfiles.update(c.files) | |
1278 | else: |
|
1280 | else: | |
1279 | # record the first changeset introducing this manifest version |
|
1281 | # record the first changeset introducing this manifest version | |
1280 | manifests.setdefault(c.manifest, x) |
|
1282 | manifests.setdefault(c.manifest, x) | |
1281 | # Record a complete list of potentially-changed files in |
|
1283 | # Record a complete list of potentially-changed files in | |
1282 | # this manifest. |
|
1284 | # this manifest. | |
1283 | changedfiles.update(c.files) |
|
1285 | changedfiles.update(c.files) | |
1284 |
|
1286 | |||
1285 | return x |
|
1287 | return x | |
1286 |
|
1288 | |||
1287 | gen = deltagroup( |
|
1289 | gen = deltagroup( | |
1288 | self._repo, |
|
1290 | self._repo, | |
1289 | cl, |
|
1291 | cl, | |
1290 | nodes, |
|
1292 | nodes, | |
1291 | True, |
|
1293 | True, | |
1292 | lookupcl, |
|
1294 | lookupcl, | |
1293 | self._forcedeltaparentprev, |
|
1295 | self._forcedeltaparentprev, | |
1294 | ellipses=self._ellipses, |
|
1296 | ellipses=self._ellipses, | |
1295 | topic=_(b'changesets'), |
|
1297 | topic=_(b'changesets'), | |
1296 | clrevtolocalrev={}, |
|
1298 | clrevtolocalrev={}, | |
1297 | fullclnodes=self._fullclnodes, |
|
1299 | fullclnodes=self._fullclnodes, | |
1298 | precomputedellipsis=self._precomputedellipsis, |
|
1300 | precomputedellipsis=self._precomputedellipsis, | |
1299 | sidedata_helpers=sidedata_helpers, |
|
1301 | sidedata_helpers=sidedata_helpers, | |
1300 | ) |
|
1302 | ) | |
1301 |
|
1303 | |||
1302 | return state, gen |
|
1304 | return state, gen | |
1303 |
|
1305 | |||
1304 | def generatemanifests( |
|
1306 | def generatemanifests( | |
1305 | self, |
|
1307 | self, | |
1306 | commonrevs, |
|
1308 | commonrevs, | |
1307 | clrevorder, |
|
1309 | clrevorder, | |
1308 | fastpathlinkrev, |
|
1310 | fastpathlinkrev, | |
1309 | manifests, |
|
1311 | manifests, | |
1310 | fnodes, |
|
1312 | fnodes, | |
1311 | source, |
|
1313 | source, | |
1312 | clrevtolocalrev, |
|
1314 | clrevtolocalrev, | |
1313 | sidedata_helpers=None, |
|
1315 | sidedata_helpers=None, | |
1314 | ): |
|
1316 | ): | |
1315 | """Returns an iterator of changegroup chunks containing manifests. |
|
1317 | """Returns an iterator of changegroup chunks containing manifests. | |
1316 |
|
1318 | |||
1317 | `source` is unused here, but is used by extensions like remotefilelog to |
|
1319 | `source` is unused here, but is used by extensions like remotefilelog to | |
1318 | change what is sent based in pulls vs pushes, etc. |
|
1320 | change what is sent based in pulls vs pushes, etc. | |
1319 |
|
1321 | |||
1320 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. |
|
1322 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
|
1323 | `sidedata_helpers`. | |||
1321 | """ |
|
1324 | """ | |
1322 | repo = self._repo |
|
1325 | repo = self._repo | |
1323 | mfl = repo.manifestlog |
|
1326 | mfl = repo.manifestlog | |
1324 | tmfnodes = {b'': manifests} |
|
1327 | tmfnodes = {b'': manifests} | |
1325 |
|
1328 | |||
1326 | # Callback for the manifest, used to collect linkrevs for filelog |
|
1329 | # Callback for the manifest, used to collect linkrevs for filelog | |
1327 | # revisions. |
|
1330 | # revisions. | |
1328 | # Returns the linkrev node (collected in lookupcl). |
|
1331 | # Returns the linkrev node (collected in lookupcl). | |
1329 | def makelookupmflinknode(tree, nodes): |
|
1332 | def makelookupmflinknode(tree, nodes): | |
1330 | if fastpathlinkrev: |
|
1333 | if fastpathlinkrev: | |
1331 | assert not tree |
|
1334 | assert not tree | |
1332 |
|
1335 | |||
1333 | # pytype: disable=unsupported-operands |
|
1336 | # pytype: disable=unsupported-operands | |
1334 | return manifests.__getitem__ |
|
1337 | return manifests.__getitem__ | |
1335 | # pytype: enable=unsupported-operands |
|
1338 | # pytype: enable=unsupported-operands | |
1336 |
|
1339 | |||
1337 | def lookupmflinknode(x): |
|
1340 | def lookupmflinknode(x): | |
1338 | """Callback for looking up the linknode for manifests. |
|
1341 | """Callback for looking up the linknode for manifests. | |
1339 |
|
1342 | |||
1340 | Returns the linkrev node for the specified manifest. |
|
1343 | Returns the linkrev node for the specified manifest. | |
1341 |
|
1344 | |||
1342 | SIDE EFFECT: |
|
1345 | SIDE EFFECT: | |
1343 |
|
1346 | |||
1344 | 1) fclnodes gets populated with the list of relevant |
|
1347 | 1) fclnodes gets populated with the list of relevant | |
1345 | file nodes if we're not using fastpathlinkrev |
|
1348 | file nodes if we're not using fastpathlinkrev | |
1346 | 2) When treemanifests are in use, collects treemanifest nodes |
|
1349 | 2) When treemanifests are in use, collects treemanifest nodes | |
1347 | to send |
|
1350 | to send | |
1348 |
|
1351 | |||
1349 | Note that this means manifests must be completely sent to |
|
1352 | Note that this means manifests must be completely sent to | |
1350 | the client before you can trust the list of files and |
|
1353 | the client before you can trust the list of files and | |
1351 | treemanifests to send. |
|
1354 | treemanifests to send. | |
1352 | """ |
|
1355 | """ | |
1353 | clnode = nodes[x] |
|
1356 | clnode = nodes[x] | |
1354 | mdata = mfl.get(tree, x).readfast(shallow=True) |
|
1357 | mdata = mfl.get(tree, x).readfast(shallow=True) | |
1355 | for p, n, fl in mdata.iterentries(): |
|
1358 | for p, n, fl in mdata.iterentries(): | |
1356 | if fl == b't': # subdirectory manifest |
|
1359 | if fl == b't': # subdirectory manifest | |
1357 | subtree = tree + p + b'/' |
|
1360 | subtree = tree + p + b'/' | |
1358 | tmfclnodes = tmfnodes.setdefault(subtree, {}) |
|
1361 | tmfclnodes = tmfnodes.setdefault(subtree, {}) | |
1359 | tmfclnode = tmfclnodes.setdefault(n, clnode) |
|
1362 | tmfclnode = tmfclnodes.setdefault(n, clnode) | |
1360 | if clrevorder[clnode] < clrevorder[tmfclnode]: |
|
1363 | if clrevorder[clnode] < clrevorder[tmfclnode]: | |
1361 | tmfclnodes[n] = clnode |
|
1364 | tmfclnodes[n] = clnode | |
1362 | else: |
|
1365 | else: | |
1363 | f = tree + p |
|
1366 | f = tree + p | |
1364 | fclnodes = fnodes.setdefault(f, {}) |
|
1367 | fclnodes = fnodes.setdefault(f, {}) | |
1365 | fclnode = fclnodes.setdefault(n, clnode) |
|
1368 | fclnode = fclnodes.setdefault(n, clnode) | |
1366 | if clrevorder[clnode] < clrevorder[fclnode]: |
|
1369 | if clrevorder[clnode] < clrevorder[fclnode]: | |
1367 | fclnodes[n] = clnode |
|
1370 | fclnodes[n] = clnode | |
1368 | return clnode |
|
1371 | return clnode | |
1369 |
|
1372 | |||
1370 | return lookupmflinknode |
|
1373 | return lookupmflinknode | |
1371 |
|
1374 | |||
1372 | while tmfnodes: |
|
1375 | while tmfnodes: | |
1373 | tree, nodes = tmfnodes.popitem() |
|
1376 | tree, nodes = tmfnodes.popitem() | |
1374 |
|
1377 | |||
1375 | should_visit = self._matcher.visitdir(tree[:-1]) |
|
1378 | should_visit = self._matcher.visitdir(tree[:-1]) | |
1376 | if tree and not should_visit: |
|
1379 | if tree and not should_visit: | |
1377 | continue |
|
1380 | continue | |
1378 |
|
1381 | |||
1379 | store = mfl.getstorage(tree) |
|
1382 | store = mfl.getstorage(tree) | |
1380 |
|
1383 | |||
1381 | if not should_visit: |
|
1384 | if not should_visit: | |
1382 | # No nodes to send because this directory is out of |
|
1385 | # No nodes to send because this directory is out of | |
1383 | # the client's view of the repository (probably |
|
1386 | # the client's view of the repository (probably | |
1384 | # because of narrow clones). Do this even for the root |
|
1387 | # because of narrow clones). Do this even for the root | |
1385 | # directory (tree=='') |
|
1388 | # directory (tree=='') | |
1386 | prunednodes = [] |
|
1389 | prunednodes = [] | |
1387 | else: |
|
1390 | else: | |
1388 | # Avoid sending any manifest nodes we can prove the |
|
1391 | # Avoid sending any manifest nodes we can prove the | |
1389 | # client already has by checking linkrevs. See the |
|
1392 | # client already has by checking linkrevs. See the | |
1390 | # related comment in generatefiles(). |
|
1393 | # related comment in generatefiles(). | |
1391 | prunednodes = self._prunemanifests(store, nodes, commonrevs) |
|
1394 | prunednodes = self._prunemanifests(store, nodes, commonrevs) | |
1392 |
|
1395 | |||
1393 | if tree and not prunednodes: |
|
1396 | if tree and not prunednodes: | |
1394 | continue |
|
1397 | continue | |
1395 |
|
1398 | |||
1396 | lookupfn = makelookupmflinknode(tree, nodes) |
|
1399 | lookupfn = makelookupmflinknode(tree, nodes) | |
1397 |
|
1400 | |||
1398 | deltas = deltagroup( |
|
1401 | deltas = deltagroup( | |
1399 | self._repo, |
|
1402 | self._repo, | |
1400 | store, |
|
1403 | store, | |
1401 | prunednodes, |
|
1404 | prunednodes, | |
1402 | False, |
|
1405 | False, | |
1403 | lookupfn, |
|
1406 | lookupfn, | |
1404 | self._forcedeltaparentprev, |
|
1407 | self._forcedeltaparentprev, | |
1405 | ellipses=self._ellipses, |
|
1408 | ellipses=self._ellipses, | |
1406 | topic=_(b'manifests'), |
|
1409 | topic=_(b'manifests'), | |
1407 | clrevtolocalrev=clrevtolocalrev, |
|
1410 | clrevtolocalrev=clrevtolocalrev, | |
1408 | fullclnodes=self._fullclnodes, |
|
1411 | fullclnodes=self._fullclnodes, | |
1409 | precomputedellipsis=self._precomputedellipsis, |
|
1412 | precomputedellipsis=self._precomputedellipsis, | |
1410 | sidedata_helpers=sidedata_helpers, |
|
1413 | sidedata_helpers=sidedata_helpers, | |
1411 | ) |
|
1414 | ) | |
1412 |
|
1415 | |||
1413 | if not self._oldmatcher.visitdir(store.tree[:-1]): |
|
1416 | if not self._oldmatcher.visitdir(store.tree[:-1]): | |
1414 | yield tree, deltas |
|
1417 | yield tree, deltas | |
1415 | else: |
|
1418 | else: | |
1416 | # 'deltas' is a generator and we need to consume it even if |
|
1419 | # 'deltas' is a generator and we need to consume it even if | |
1417 | # we are not going to send it because a side-effect is that |
|
1420 | # we are not going to send it because a side-effect is that | |
1418 | # it updates tmdnodes (via lookupfn) |
|
1421 | # it updates tmdnodes (via lookupfn) | |
1419 | for d in deltas: |
|
1422 | for d in deltas: | |
1420 | pass |
|
1423 | pass | |
1421 | if not tree: |
|
1424 | if not tree: | |
1422 | yield tree, [] |
|
1425 | yield tree, [] | |
1423 |
|
1426 | |||
1424 | def _prunemanifests(self, store, nodes, commonrevs): |
|
1427 | def _prunemanifests(self, store, nodes, commonrevs): | |
1425 | if not self._ellipses: |
|
1428 | if not self._ellipses: | |
1426 | # In non-ellipses case and large repositories, it is better to |
|
1429 | # In non-ellipses case and large repositories, it is better to | |
1427 | # prevent calling of store.rev and store.linkrev on a lot of |
|
1430 | # prevent calling of store.rev and store.linkrev on a lot of | |
1428 | # nodes as compared to sending some extra data |
|
1431 | # nodes as compared to sending some extra data | |
1429 | return nodes.copy() |
|
1432 | return nodes.copy() | |
1430 | # This is split out as a separate method to allow filtering |
|
1433 | # This is split out as a separate method to allow filtering | |
1431 | # commonrevs in extension code. |
|
1434 | # commonrevs in extension code. | |
1432 | # |
|
1435 | # | |
1433 | # TODO(augie): this shouldn't be required, instead we should |
|
1436 | # TODO(augie): this shouldn't be required, instead we should | |
1434 | # make filtering of revisions to send delegated to the store |
|
1437 | # make filtering of revisions to send delegated to the store | |
1435 | # layer. |
|
1438 | # layer. | |
1436 | frev, flr = store.rev, store.linkrev |
|
1439 | frev, flr = store.rev, store.linkrev | |
1437 | return [n for n in nodes if flr(frev(n)) not in commonrevs] |
|
1440 | return [n for n in nodes if flr(frev(n)) not in commonrevs] | |
1438 |
|
1441 | |||
1439 | # The 'source' parameter is useful for extensions |
|
1442 | # The 'source' parameter is useful for extensions | |
1440 | def generatefiles( |
|
1443 | def generatefiles( | |
1441 | self, |
|
1444 | self, | |
1442 | changedfiles, |
|
1445 | changedfiles, | |
1443 | commonrevs, |
|
1446 | commonrevs, | |
1444 | source, |
|
1447 | source, | |
1445 | mfdicts, |
|
1448 | mfdicts, | |
1446 | fastpathlinkrev, |
|
1449 | fastpathlinkrev, | |
1447 | fnodes, |
|
1450 | fnodes, | |
1448 | clrevs, |
|
1451 | clrevs, | |
1449 | sidedata_helpers=None, |
|
1452 | sidedata_helpers=None, | |
1450 | ): |
|
1453 | ): | |
1451 | changedfiles = [ |
|
1454 | changedfiles = [ | |
1452 | f |
|
1455 | f | |
1453 | for f in changedfiles |
|
1456 | for f in changedfiles | |
1454 | if self._matcher(f) and not self._oldmatcher(f) |
|
1457 | if self._matcher(f) and not self._oldmatcher(f) | |
1455 | ] |
|
1458 | ] | |
1456 |
|
1459 | |||
1457 | if not fastpathlinkrev: |
|
1460 | if not fastpathlinkrev: | |
1458 |
|
1461 | |||
1459 | def normallinknodes(unused, fname): |
|
1462 | def normallinknodes(unused, fname): | |
1460 | return fnodes.get(fname, {}) |
|
1463 | return fnodes.get(fname, {}) | |
1461 |
|
1464 | |||
1462 | else: |
|
1465 | else: | |
1463 | cln = self._repo.changelog.node |
|
1466 | cln = self._repo.changelog.node | |
1464 |
|
1467 | |||
1465 | def normallinknodes(store, fname): |
|
1468 | def normallinknodes(store, fname): | |
1466 | flinkrev = store.linkrev |
|
1469 | flinkrev = store.linkrev | |
1467 | fnode = store.node |
|
1470 | fnode = store.node | |
1468 | revs = ((r, flinkrev(r)) for r in store) |
|
1471 | revs = ((r, flinkrev(r)) for r in store) | |
1469 | return {fnode(r): cln(lr) for r, lr in revs if lr in clrevs} |
|
1472 | return {fnode(r): cln(lr) for r, lr in revs if lr in clrevs} | |
1470 |
|
1473 | |||
1471 | clrevtolocalrev = {} |
|
1474 | clrevtolocalrev = {} | |
1472 |
|
1475 | |||
1473 | if self._isshallow: |
|
1476 | if self._isshallow: | |
1474 | # In a shallow clone, the linknodes callback needs to also include |
|
1477 | # In a shallow clone, the linknodes callback needs to also include | |
1475 | # those file nodes that are in the manifests we sent but weren't |
|
1478 | # those file nodes that are in the manifests we sent but weren't | |
1476 | # introduced by those manifests. |
|
1479 | # introduced by those manifests. | |
1477 | commonctxs = [self._repo[c] for c in commonrevs] |
|
1480 | commonctxs = [self._repo[c] for c in commonrevs] | |
1478 | clrev = self._repo.changelog.rev |
|
1481 | clrev = self._repo.changelog.rev | |
1479 |
|
1482 | |||
1480 | def linknodes(flog, fname): |
|
1483 | def linknodes(flog, fname): | |
1481 | for c in commonctxs: |
|
1484 | for c in commonctxs: | |
1482 | try: |
|
1485 | try: | |
1483 | fnode = c.filenode(fname) |
|
1486 | fnode = c.filenode(fname) | |
1484 | clrevtolocalrev[c.rev()] = flog.rev(fnode) |
|
1487 | clrevtolocalrev[c.rev()] = flog.rev(fnode) | |
1485 | except error.ManifestLookupError: |
|
1488 | except error.ManifestLookupError: | |
1486 | pass |
|
1489 | pass | |
1487 | links = normallinknodes(flog, fname) |
|
1490 | links = normallinknodes(flog, fname) | |
1488 | if len(links) != len(mfdicts): |
|
1491 | if len(links) != len(mfdicts): | |
1489 | for mf, lr in mfdicts: |
|
1492 | for mf, lr in mfdicts: | |
1490 | fnode = mf.get(fname, None) |
|
1493 | fnode = mf.get(fname, None) | |
1491 | if fnode in links: |
|
1494 | if fnode in links: | |
1492 | links[fnode] = min(links[fnode], lr, key=clrev) |
|
1495 | links[fnode] = min(links[fnode], lr, key=clrev) | |
1493 | elif fnode: |
|
1496 | elif fnode: | |
1494 | links[fnode] = lr |
|
1497 | links[fnode] = lr | |
1495 | return links |
|
1498 | return links | |
1496 |
|
1499 | |||
1497 | else: |
|
1500 | else: | |
1498 | linknodes = normallinknodes |
|
1501 | linknodes = normallinknodes | |
1499 |
|
1502 | |||
1500 | repo = self._repo |
|
1503 | repo = self._repo | |
1501 | progress = repo.ui.makeprogress( |
|
1504 | progress = repo.ui.makeprogress( | |
1502 | _(b'files'), unit=_(b'files'), total=len(changedfiles) |
|
1505 | _(b'files'), unit=_(b'files'), total=len(changedfiles) | |
1503 | ) |
|
1506 | ) | |
1504 | for i, fname in enumerate(sorted(changedfiles)): |
|
1507 | for i, fname in enumerate(sorted(changedfiles)): | |
1505 | filerevlog = repo.file(fname) |
|
1508 | filerevlog = repo.file(fname) | |
1506 | if not filerevlog: |
|
1509 | if not filerevlog: | |
1507 | raise error.Abort( |
|
1510 | raise error.Abort( | |
1508 | _(b"empty or missing file data for %s") % fname |
|
1511 | _(b"empty or missing file data for %s") % fname | |
1509 | ) |
|
1512 | ) | |
1510 |
|
1513 | |||
1511 | clrevtolocalrev.clear() |
|
1514 | clrevtolocalrev.clear() | |
1512 |
|
1515 | |||
1513 | linkrevnodes = linknodes(filerevlog, fname) |
|
1516 | linkrevnodes = linknodes(filerevlog, fname) | |
1514 | # Lookup for filenodes, we collected the linkrev nodes above in the |
|
1517 | # Lookup for filenodes, we collected the linkrev nodes above in the | |
1515 | # fastpath case and with lookupmf in the slowpath case. |
|
1518 | # fastpath case and with lookupmf in the slowpath case. | |
1516 | def lookupfilelog(x): |
|
1519 | def lookupfilelog(x): | |
1517 | return linkrevnodes[x] |
|
1520 | return linkrevnodes[x] | |
1518 |
|
1521 | |||
1519 | frev, flr = filerevlog.rev, filerevlog.linkrev |
|
1522 | frev, flr = filerevlog.rev, filerevlog.linkrev | |
1520 | # Skip sending any filenode we know the client already |
|
1523 | # Skip sending any filenode we know the client already | |
1521 | # has. This avoids over-sending files relatively |
|
1524 | # has. This avoids over-sending files relatively | |
1522 | # inexpensively, so it's not a problem if we under-filter |
|
1525 | # inexpensively, so it's not a problem if we under-filter | |
1523 | # here. |
|
1526 | # here. | |
1524 | filenodes = [ |
|
1527 | filenodes = [ | |
1525 | n for n in linkrevnodes if flr(frev(n)) not in commonrevs |
|
1528 | n for n in linkrevnodes if flr(frev(n)) not in commonrevs | |
1526 | ] |
|
1529 | ] | |
1527 |
|
1530 | |||
1528 | if not filenodes: |
|
1531 | if not filenodes: | |
1529 | continue |
|
1532 | continue | |
1530 |
|
1533 | |||
1531 | progress.update(i + 1, item=fname) |
|
1534 | progress.update(i + 1, item=fname) | |
1532 |
|
1535 | |||
1533 | deltas = deltagroup( |
|
1536 | deltas = deltagroup( | |
1534 | self._repo, |
|
1537 | self._repo, | |
1535 | filerevlog, |
|
1538 | filerevlog, | |
1536 | filenodes, |
|
1539 | filenodes, | |
1537 | False, |
|
1540 | False, | |
1538 | lookupfilelog, |
|
1541 | lookupfilelog, | |
1539 | self._forcedeltaparentprev, |
|
1542 | self._forcedeltaparentprev, | |
1540 | ellipses=self._ellipses, |
|
1543 | ellipses=self._ellipses, | |
1541 | clrevtolocalrev=clrevtolocalrev, |
|
1544 | clrevtolocalrev=clrevtolocalrev, | |
1542 | fullclnodes=self._fullclnodes, |
|
1545 | fullclnodes=self._fullclnodes, | |
1543 | precomputedellipsis=self._precomputedellipsis, |
|
1546 | precomputedellipsis=self._precomputedellipsis, | |
1544 | sidedata_helpers=sidedata_helpers, |
|
1547 | sidedata_helpers=sidedata_helpers, | |
1545 | ) |
|
1548 | ) | |
1546 |
|
1549 | |||
1547 | yield fname, deltas |
|
1550 | yield fname, deltas | |
1548 |
|
1551 | |||
1549 | progress.complete() |
|
1552 | progress.complete() | |
1550 |
|
1553 | |||
1551 |
|
1554 | |||
1552 | def _makecg1packer( |
|
1555 | def _makecg1packer( | |
1553 | repo, |
|
1556 | repo, | |
1554 | oldmatcher, |
|
1557 | oldmatcher, | |
1555 | matcher, |
|
1558 | matcher, | |
1556 | bundlecaps, |
|
1559 | bundlecaps, | |
1557 | ellipses=False, |
|
1560 | ellipses=False, | |
1558 | shallow=False, |
|
1561 | shallow=False, | |
1559 | ellipsisroots=None, |
|
1562 | ellipsisroots=None, | |
1560 | fullnodes=None, |
|
1563 | fullnodes=None, | |
1561 | remote_sidedata=None, |
|
1564 | remote_sidedata=None, | |
1562 | ): |
|
1565 | ): | |
1563 | builddeltaheader = lambda d: _CHANGEGROUPV1_DELTA_HEADER.pack( |
|
1566 | builddeltaheader = lambda d: _CHANGEGROUPV1_DELTA_HEADER.pack( | |
1564 | d.node, d.p1node, d.p2node, d.linknode |
|
1567 | d.node, d.p1node, d.p2node, d.linknode | |
1565 | ) |
|
1568 | ) | |
1566 |
|
1569 | |||
1567 | return cgpacker( |
|
1570 | return cgpacker( | |
1568 | repo, |
|
1571 | repo, | |
1569 | oldmatcher, |
|
1572 | oldmatcher, | |
1570 | matcher, |
|
1573 | matcher, | |
1571 | b'01', |
|
1574 | b'01', | |
1572 | builddeltaheader=builddeltaheader, |
|
1575 | builddeltaheader=builddeltaheader, | |
1573 | manifestsend=b'', |
|
1576 | manifestsend=b'', | |
1574 | forcedeltaparentprev=True, |
|
1577 | forcedeltaparentprev=True, | |
1575 | bundlecaps=bundlecaps, |
|
1578 | bundlecaps=bundlecaps, | |
1576 | ellipses=ellipses, |
|
1579 | ellipses=ellipses, | |
1577 | shallow=shallow, |
|
1580 | shallow=shallow, | |
1578 | ellipsisroots=ellipsisroots, |
|
1581 | ellipsisroots=ellipsisroots, | |
1579 | fullnodes=fullnodes, |
|
1582 | fullnodes=fullnodes, | |
1580 | ) |
|
1583 | ) | |
1581 |
|
1584 | |||
1582 |
|
1585 | |||
1583 | def _makecg2packer( |
|
1586 | def _makecg2packer( | |
1584 | repo, |
|
1587 | repo, | |
1585 | oldmatcher, |
|
1588 | oldmatcher, | |
1586 | matcher, |
|
1589 | matcher, | |
1587 | bundlecaps, |
|
1590 | bundlecaps, | |
1588 | ellipses=False, |
|
1591 | ellipses=False, | |
1589 | shallow=False, |
|
1592 | shallow=False, | |
1590 | ellipsisroots=None, |
|
1593 | ellipsisroots=None, | |
1591 | fullnodes=None, |
|
1594 | fullnodes=None, | |
1592 | remote_sidedata=None, |
|
1595 | remote_sidedata=None, | |
1593 | ): |
|
1596 | ): | |
1594 | builddeltaheader = lambda d: _CHANGEGROUPV2_DELTA_HEADER.pack( |
|
1597 | builddeltaheader = lambda d: _CHANGEGROUPV2_DELTA_HEADER.pack( | |
1595 | d.node, d.p1node, d.p2node, d.basenode, d.linknode |
|
1598 | d.node, d.p1node, d.p2node, d.basenode, d.linknode | |
1596 | ) |
|
1599 | ) | |
1597 |
|
1600 | |||
1598 | return cgpacker( |
|
1601 | return cgpacker( | |
1599 | repo, |
|
1602 | repo, | |
1600 | oldmatcher, |
|
1603 | oldmatcher, | |
1601 | matcher, |
|
1604 | matcher, | |
1602 | b'02', |
|
1605 | b'02', | |
1603 | builddeltaheader=builddeltaheader, |
|
1606 | builddeltaheader=builddeltaheader, | |
1604 | manifestsend=b'', |
|
1607 | manifestsend=b'', | |
1605 | bundlecaps=bundlecaps, |
|
1608 | bundlecaps=bundlecaps, | |
1606 | ellipses=ellipses, |
|
1609 | ellipses=ellipses, | |
1607 | shallow=shallow, |
|
1610 | shallow=shallow, | |
1608 | ellipsisroots=ellipsisroots, |
|
1611 | ellipsisroots=ellipsisroots, | |
1609 | fullnodes=fullnodes, |
|
1612 | fullnodes=fullnodes, | |
1610 | ) |
|
1613 | ) | |
1611 |
|
1614 | |||
1612 |
|
1615 | |||
1613 | def _makecg3packer( |
|
1616 | def _makecg3packer( | |
1614 | repo, |
|
1617 | repo, | |
1615 | oldmatcher, |
|
1618 | oldmatcher, | |
1616 | matcher, |
|
1619 | matcher, | |
1617 | bundlecaps, |
|
1620 | bundlecaps, | |
1618 | ellipses=False, |
|
1621 | ellipses=False, | |
1619 | shallow=False, |
|
1622 | shallow=False, | |
1620 | ellipsisroots=None, |
|
1623 | ellipsisroots=None, | |
1621 | fullnodes=None, |
|
1624 | fullnodes=None, | |
1622 | remote_sidedata=None, |
|
1625 | remote_sidedata=None, | |
1623 | ): |
|
1626 | ): | |
1624 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( |
|
1627 | builddeltaheader = lambda d: _CHANGEGROUPV3_DELTA_HEADER.pack( | |
1625 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags |
|
1628 | d.node, d.p1node, d.p2node, d.basenode, d.linknode, d.flags | |
1626 | ) |
|
1629 | ) | |
1627 |
|
1630 | |||
1628 | return cgpacker( |
|
1631 | return cgpacker( | |
1629 | repo, |
|
1632 | repo, | |
1630 | oldmatcher, |
|
1633 | oldmatcher, | |
1631 | matcher, |
|
1634 | matcher, | |
1632 | b'03', |
|
1635 | b'03', | |
1633 | builddeltaheader=builddeltaheader, |
|
1636 | builddeltaheader=builddeltaheader, | |
1634 | manifestsend=closechunk(), |
|
1637 | manifestsend=closechunk(), | |
1635 | bundlecaps=bundlecaps, |
|
1638 | bundlecaps=bundlecaps, | |
1636 | ellipses=ellipses, |
|
1639 | ellipses=ellipses, | |
1637 | shallow=shallow, |
|
1640 | shallow=shallow, | |
1638 | ellipsisroots=ellipsisroots, |
|
1641 | ellipsisroots=ellipsisroots, | |
1639 | fullnodes=fullnodes, |
|
1642 | fullnodes=fullnodes, | |
1640 | ) |
|
1643 | ) | |
1641 |
|
1644 | |||
1642 |
|
1645 | |||
1643 | def _makecg4packer( |
|
1646 | def _makecg4packer( | |
1644 | repo, |
|
1647 | repo, | |
1645 | oldmatcher, |
|
1648 | oldmatcher, | |
1646 | matcher, |
|
1649 | matcher, | |
1647 | bundlecaps, |
|
1650 | bundlecaps, | |
1648 | ellipses=False, |
|
1651 | ellipses=False, | |
1649 | shallow=False, |
|
1652 | shallow=False, | |
1650 | ellipsisroots=None, |
|
1653 | ellipsisroots=None, | |
1651 | fullnodes=None, |
|
1654 | fullnodes=None, | |
1652 | remote_sidedata=None, |
|
1655 | remote_sidedata=None, | |
1653 | ): |
|
1656 | ): | |
1654 | # Sidedata is in a separate chunk from the delta to differentiate |
|
1657 | # Sidedata is in a separate chunk from the delta to differentiate | |
1655 | # "raw delta" and sidedata. |
|
1658 | # "raw delta" and sidedata. | |
1656 | def builddeltaheader(d): |
|
1659 | def builddeltaheader(d): | |
1657 | return _CHANGEGROUPV4_DELTA_HEADER.pack( |
|
1660 | return _CHANGEGROUPV4_DELTA_HEADER.pack( | |
1658 | d.protocol_flags, |
|
1661 | d.protocol_flags, | |
1659 | d.node, |
|
1662 | d.node, | |
1660 | d.p1node, |
|
1663 | d.p1node, | |
1661 | d.p2node, |
|
1664 | d.p2node, | |
1662 | d.basenode, |
|
1665 | d.basenode, | |
1663 | d.linknode, |
|
1666 | d.linknode, | |
1664 | d.flags, |
|
1667 | d.flags, | |
1665 | ) |
|
1668 | ) | |
1666 |
|
1669 | |||
1667 | return cgpacker( |
|
1670 | return cgpacker( | |
1668 | repo, |
|
1671 | repo, | |
1669 | oldmatcher, |
|
1672 | oldmatcher, | |
1670 | matcher, |
|
1673 | matcher, | |
1671 | b'04', |
|
1674 | b'04', | |
1672 | builddeltaheader=builddeltaheader, |
|
1675 | builddeltaheader=builddeltaheader, | |
1673 | manifestsend=closechunk(), |
|
1676 | manifestsend=closechunk(), | |
1674 | bundlecaps=bundlecaps, |
|
1677 | bundlecaps=bundlecaps, | |
1675 | ellipses=ellipses, |
|
1678 | ellipses=ellipses, | |
1676 | shallow=shallow, |
|
1679 | shallow=shallow, | |
1677 | ellipsisroots=ellipsisroots, |
|
1680 | ellipsisroots=ellipsisroots, | |
1678 | fullnodes=fullnodes, |
|
1681 | fullnodes=fullnodes, | |
1679 | remote_sidedata=remote_sidedata, |
|
1682 | remote_sidedata=remote_sidedata, | |
1680 | ) |
|
1683 | ) | |
1681 |
|
1684 | |||
1682 |
|
1685 | |||
1683 | _packermap = { |
|
1686 | _packermap = { | |
1684 | b'01': (_makecg1packer, cg1unpacker), |
|
1687 | b'01': (_makecg1packer, cg1unpacker), | |
1685 | # cg2 adds support for exchanging generaldelta |
|
1688 | # cg2 adds support for exchanging generaldelta | |
1686 | b'02': (_makecg2packer, cg2unpacker), |
|
1689 | b'02': (_makecg2packer, cg2unpacker), | |
1687 | # cg3 adds support for exchanging revlog flags and treemanifests |
|
1690 | # cg3 adds support for exchanging revlog flags and treemanifests | |
1688 | b'03': (_makecg3packer, cg3unpacker), |
|
1691 | b'03': (_makecg3packer, cg3unpacker), | |
1689 | # ch4 adds support for exchanging sidedata |
|
1692 | # ch4 adds support for exchanging sidedata | |
1690 | b'04': (_makecg4packer, cg4unpacker), |
|
1693 | b'04': (_makecg4packer, cg4unpacker), | |
1691 | } |
|
1694 | } | |
1692 |
|
1695 | |||
1693 |
|
1696 | |||
1694 | def allsupportedversions(repo): |
|
1697 | def allsupportedversions(repo): | |
1695 | versions = set(_packermap.keys()) |
|
1698 | versions = set(_packermap.keys()) | |
1696 | needv03 = False |
|
1699 | needv03 = False | |
1697 | if ( |
|
1700 | if ( | |
1698 | repo.ui.configbool(b'experimental', b'changegroup3') |
|
1701 | repo.ui.configbool(b'experimental', b'changegroup3') | |
1699 | or repo.ui.configbool(b'experimental', b'treemanifest') |
|
1702 | or repo.ui.configbool(b'experimental', b'treemanifest') | |
1700 | or scmutil.istreemanifest(repo) |
|
1703 | or scmutil.istreemanifest(repo) | |
1701 | ): |
|
1704 | ): | |
1702 | # we keep version 03 because we need to to exchange treemanifest data |
|
1705 | # we keep version 03 because we need to to exchange treemanifest data | |
1703 | # |
|
1706 | # | |
1704 | # we also keep vresion 01 and 02, because it is possible for repo to |
|
1707 | # we also keep vresion 01 and 02, because it is possible for repo to | |
1705 | # contains both normal and tree manifest at the same time. so using |
|
1708 | # contains both normal and tree manifest at the same time. so using | |
1706 | # older version to pull data is viable |
|
1709 | # older version to pull data is viable | |
1707 | # |
|
1710 | # | |
1708 | # (or even to push subset of history) |
|
1711 | # (or even to push subset of history) | |
1709 | needv03 = True |
|
1712 | needv03 = True | |
1710 | if not needv03: |
|
1713 | if not needv03: | |
1711 | versions.discard(b'03') |
|
1714 | versions.discard(b'03') | |
1712 | want_v4 = ( |
|
1715 | want_v4 = ( | |
1713 | repo.ui.configbool(b'experimental', b'changegroup4') |
|
1716 | repo.ui.configbool(b'experimental', b'changegroup4') | |
1714 | or requirements.REVLOGV2_REQUIREMENT in repo.requirements |
|
1717 | or requirements.REVLOGV2_REQUIREMENT in repo.requirements | |
1715 | ) |
|
1718 | ) | |
1716 | if not want_v4: |
|
1719 | if not want_v4: | |
1717 | versions.discard(b'04') |
|
1720 | versions.discard(b'04') | |
1718 | return versions |
|
1721 | return versions | |
1719 |
|
1722 | |||
1720 |
|
1723 | |||
1721 | # Changegroup versions that can be applied to the repo |
|
1724 | # Changegroup versions that can be applied to the repo | |
1722 | def supportedincomingversions(repo): |
|
1725 | def supportedincomingversions(repo): | |
1723 | return allsupportedversions(repo) |
|
1726 | return allsupportedversions(repo) | |
1724 |
|
1727 | |||
1725 |
|
1728 | |||
1726 | # Changegroup versions that can be created from the repo |
|
1729 | # Changegroup versions that can be created from the repo | |
1727 | def supportedoutgoingversions(repo): |
|
1730 | def supportedoutgoingversions(repo): | |
1728 | versions = allsupportedversions(repo) |
|
1731 | versions = allsupportedversions(repo) | |
1729 | if scmutil.istreemanifest(repo): |
|
1732 | if scmutil.istreemanifest(repo): | |
1730 | # Versions 01 and 02 support only flat manifests and it's just too |
|
1733 | # Versions 01 and 02 support only flat manifests and it's just too | |
1731 | # expensive to convert between the flat manifest and tree manifest on |
|
1734 | # expensive to convert between the flat manifest and tree manifest on | |
1732 | # the fly. Since tree manifests are hashed differently, all of history |
|
1735 | # the fly. Since tree manifests are hashed differently, all of history | |
1733 | # would have to be converted. Instead, we simply don't even pretend to |
|
1736 | # would have to be converted. Instead, we simply don't even pretend to | |
1734 | # support versions 01 and 02. |
|
1737 | # support versions 01 and 02. | |
1735 | versions.discard(b'01') |
|
1738 | versions.discard(b'01') | |
1736 | versions.discard(b'02') |
|
1739 | versions.discard(b'02') | |
1737 | if requirements.NARROW_REQUIREMENT in repo.requirements: |
|
1740 | if requirements.NARROW_REQUIREMENT in repo.requirements: | |
1738 | # Versions 01 and 02 don't support revlog flags, and we need to |
|
1741 | # Versions 01 and 02 don't support revlog flags, and we need to | |
1739 | # support that for stripping and unbundling to work. |
|
1742 | # support that for stripping and unbundling to work. | |
1740 | versions.discard(b'01') |
|
1743 | versions.discard(b'01') | |
1741 | versions.discard(b'02') |
|
1744 | versions.discard(b'02') | |
1742 | if LFS_REQUIREMENT in repo.requirements: |
|
1745 | if LFS_REQUIREMENT in repo.requirements: | |
1743 | # Versions 01 and 02 don't support revlog flags, and we need to |
|
1746 | # Versions 01 and 02 don't support revlog flags, and we need to | |
1744 | # mark LFS entries with REVIDX_EXTSTORED. |
|
1747 | # mark LFS entries with REVIDX_EXTSTORED. | |
1745 | versions.discard(b'01') |
|
1748 | versions.discard(b'01') | |
1746 | versions.discard(b'02') |
|
1749 | versions.discard(b'02') | |
1747 |
|
1750 | |||
1748 | return versions |
|
1751 | return versions | |
1749 |
|
1752 | |||
1750 |
|
1753 | |||
1751 | def localversion(repo): |
|
1754 | def localversion(repo): | |
1752 | # Finds the best version to use for bundles that are meant to be used |
|
1755 | # Finds the best version to use for bundles that are meant to be used | |
1753 | # locally, such as those from strip and shelve, and temporary bundles. |
|
1756 | # locally, such as those from strip and shelve, and temporary bundles. | |
1754 | return max(supportedoutgoingversions(repo)) |
|
1757 | return max(supportedoutgoingversions(repo)) | |
1755 |
|
1758 | |||
1756 |
|
1759 | |||
1757 | def safeversion(repo): |
|
1760 | def safeversion(repo): | |
1758 | # Finds the smallest version that it's safe to assume clients of the repo |
|
1761 | # Finds the smallest version that it's safe to assume clients of the repo | |
1759 | # will support. For example, all hg versions that support generaldelta also |
|
1762 | # will support. For example, all hg versions that support generaldelta also | |
1760 | # support changegroup 02. |
|
1763 | # support changegroup 02. | |
1761 | versions = supportedoutgoingversions(repo) |
|
1764 | versions = supportedoutgoingversions(repo) | |
1762 | if requirements.GENERALDELTA_REQUIREMENT in repo.requirements: |
|
1765 | if requirements.GENERALDELTA_REQUIREMENT in repo.requirements: | |
1763 | versions.discard(b'01') |
|
1766 | versions.discard(b'01') | |
1764 | assert versions |
|
1767 | assert versions | |
1765 | return min(versions) |
|
1768 | return min(versions) | |
1766 |
|
1769 | |||
1767 |
|
1770 | |||
1768 | def getbundler( |
|
1771 | def getbundler( | |
1769 | version, |
|
1772 | version, | |
1770 | repo, |
|
1773 | repo, | |
1771 | bundlecaps=None, |
|
1774 | bundlecaps=None, | |
1772 | oldmatcher=None, |
|
1775 | oldmatcher=None, | |
1773 | matcher=None, |
|
1776 | matcher=None, | |
1774 | ellipses=False, |
|
1777 | ellipses=False, | |
1775 | shallow=False, |
|
1778 | shallow=False, | |
1776 | ellipsisroots=None, |
|
1779 | ellipsisroots=None, | |
1777 | fullnodes=None, |
|
1780 | fullnodes=None, | |
1778 | remote_sidedata=None, |
|
1781 | remote_sidedata=None, | |
1779 | ): |
|
1782 | ): | |
1780 | assert version in supportedoutgoingversions(repo) |
|
1783 | assert version in supportedoutgoingversions(repo) | |
1781 |
|
1784 | |||
1782 | if matcher is None: |
|
1785 | if matcher is None: | |
1783 | matcher = matchmod.always() |
|
1786 | matcher = matchmod.always() | |
1784 | if oldmatcher is None: |
|
1787 | if oldmatcher is None: | |
1785 | oldmatcher = matchmod.never() |
|
1788 | oldmatcher = matchmod.never() | |
1786 |
|
1789 | |||
1787 | if version == b'01' and not matcher.always(): |
|
1790 | if version == b'01' and not matcher.always(): | |
1788 | raise error.ProgrammingError( |
|
1791 | raise error.ProgrammingError( | |
1789 | b'version 01 changegroups do not support sparse file matchers' |
|
1792 | b'version 01 changegroups do not support sparse file matchers' | |
1790 | ) |
|
1793 | ) | |
1791 |
|
1794 | |||
1792 | if ellipses and version in (b'01', b'02'): |
|
1795 | if ellipses and version in (b'01', b'02'): | |
1793 | raise error.Abort( |
|
1796 | raise error.Abort( | |
1794 | _( |
|
1797 | _( | |
1795 | b'ellipsis nodes require at least cg3 on client and server, ' |
|
1798 | b'ellipsis nodes require at least cg3 on client and server, ' | |
1796 | b'but negotiated version %s' |
|
1799 | b'but negotiated version %s' | |
1797 | ) |
|
1800 | ) | |
1798 | % version |
|
1801 | % version | |
1799 | ) |
|
1802 | ) | |
1800 |
|
1803 | |||
1801 | # Requested files could include files not in the local store. So |
|
1804 | # Requested files could include files not in the local store. So | |
1802 | # filter those out. |
|
1805 | # filter those out. | |
1803 | matcher = repo.narrowmatch(matcher) |
|
1806 | matcher = repo.narrowmatch(matcher) | |
1804 |
|
1807 | |||
1805 | fn = _packermap[version][0] |
|
1808 | fn = _packermap[version][0] | |
1806 | return fn( |
|
1809 | return fn( | |
1807 | repo, |
|
1810 | repo, | |
1808 | oldmatcher, |
|
1811 | oldmatcher, | |
1809 | matcher, |
|
1812 | matcher, | |
1810 | bundlecaps, |
|
1813 | bundlecaps, | |
1811 | ellipses=ellipses, |
|
1814 | ellipses=ellipses, | |
1812 | shallow=shallow, |
|
1815 | shallow=shallow, | |
1813 | ellipsisroots=ellipsisroots, |
|
1816 | ellipsisroots=ellipsisroots, | |
1814 | fullnodes=fullnodes, |
|
1817 | fullnodes=fullnodes, | |
1815 | remote_sidedata=remote_sidedata, |
|
1818 | remote_sidedata=remote_sidedata, | |
1816 | ) |
|
1819 | ) | |
1817 |
|
1820 | |||
1818 |
|
1821 | |||
1819 | def getunbundler(version, fh, alg, extras=None): |
|
1822 | def getunbundler(version, fh, alg, extras=None): | |
1820 | return _packermap[version][1](fh, alg, extras=extras) |
|
1823 | return _packermap[version][1](fh, alg, extras=extras) | |
1821 |
|
1824 | |||
1822 |
|
1825 | |||
1823 | def _changegroupinfo(repo, nodes, source): |
|
1826 | def _changegroupinfo(repo, nodes, source): | |
1824 | if repo.ui.verbose or source == b'bundle': |
|
1827 | if repo.ui.verbose or source == b'bundle': | |
1825 | repo.ui.status(_(b"%d changesets found\n") % len(nodes)) |
|
1828 | repo.ui.status(_(b"%d changesets found\n") % len(nodes)) | |
1826 | if repo.ui.debugflag: |
|
1829 | if repo.ui.debugflag: | |
1827 | repo.ui.debug(b"list of changesets:\n") |
|
1830 | repo.ui.debug(b"list of changesets:\n") | |
1828 | for node in nodes: |
|
1831 | for node in nodes: | |
1829 | repo.ui.debug(b"%s\n" % hex(node)) |
|
1832 | repo.ui.debug(b"%s\n" % hex(node)) | |
1830 |
|
1833 | |||
1831 |
|
1834 | |||
1832 | def makechangegroup( |
|
1835 | def makechangegroup( | |
1833 | repo, outgoing, version, source, fastpath=False, bundlecaps=None |
|
1836 | repo, outgoing, version, source, fastpath=False, bundlecaps=None | |
1834 | ): |
|
1837 | ): | |
1835 | cgstream = makestream( |
|
1838 | cgstream = makestream( | |
1836 | repo, |
|
1839 | repo, | |
1837 | outgoing, |
|
1840 | outgoing, | |
1838 | version, |
|
1841 | version, | |
1839 | source, |
|
1842 | source, | |
1840 | fastpath=fastpath, |
|
1843 | fastpath=fastpath, | |
1841 | bundlecaps=bundlecaps, |
|
1844 | bundlecaps=bundlecaps, | |
1842 | ) |
|
1845 | ) | |
1843 | return getunbundler( |
|
1846 | return getunbundler( | |
1844 | version, |
|
1847 | version, | |
1845 | util.chunkbuffer(cgstream), |
|
1848 | util.chunkbuffer(cgstream), | |
1846 | None, |
|
1849 | None, | |
1847 | {b'clcount': len(outgoing.missing)}, |
|
1850 | {b'clcount': len(outgoing.missing)}, | |
1848 | ) |
|
1851 | ) | |
1849 |
|
1852 | |||
1850 |
|
1853 | |||
1851 | def makestream( |
|
1854 | def makestream( | |
1852 | repo, |
|
1855 | repo, | |
1853 | outgoing, |
|
1856 | outgoing, | |
1854 | version, |
|
1857 | version, | |
1855 | source, |
|
1858 | source, | |
1856 | fastpath=False, |
|
1859 | fastpath=False, | |
1857 | bundlecaps=None, |
|
1860 | bundlecaps=None, | |
1858 | matcher=None, |
|
1861 | matcher=None, | |
1859 | remote_sidedata=None, |
|
1862 | remote_sidedata=None, | |
1860 | ): |
|
1863 | ): | |
1861 | bundler = getbundler( |
|
1864 | bundler = getbundler( | |
1862 | version, |
|
1865 | version, | |
1863 | repo, |
|
1866 | repo, | |
1864 | bundlecaps=bundlecaps, |
|
1867 | bundlecaps=bundlecaps, | |
1865 | matcher=matcher, |
|
1868 | matcher=matcher, | |
1866 | remote_sidedata=remote_sidedata, |
|
1869 | remote_sidedata=remote_sidedata, | |
1867 | ) |
|
1870 | ) | |
1868 |
|
1871 | |||
1869 | repo = repo.unfiltered() |
|
1872 | repo = repo.unfiltered() | |
1870 | commonrevs = outgoing.common |
|
1873 | commonrevs = outgoing.common | |
1871 | csets = outgoing.missing |
|
1874 | csets = outgoing.missing | |
1872 | heads = outgoing.ancestorsof |
|
1875 | heads = outgoing.ancestorsof | |
1873 | # We go through the fast path if we get told to, or if all (unfiltered |
|
1876 | # We go through the fast path if we get told to, or if all (unfiltered | |
1874 | # heads have been requested (since we then know there all linkrevs will |
|
1877 | # heads have been requested (since we then know there all linkrevs will | |
1875 | # be pulled by the client). |
|
1878 | # be pulled by the client). | |
1876 | heads.sort() |
|
1879 | heads.sort() | |
1877 | fastpathlinkrev = fastpath or ( |
|
1880 | fastpathlinkrev = fastpath or ( | |
1878 | repo.filtername is None and heads == sorted(repo.heads()) |
|
1881 | repo.filtername is None and heads == sorted(repo.heads()) | |
1879 | ) |
|
1882 | ) | |
1880 |
|
1883 | |||
1881 | repo.hook(b'preoutgoing', throw=True, source=source) |
|
1884 | repo.hook(b'preoutgoing', throw=True, source=source) | |
1882 | _changegroupinfo(repo, csets, source) |
|
1885 | _changegroupinfo(repo, csets, source) | |
1883 | return bundler.generate(commonrevs, csets, fastpathlinkrev, source) |
|
1886 | return bundler.generate(commonrevs, csets, fastpathlinkrev, source) | |
1884 |
|
1887 | |||
1885 |
|
1888 | |||
1886 | def _addchangegroupfiles( |
|
1889 | def _addchangegroupfiles( | |
1887 | repo, |
|
1890 | repo, | |
1888 | source, |
|
1891 | source, | |
1889 | revmap, |
|
1892 | revmap, | |
1890 | trp, |
|
1893 | trp, | |
1891 | expectedfiles, |
|
1894 | expectedfiles, | |
1892 | needfiles, |
|
1895 | needfiles, | |
1893 | addrevisioncb=None, |
|
1896 | addrevisioncb=None, | |
1894 | ): |
|
1897 | ): | |
1895 | revisions = 0 |
|
1898 | revisions = 0 | |
1896 | files = 0 |
|
1899 | files = 0 | |
1897 | progress = repo.ui.makeprogress( |
|
1900 | progress = repo.ui.makeprogress( | |
1898 | _(b'files'), unit=_(b'files'), total=expectedfiles |
|
1901 | _(b'files'), unit=_(b'files'), total=expectedfiles | |
1899 | ) |
|
1902 | ) | |
1900 | for chunkdata in iter(source.filelogheader, {}): |
|
1903 | for chunkdata in iter(source.filelogheader, {}): | |
1901 | files += 1 |
|
1904 | files += 1 | |
1902 | f = chunkdata[b"filename"] |
|
1905 | f = chunkdata[b"filename"] | |
1903 | repo.ui.debug(b"adding %s revisions\n" % f) |
|
1906 | repo.ui.debug(b"adding %s revisions\n" % f) | |
1904 | progress.increment() |
|
1907 | progress.increment() | |
1905 | fl = repo.file(f) |
|
1908 | fl = repo.file(f) | |
1906 | o = len(fl) |
|
1909 | o = len(fl) | |
1907 | try: |
|
1910 | try: | |
1908 | deltas = source.deltaiter() |
|
1911 | deltas = source.deltaiter() | |
1909 | added = fl.addgroup( |
|
1912 | added = fl.addgroup( | |
1910 | deltas, |
|
1913 | deltas, | |
1911 | revmap, |
|
1914 | revmap, | |
1912 | trp, |
|
1915 | trp, | |
1913 | addrevisioncb=addrevisioncb, |
|
1916 | addrevisioncb=addrevisioncb, | |
1914 | ) |
|
1917 | ) | |
1915 | if not added: |
|
1918 | if not added: | |
1916 | raise error.Abort(_(b"received file revlog group is empty")) |
|
1919 | raise error.Abort(_(b"received file revlog group is empty")) | |
1917 | except error.CensoredBaseError as e: |
|
1920 | except error.CensoredBaseError as e: | |
1918 | raise error.Abort(_(b"received delta base is censored: %s") % e) |
|
1921 | raise error.Abort(_(b"received delta base is censored: %s") % e) | |
1919 | revisions += len(fl) - o |
|
1922 | revisions += len(fl) - o | |
1920 | if f in needfiles: |
|
1923 | if f in needfiles: | |
1921 | needs = needfiles[f] |
|
1924 | needs = needfiles[f] | |
1922 | for new in pycompat.xrange(o, len(fl)): |
|
1925 | for new in pycompat.xrange(o, len(fl)): | |
1923 | n = fl.node(new) |
|
1926 | n = fl.node(new) | |
1924 | if n in needs: |
|
1927 | if n in needs: | |
1925 | needs.remove(n) |
|
1928 | needs.remove(n) | |
1926 | else: |
|
1929 | else: | |
1927 | raise error.Abort(_(b"received spurious file revlog entry")) |
|
1930 | raise error.Abort(_(b"received spurious file revlog entry")) | |
1928 | if not needs: |
|
1931 | if not needs: | |
1929 | del needfiles[f] |
|
1932 | del needfiles[f] | |
1930 | progress.complete() |
|
1933 | progress.complete() | |
1931 |
|
1934 | |||
1932 | for f, needs in pycompat.iteritems(needfiles): |
|
1935 | for f, needs in pycompat.iteritems(needfiles): | |
1933 | fl = repo.file(f) |
|
1936 | fl = repo.file(f) | |
1934 | for n in needs: |
|
1937 | for n in needs: | |
1935 | try: |
|
1938 | try: | |
1936 | fl.rev(n) |
|
1939 | fl.rev(n) | |
1937 | except error.LookupError: |
|
1940 | except error.LookupError: | |
1938 | raise error.Abort( |
|
1941 | raise error.Abort( | |
1939 | _(b'missing file data for %s:%s - run hg verify') |
|
1942 | _(b'missing file data for %s:%s - run hg verify') | |
1940 | % (f, hex(n)) |
|
1943 | % (f, hex(n)) | |
1941 | ) |
|
1944 | ) | |
1942 |
|
1945 | |||
1943 | return revisions, files |
|
1946 | return revisions, files |
@@ -1,3129 +1,3130 b'' | |||||
1 | # revlog.py - storage back-end for mercurial |
|
1 | # revlog.py - storage back-end for mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> |
|
3 | # Copyright 2005-2007 Olivia Mackall <olivia@selenic.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | """Storage back-end for Mercurial. |
|
8 | """Storage back-end for Mercurial. | |
9 |
|
9 | |||
10 | This provides efficient delta storage with O(1) retrieve and append |
|
10 | This provides efficient delta storage with O(1) retrieve and append | |
11 | and O(changes) merge between branches. |
|
11 | and O(changes) merge between branches. | |
12 | """ |
|
12 | """ | |
13 |
|
13 | |||
14 | from __future__ import absolute_import |
|
14 | from __future__ import absolute_import | |
15 |
|
15 | |||
16 | import binascii |
|
16 | import binascii | |
17 | import collections |
|
17 | import collections | |
18 | import contextlib |
|
18 | import contextlib | |
19 | import errno |
|
19 | import errno | |
20 | import io |
|
20 | import io | |
21 | import os |
|
21 | import os | |
22 | import struct |
|
22 | import struct | |
23 | import zlib |
|
23 | import zlib | |
24 |
|
24 | |||
25 | # import stuff from node for others to import from revlog |
|
25 | # import stuff from node for others to import from revlog | |
26 | from .node import ( |
|
26 | from .node import ( | |
27 | bin, |
|
27 | bin, | |
28 | hex, |
|
28 | hex, | |
29 | nullrev, |
|
29 | nullrev, | |
30 | sha1nodeconstants, |
|
30 | sha1nodeconstants, | |
31 | short, |
|
31 | short, | |
32 | wdirrev, |
|
32 | wdirrev, | |
33 | ) |
|
33 | ) | |
34 | from .i18n import _ |
|
34 | from .i18n import _ | |
35 | from .pycompat import getattr |
|
35 | from .pycompat import getattr | |
36 | from .revlogutils.constants import ( |
|
36 | from .revlogutils.constants import ( | |
37 | ALL_KINDS, |
|
37 | ALL_KINDS, | |
38 | FLAG_GENERALDELTA, |
|
38 | FLAG_GENERALDELTA, | |
39 | FLAG_INLINE_DATA, |
|
39 | FLAG_INLINE_DATA, | |
40 | INDEX_HEADER, |
|
40 | INDEX_HEADER, | |
41 | REVLOGV0, |
|
41 | REVLOGV0, | |
42 | REVLOGV1, |
|
42 | REVLOGV1, | |
43 | REVLOGV1_FLAGS, |
|
43 | REVLOGV1_FLAGS, | |
44 | REVLOGV2, |
|
44 | REVLOGV2, | |
45 | REVLOGV2_FLAGS, |
|
45 | REVLOGV2_FLAGS, | |
46 | REVLOG_DEFAULT_FLAGS, |
|
46 | REVLOG_DEFAULT_FLAGS, | |
47 | REVLOG_DEFAULT_FORMAT, |
|
47 | REVLOG_DEFAULT_FORMAT, | |
48 | REVLOG_DEFAULT_VERSION, |
|
48 | REVLOG_DEFAULT_VERSION, | |
49 | ) |
|
49 | ) | |
50 | from .revlogutils.flagutil import ( |
|
50 | from .revlogutils.flagutil import ( | |
51 | REVIDX_DEFAULT_FLAGS, |
|
51 | REVIDX_DEFAULT_FLAGS, | |
52 | REVIDX_ELLIPSIS, |
|
52 | REVIDX_ELLIPSIS, | |
53 | REVIDX_EXTSTORED, |
|
53 | REVIDX_EXTSTORED, | |
54 | REVIDX_FLAGS_ORDER, |
|
54 | REVIDX_FLAGS_ORDER, | |
55 | REVIDX_HASCOPIESINFO, |
|
55 | REVIDX_HASCOPIESINFO, | |
56 | REVIDX_ISCENSORED, |
|
56 | REVIDX_ISCENSORED, | |
57 | REVIDX_RAWTEXT_CHANGING_FLAGS, |
|
57 | REVIDX_RAWTEXT_CHANGING_FLAGS, | |
58 | ) |
|
58 | ) | |
59 | from .thirdparty import attr |
|
59 | from .thirdparty import attr | |
60 | from . import ( |
|
60 | from . import ( | |
61 | ancestor, |
|
61 | ancestor, | |
62 | dagop, |
|
62 | dagop, | |
63 | error, |
|
63 | error, | |
64 | mdiff, |
|
64 | mdiff, | |
65 | policy, |
|
65 | policy, | |
66 | pycompat, |
|
66 | pycompat, | |
67 | templatefilters, |
|
67 | templatefilters, | |
68 | util, |
|
68 | util, | |
69 | ) |
|
69 | ) | |
70 | from .interfaces import ( |
|
70 | from .interfaces import ( | |
71 | repository, |
|
71 | repository, | |
72 | util as interfaceutil, |
|
72 | util as interfaceutil, | |
73 | ) |
|
73 | ) | |
74 | from .revlogutils import ( |
|
74 | from .revlogutils import ( | |
75 | deltas as deltautil, |
|
75 | deltas as deltautil, | |
76 | flagutil, |
|
76 | flagutil, | |
77 | nodemap as nodemaputil, |
|
77 | nodemap as nodemaputil, | |
78 | revlogv0, |
|
78 | revlogv0, | |
79 | sidedata as sidedatautil, |
|
79 | sidedata as sidedatautil, | |
80 | ) |
|
80 | ) | |
81 | from .utils import ( |
|
81 | from .utils import ( | |
82 | storageutil, |
|
82 | storageutil, | |
83 | stringutil, |
|
83 | stringutil, | |
84 | ) |
|
84 | ) | |
85 |
|
85 | |||
86 | # blanked usage of all the name to prevent pyflakes constraints |
|
86 | # blanked usage of all the name to prevent pyflakes constraints | |
87 | # We need these name available in the module for extensions. |
|
87 | # We need these name available in the module for extensions. | |
88 |
|
88 | |||
89 | REVLOGV0 |
|
89 | REVLOGV0 | |
90 | REVLOGV1 |
|
90 | REVLOGV1 | |
91 | REVLOGV2 |
|
91 | REVLOGV2 | |
92 | FLAG_INLINE_DATA |
|
92 | FLAG_INLINE_DATA | |
93 | FLAG_GENERALDELTA |
|
93 | FLAG_GENERALDELTA | |
94 | REVLOG_DEFAULT_FLAGS |
|
94 | REVLOG_DEFAULT_FLAGS | |
95 | REVLOG_DEFAULT_FORMAT |
|
95 | REVLOG_DEFAULT_FORMAT | |
96 | REVLOG_DEFAULT_VERSION |
|
96 | REVLOG_DEFAULT_VERSION | |
97 | REVLOGV1_FLAGS |
|
97 | REVLOGV1_FLAGS | |
98 | REVLOGV2_FLAGS |
|
98 | REVLOGV2_FLAGS | |
99 | REVIDX_ISCENSORED |
|
99 | REVIDX_ISCENSORED | |
100 | REVIDX_ELLIPSIS |
|
100 | REVIDX_ELLIPSIS | |
101 | REVIDX_HASCOPIESINFO |
|
101 | REVIDX_HASCOPIESINFO | |
102 | REVIDX_EXTSTORED |
|
102 | REVIDX_EXTSTORED | |
103 | REVIDX_DEFAULT_FLAGS |
|
103 | REVIDX_DEFAULT_FLAGS | |
104 | REVIDX_FLAGS_ORDER |
|
104 | REVIDX_FLAGS_ORDER | |
105 | REVIDX_RAWTEXT_CHANGING_FLAGS |
|
105 | REVIDX_RAWTEXT_CHANGING_FLAGS | |
106 |
|
106 | |||
107 | parsers = policy.importmod('parsers') |
|
107 | parsers = policy.importmod('parsers') | |
108 | rustancestor = policy.importrust('ancestor') |
|
108 | rustancestor = policy.importrust('ancestor') | |
109 | rustdagop = policy.importrust('dagop') |
|
109 | rustdagop = policy.importrust('dagop') | |
110 | rustrevlog = policy.importrust('revlog') |
|
110 | rustrevlog = policy.importrust('revlog') | |
111 |
|
111 | |||
112 | # Aliased for performance. |
|
112 | # Aliased for performance. | |
113 | _zlibdecompress = zlib.decompress |
|
113 | _zlibdecompress = zlib.decompress | |
114 |
|
114 | |||
115 | # max size of revlog with inline data |
|
115 | # max size of revlog with inline data | |
116 | _maxinline = 131072 |
|
116 | _maxinline = 131072 | |
117 | _chunksize = 1048576 |
|
117 | _chunksize = 1048576 | |
118 |
|
118 | |||
119 | # Flag processors for REVIDX_ELLIPSIS. |
|
119 | # Flag processors for REVIDX_ELLIPSIS. | |
120 | def ellipsisreadprocessor(rl, text): |
|
120 | def ellipsisreadprocessor(rl, text): | |
121 | return text, False |
|
121 | return text, False | |
122 |
|
122 | |||
123 |
|
123 | |||
124 | def ellipsiswriteprocessor(rl, text): |
|
124 | def ellipsiswriteprocessor(rl, text): | |
125 | return text, False |
|
125 | return text, False | |
126 |
|
126 | |||
127 |
|
127 | |||
128 | def ellipsisrawprocessor(rl, text): |
|
128 | def ellipsisrawprocessor(rl, text): | |
129 | return False |
|
129 | return False | |
130 |
|
130 | |||
131 |
|
131 | |||
132 | ellipsisprocessor = ( |
|
132 | ellipsisprocessor = ( | |
133 | ellipsisreadprocessor, |
|
133 | ellipsisreadprocessor, | |
134 | ellipsiswriteprocessor, |
|
134 | ellipsiswriteprocessor, | |
135 | ellipsisrawprocessor, |
|
135 | ellipsisrawprocessor, | |
136 | ) |
|
136 | ) | |
137 |
|
137 | |||
138 |
|
138 | |||
139 | def offset_type(offset, type): |
|
139 | def offset_type(offset, type): | |
140 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: |
|
140 | if (type & ~flagutil.REVIDX_KNOWN_FLAGS) != 0: | |
141 | raise ValueError(b'unknown revlog index flags') |
|
141 | raise ValueError(b'unknown revlog index flags') | |
142 | return int(int(offset) << 16 | type) |
|
142 | return int(int(offset) << 16 | type) | |
143 |
|
143 | |||
144 |
|
144 | |||
145 | def _verify_revision(rl, skipflags, state, node): |
|
145 | def _verify_revision(rl, skipflags, state, node): | |
146 | """Verify the integrity of the given revlog ``node`` while providing a hook |
|
146 | """Verify the integrity of the given revlog ``node`` while providing a hook | |
147 | point for extensions to influence the operation.""" |
|
147 | point for extensions to influence the operation.""" | |
148 | if skipflags: |
|
148 | if skipflags: | |
149 | state[b'skipread'].add(node) |
|
149 | state[b'skipread'].add(node) | |
150 | else: |
|
150 | else: | |
151 | # Side-effect: read content and verify hash. |
|
151 | # Side-effect: read content and verify hash. | |
152 | rl.revision(node) |
|
152 | rl.revision(node) | |
153 |
|
153 | |||
154 |
|
154 | |||
155 | # True if a fast implementation for persistent-nodemap is available |
|
155 | # True if a fast implementation for persistent-nodemap is available | |
156 | # |
|
156 | # | |
157 | # We also consider we have a "fast" implementation in "pure" python because |
|
157 | # We also consider we have a "fast" implementation in "pure" python because | |
158 | # people using pure don't really have performance consideration (and a |
|
158 | # people using pure don't really have performance consideration (and a | |
159 | # wheelbarrow of other slowness source) |
|
159 | # wheelbarrow of other slowness source) | |
160 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( |
|
160 | HAS_FAST_PERSISTENT_NODEMAP = rustrevlog is not None or util.safehasattr( | |
161 | parsers, 'BaseIndexObject' |
|
161 | parsers, 'BaseIndexObject' | |
162 | ) |
|
162 | ) | |
163 |
|
163 | |||
164 |
|
164 | |||
165 | @attr.s(slots=True, frozen=True) |
|
165 | @attr.s(slots=True, frozen=True) | |
166 | class _revisioninfo(object): |
|
166 | class _revisioninfo(object): | |
167 | """Information about a revision that allows building its fulltext |
|
167 | """Information about a revision that allows building its fulltext | |
168 | node: expected hash of the revision |
|
168 | node: expected hash of the revision | |
169 | p1, p2: parent revs of the revision |
|
169 | p1, p2: parent revs of the revision | |
170 | btext: built text cache consisting of a one-element list |
|
170 | btext: built text cache consisting of a one-element list | |
171 | cachedelta: (baserev, uncompressed_delta) or None |
|
171 | cachedelta: (baserev, uncompressed_delta) or None | |
172 | flags: flags associated to the revision storage |
|
172 | flags: flags associated to the revision storage | |
173 |
|
173 | |||
174 | One of btext[0] or cachedelta must be set. |
|
174 | One of btext[0] or cachedelta must be set. | |
175 | """ |
|
175 | """ | |
176 |
|
176 | |||
177 | node = attr.ib() |
|
177 | node = attr.ib() | |
178 | p1 = attr.ib() |
|
178 | p1 = attr.ib() | |
179 | p2 = attr.ib() |
|
179 | p2 = attr.ib() | |
180 | btext = attr.ib() |
|
180 | btext = attr.ib() | |
181 | textlen = attr.ib() |
|
181 | textlen = attr.ib() | |
182 | cachedelta = attr.ib() |
|
182 | cachedelta = attr.ib() | |
183 | flags = attr.ib() |
|
183 | flags = attr.ib() | |
184 |
|
184 | |||
185 |
|
185 | |||
186 | @interfaceutil.implementer(repository.irevisiondelta) |
|
186 | @interfaceutil.implementer(repository.irevisiondelta) | |
187 | @attr.s(slots=True) |
|
187 | @attr.s(slots=True) | |
188 | class revlogrevisiondelta(object): |
|
188 | class revlogrevisiondelta(object): | |
189 | node = attr.ib() |
|
189 | node = attr.ib() | |
190 | p1node = attr.ib() |
|
190 | p1node = attr.ib() | |
191 | p2node = attr.ib() |
|
191 | p2node = attr.ib() | |
192 | basenode = attr.ib() |
|
192 | basenode = attr.ib() | |
193 | flags = attr.ib() |
|
193 | flags = attr.ib() | |
194 | baserevisionsize = attr.ib() |
|
194 | baserevisionsize = attr.ib() | |
195 | revision = attr.ib() |
|
195 | revision = attr.ib() | |
196 | delta = attr.ib() |
|
196 | delta = attr.ib() | |
197 | sidedata = attr.ib() |
|
197 | sidedata = attr.ib() | |
198 | protocol_flags = attr.ib() |
|
198 | protocol_flags = attr.ib() | |
199 | linknode = attr.ib(default=None) |
|
199 | linknode = attr.ib(default=None) | |
200 |
|
200 | |||
201 |
|
201 | |||
202 | @interfaceutil.implementer(repository.iverifyproblem) |
|
202 | @interfaceutil.implementer(repository.iverifyproblem) | |
203 | @attr.s(frozen=True) |
|
203 | @attr.s(frozen=True) | |
204 | class revlogproblem(object): |
|
204 | class revlogproblem(object): | |
205 | warning = attr.ib(default=None) |
|
205 | warning = attr.ib(default=None) | |
206 | error = attr.ib(default=None) |
|
206 | error = attr.ib(default=None) | |
207 | node = attr.ib(default=None) |
|
207 | node = attr.ib(default=None) | |
208 |
|
208 | |||
209 |
|
209 | |||
210 | def parse_index_v1(data, inline): |
|
210 | def parse_index_v1(data, inline): | |
211 | # call the C implementation to parse the index data |
|
211 | # call the C implementation to parse the index data | |
212 | index, cache = parsers.parse_index2(data, inline) |
|
212 | index, cache = parsers.parse_index2(data, inline) | |
213 | return index, cache |
|
213 | return index, cache | |
214 |
|
214 | |||
215 |
|
215 | |||
216 | def parse_index_v2(data, inline): |
|
216 | def parse_index_v2(data, inline): | |
217 | # call the C implementation to parse the index data |
|
217 | # call the C implementation to parse the index data | |
218 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) |
|
218 | index, cache = parsers.parse_index2(data, inline, revlogv2=True) | |
219 | return index, cache |
|
219 | return index, cache | |
220 |
|
220 | |||
221 |
|
221 | |||
222 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): |
|
222 | if util.safehasattr(parsers, 'parse_index_devel_nodemap'): | |
223 |
|
223 | |||
224 | def parse_index_v1_nodemap(data, inline): |
|
224 | def parse_index_v1_nodemap(data, inline): | |
225 | index, cache = parsers.parse_index_devel_nodemap(data, inline) |
|
225 | index, cache = parsers.parse_index_devel_nodemap(data, inline) | |
226 | return index, cache |
|
226 | return index, cache | |
227 |
|
227 | |||
228 |
|
228 | |||
229 | else: |
|
229 | else: | |
230 | parse_index_v1_nodemap = None |
|
230 | parse_index_v1_nodemap = None | |
231 |
|
231 | |||
232 |
|
232 | |||
233 | def parse_index_v1_mixed(data, inline): |
|
233 | def parse_index_v1_mixed(data, inline): | |
234 | index, cache = parse_index_v1(data, inline) |
|
234 | index, cache = parse_index_v1(data, inline) | |
235 | return rustrevlog.MixedIndex(index), cache |
|
235 | return rustrevlog.MixedIndex(index), cache | |
236 |
|
236 | |||
237 |
|
237 | |||
238 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte |
|
238 | # corresponds to uncompressed length of indexformatng (2 gigs, 4-byte | |
239 | # signed integer) |
|
239 | # signed integer) | |
240 | _maxentrysize = 0x7FFFFFFF |
|
240 | _maxentrysize = 0x7FFFFFFF | |
241 |
|
241 | |||
242 |
|
242 | |||
243 | class revlog(object): |
|
243 | class revlog(object): | |
244 | """ |
|
244 | """ | |
245 | the underlying revision storage object |
|
245 | the underlying revision storage object | |
246 |
|
246 | |||
247 | A revlog consists of two parts, an index and the revision data. |
|
247 | A revlog consists of two parts, an index and the revision data. | |
248 |
|
248 | |||
249 | The index is a file with a fixed record size containing |
|
249 | The index is a file with a fixed record size containing | |
250 | information on each revision, including its nodeid (hash), the |
|
250 | information on each revision, including its nodeid (hash), the | |
251 | nodeids of its parents, the position and offset of its data within |
|
251 | nodeids of its parents, the position and offset of its data within | |
252 | the data file, and the revision it's based on. Finally, each entry |
|
252 | the data file, and the revision it's based on. Finally, each entry | |
253 | contains a linkrev entry that can serve as a pointer to external |
|
253 | contains a linkrev entry that can serve as a pointer to external | |
254 | data. |
|
254 | data. | |
255 |
|
255 | |||
256 | The revision data itself is a linear collection of data chunks. |
|
256 | The revision data itself is a linear collection of data chunks. | |
257 | Each chunk represents a revision and is usually represented as a |
|
257 | Each chunk represents a revision and is usually represented as a | |
258 | delta against the previous chunk. To bound lookup time, runs of |
|
258 | delta against the previous chunk. To bound lookup time, runs of | |
259 | deltas are limited to about 2 times the length of the original |
|
259 | deltas are limited to about 2 times the length of the original | |
260 | version data. This makes retrieval of a version proportional to |
|
260 | version data. This makes retrieval of a version proportional to | |
261 | its size, or O(1) relative to the number of revisions. |
|
261 | its size, or O(1) relative to the number of revisions. | |
262 |
|
262 | |||
263 | Both pieces of the revlog are written to in an append-only |
|
263 | Both pieces of the revlog are written to in an append-only | |
264 | fashion, which means we never need to rewrite a file to insert or |
|
264 | fashion, which means we never need to rewrite a file to insert or | |
265 | remove data, and can use some simple techniques to avoid the need |
|
265 | remove data, and can use some simple techniques to avoid the need | |
266 | for locking while reading. |
|
266 | for locking while reading. | |
267 |
|
267 | |||
268 | If checkambig, indexfile is opened with checkambig=True at |
|
268 | If checkambig, indexfile is opened with checkambig=True at | |
269 | writing, to avoid file stat ambiguity. |
|
269 | writing, to avoid file stat ambiguity. | |
270 |
|
270 | |||
271 | If mmaplargeindex is True, and an mmapindexthreshold is set, the |
|
271 | If mmaplargeindex is True, and an mmapindexthreshold is set, the | |
272 | index will be mmapped rather than read if it is larger than the |
|
272 | index will be mmapped rather than read if it is larger than the | |
273 | configured threshold. |
|
273 | configured threshold. | |
274 |
|
274 | |||
275 | If censorable is True, the revlog can have censored revisions. |
|
275 | If censorable is True, the revlog can have censored revisions. | |
276 |
|
276 | |||
277 | If `upperboundcomp` is not None, this is the expected maximal gain from |
|
277 | If `upperboundcomp` is not None, this is the expected maximal gain from | |
278 | compression for the data content. |
|
278 | compression for the data content. | |
279 |
|
279 | |||
280 | `concurrencychecker` is an optional function that receives 3 arguments: a |
|
280 | `concurrencychecker` is an optional function that receives 3 arguments: a | |
281 | file handle, a filename, and an expected position. It should check whether |
|
281 | file handle, a filename, and an expected position. It should check whether | |
282 | the current position in the file handle is valid, and log/warn/fail (by |
|
282 | the current position in the file handle is valid, and log/warn/fail (by | |
283 | raising). |
|
283 | raising). | |
284 | """ |
|
284 | """ | |
285 |
|
285 | |||
286 | _flagserrorclass = error.RevlogError |
|
286 | _flagserrorclass = error.RevlogError | |
287 |
|
287 | |||
288 | def __init__( |
|
288 | def __init__( | |
289 | self, |
|
289 | self, | |
290 | opener, |
|
290 | opener, | |
291 | target, |
|
291 | target, | |
292 | indexfile=None, |
|
292 | indexfile=None, | |
293 | datafile=None, |
|
293 | datafile=None, | |
294 | checkambig=False, |
|
294 | checkambig=False, | |
295 | mmaplargeindex=False, |
|
295 | mmaplargeindex=False, | |
296 | censorable=False, |
|
296 | censorable=False, | |
297 | upperboundcomp=None, |
|
297 | upperboundcomp=None, | |
298 | persistentnodemap=False, |
|
298 | persistentnodemap=False, | |
299 | concurrencychecker=None, |
|
299 | concurrencychecker=None, | |
300 | ): |
|
300 | ): | |
301 | """ |
|
301 | """ | |
302 | create a revlog object |
|
302 | create a revlog object | |
303 |
|
303 | |||
304 | opener is a function that abstracts the file opening operation |
|
304 | opener is a function that abstracts the file opening operation | |
305 | and can be used to implement COW semantics or the like. |
|
305 | and can be used to implement COW semantics or the like. | |
306 |
|
306 | |||
307 | `target`: a (KIND, ID) tuple that identify the content stored in |
|
307 | `target`: a (KIND, ID) tuple that identify the content stored in | |
308 | this revlog. It help the rest of the code to understand what the revlog |
|
308 | this revlog. It help the rest of the code to understand what the revlog | |
309 | is about without having to resort to heuristic and index filename |
|
309 | is about without having to resort to heuristic and index filename | |
310 | analysis. Note: that this must be reliably be set by normal code, but |
|
310 | analysis. Note: that this must be reliably be set by normal code, but | |
311 | that test, debug, or performance measurement code might not set this to |
|
311 | that test, debug, or performance measurement code might not set this to | |
312 | accurate value. |
|
312 | accurate value. | |
313 | """ |
|
313 | """ | |
314 | self.upperboundcomp = upperboundcomp |
|
314 | self.upperboundcomp = upperboundcomp | |
315 | self.indexfile = indexfile |
|
315 | self.indexfile = indexfile | |
316 | self.datafile = datafile or (indexfile[:-2] + b".d") |
|
316 | self.datafile = datafile or (indexfile[:-2] + b".d") | |
317 | self.nodemap_file = None |
|
317 | self.nodemap_file = None | |
318 | if persistentnodemap: |
|
318 | if persistentnodemap: | |
319 | self.nodemap_file = nodemaputil.get_nodemap_file( |
|
319 | self.nodemap_file = nodemaputil.get_nodemap_file( | |
320 | opener, self.indexfile |
|
320 | opener, self.indexfile | |
321 | ) |
|
321 | ) | |
322 |
|
322 | |||
323 | self.opener = opener |
|
323 | self.opener = opener | |
324 | assert target[0] in ALL_KINDS |
|
324 | assert target[0] in ALL_KINDS | |
325 | assert len(target) == 2 |
|
325 | assert len(target) == 2 | |
326 | self.target = target |
|
326 | self.target = target | |
327 | # When True, indexfile is opened with checkambig=True at writing, to |
|
327 | # When True, indexfile is opened with checkambig=True at writing, to | |
328 | # avoid file stat ambiguity. |
|
328 | # avoid file stat ambiguity. | |
329 | self._checkambig = checkambig |
|
329 | self._checkambig = checkambig | |
330 | self._mmaplargeindex = mmaplargeindex |
|
330 | self._mmaplargeindex = mmaplargeindex | |
331 | self._censorable = censorable |
|
331 | self._censorable = censorable | |
332 | # 3-tuple of (node, rev, text) for a raw revision. |
|
332 | # 3-tuple of (node, rev, text) for a raw revision. | |
333 | self._revisioncache = None |
|
333 | self._revisioncache = None | |
334 | # Maps rev to chain base rev. |
|
334 | # Maps rev to chain base rev. | |
335 | self._chainbasecache = util.lrucachedict(100) |
|
335 | self._chainbasecache = util.lrucachedict(100) | |
336 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. |
|
336 | # 2-tuple of (offset, data) of raw data from the revlog at an offset. | |
337 | self._chunkcache = (0, b'') |
|
337 | self._chunkcache = (0, b'') | |
338 | # How much data to read and cache into the raw revlog data cache. |
|
338 | # How much data to read and cache into the raw revlog data cache. | |
339 | self._chunkcachesize = 65536 |
|
339 | self._chunkcachesize = 65536 | |
340 | self._maxchainlen = None |
|
340 | self._maxchainlen = None | |
341 | self._deltabothparents = True |
|
341 | self._deltabothparents = True | |
342 | self.index = None |
|
342 | self.index = None | |
343 | self._nodemap_docket = None |
|
343 | self._nodemap_docket = None | |
344 | # Mapping of partial identifiers to full nodes. |
|
344 | # Mapping of partial identifiers to full nodes. | |
345 | self._pcache = {} |
|
345 | self._pcache = {} | |
346 | # Mapping of revision integer to full node. |
|
346 | # Mapping of revision integer to full node. | |
347 | self._compengine = b'zlib' |
|
347 | self._compengine = b'zlib' | |
348 | self._compengineopts = {} |
|
348 | self._compengineopts = {} | |
349 | self._maxdeltachainspan = -1 |
|
349 | self._maxdeltachainspan = -1 | |
350 | self._withsparseread = False |
|
350 | self._withsparseread = False | |
351 | self._sparserevlog = False |
|
351 | self._sparserevlog = False | |
352 | self._srdensitythreshold = 0.50 |
|
352 | self._srdensitythreshold = 0.50 | |
353 | self._srmingapsize = 262144 |
|
353 | self._srmingapsize = 262144 | |
354 |
|
354 | |||
355 | # Make copy of flag processors so each revlog instance can support |
|
355 | # Make copy of flag processors so each revlog instance can support | |
356 | # custom flags. |
|
356 | # custom flags. | |
357 | self._flagprocessors = dict(flagutil.flagprocessors) |
|
357 | self._flagprocessors = dict(flagutil.flagprocessors) | |
358 |
|
358 | |||
359 | # 2-tuple of file handles being used for active writing. |
|
359 | # 2-tuple of file handles being used for active writing. | |
360 | self._writinghandles = None |
|
360 | self._writinghandles = None | |
361 |
|
361 | |||
362 | self._loadindex() |
|
362 | self._loadindex() | |
363 |
|
363 | |||
364 | self._concurrencychecker = concurrencychecker |
|
364 | self._concurrencychecker = concurrencychecker | |
365 |
|
365 | |||
366 | def _loadindex(self): |
|
366 | def _loadindex(self): | |
367 | mmapindexthreshold = None |
|
367 | mmapindexthreshold = None | |
368 | opts = self.opener.options |
|
368 | opts = self.opener.options | |
369 |
|
369 | |||
370 | if b'revlogv2' in opts: |
|
370 | if b'revlogv2' in opts: | |
371 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA |
|
371 | newversionflags = REVLOGV2 | FLAG_INLINE_DATA | |
372 | elif b'revlogv1' in opts: |
|
372 | elif b'revlogv1' in opts: | |
373 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA |
|
373 | newversionflags = REVLOGV1 | FLAG_INLINE_DATA | |
374 | if b'generaldelta' in opts: |
|
374 | if b'generaldelta' in opts: | |
375 | newversionflags |= FLAG_GENERALDELTA |
|
375 | newversionflags |= FLAG_GENERALDELTA | |
376 | elif b'revlogv0' in self.opener.options: |
|
376 | elif b'revlogv0' in self.opener.options: | |
377 | newversionflags = REVLOGV0 |
|
377 | newversionflags = REVLOGV0 | |
378 | else: |
|
378 | else: | |
379 | newversionflags = REVLOG_DEFAULT_VERSION |
|
379 | newversionflags = REVLOG_DEFAULT_VERSION | |
380 |
|
380 | |||
381 | if b'chunkcachesize' in opts: |
|
381 | if b'chunkcachesize' in opts: | |
382 | self._chunkcachesize = opts[b'chunkcachesize'] |
|
382 | self._chunkcachesize = opts[b'chunkcachesize'] | |
383 | if b'maxchainlen' in opts: |
|
383 | if b'maxchainlen' in opts: | |
384 | self._maxchainlen = opts[b'maxchainlen'] |
|
384 | self._maxchainlen = opts[b'maxchainlen'] | |
385 | if b'deltabothparents' in opts: |
|
385 | if b'deltabothparents' in opts: | |
386 | self._deltabothparents = opts[b'deltabothparents'] |
|
386 | self._deltabothparents = opts[b'deltabothparents'] | |
387 | self._lazydelta = bool(opts.get(b'lazydelta', True)) |
|
387 | self._lazydelta = bool(opts.get(b'lazydelta', True)) | |
388 | self._lazydeltabase = False |
|
388 | self._lazydeltabase = False | |
389 | if self._lazydelta: |
|
389 | if self._lazydelta: | |
390 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) |
|
390 | self._lazydeltabase = bool(opts.get(b'lazydeltabase', False)) | |
391 | if b'compengine' in opts: |
|
391 | if b'compengine' in opts: | |
392 | self._compengine = opts[b'compengine'] |
|
392 | self._compengine = opts[b'compengine'] | |
393 | if b'zlib.level' in opts: |
|
393 | if b'zlib.level' in opts: | |
394 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] |
|
394 | self._compengineopts[b'zlib.level'] = opts[b'zlib.level'] | |
395 | if b'zstd.level' in opts: |
|
395 | if b'zstd.level' in opts: | |
396 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] |
|
396 | self._compengineopts[b'zstd.level'] = opts[b'zstd.level'] | |
397 | if b'maxdeltachainspan' in opts: |
|
397 | if b'maxdeltachainspan' in opts: | |
398 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] |
|
398 | self._maxdeltachainspan = opts[b'maxdeltachainspan'] | |
399 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: |
|
399 | if self._mmaplargeindex and b'mmapindexthreshold' in opts: | |
400 | mmapindexthreshold = opts[b'mmapindexthreshold'] |
|
400 | mmapindexthreshold = opts[b'mmapindexthreshold'] | |
401 | self.hassidedata = bool(opts.get(b'side-data', False)) |
|
401 | self.hassidedata = bool(opts.get(b'side-data', False)) | |
402 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) |
|
402 | self._sparserevlog = bool(opts.get(b'sparse-revlog', False)) | |
403 | withsparseread = bool(opts.get(b'with-sparse-read', False)) |
|
403 | withsparseread = bool(opts.get(b'with-sparse-read', False)) | |
404 | # sparse-revlog forces sparse-read |
|
404 | # sparse-revlog forces sparse-read | |
405 | self._withsparseread = self._sparserevlog or withsparseread |
|
405 | self._withsparseread = self._sparserevlog or withsparseread | |
406 | if b'sparse-read-density-threshold' in opts: |
|
406 | if b'sparse-read-density-threshold' in opts: | |
407 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] |
|
407 | self._srdensitythreshold = opts[b'sparse-read-density-threshold'] | |
408 | if b'sparse-read-min-gap-size' in opts: |
|
408 | if b'sparse-read-min-gap-size' in opts: | |
409 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] |
|
409 | self._srmingapsize = opts[b'sparse-read-min-gap-size'] | |
410 | if opts.get(b'enableellipsis'): |
|
410 | if opts.get(b'enableellipsis'): | |
411 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor |
|
411 | self._flagprocessors[REVIDX_ELLIPSIS] = ellipsisprocessor | |
412 |
|
412 | |||
413 | # revlog v0 doesn't have flag processors |
|
413 | # revlog v0 doesn't have flag processors | |
414 | for flag, processor in pycompat.iteritems( |
|
414 | for flag, processor in pycompat.iteritems( | |
415 | opts.get(b'flagprocessors', {}) |
|
415 | opts.get(b'flagprocessors', {}) | |
416 | ): |
|
416 | ): | |
417 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) |
|
417 | flagutil.insertflagprocessor(flag, processor, self._flagprocessors) | |
418 |
|
418 | |||
419 | if self._chunkcachesize <= 0: |
|
419 | if self._chunkcachesize <= 0: | |
420 | raise error.RevlogError( |
|
420 | raise error.RevlogError( | |
421 | _(b'revlog chunk cache size %r is not greater than 0') |
|
421 | _(b'revlog chunk cache size %r is not greater than 0') | |
422 | % self._chunkcachesize |
|
422 | % self._chunkcachesize | |
423 | ) |
|
423 | ) | |
424 | elif self._chunkcachesize & (self._chunkcachesize - 1): |
|
424 | elif self._chunkcachesize & (self._chunkcachesize - 1): | |
425 | raise error.RevlogError( |
|
425 | raise error.RevlogError( | |
426 | _(b'revlog chunk cache size %r is not a power of 2') |
|
426 | _(b'revlog chunk cache size %r is not a power of 2') | |
427 | % self._chunkcachesize |
|
427 | % self._chunkcachesize | |
428 | ) |
|
428 | ) | |
429 |
|
429 | |||
430 | indexdata = b'' |
|
430 | indexdata = b'' | |
431 | self._initempty = True |
|
431 | self._initempty = True | |
432 | try: |
|
432 | try: | |
433 | with self._indexfp() as f: |
|
433 | with self._indexfp() as f: | |
434 | if ( |
|
434 | if ( | |
435 | mmapindexthreshold is not None |
|
435 | mmapindexthreshold is not None | |
436 | and self.opener.fstat(f).st_size >= mmapindexthreshold |
|
436 | and self.opener.fstat(f).st_size >= mmapindexthreshold | |
437 | ): |
|
437 | ): | |
438 | # TODO: should .close() to release resources without |
|
438 | # TODO: should .close() to release resources without | |
439 | # relying on Python GC |
|
439 | # relying on Python GC | |
440 | indexdata = util.buffer(util.mmapread(f)) |
|
440 | indexdata = util.buffer(util.mmapread(f)) | |
441 | else: |
|
441 | else: | |
442 | indexdata = f.read() |
|
442 | indexdata = f.read() | |
443 | if len(indexdata) > 0: |
|
443 | if len(indexdata) > 0: | |
444 | versionflags = INDEX_HEADER.unpack(indexdata[:4])[0] |
|
444 | versionflags = INDEX_HEADER.unpack(indexdata[:4])[0] | |
445 | self._initempty = False |
|
445 | self._initempty = False | |
446 | else: |
|
446 | else: | |
447 | versionflags = newversionflags |
|
447 | versionflags = newversionflags | |
448 | except IOError as inst: |
|
448 | except IOError as inst: | |
449 | if inst.errno != errno.ENOENT: |
|
449 | if inst.errno != errno.ENOENT: | |
450 | raise |
|
450 | raise | |
451 |
|
451 | |||
452 | versionflags = newversionflags |
|
452 | versionflags = newversionflags | |
453 |
|
453 | |||
454 | self.version = versionflags |
|
454 | self.version = versionflags | |
455 |
|
455 | |||
456 | flags = versionflags & ~0xFFFF |
|
456 | flags = versionflags & ~0xFFFF | |
457 | fmt = versionflags & 0xFFFF |
|
457 | fmt = versionflags & 0xFFFF | |
458 |
|
458 | |||
459 | if fmt == REVLOGV0: |
|
459 | if fmt == REVLOGV0: | |
460 | if flags: |
|
460 | if flags: | |
461 | raise error.RevlogError( |
|
461 | raise error.RevlogError( | |
462 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
462 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
463 | % (flags >> 16, fmt, self.indexfile) |
|
463 | % (flags >> 16, fmt, self.indexfile) | |
464 | ) |
|
464 | ) | |
465 |
|
465 | |||
466 | self._inline = False |
|
466 | self._inline = False | |
467 | self._generaldelta = False |
|
467 | self._generaldelta = False | |
468 |
|
468 | |||
469 | elif fmt == REVLOGV1: |
|
469 | elif fmt == REVLOGV1: | |
470 | if flags & ~REVLOGV1_FLAGS: |
|
470 | if flags & ~REVLOGV1_FLAGS: | |
471 | raise error.RevlogError( |
|
471 | raise error.RevlogError( | |
472 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
472 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
473 | % (flags >> 16, fmt, self.indexfile) |
|
473 | % (flags >> 16, fmt, self.indexfile) | |
474 | ) |
|
474 | ) | |
475 |
|
475 | |||
476 | self._inline = versionflags & FLAG_INLINE_DATA |
|
476 | self._inline = versionflags & FLAG_INLINE_DATA | |
477 | self._generaldelta = versionflags & FLAG_GENERALDELTA |
|
477 | self._generaldelta = versionflags & FLAG_GENERALDELTA | |
478 |
|
478 | |||
479 | elif fmt == REVLOGV2: |
|
479 | elif fmt == REVLOGV2: | |
480 | if flags & ~REVLOGV2_FLAGS: |
|
480 | if flags & ~REVLOGV2_FLAGS: | |
481 | raise error.RevlogError( |
|
481 | raise error.RevlogError( | |
482 | _(b'unknown flags (%#04x) in version %d revlog %s') |
|
482 | _(b'unknown flags (%#04x) in version %d revlog %s') | |
483 | % (flags >> 16, fmt, self.indexfile) |
|
483 | % (flags >> 16, fmt, self.indexfile) | |
484 | ) |
|
484 | ) | |
485 |
|
485 | |||
486 | # There is a bug in the transaction handling when going from an |
|
486 | # There is a bug in the transaction handling when going from an | |
487 | # inline revlog to a separate index and data file. Turn it off until |
|
487 | # inline revlog to a separate index and data file. Turn it off until | |
488 | # it's fixed, since v2 revlogs sometimes get rewritten on exchange. |
|
488 | # it's fixed, since v2 revlogs sometimes get rewritten on exchange. | |
489 | # See issue6485 |
|
489 | # See issue6485 | |
490 | self._inline = False |
|
490 | self._inline = False | |
491 | # generaldelta implied by version 2 revlogs. |
|
491 | # generaldelta implied by version 2 revlogs. | |
492 | self._generaldelta = True |
|
492 | self._generaldelta = True | |
493 |
|
493 | |||
494 | else: |
|
494 | else: | |
495 | raise error.RevlogError( |
|
495 | raise error.RevlogError( | |
496 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) |
|
496 | _(b'unknown version (%d) in revlog %s') % (fmt, self.indexfile) | |
497 | ) |
|
497 | ) | |
498 |
|
498 | |||
499 | self.nodeconstants = sha1nodeconstants |
|
499 | self.nodeconstants = sha1nodeconstants | |
500 | self.nullid = self.nodeconstants.nullid |
|
500 | self.nullid = self.nodeconstants.nullid | |
501 |
|
501 | |||
502 | # sparse-revlog can't be on without general-delta (issue6056) |
|
502 | # sparse-revlog can't be on without general-delta (issue6056) | |
503 | if not self._generaldelta: |
|
503 | if not self._generaldelta: | |
504 | self._sparserevlog = False |
|
504 | self._sparserevlog = False | |
505 |
|
505 | |||
506 | self._storedeltachains = True |
|
506 | self._storedeltachains = True | |
507 |
|
507 | |||
508 | devel_nodemap = ( |
|
508 | devel_nodemap = ( | |
509 | self.nodemap_file |
|
509 | self.nodemap_file | |
510 | and opts.get(b'devel-force-nodemap', False) |
|
510 | and opts.get(b'devel-force-nodemap', False) | |
511 | and parse_index_v1_nodemap is not None |
|
511 | and parse_index_v1_nodemap is not None | |
512 | ) |
|
512 | ) | |
513 |
|
513 | |||
514 | use_rust_index = False |
|
514 | use_rust_index = False | |
515 | if rustrevlog is not None: |
|
515 | if rustrevlog is not None: | |
516 | if self.nodemap_file is not None: |
|
516 | if self.nodemap_file is not None: | |
517 | use_rust_index = True |
|
517 | use_rust_index = True | |
518 | else: |
|
518 | else: | |
519 | use_rust_index = self.opener.options.get(b'rust.index') |
|
519 | use_rust_index = self.opener.options.get(b'rust.index') | |
520 |
|
520 | |||
521 | self._parse_index = parse_index_v1 |
|
521 | self._parse_index = parse_index_v1 | |
522 | if self.version == REVLOGV0: |
|
522 | if self.version == REVLOGV0: | |
523 | self._parse_index = revlogv0.parse_index_v0 |
|
523 | self._parse_index = revlogv0.parse_index_v0 | |
524 | elif fmt == REVLOGV2: |
|
524 | elif fmt == REVLOGV2: | |
525 | self._parse_index = parse_index_v2 |
|
525 | self._parse_index = parse_index_v2 | |
526 | elif devel_nodemap: |
|
526 | elif devel_nodemap: | |
527 | self._parse_index = parse_index_v1_nodemap |
|
527 | self._parse_index = parse_index_v1_nodemap | |
528 | elif use_rust_index: |
|
528 | elif use_rust_index: | |
529 | self._parse_index = parse_index_v1_mixed |
|
529 | self._parse_index = parse_index_v1_mixed | |
530 | try: |
|
530 | try: | |
531 | d = self._parse_index(indexdata, self._inline) |
|
531 | d = self._parse_index(indexdata, self._inline) | |
532 | index, _chunkcache = d |
|
532 | index, _chunkcache = d | |
533 | use_nodemap = ( |
|
533 | use_nodemap = ( | |
534 | not self._inline |
|
534 | not self._inline | |
535 | and self.nodemap_file is not None |
|
535 | and self.nodemap_file is not None | |
536 | and util.safehasattr(index, 'update_nodemap_data') |
|
536 | and util.safehasattr(index, 'update_nodemap_data') | |
537 | ) |
|
537 | ) | |
538 | if use_nodemap: |
|
538 | if use_nodemap: | |
539 | nodemap_data = nodemaputil.persisted_data(self) |
|
539 | nodemap_data = nodemaputil.persisted_data(self) | |
540 | if nodemap_data is not None: |
|
540 | if nodemap_data is not None: | |
541 | docket = nodemap_data[0] |
|
541 | docket = nodemap_data[0] | |
542 | if ( |
|
542 | if ( | |
543 | len(d[0]) > docket.tip_rev |
|
543 | len(d[0]) > docket.tip_rev | |
544 | and d[0][docket.tip_rev][7] == docket.tip_node |
|
544 | and d[0][docket.tip_rev][7] == docket.tip_node | |
545 | ): |
|
545 | ): | |
546 | # no changelog tampering |
|
546 | # no changelog tampering | |
547 | self._nodemap_docket = docket |
|
547 | self._nodemap_docket = docket | |
548 | index.update_nodemap_data(*nodemap_data) |
|
548 | index.update_nodemap_data(*nodemap_data) | |
549 | except (ValueError, IndexError): |
|
549 | except (ValueError, IndexError): | |
550 | raise error.RevlogError( |
|
550 | raise error.RevlogError( | |
551 | _(b"index %s is corrupted") % self.indexfile |
|
551 | _(b"index %s is corrupted") % self.indexfile | |
552 | ) |
|
552 | ) | |
553 | self.index, self._chunkcache = d |
|
553 | self.index, self._chunkcache = d | |
554 | if not self._chunkcache: |
|
554 | if not self._chunkcache: | |
555 | self._chunkclear() |
|
555 | self._chunkclear() | |
556 | # revnum -> (chain-length, sum-delta-length) |
|
556 | # revnum -> (chain-length, sum-delta-length) | |
557 | self._chaininfocache = util.lrucachedict(500) |
|
557 | self._chaininfocache = util.lrucachedict(500) | |
558 | # revlog header -> revlog compressor |
|
558 | # revlog header -> revlog compressor | |
559 | self._decompressors = {} |
|
559 | self._decompressors = {} | |
560 |
|
560 | |||
561 | @util.propertycache |
|
561 | @util.propertycache | |
562 | def revlog_kind(self): |
|
562 | def revlog_kind(self): | |
563 | return self.target[0] |
|
563 | return self.target[0] | |
564 |
|
564 | |||
565 | @util.propertycache |
|
565 | @util.propertycache | |
566 | def _compressor(self): |
|
566 | def _compressor(self): | |
567 | engine = util.compengines[self._compengine] |
|
567 | engine = util.compengines[self._compengine] | |
568 | return engine.revlogcompressor(self._compengineopts) |
|
568 | return engine.revlogcompressor(self._compengineopts) | |
569 |
|
569 | |||
570 | def _indexfp(self, mode=b'r'): |
|
570 | def _indexfp(self, mode=b'r'): | |
571 | """file object for the revlog's index file""" |
|
571 | """file object for the revlog's index file""" | |
572 | args = {'mode': mode} |
|
572 | args = {'mode': mode} | |
573 | if mode != b'r': |
|
573 | if mode != b'r': | |
574 | args['checkambig'] = self._checkambig |
|
574 | args['checkambig'] = self._checkambig | |
575 | if mode == b'w': |
|
575 | if mode == b'w': | |
576 | args['atomictemp'] = True |
|
576 | args['atomictemp'] = True | |
577 | return self.opener(self.indexfile, **args) |
|
577 | return self.opener(self.indexfile, **args) | |
578 |
|
578 | |||
579 | def _datafp(self, mode=b'r'): |
|
579 | def _datafp(self, mode=b'r'): | |
580 | """file object for the revlog's data file""" |
|
580 | """file object for the revlog's data file""" | |
581 | return self.opener(self.datafile, mode=mode) |
|
581 | return self.opener(self.datafile, mode=mode) | |
582 |
|
582 | |||
583 | @contextlib.contextmanager |
|
583 | @contextlib.contextmanager | |
584 | def _datareadfp(self, existingfp=None): |
|
584 | def _datareadfp(self, existingfp=None): | |
585 | """file object suitable to read data""" |
|
585 | """file object suitable to read data""" | |
586 | # Use explicit file handle, if given. |
|
586 | # Use explicit file handle, if given. | |
587 | if existingfp is not None: |
|
587 | if existingfp is not None: | |
588 | yield existingfp |
|
588 | yield existingfp | |
589 |
|
589 | |||
590 | # Use a file handle being actively used for writes, if available. |
|
590 | # Use a file handle being actively used for writes, if available. | |
591 | # There is some danger to doing this because reads will seek the |
|
591 | # There is some danger to doing this because reads will seek the | |
592 | # file. However, _writeentry() performs a SEEK_END before all writes, |
|
592 | # file. However, _writeentry() performs a SEEK_END before all writes, | |
593 | # so we should be safe. |
|
593 | # so we should be safe. | |
594 | elif self._writinghandles: |
|
594 | elif self._writinghandles: | |
595 | if self._inline: |
|
595 | if self._inline: | |
596 | yield self._writinghandles[0] |
|
596 | yield self._writinghandles[0] | |
597 | else: |
|
597 | else: | |
598 | yield self._writinghandles[1] |
|
598 | yield self._writinghandles[1] | |
599 |
|
599 | |||
600 | # Otherwise open a new file handle. |
|
600 | # Otherwise open a new file handle. | |
601 | else: |
|
601 | else: | |
602 | if self._inline: |
|
602 | if self._inline: | |
603 | func = self._indexfp |
|
603 | func = self._indexfp | |
604 | else: |
|
604 | else: | |
605 | func = self._datafp |
|
605 | func = self._datafp | |
606 | with func() as fp: |
|
606 | with func() as fp: | |
607 | yield fp |
|
607 | yield fp | |
608 |
|
608 | |||
609 | def tiprev(self): |
|
609 | def tiprev(self): | |
610 | return len(self.index) - 1 |
|
610 | return len(self.index) - 1 | |
611 |
|
611 | |||
612 | def tip(self): |
|
612 | def tip(self): | |
613 | return self.node(self.tiprev()) |
|
613 | return self.node(self.tiprev()) | |
614 |
|
614 | |||
615 | def __contains__(self, rev): |
|
615 | def __contains__(self, rev): | |
616 | return 0 <= rev < len(self) |
|
616 | return 0 <= rev < len(self) | |
617 |
|
617 | |||
618 | def __len__(self): |
|
618 | def __len__(self): | |
619 | return len(self.index) |
|
619 | return len(self.index) | |
620 |
|
620 | |||
621 | def __iter__(self): |
|
621 | def __iter__(self): | |
622 | return iter(pycompat.xrange(len(self))) |
|
622 | return iter(pycompat.xrange(len(self))) | |
623 |
|
623 | |||
624 | def revs(self, start=0, stop=None): |
|
624 | def revs(self, start=0, stop=None): | |
625 | """iterate over all rev in this revlog (from start to stop)""" |
|
625 | """iterate over all rev in this revlog (from start to stop)""" | |
626 | return storageutil.iterrevs(len(self), start=start, stop=stop) |
|
626 | return storageutil.iterrevs(len(self), start=start, stop=stop) | |
627 |
|
627 | |||
628 | @property |
|
628 | @property | |
629 | def nodemap(self): |
|
629 | def nodemap(self): | |
630 | msg = ( |
|
630 | msg = ( | |
631 | b"revlog.nodemap is deprecated, " |
|
631 | b"revlog.nodemap is deprecated, " | |
632 | b"use revlog.index.[has_node|rev|get_rev]" |
|
632 | b"use revlog.index.[has_node|rev|get_rev]" | |
633 | ) |
|
633 | ) | |
634 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
634 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
635 | return self.index.nodemap |
|
635 | return self.index.nodemap | |
636 |
|
636 | |||
637 | @property |
|
637 | @property | |
638 | def _nodecache(self): |
|
638 | def _nodecache(self): | |
639 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" |
|
639 | msg = b"revlog._nodecache is deprecated, use revlog.index.nodemap" | |
640 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) |
|
640 | util.nouideprecwarn(msg, b'5.3', stacklevel=2) | |
641 | return self.index.nodemap |
|
641 | return self.index.nodemap | |
642 |
|
642 | |||
643 | def hasnode(self, node): |
|
643 | def hasnode(self, node): | |
644 | try: |
|
644 | try: | |
645 | self.rev(node) |
|
645 | self.rev(node) | |
646 | return True |
|
646 | return True | |
647 | except KeyError: |
|
647 | except KeyError: | |
648 | return False |
|
648 | return False | |
649 |
|
649 | |||
650 | def candelta(self, baserev, rev): |
|
650 | def candelta(self, baserev, rev): | |
651 | """whether two revisions (baserev, rev) can be delta-ed or not""" |
|
651 | """whether two revisions (baserev, rev) can be delta-ed or not""" | |
652 | # Disable delta if either rev requires a content-changing flag |
|
652 | # Disable delta if either rev requires a content-changing flag | |
653 | # processor (ex. LFS). This is because such flag processor can alter |
|
653 | # processor (ex. LFS). This is because such flag processor can alter | |
654 | # the rawtext content that the delta will be based on, and two clients |
|
654 | # the rawtext content that the delta will be based on, and two clients | |
655 | # could have a same revlog node with different flags (i.e. different |
|
655 | # could have a same revlog node with different flags (i.e. different | |
656 | # rawtext contents) and the delta could be incompatible. |
|
656 | # rawtext contents) and the delta could be incompatible. | |
657 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( |
|
657 | if (self.flags(baserev) & REVIDX_RAWTEXT_CHANGING_FLAGS) or ( | |
658 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS |
|
658 | self.flags(rev) & REVIDX_RAWTEXT_CHANGING_FLAGS | |
659 | ): |
|
659 | ): | |
660 | return False |
|
660 | return False | |
661 | return True |
|
661 | return True | |
662 |
|
662 | |||
663 | def update_caches(self, transaction): |
|
663 | def update_caches(self, transaction): | |
664 | if self.nodemap_file is not None: |
|
664 | if self.nodemap_file is not None: | |
665 | if transaction is None: |
|
665 | if transaction is None: | |
666 | nodemaputil.update_persistent_nodemap(self) |
|
666 | nodemaputil.update_persistent_nodemap(self) | |
667 | else: |
|
667 | else: | |
668 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
668 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
669 |
|
669 | |||
670 | def clearcaches(self): |
|
670 | def clearcaches(self): | |
671 | self._revisioncache = None |
|
671 | self._revisioncache = None | |
672 | self._chainbasecache.clear() |
|
672 | self._chainbasecache.clear() | |
673 | self._chunkcache = (0, b'') |
|
673 | self._chunkcache = (0, b'') | |
674 | self._pcache = {} |
|
674 | self._pcache = {} | |
675 | self._nodemap_docket = None |
|
675 | self._nodemap_docket = None | |
676 | self.index.clearcaches() |
|
676 | self.index.clearcaches() | |
677 | # The python code is the one responsible for validating the docket, we |
|
677 | # The python code is the one responsible for validating the docket, we | |
678 | # end up having to refresh it here. |
|
678 | # end up having to refresh it here. | |
679 | use_nodemap = ( |
|
679 | use_nodemap = ( | |
680 | not self._inline |
|
680 | not self._inline | |
681 | and self.nodemap_file is not None |
|
681 | and self.nodemap_file is not None | |
682 | and util.safehasattr(self.index, 'update_nodemap_data') |
|
682 | and util.safehasattr(self.index, 'update_nodemap_data') | |
683 | ) |
|
683 | ) | |
684 | if use_nodemap: |
|
684 | if use_nodemap: | |
685 | nodemap_data = nodemaputil.persisted_data(self) |
|
685 | nodemap_data = nodemaputil.persisted_data(self) | |
686 | if nodemap_data is not None: |
|
686 | if nodemap_data is not None: | |
687 | self._nodemap_docket = nodemap_data[0] |
|
687 | self._nodemap_docket = nodemap_data[0] | |
688 | self.index.update_nodemap_data(*nodemap_data) |
|
688 | self.index.update_nodemap_data(*nodemap_data) | |
689 |
|
689 | |||
690 | def rev(self, node): |
|
690 | def rev(self, node): | |
691 | try: |
|
691 | try: | |
692 | return self.index.rev(node) |
|
692 | return self.index.rev(node) | |
693 | except TypeError: |
|
693 | except TypeError: | |
694 | raise |
|
694 | raise | |
695 | except error.RevlogError: |
|
695 | except error.RevlogError: | |
696 | # parsers.c radix tree lookup failed |
|
696 | # parsers.c radix tree lookup failed | |
697 | if ( |
|
697 | if ( | |
698 | node == self.nodeconstants.wdirid |
|
698 | node == self.nodeconstants.wdirid | |
699 | or node in self.nodeconstants.wdirfilenodeids |
|
699 | or node in self.nodeconstants.wdirfilenodeids | |
700 | ): |
|
700 | ): | |
701 | raise error.WdirUnsupported |
|
701 | raise error.WdirUnsupported | |
702 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
702 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
703 |
|
703 | |||
704 | # Accessors for index entries. |
|
704 | # Accessors for index entries. | |
705 |
|
705 | |||
706 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes |
|
706 | # First tuple entry is 8 bytes. First 6 bytes are offset. Last 2 bytes | |
707 | # are flags. |
|
707 | # are flags. | |
708 | def start(self, rev): |
|
708 | def start(self, rev): | |
709 | return int(self.index[rev][0] >> 16) |
|
709 | return int(self.index[rev][0] >> 16) | |
710 |
|
710 | |||
711 | def flags(self, rev): |
|
711 | def flags(self, rev): | |
712 | return self.index[rev][0] & 0xFFFF |
|
712 | return self.index[rev][0] & 0xFFFF | |
713 |
|
713 | |||
714 | def length(self, rev): |
|
714 | def length(self, rev): | |
715 | return self.index[rev][1] |
|
715 | return self.index[rev][1] | |
716 |
|
716 | |||
717 | def sidedata_length(self, rev): |
|
717 | def sidedata_length(self, rev): | |
718 | if self.version & 0xFFFF != REVLOGV2: |
|
718 | if self.version & 0xFFFF != REVLOGV2: | |
719 | return 0 |
|
719 | return 0 | |
720 | return self.index[rev][9] |
|
720 | return self.index[rev][9] | |
721 |
|
721 | |||
722 | def rawsize(self, rev): |
|
722 | def rawsize(self, rev): | |
723 | """return the length of the uncompressed text for a given revision""" |
|
723 | """return the length of the uncompressed text for a given revision""" | |
724 | l = self.index[rev][2] |
|
724 | l = self.index[rev][2] | |
725 | if l >= 0: |
|
725 | if l >= 0: | |
726 | return l |
|
726 | return l | |
727 |
|
727 | |||
728 | t = self.rawdata(rev) |
|
728 | t = self.rawdata(rev) | |
729 | return len(t) |
|
729 | return len(t) | |
730 |
|
730 | |||
731 | def size(self, rev): |
|
731 | def size(self, rev): | |
732 | """length of non-raw text (processed by a "read" flag processor)""" |
|
732 | """length of non-raw text (processed by a "read" flag processor)""" | |
733 | # fast path: if no "read" flag processor could change the content, |
|
733 | # fast path: if no "read" flag processor could change the content, | |
734 | # size is rawsize. note: ELLIPSIS is known to not change the content. |
|
734 | # size is rawsize. note: ELLIPSIS is known to not change the content. | |
735 | flags = self.flags(rev) |
|
735 | flags = self.flags(rev) | |
736 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: |
|
736 | if flags & (flagutil.REVIDX_KNOWN_FLAGS ^ REVIDX_ELLIPSIS) == 0: | |
737 | return self.rawsize(rev) |
|
737 | return self.rawsize(rev) | |
738 |
|
738 | |||
739 | return len(self.revision(rev, raw=False)) |
|
739 | return len(self.revision(rev, raw=False)) | |
740 |
|
740 | |||
741 | def chainbase(self, rev): |
|
741 | def chainbase(self, rev): | |
742 | base = self._chainbasecache.get(rev) |
|
742 | base = self._chainbasecache.get(rev) | |
743 | if base is not None: |
|
743 | if base is not None: | |
744 | return base |
|
744 | return base | |
745 |
|
745 | |||
746 | index = self.index |
|
746 | index = self.index | |
747 | iterrev = rev |
|
747 | iterrev = rev | |
748 | base = index[iterrev][3] |
|
748 | base = index[iterrev][3] | |
749 | while base != iterrev: |
|
749 | while base != iterrev: | |
750 | iterrev = base |
|
750 | iterrev = base | |
751 | base = index[iterrev][3] |
|
751 | base = index[iterrev][3] | |
752 |
|
752 | |||
753 | self._chainbasecache[rev] = base |
|
753 | self._chainbasecache[rev] = base | |
754 | return base |
|
754 | return base | |
755 |
|
755 | |||
756 | def linkrev(self, rev): |
|
756 | def linkrev(self, rev): | |
757 | return self.index[rev][4] |
|
757 | return self.index[rev][4] | |
758 |
|
758 | |||
759 | def parentrevs(self, rev): |
|
759 | def parentrevs(self, rev): | |
760 | try: |
|
760 | try: | |
761 | entry = self.index[rev] |
|
761 | entry = self.index[rev] | |
762 | except IndexError: |
|
762 | except IndexError: | |
763 | if rev == wdirrev: |
|
763 | if rev == wdirrev: | |
764 | raise error.WdirUnsupported |
|
764 | raise error.WdirUnsupported | |
765 | raise |
|
765 | raise | |
766 | if entry[5] == nullrev: |
|
766 | if entry[5] == nullrev: | |
767 | return entry[6], entry[5] |
|
767 | return entry[6], entry[5] | |
768 | else: |
|
768 | else: | |
769 | return entry[5], entry[6] |
|
769 | return entry[5], entry[6] | |
770 |
|
770 | |||
771 | # fast parentrevs(rev) where rev isn't filtered |
|
771 | # fast parentrevs(rev) where rev isn't filtered | |
772 | _uncheckedparentrevs = parentrevs |
|
772 | _uncheckedparentrevs = parentrevs | |
773 |
|
773 | |||
774 | def node(self, rev): |
|
774 | def node(self, rev): | |
775 | try: |
|
775 | try: | |
776 | return self.index[rev][7] |
|
776 | return self.index[rev][7] | |
777 | except IndexError: |
|
777 | except IndexError: | |
778 | if rev == wdirrev: |
|
778 | if rev == wdirrev: | |
779 | raise error.WdirUnsupported |
|
779 | raise error.WdirUnsupported | |
780 | raise |
|
780 | raise | |
781 |
|
781 | |||
782 | # Derived from index values. |
|
782 | # Derived from index values. | |
783 |
|
783 | |||
784 | def end(self, rev): |
|
784 | def end(self, rev): | |
785 | return self.start(rev) + self.length(rev) |
|
785 | return self.start(rev) + self.length(rev) | |
786 |
|
786 | |||
787 | def parents(self, node): |
|
787 | def parents(self, node): | |
788 | i = self.index |
|
788 | i = self.index | |
789 | d = i[self.rev(node)] |
|
789 | d = i[self.rev(node)] | |
790 | # inline node() to avoid function call overhead |
|
790 | # inline node() to avoid function call overhead | |
791 | if d[5] == self.nullid: |
|
791 | if d[5] == self.nullid: | |
792 | return i[d[6]][7], i[d[5]][7] |
|
792 | return i[d[6]][7], i[d[5]][7] | |
793 | else: |
|
793 | else: | |
794 | return i[d[5]][7], i[d[6]][7] |
|
794 | return i[d[5]][7], i[d[6]][7] | |
795 |
|
795 | |||
796 | def chainlen(self, rev): |
|
796 | def chainlen(self, rev): | |
797 | return self._chaininfo(rev)[0] |
|
797 | return self._chaininfo(rev)[0] | |
798 |
|
798 | |||
799 | def _chaininfo(self, rev): |
|
799 | def _chaininfo(self, rev): | |
800 | chaininfocache = self._chaininfocache |
|
800 | chaininfocache = self._chaininfocache | |
801 | if rev in chaininfocache: |
|
801 | if rev in chaininfocache: | |
802 | return chaininfocache[rev] |
|
802 | return chaininfocache[rev] | |
803 | index = self.index |
|
803 | index = self.index | |
804 | generaldelta = self._generaldelta |
|
804 | generaldelta = self._generaldelta | |
805 | iterrev = rev |
|
805 | iterrev = rev | |
806 | e = index[iterrev] |
|
806 | e = index[iterrev] | |
807 | clen = 0 |
|
807 | clen = 0 | |
808 | compresseddeltalen = 0 |
|
808 | compresseddeltalen = 0 | |
809 | while iterrev != e[3]: |
|
809 | while iterrev != e[3]: | |
810 | clen += 1 |
|
810 | clen += 1 | |
811 | compresseddeltalen += e[1] |
|
811 | compresseddeltalen += e[1] | |
812 | if generaldelta: |
|
812 | if generaldelta: | |
813 | iterrev = e[3] |
|
813 | iterrev = e[3] | |
814 | else: |
|
814 | else: | |
815 | iterrev -= 1 |
|
815 | iterrev -= 1 | |
816 | if iterrev in chaininfocache: |
|
816 | if iterrev in chaininfocache: | |
817 | t = chaininfocache[iterrev] |
|
817 | t = chaininfocache[iterrev] | |
818 | clen += t[0] |
|
818 | clen += t[0] | |
819 | compresseddeltalen += t[1] |
|
819 | compresseddeltalen += t[1] | |
820 | break |
|
820 | break | |
821 | e = index[iterrev] |
|
821 | e = index[iterrev] | |
822 | else: |
|
822 | else: | |
823 | # Add text length of base since decompressing that also takes |
|
823 | # Add text length of base since decompressing that also takes | |
824 | # work. For cache hits the length is already included. |
|
824 | # work. For cache hits the length is already included. | |
825 | compresseddeltalen += e[1] |
|
825 | compresseddeltalen += e[1] | |
826 | r = (clen, compresseddeltalen) |
|
826 | r = (clen, compresseddeltalen) | |
827 | chaininfocache[rev] = r |
|
827 | chaininfocache[rev] = r | |
828 | return r |
|
828 | return r | |
829 |
|
829 | |||
830 | def _deltachain(self, rev, stoprev=None): |
|
830 | def _deltachain(self, rev, stoprev=None): | |
831 | """Obtain the delta chain for a revision. |
|
831 | """Obtain the delta chain for a revision. | |
832 |
|
832 | |||
833 | ``stoprev`` specifies a revision to stop at. If not specified, we |
|
833 | ``stoprev`` specifies a revision to stop at. If not specified, we | |
834 | stop at the base of the chain. |
|
834 | stop at the base of the chain. | |
835 |
|
835 | |||
836 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of |
|
836 | Returns a 2-tuple of (chain, stopped) where ``chain`` is a list of | |
837 | revs in ascending order and ``stopped`` is a bool indicating whether |
|
837 | revs in ascending order and ``stopped`` is a bool indicating whether | |
838 | ``stoprev`` was hit. |
|
838 | ``stoprev`` was hit. | |
839 | """ |
|
839 | """ | |
840 | # Try C implementation. |
|
840 | # Try C implementation. | |
841 | try: |
|
841 | try: | |
842 | return self.index.deltachain(rev, stoprev, self._generaldelta) |
|
842 | return self.index.deltachain(rev, stoprev, self._generaldelta) | |
843 | except AttributeError: |
|
843 | except AttributeError: | |
844 | pass |
|
844 | pass | |
845 |
|
845 | |||
846 | chain = [] |
|
846 | chain = [] | |
847 |
|
847 | |||
848 | # Alias to prevent attribute lookup in tight loop. |
|
848 | # Alias to prevent attribute lookup in tight loop. | |
849 | index = self.index |
|
849 | index = self.index | |
850 | generaldelta = self._generaldelta |
|
850 | generaldelta = self._generaldelta | |
851 |
|
851 | |||
852 | iterrev = rev |
|
852 | iterrev = rev | |
853 | e = index[iterrev] |
|
853 | e = index[iterrev] | |
854 | while iterrev != e[3] and iterrev != stoprev: |
|
854 | while iterrev != e[3] and iterrev != stoprev: | |
855 | chain.append(iterrev) |
|
855 | chain.append(iterrev) | |
856 | if generaldelta: |
|
856 | if generaldelta: | |
857 | iterrev = e[3] |
|
857 | iterrev = e[3] | |
858 | else: |
|
858 | else: | |
859 | iterrev -= 1 |
|
859 | iterrev -= 1 | |
860 | e = index[iterrev] |
|
860 | e = index[iterrev] | |
861 |
|
861 | |||
862 | if iterrev == stoprev: |
|
862 | if iterrev == stoprev: | |
863 | stopped = True |
|
863 | stopped = True | |
864 | else: |
|
864 | else: | |
865 | chain.append(iterrev) |
|
865 | chain.append(iterrev) | |
866 | stopped = False |
|
866 | stopped = False | |
867 |
|
867 | |||
868 | chain.reverse() |
|
868 | chain.reverse() | |
869 | return chain, stopped |
|
869 | return chain, stopped | |
870 |
|
870 | |||
871 | def ancestors(self, revs, stoprev=0, inclusive=False): |
|
871 | def ancestors(self, revs, stoprev=0, inclusive=False): | |
872 | """Generate the ancestors of 'revs' in reverse revision order. |
|
872 | """Generate the ancestors of 'revs' in reverse revision order. | |
873 | Does not generate revs lower than stoprev. |
|
873 | Does not generate revs lower than stoprev. | |
874 |
|
874 | |||
875 | See the documentation for ancestor.lazyancestors for more details.""" |
|
875 | See the documentation for ancestor.lazyancestors for more details.""" | |
876 |
|
876 | |||
877 | # first, make sure start revisions aren't filtered |
|
877 | # first, make sure start revisions aren't filtered | |
878 | revs = list(revs) |
|
878 | revs = list(revs) | |
879 | checkrev = self.node |
|
879 | checkrev = self.node | |
880 | for r in revs: |
|
880 | for r in revs: | |
881 | checkrev(r) |
|
881 | checkrev(r) | |
882 | # and we're sure ancestors aren't filtered as well |
|
882 | # and we're sure ancestors aren't filtered as well | |
883 |
|
883 | |||
884 | if rustancestor is not None: |
|
884 | if rustancestor is not None: | |
885 | lazyancestors = rustancestor.LazyAncestors |
|
885 | lazyancestors = rustancestor.LazyAncestors | |
886 | arg = self.index |
|
886 | arg = self.index | |
887 | else: |
|
887 | else: | |
888 | lazyancestors = ancestor.lazyancestors |
|
888 | lazyancestors = ancestor.lazyancestors | |
889 | arg = self._uncheckedparentrevs |
|
889 | arg = self._uncheckedparentrevs | |
890 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) |
|
890 | return lazyancestors(arg, revs, stoprev=stoprev, inclusive=inclusive) | |
891 |
|
891 | |||
892 | def descendants(self, revs): |
|
892 | def descendants(self, revs): | |
893 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) |
|
893 | return dagop.descendantrevs(revs, self.revs, self.parentrevs) | |
894 |
|
894 | |||
895 | def findcommonmissing(self, common=None, heads=None): |
|
895 | def findcommonmissing(self, common=None, heads=None): | |
896 | """Return a tuple of the ancestors of common and the ancestors of heads |
|
896 | """Return a tuple of the ancestors of common and the ancestors of heads | |
897 | that are not ancestors of common. In revset terminology, we return the |
|
897 | that are not ancestors of common. In revset terminology, we return the | |
898 | tuple: |
|
898 | tuple: | |
899 |
|
899 | |||
900 | ::common, (::heads) - (::common) |
|
900 | ::common, (::heads) - (::common) | |
901 |
|
901 | |||
902 | The list is sorted by revision number, meaning it is |
|
902 | The list is sorted by revision number, meaning it is | |
903 | topologically sorted. |
|
903 | topologically sorted. | |
904 |
|
904 | |||
905 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
905 | 'heads' and 'common' are both lists of node IDs. If heads is | |
906 | not supplied, uses all of the revlog's heads. If common is not |
|
906 | not supplied, uses all of the revlog's heads. If common is not | |
907 | supplied, uses nullid.""" |
|
907 | supplied, uses nullid.""" | |
908 | if common is None: |
|
908 | if common is None: | |
909 | common = [self.nullid] |
|
909 | common = [self.nullid] | |
910 | if heads is None: |
|
910 | if heads is None: | |
911 | heads = self.heads() |
|
911 | heads = self.heads() | |
912 |
|
912 | |||
913 | common = [self.rev(n) for n in common] |
|
913 | common = [self.rev(n) for n in common] | |
914 | heads = [self.rev(n) for n in heads] |
|
914 | heads = [self.rev(n) for n in heads] | |
915 |
|
915 | |||
916 | # we want the ancestors, but inclusive |
|
916 | # we want the ancestors, but inclusive | |
917 | class lazyset(object): |
|
917 | class lazyset(object): | |
918 | def __init__(self, lazyvalues): |
|
918 | def __init__(self, lazyvalues): | |
919 | self.addedvalues = set() |
|
919 | self.addedvalues = set() | |
920 | self.lazyvalues = lazyvalues |
|
920 | self.lazyvalues = lazyvalues | |
921 |
|
921 | |||
922 | def __contains__(self, value): |
|
922 | def __contains__(self, value): | |
923 | return value in self.addedvalues or value in self.lazyvalues |
|
923 | return value in self.addedvalues or value in self.lazyvalues | |
924 |
|
924 | |||
925 | def __iter__(self): |
|
925 | def __iter__(self): | |
926 | added = self.addedvalues |
|
926 | added = self.addedvalues | |
927 | for r in added: |
|
927 | for r in added: | |
928 | yield r |
|
928 | yield r | |
929 | for r in self.lazyvalues: |
|
929 | for r in self.lazyvalues: | |
930 | if not r in added: |
|
930 | if not r in added: | |
931 | yield r |
|
931 | yield r | |
932 |
|
932 | |||
933 | def add(self, value): |
|
933 | def add(self, value): | |
934 | self.addedvalues.add(value) |
|
934 | self.addedvalues.add(value) | |
935 |
|
935 | |||
936 | def update(self, values): |
|
936 | def update(self, values): | |
937 | self.addedvalues.update(values) |
|
937 | self.addedvalues.update(values) | |
938 |
|
938 | |||
939 | has = lazyset(self.ancestors(common)) |
|
939 | has = lazyset(self.ancestors(common)) | |
940 | has.add(nullrev) |
|
940 | has.add(nullrev) | |
941 | has.update(common) |
|
941 | has.update(common) | |
942 |
|
942 | |||
943 | # take all ancestors from heads that aren't in has |
|
943 | # take all ancestors from heads that aren't in has | |
944 | missing = set() |
|
944 | missing = set() | |
945 | visit = collections.deque(r for r in heads if r not in has) |
|
945 | visit = collections.deque(r for r in heads if r not in has) | |
946 | while visit: |
|
946 | while visit: | |
947 | r = visit.popleft() |
|
947 | r = visit.popleft() | |
948 | if r in missing: |
|
948 | if r in missing: | |
949 | continue |
|
949 | continue | |
950 | else: |
|
950 | else: | |
951 | missing.add(r) |
|
951 | missing.add(r) | |
952 | for p in self.parentrevs(r): |
|
952 | for p in self.parentrevs(r): | |
953 | if p not in has: |
|
953 | if p not in has: | |
954 | visit.append(p) |
|
954 | visit.append(p) | |
955 | missing = list(missing) |
|
955 | missing = list(missing) | |
956 | missing.sort() |
|
956 | missing.sort() | |
957 | return has, [self.node(miss) for miss in missing] |
|
957 | return has, [self.node(miss) for miss in missing] | |
958 |
|
958 | |||
959 | def incrementalmissingrevs(self, common=None): |
|
959 | def incrementalmissingrevs(self, common=None): | |
960 | """Return an object that can be used to incrementally compute the |
|
960 | """Return an object that can be used to incrementally compute the | |
961 | revision numbers of the ancestors of arbitrary sets that are not |
|
961 | revision numbers of the ancestors of arbitrary sets that are not | |
962 | ancestors of common. This is an ancestor.incrementalmissingancestors |
|
962 | ancestors of common. This is an ancestor.incrementalmissingancestors | |
963 | object. |
|
963 | object. | |
964 |
|
964 | |||
965 | 'common' is a list of revision numbers. If common is not supplied, uses |
|
965 | 'common' is a list of revision numbers. If common is not supplied, uses | |
966 | nullrev. |
|
966 | nullrev. | |
967 | """ |
|
967 | """ | |
968 | if common is None: |
|
968 | if common is None: | |
969 | common = [nullrev] |
|
969 | common = [nullrev] | |
970 |
|
970 | |||
971 | if rustancestor is not None: |
|
971 | if rustancestor is not None: | |
972 | return rustancestor.MissingAncestors(self.index, common) |
|
972 | return rustancestor.MissingAncestors(self.index, common) | |
973 | return ancestor.incrementalmissingancestors(self.parentrevs, common) |
|
973 | return ancestor.incrementalmissingancestors(self.parentrevs, common) | |
974 |
|
974 | |||
975 | def findmissingrevs(self, common=None, heads=None): |
|
975 | def findmissingrevs(self, common=None, heads=None): | |
976 | """Return the revision numbers of the ancestors of heads that |
|
976 | """Return the revision numbers of the ancestors of heads that | |
977 | are not ancestors of common. |
|
977 | are not ancestors of common. | |
978 |
|
978 | |||
979 | More specifically, return a list of revision numbers corresponding to |
|
979 | More specifically, return a list of revision numbers corresponding to | |
980 | nodes N such that every N satisfies the following constraints: |
|
980 | nodes N such that every N satisfies the following constraints: | |
981 |
|
981 | |||
982 | 1. N is an ancestor of some node in 'heads' |
|
982 | 1. N is an ancestor of some node in 'heads' | |
983 | 2. N is not an ancestor of any node in 'common' |
|
983 | 2. N is not an ancestor of any node in 'common' | |
984 |
|
984 | |||
985 | The list is sorted by revision number, meaning it is |
|
985 | The list is sorted by revision number, meaning it is | |
986 | topologically sorted. |
|
986 | topologically sorted. | |
987 |
|
987 | |||
988 | 'heads' and 'common' are both lists of revision numbers. If heads is |
|
988 | 'heads' and 'common' are both lists of revision numbers. If heads is | |
989 | not supplied, uses all of the revlog's heads. If common is not |
|
989 | not supplied, uses all of the revlog's heads. If common is not | |
990 | supplied, uses nullid.""" |
|
990 | supplied, uses nullid.""" | |
991 | if common is None: |
|
991 | if common is None: | |
992 | common = [nullrev] |
|
992 | common = [nullrev] | |
993 | if heads is None: |
|
993 | if heads is None: | |
994 | heads = self.headrevs() |
|
994 | heads = self.headrevs() | |
995 |
|
995 | |||
996 | inc = self.incrementalmissingrevs(common=common) |
|
996 | inc = self.incrementalmissingrevs(common=common) | |
997 | return inc.missingancestors(heads) |
|
997 | return inc.missingancestors(heads) | |
998 |
|
998 | |||
999 | def findmissing(self, common=None, heads=None): |
|
999 | def findmissing(self, common=None, heads=None): | |
1000 | """Return the ancestors of heads that are not ancestors of common. |
|
1000 | """Return the ancestors of heads that are not ancestors of common. | |
1001 |
|
1001 | |||
1002 | More specifically, return a list of nodes N such that every N |
|
1002 | More specifically, return a list of nodes N such that every N | |
1003 | satisfies the following constraints: |
|
1003 | satisfies the following constraints: | |
1004 |
|
1004 | |||
1005 | 1. N is an ancestor of some node in 'heads' |
|
1005 | 1. N is an ancestor of some node in 'heads' | |
1006 | 2. N is not an ancestor of any node in 'common' |
|
1006 | 2. N is not an ancestor of any node in 'common' | |
1007 |
|
1007 | |||
1008 | The list is sorted by revision number, meaning it is |
|
1008 | The list is sorted by revision number, meaning it is | |
1009 | topologically sorted. |
|
1009 | topologically sorted. | |
1010 |
|
1010 | |||
1011 | 'heads' and 'common' are both lists of node IDs. If heads is |
|
1011 | 'heads' and 'common' are both lists of node IDs. If heads is | |
1012 | not supplied, uses all of the revlog's heads. If common is not |
|
1012 | not supplied, uses all of the revlog's heads. If common is not | |
1013 | supplied, uses nullid.""" |
|
1013 | supplied, uses nullid.""" | |
1014 | if common is None: |
|
1014 | if common is None: | |
1015 | common = [self.nullid] |
|
1015 | common = [self.nullid] | |
1016 | if heads is None: |
|
1016 | if heads is None: | |
1017 | heads = self.heads() |
|
1017 | heads = self.heads() | |
1018 |
|
1018 | |||
1019 | common = [self.rev(n) for n in common] |
|
1019 | common = [self.rev(n) for n in common] | |
1020 | heads = [self.rev(n) for n in heads] |
|
1020 | heads = [self.rev(n) for n in heads] | |
1021 |
|
1021 | |||
1022 | inc = self.incrementalmissingrevs(common=common) |
|
1022 | inc = self.incrementalmissingrevs(common=common) | |
1023 | return [self.node(r) for r in inc.missingancestors(heads)] |
|
1023 | return [self.node(r) for r in inc.missingancestors(heads)] | |
1024 |
|
1024 | |||
1025 | def nodesbetween(self, roots=None, heads=None): |
|
1025 | def nodesbetween(self, roots=None, heads=None): | |
1026 | """Return a topological path from 'roots' to 'heads'. |
|
1026 | """Return a topological path from 'roots' to 'heads'. | |
1027 |
|
1027 | |||
1028 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a |
|
1028 | Return a tuple (nodes, outroots, outheads) where 'nodes' is a | |
1029 | topologically sorted list of all nodes N that satisfy both of |
|
1029 | topologically sorted list of all nodes N that satisfy both of | |
1030 | these constraints: |
|
1030 | these constraints: | |
1031 |
|
1031 | |||
1032 | 1. N is a descendant of some node in 'roots' |
|
1032 | 1. N is a descendant of some node in 'roots' | |
1033 | 2. N is an ancestor of some node in 'heads' |
|
1033 | 2. N is an ancestor of some node in 'heads' | |
1034 |
|
1034 | |||
1035 | Every node is considered to be both a descendant and an ancestor |
|
1035 | Every node is considered to be both a descendant and an ancestor | |
1036 | of itself, so every reachable node in 'roots' and 'heads' will be |
|
1036 | of itself, so every reachable node in 'roots' and 'heads' will be | |
1037 | included in 'nodes'. |
|
1037 | included in 'nodes'. | |
1038 |
|
1038 | |||
1039 | 'outroots' is the list of reachable nodes in 'roots', i.e., the |
|
1039 | 'outroots' is the list of reachable nodes in 'roots', i.e., the | |
1040 | subset of 'roots' that is returned in 'nodes'. Likewise, |
|
1040 | subset of 'roots' that is returned in 'nodes'. Likewise, | |
1041 | 'outheads' is the subset of 'heads' that is also in 'nodes'. |
|
1041 | 'outheads' is the subset of 'heads' that is also in 'nodes'. | |
1042 |
|
1042 | |||
1043 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is |
|
1043 | 'roots' and 'heads' are both lists of node IDs. If 'roots' is | |
1044 | unspecified, uses nullid as the only root. If 'heads' is |
|
1044 | unspecified, uses nullid as the only root. If 'heads' is | |
1045 | unspecified, uses list of all of the revlog's heads.""" |
|
1045 | unspecified, uses list of all of the revlog's heads.""" | |
1046 | nonodes = ([], [], []) |
|
1046 | nonodes = ([], [], []) | |
1047 | if roots is not None: |
|
1047 | if roots is not None: | |
1048 | roots = list(roots) |
|
1048 | roots = list(roots) | |
1049 | if not roots: |
|
1049 | if not roots: | |
1050 | return nonodes |
|
1050 | return nonodes | |
1051 | lowestrev = min([self.rev(n) for n in roots]) |
|
1051 | lowestrev = min([self.rev(n) for n in roots]) | |
1052 | else: |
|
1052 | else: | |
1053 | roots = [self.nullid] # Everybody's a descendant of nullid |
|
1053 | roots = [self.nullid] # Everybody's a descendant of nullid | |
1054 | lowestrev = nullrev |
|
1054 | lowestrev = nullrev | |
1055 | if (lowestrev == nullrev) and (heads is None): |
|
1055 | if (lowestrev == nullrev) and (heads is None): | |
1056 | # We want _all_ the nodes! |
|
1056 | # We want _all_ the nodes! | |
1057 | return ( |
|
1057 | return ( | |
1058 | [self.node(r) for r in self], |
|
1058 | [self.node(r) for r in self], | |
1059 | [self.nullid], |
|
1059 | [self.nullid], | |
1060 | list(self.heads()), |
|
1060 | list(self.heads()), | |
1061 | ) |
|
1061 | ) | |
1062 | if heads is None: |
|
1062 | if heads is None: | |
1063 | # All nodes are ancestors, so the latest ancestor is the last |
|
1063 | # All nodes are ancestors, so the latest ancestor is the last | |
1064 | # node. |
|
1064 | # node. | |
1065 | highestrev = len(self) - 1 |
|
1065 | highestrev = len(self) - 1 | |
1066 | # Set ancestors to None to signal that every node is an ancestor. |
|
1066 | # Set ancestors to None to signal that every node is an ancestor. | |
1067 | ancestors = None |
|
1067 | ancestors = None | |
1068 | # Set heads to an empty dictionary for later discovery of heads |
|
1068 | # Set heads to an empty dictionary for later discovery of heads | |
1069 | heads = {} |
|
1069 | heads = {} | |
1070 | else: |
|
1070 | else: | |
1071 | heads = list(heads) |
|
1071 | heads = list(heads) | |
1072 | if not heads: |
|
1072 | if not heads: | |
1073 | return nonodes |
|
1073 | return nonodes | |
1074 | ancestors = set() |
|
1074 | ancestors = set() | |
1075 | # Turn heads into a dictionary so we can remove 'fake' heads. |
|
1075 | # Turn heads into a dictionary so we can remove 'fake' heads. | |
1076 | # Also, later we will be using it to filter out the heads we can't |
|
1076 | # Also, later we will be using it to filter out the heads we can't | |
1077 | # find from roots. |
|
1077 | # find from roots. | |
1078 | heads = dict.fromkeys(heads, False) |
|
1078 | heads = dict.fromkeys(heads, False) | |
1079 | # Start at the top and keep marking parents until we're done. |
|
1079 | # Start at the top and keep marking parents until we're done. | |
1080 | nodestotag = set(heads) |
|
1080 | nodestotag = set(heads) | |
1081 | # Remember where the top was so we can use it as a limit later. |
|
1081 | # Remember where the top was so we can use it as a limit later. | |
1082 | highestrev = max([self.rev(n) for n in nodestotag]) |
|
1082 | highestrev = max([self.rev(n) for n in nodestotag]) | |
1083 | while nodestotag: |
|
1083 | while nodestotag: | |
1084 | # grab a node to tag |
|
1084 | # grab a node to tag | |
1085 | n = nodestotag.pop() |
|
1085 | n = nodestotag.pop() | |
1086 | # Never tag nullid |
|
1086 | # Never tag nullid | |
1087 | if n == self.nullid: |
|
1087 | if n == self.nullid: | |
1088 | continue |
|
1088 | continue | |
1089 | # A node's revision number represents its place in a |
|
1089 | # A node's revision number represents its place in a | |
1090 | # topologically sorted list of nodes. |
|
1090 | # topologically sorted list of nodes. | |
1091 | r = self.rev(n) |
|
1091 | r = self.rev(n) | |
1092 | if r >= lowestrev: |
|
1092 | if r >= lowestrev: | |
1093 | if n not in ancestors: |
|
1093 | if n not in ancestors: | |
1094 | # If we are possibly a descendant of one of the roots |
|
1094 | # If we are possibly a descendant of one of the roots | |
1095 | # and we haven't already been marked as an ancestor |
|
1095 | # and we haven't already been marked as an ancestor | |
1096 | ancestors.add(n) # Mark as ancestor |
|
1096 | ancestors.add(n) # Mark as ancestor | |
1097 | # Add non-nullid parents to list of nodes to tag. |
|
1097 | # Add non-nullid parents to list of nodes to tag. | |
1098 | nodestotag.update( |
|
1098 | nodestotag.update( | |
1099 | [p for p in self.parents(n) if p != self.nullid] |
|
1099 | [p for p in self.parents(n) if p != self.nullid] | |
1100 | ) |
|
1100 | ) | |
1101 | elif n in heads: # We've seen it before, is it a fake head? |
|
1101 | elif n in heads: # We've seen it before, is it a fake head? | |
1102 | # So it is, real heads should not be the ancestors of |
|
1102 | # So it is, real heads should not be the ancestors of | |
1103 | # any other heads. |
|
1103 | # any other heads. | |
1104 | heads.pop(n) |
|
1104 | heads.pop(n) | |
1105 | if not ancestors: |
|
1105 | if not ancestors: | |
1106 | return nonodes |
|
1106 | return nonodes | |
1107 | # Now that we have our set of ancestors, we want to remove any |
|
1107 | # Now that we have our set of ancestors, we want to remove any | |
1108 | # roots that are not ancestors. |
|
1108 | # roots that are not ancestors. | |
1109 |
|
1109 | |||
1110 | # If one of the roots was nullid, everything is included anyway. |
|
1110 | # If one of the roots was nullid, everything is included anyway. | |
1111 | if lowestrev > nullrev: |
|
1111 | if lowestrev > nullrev: | |
1112 | # But, since we weren't, let's recompute the lowest rev to not |
|
1112 | # But, since we weren't, let's recompute the lowest rev to not | |
1113 | # include roots that aren't ancestors. |
|
1113 | # include roots that aren't ancestors. | |
1114 |
|
1114 | |||
1115 | # Filter out roots that aren't ancestors of heads |
|
1115 | # Filter out roots that aren't ancestors of heads | |
1116 | roots = [root for root in roots if root in ancestors] |
|
1116 | roots = [root for root in roots if root in ancestors] | |
1117 | # Recompute the lowest revision |
|
1117 | # Recompute the lowest revision | |
1118 | if roots: |
|
1118 | if roots: | |
1119 | lowestrev = min([self.rev(root) for root in roots]) |
|
1119 | lowestrev = min([self.rev(root) for root in roots]) | |
1120 | else: |
|
1120 | else: | |
1121 | # No more roots? Return empty list |
|
1121 | # No more roots? Return empty list | |
1122 | return nonodes |
|
1122 | return nonodes | |
1123 | else: |
|
1123 | else: | |
1124 | # We are descending from nullid, and don't need to care about |
|
1124 | # We are descending from nullid, and don't need to care about | |
1125 | # any other roots. |
|
1125 | # any other roots. | |
1126 | lowestrev = nullrev |
|
1126 | lowestrev = nullrev | |
1127 | roots = [self.nullid] |
|
1127 | roots = [self.nullid] | |
1128 | # Transform our roots list into a set. |
|
1128 | # Transform our roots list into a set. | |
1129 | descendants = set(roots) |
|
1129 | descendants = set(roots) | |
1130 | # Also, keep the original roots so we can filter out roots that aren't |
|
1130 | # Also, keep the original roots so we can filter out roots that aren't | |
1131 | # 'real' roots (i.e. are descended from other roots). |
|
1131 | # 'real' roots (i.e. are descended from other roots). | |
1132 | roots = descendants.copy() |
|
1132 | roots = descendants.copy() | |
1133 | # Our topologically sorted list of output nodes. |
|
1133 | # Our topologically sorted list of output nodes. | |
1134 | orderedout = [] |
|
1134 | orderedout = [] | |
1135 | # Don't start at nullid since we don't want nullid in our output list, |
|
1135 | # Don't start at nullid since we don't want nullid in our output list, | |
1136 | # and if nullid shows up in descendants, empty parents will look like |
|
1136 | # and if nullid shows up in descendants, empty parents will look like | |
1137 | # they're descendants. |
|
1137 | # they're descendants. | |
1138 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): |
|
1138 | for r in self.revs(start=max(lowestrev, 0), stop=highestrev + 1): | |
1139 | n = self.node(r) |
|
1139 | n = self.node(r) | |
1140 | isdescendant = False |
|
1140 | isdescendant = False | |
1141 | if lowestrev == nullrev: # Everybody is a descendant of nullid |
|
1141 | if lowestrev == nullrev: # Everybody is a descendant of nullid | |
1142 | isdescendant = True |
|
1142 | isdescendant = True | |
1143 | elif n in descendants: |
|
1143 | elif n in descendants: | |
1144 | # n is already a descendant |
|
1144 | # n is already a descendant | |
1145 | isdescendant = True |
|
1145 | isdescendant = True | |
1146 | # This check only needs to be done here because all the roots |
|
1146 | # This check only needs to be done here because all the roots | |
1147 | # will start being marked is descendants before the loop. |
|
1147 | # will start being marked is descendants before the loop. | |
1148 | if n in roots: |
|
1148 | if n in roots: | |
1149 | # If n was a root, check if it's a 'real' root. |
|
1149 | # If n was a root, check if it's a 'real' root. | |
1150 | p = tuple(self.parents(n)) |
|
1150 | p = tuple(self.parents(n)) | |
1151 | # If any of its parents are descendants, it's not a root. |
|
1151 | # If any of its parents are descendants, it's not a root. | |
1152 | if (p[0] in descendants) or (p[1] in descendants): |
|
1152 | if (p[0] in descendants) or (p[1] in descendants): | |
1153 | roots.remove(n) |
|
1153 | roots.remove(n) | |
1154 | else: |
|
1154 | else: | |
1155 | p = tuple(self.parents(n)) |
|
1155 | p = tuple(self.parents(n)) | |
1156 | # A node is a descendant if either of its parents are |
|
1156 | # A node is a descendant if either of its parents are | |
1157 | # descendants. (We seeded the dependents list with the roots |
|
1157 | # descendants. (We seeded the dependents list with the roots | |
1158 | # up there, remember?) |
|
1158 | # up there, remember?) | |
1159 | if (p[0] in descendants) or (p[1] in descendants): |
|
1159 | if (p[0] in descendants) or (p[1] in descendants): | |
1160 | descendants.add(n) |
|
1160 | descendants.add(n) | |
1161 | isdescendant = True |
|
1161 | isdescendant = True | |
1162 | if isdescendant and ((ancestors is None) or (n in ancestors)): |
|
1162 | if isdescendant and ((ancestors is None) or (n in ancestors)): | |
1163 | # Only include nodes that are both descendants and ancestors. |
|
1163 | # Only include nodes that are both descendants and ancestors. | |
1164 | orderedout.append(n) |
|
1164 | orderedout.append(n) | |
1165 | if (ancestors is not None) and (n in heads): |
|
1165 | if (ancestors is not None) and (n in heads): | |
1166 | # We're trying to figure out which heads are reachable |
|
1166 | # We're trying to figure out which heads are reachable | |
1167 | # from roots. |
|
1167 | # from roots. | |
1168 | # Mark this head as having been reached |
|
1168 | # Mark this head as having been reached | |
1169 | heads[n] = True |
|
1169 | heads[n] = True | |
1170 | elif ancestors is None: |
|
1170 | elif ancestors is None: | |
1171 | # Otherwise, we're trying to discover the heads. |
|
1171 | # Otherwise, we're trying to discover the heads. | |
1172 | # Assume this is a head because if it isn't, the next step |
|
1172 | # Assume this is a head because if it isn't, the next step | |
1173 | # will eventually remove it. |
|
1173 | # will eventually remove it. | |
1174 | heads[n] = True |
|
1174 | heads[n] = True | |
1175 | # But, obviously its parents aren't. |
|
1175 | # But, obviously its parents aren't. | |
1176 | for p in self.parents(n): |
|
1176 | for p in self.parents(n): | |
1177 | heads.pop(p, None) |
|
1177 | heads.pop(p, None) | |
1178 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] |
|
1178 | heads = [head for head, flag in pycompat.iteritems(heads) if flag] | |
1179 | roots = list(roots) |
|
1179 | roots = list(roots) | |
1180 | assert orderedout |
|
1180 | assert orderedout | |
1181 | assert roots |
|
1181 | assert roots | |
1182 | assert heads |
|
1182 | assert heads | |
1183 | return (orderedout, roots, heads) |
|
1183 | return (orderedout, roots, heads) | |
1184 |
|
1184 | |||
1185 | def headrevs(self, revs=None): |
|
1185 | def headrevs(self, revs=None): | |
1186 | if revs is None: |
|
1186 | if revs is None: | |
1187 | try: |
|
1187 | try: | |
1188 | return self.index.headrevs() |
|
1188 | return self.index.headrevs() | |
1189 | except AttributeError: |
|
1189 | except AttributeError: | |
1190 | return self._headrevs() |
|
1190 | return self._headrevs() | |
1191 | if rustdagop is not None: |
|
1191 | if rustdagop is not None: | |
1192 | return rustdagop.headrevs(self.index, revs) |
|
1192 | return rustdagop.headrevs(self.index, revs) | |
1193 | return dagop.headrevs(revs, self._uncheckedparentrevs) |
|
1193 | return dagop.headrevs(revs, self._uncheckedparentrevs) | |
1194 |
|
1194 | |||
1195 | def computephases(self, roots): |
|
1195 | def computephases(self, roots): | |
1196 | return self.index.computephasesmapsets(roots) |
|
1196 | return self.index.computephasesmapsets(roots) | |
1197 |
|
1197 | |||
1198 | def _headrevs(self): |
|
1198 | def _headrevs(self): | |
1199 | count = len(self) |
|
1199 | count = len(self) | |
1200 | if not count: |
|
1200 | if not count: | |
1201 | return [nullrev] |
|
1201 | return [nullrev] | |
1202 | # we won't iter over filtered rev so nobody is a head at start |
|
1202 | # we won't iter over filtered rev so nobody is a head at start | |
1203 | ishead = [0] * (count + 1) |
|
1203 | ishead = [0] * (count + 1) | |
1204 | index = self.index |
|
1204 | index = self.index | |
1205 | for r in self: |
|
1205 | for r in self: | |
1206 | ishead[r] = 1 # I may be an head |
|
1206 | ishead[r] = 1 # I may be an head | |
1207 | e = index[r] |
|
1207 | e = index[r] | |
1208 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not |
|
1208 | ishead[e[5]] = ishead[e[6]] = 0 # my parent are not | |
1209 | return [r for r, val in enumerate(ishead) if val] |
|
1209 | return [r for r, val in enumerate(ishead) if val] | |
1210 |
|
1210 | |||
1211 | def heads(self, start=None, stop=None): |
|
1211 | def heads(self, start=None, stop=None): | |
1212 | """return the list of all nodes that have no children |
|
1212 | """return the list of all nodes that have no children | |
1213 |
|
1213 | |||
1214 | if start is specified, only heads that are descendants of |
|
1214 | if start is specified, only heads that are descendants of | |
1215 | start will be returned |
|
1215 | start will be returned | |
1216 | if stop is specified, it will consider all the revs from stop |
|
1216 | if stop is specified, it will consider all the revs from stop | |
1217 | as if they had no children |
|
1217 | as if they had no children | |
1218 | """ |
|
1218 | """ | |
1219 | if start is None and stop is None: |
|
1219 | if start is None and stop is None: | |
1220 | if not len(self): |
|
1220 | if not len(self): | |
1221 | return [self.nullid] |
|
1221 | return [self.nullid] | |
1222 | return [self.node(r) for r in self.headrevs()] |
|
1222 | return [self.node(r) for r in self.headrevs()] | |
1223 |
|
1223 | |||
1224 | if start is None: |
|
1224 | if start is None: | |
1225 | start = nullrev |
|
1225 | start = nullrev | |
1226 | else: |
|
1226 | else: | |
1227 | start = self.rev(start) |
|
1227 | start = self.rev(start) | |
1228 |
|
1228 | |||
1229 | stoprevs = {self.rev(n) for n in stop or []} |
|
1229 | stoprevs = {self.rev(n) for n in stop or []} | |
1230 |
|
1230 | |||
1231 | revs = dagop.headrevssubset( |
|
1231 | revs = dagop.headrevssubset( | |
1232 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs |
|
1232 | self.revs, self.parentrevs, startrev=start, stoprevs=stoprevs | |
1233 | ) |
|
1233 | ) | |
1234 |
|
1234 | |||
1235 | return [self.node(rev) for rev in revs] |
|
1235 | return [self.node(rev) for rev in revs] | |
1236 |
|
1236 | |||
1237 | def children(self, node): |
|
1237 | def children(self, node): | |
1238 | """find the children of a given node""" |
|
1238 | """find the children of a given node""" | |
1239 | c = [] |
|
1239 | c = [] | |
1240 | p = self.rev(node) |
|
1240 | p = self.rev(node) | |
1241 | for r in self.revs(start=p + 1): |
|
1241 | for r in self.revs(start=p + 1): | |
1242 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] |
|
1242 | prevs = [pr for pr in self.parentrevs(r) if pr != nullrev] | |
1243 | if prevs: |
|
1243 | if prevs: | |
1244 | for pr in prevs: |
|
1244 | for pr in prevs: | |
1245 | if pr == p: |
|
1245 | if pr == p: | |
1246 | c.append(self.node(r)) |
|
1246 | c.append(self.node(r)) | |
1247 | elif p == nullrev: |
|
1247 | elif p == nullrev: | |
1248 | c.append(self.node(r)) |
|
1248 | c.append(self.node(r)) | |
1249 | return c |
|
1249 | return c | |
1250 |
|
1250 | |||
1251 | def commonancestorsheads(self, a, b): |
|
1251 | def commonancestorsheads(self, a, b): | |
1252 | """calculate all the heads of the common ancestors of nodes a and b""" |
|
1252 | """calculate all the heads of the common ancestors of nodes a and b""" | |
1253 | a, b = self.rev(a), self.rev(b) |
|
1253 | a, b = self.rev(a), self.rev(b) | |
1254 | ancs = self._commonancestorsheads(a, b) |
|
1254 | ancs = self._commonancestorsheads(a, b) | |
1255 | return pycompat.maplist(self.node, ancs) |
|
1255 | return pycompat.maplist(self.node, ancs) | |
1256 |
|
1256 | |||
1257 | def _commonancestorsheads(self, *revs): |
|
1257 | def _commonancestorsheads(self, *revs): | |
1258 | """calculate all the heads of the common ancestors of revs""" |
|
1258 | """calculate all the heads of the common ancestors of revs""" | |
1259 | try: |
|
1259 | try: | |
1260 | ancs = self.index.commonancestorsheads(*revs) |
|
1260 | ancs = self.index.commonancestorsheads(*revs) | |
1261 | except (AttributeError, OverflowError): # C implementation failed |
|
1261 | except (AttributeError, OverflowError): # C implementation failed | |
1262 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) |
|
1262 | ancs = ancestor.commonancestorsheads(self.parentrevs, *revs) | |
1263 | return ancs |
|
1263 | return ancs | |
1264 |
|
1264 | |||
1265 | def isancestor(self, a, b): |
|
1265 | def isancestor(self, a, b): | |
1266 | """return True if node a is an ancestor of node b |
|
1266 | """return True if node a is an ancestor of node b | |
1267 |
|
1267 | |||
1268 | A revision is considered an ancestor of itself.""" |
|
1268 | A revision is considered an ancestor of itself.""" | |
1269 | a, b = self.rev(a), self.rev(b) |
|
1269 | a, b = self.rev(a), self.rev(b) | |
1270 | return self.isancestorrev(a, b) |
|
1270 | return self.isancestorrev(a, b) | |
1271 |
|
1271 | |||
1272 | def isancestorrev(self, a, b): |
|
1272 | def isancestorrev(self, a, b): | |
1273 | """return True if revision a is an ancestor of revision b |
|
1273 | """return True if revision a is an ancestor of revision b | |
1274 |
|
1274 | |||
1275 | A revision is considered an ancestor of itself. |
|
1275 | A revision is considered an ancestor of itself. | |
1276 |
|
1276 | |||
1277 | The implementation of this is trivial but the use of |
|
1277 | The implementation of this is trivial but the use of | |
1278 | reachableroots is not.""" |
|
1278 | reachableroots is not.""" | |
1279 | if a == nullrev: |
|
1279 | if a == nullrev: | |
1280 | return True |
|
1280 | return True | |
1281 | elif a == b: |
|
1281 | elif a == b: | |
1282 | return True |
|
1282 | return True | |
1283 | elif a > b: |
|
1283 | elif a > b: | |
1284 | return False |
|
1284 | return False | |
1285 | return bool(self.reachableroots(a, [b], [a], includepath=False)) |
|
1285 | return bool(self.reachableroots(a, [b], [a], includepath=False)) | |
1286 |
|
1286 | |||
1287 | def reachableroots(self, minroot, heads, roots, includepath=False): |
|
1287 | def reachableroots(self, minroot, heads, roots, includepath=False): | |
1288 | """return (heads(::(<roots> and <roots>::<heads>))) |
|
1288 | """return (heads(::(<roots> and <roots>::<heads>))) | |
1289 |
|
1289 | |||
1290 | If includepath is True, return (<roots>::<heads>).""" |
|
1290 | If includepath is True, return (<roots>::<heads>).""" | |
1291 | try: |
|
1291 | try: | |
1292 | return self.index.reachableroots2( |
|
1292 | return self.index.reachableroots2( | |
1293 | minroot, heads, roots, includepath |
|
1293 | minroot, heads, roots, includepath | |
1294 | ) |
|
1294 | ) | |
1295 | except AttributeError: |
|
1295 | except AttributeError: | |
1296 | return dagop._reachablerootspure( |
|
1296 | return dagop._reachablerootspure( | |
1297 | self.parentrevs, minroot, roots, heads, includepath |
|
1297 | self.parentrevs, minroot, roots, heads, includepath | |
1298 | ) |
|
1298 | ) | |
1299 |
|
1299 | |||
1300 | def ancestor(self, a, b): |
|
1300 | def ancestor(self, a, b): | |
1301 | """calculate the "best" common ancestor of nodes a and b""" |
|
1301 | """calculate the "best" common ancestor of nodes a and b""" | |
1302 |
|
1302 | |||
1303 | a, b = self.rev(a), self.rev(b) |
|
1303 | a, b = self.rev(a), self.rev(b) | |
1304 | try: |
|
1304 | try: | |
1305 | ancs = self.index.ancestors(a, b) |
|
1305 | ancs = self.index.ancestors(a, b) | |
1306 | except (AttributeError, OverflowError): |
|
1306 | except (AttributeError, OverflowError): | |
1307 | ancs = ancestor.ancestors(self.parentrevs, a, b) |
|
1307 | ancs = ancestor.ancestors(self.parentrevs, a, b) | |
1308 | if ancs: |
|
1308 | if ancs: | |
1309 | # choose a consistent winner when there's a tie |
|
1309 | # choose a consistent winner when there's a tie | |
1310 | return min(map(self.node, ancs)) |
|
1310 | return min(map(self.node, ancs)) | |
1311 | return self.nullid |
|
1311 | return self.nullid | |
1312 |
|
1312 | |||
1313 | def _match(self, id): |
|
1313 | def _match(self, id): | |
1314 | if isinstance(id, int): |
|
1314 | if isinstance(id, int): | |
1315 | # rev |
|
1315 | # rev | |
1316 | return self.node(id) |
|
1316 | return self.node(id) | |
1317 | if len(id) == self.nodeconstants.nodelen: |
|
1317 | if len(id) == self.nodeconstants.nodelen: | |
1318 | # possibly a binary node |
|
1318 | # possibly a binary node | |
1319 | # odds of a binary node being all hex in ASCII are 1 in 10**25 |
|
1319 | # odds of a binary node being all hex in ASCII are 1 in 10**25 | |
1320 | try: |
|
1320 | try: | |
1321 | node = id |
|
1321 | node = id | |
1322 | self.rev(node) # quick search the index |
|
1322 | self.rev(node) # quick search the index | |
1323 | return node |
|
1323 | return node | |
1324 | except error.LookupError: |
|
1324 | except error.LookupError: | |
1325 | pass # may be partial hex id |
|
1325 | pass # may be partial hex id | |
1326 | try: |
|
1326 | try: | |
1327 | # str(rev) |
|
1327 | # str(rev) | |
1328 | rev = int(id) |
|
1328 | rev = int(id) | |
1329 | if b"%d" % rev != id: |
|
1329 | if b"%d" % rev != id: | |
1330 | raise ValueError |
|
1330 | raise ValueError | |
1331 | if rev < 0: |
|
1331 | if rev < 0: | |
1332 | rev = len(self) + rev |
|
1332 | rev = len(self) + rev | |
1333 | if rev < 0 or rev >= len(self): |
|
1333 | if rev < 0 or rev >= len(self): | |
1334 | raise ValueError |
|
1334 | raise ValueError | |
1335 | return self.node(rev) |
|
1335 | return self.node(rev) | |
1336 | except (ValueError, OverflowError): |
|
1336 | except (ValueError, OverflowError): | |
1337 | pass |
|
1337 | pass | |
1338 | if len(id) == 2 * self.nodeconstants.nodelen: |
|
1338 | if len(id) == 2 * self.nodeconstants.nodelen: | |
1339 | try: |
|
1339 | try: | |
1340 | # a full hex nodeid? |
|
1340 | # a full hex nodeid? | |
1341 | node = bin(id) |
|
1341 | node = bin(id) | |
1342 | self.rev(node) |
|
1342 | self.rev(node) | |
1343 | return node |
|
1343 | return node | |
1344 | except (TypeError, error.LookupError): |
|
1344 | except (TypeError, error.LookupError): | |
1345 | pass |
|
1345 | pass | |
1346 |
|
1346 | |||
1347 | def _partialmatch(self, id): |
|
1347 | def _partialmatch(self, id): | |
1348 | # we don't care wdirfilenodeids as they should be always full hash |
|
1348 | # we don't care wdirfilenodeids as they should be always full hash | |
1349 | maybewdir = self.nodeconstants.wdirhex.startswith(id) |
|
1349 | maybewdir = self.nodeconstants.wdirhex.startswith(id) | |
1350 | try: |
|
1350 | try: | |
1351 | partial = self.index.partialmatch(id) |
|
1351 | partial = self.index.partialmatch(id) | |
1352 | if partial and self.hasnode(partial): |
|
1352 | if partial and self.hasnode(partial): | |
1353 | if maybewdir: |
|
1353 | if maybewdir: | |
1354 | # single 'ff...' match in radix tree, ambiguous with wdir |
|
1354 | # single 'ff...' match in radix tree, ambiguous with wdir | |
1355 | raise error.RevlogError |
|
1355 | raise error.RevlogError | |
1356 | return partial |
|
1356 | return partial | |
1357 | if maybewdir: |
|
1357 | if maybewdir: | |
1358 | # no 'ff...' match in radix tree, wdir identified |
|
1358 | # no 'ff...' match in radix tree, wdir identified | |
1359 | raise error.WdirUnsupported |
|
1359 | raise error.WdirUnsupported | |
1360 | return None |
|
1360 | return None | |
1361 | except error.RevlogError: |
|
1361 | except error.RevlogError: | |
1362 | # parsers.c radix tree lookup gave multiple matches |
|
1362 | # parsers.c radix tree lookup gave multiple matches | |
1363 | # fast path: for unfiltered changelog, radix tree is accurate |
|
1363 | # fast path: for unfiltered changelog, radix tree is accurate | |
1364 | if not getattr(self, 'filteredrevs', None): |
|
1364 | if not getattr(self, 'filteredrevs', None): | |
1365 | raise error.AmbiguousPrefixLookupError( |
|
1365 | raise error.AmbiguousPrefixLookupError( | |
1366 | id, self.indexfile, _(b'ambiguous identifier') |
|
1366 | id, self.indexfile, _(b'ambiguous identifier') | |
1367 | ) |
|
1367 | ) | |
1368 | # fall through to slow path that filters hidden revisions |
|
1368 | # fall through to slow path that filters hidden revisions | |
1369 | except (AttributeError, ValueError): |
|
1369 | except (AttributeError, ValueError): | |
1370 | # we are pure python, or key was too short to search radix tree |
|
1370 | # we are pure python, or key was too short to search radix tree | |
1371 | pass |
|
1371 | pass | |
1372 |
|
1372 | |||
1373 | if id in self._pcache: |
|
1373 | if id in self._pcache: | |
1374 | return self._pcache[id] |
|
1374 | return self._pcache[id] | |
1375 |
|
1375 | |||
1376 | if len(id) <= 40: |
|
1376 | if len(id) <= 40: | |
1377 | try: |
|
1377 | try: | |
1378 | # hex(node)[:...] |
|
1378 | # hex(node)[:...] | |
1379 | l = len(id) // 2 # grab an even number of digits |
|
1379 | l = len(id) // 2 # grab an even number of digits | |
1380 | prefix = bin(id[: l * 2]) |
|
1380 | prefix = bin(id[: l * 2]) | |
1381 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] |
|
1381 | nl = [e[7] for e in self.index if e[7].startswith(prefix)] | |
1382 | nl = [ |
|
1382 | nl = [ | |
1383 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) |
|
1383 | n for n in nl if hex(n).startswith(id) and self.hasnode(n) | |
1384 | ] |
|
1384 | ] | |
1385 | if self.nodeconstants.nullhex.startswith(id): |
|
1385 | if self.nodeconstants.nullhex.startswith(id): | |
1386 | nl.append(self.nullid) |
|
1386 | nl.append(self.nullid) | |
1387 | if len(nl) > 0: |
|
1387 | if len(nl) > 0: | |
1388 | if len(nl) == 1 and not maybewdir: |
|
1388 | if len(nl) == 1 and not maybewdir: | |
1389 | self._pcache[id] = nl[0] |
|
1389 | self._pcache[id] = nl[0] | |
1390 | return nl[0] |
|
1390 | return nl[0] | |
1391 | raise error.AmbiguousPrefixLookupError( |
|
1391 | raise error.AmbiguousPrefixLookupError( | |
1392 | id, self.indexfile, _(b'ambiguous identifier') |
|
1392 | id, self.indexfile, _(b'ambiguous identifier') | |
1393 | ) |
|
1393 | ) | |
1394 | if maybewdir: |
|
1394 | if maybewdir: | |
1395 | raise error.WdirUnsupported |
|
1395 | raise error.WdirUnsupported | |
1396 | return None |
|
1396 | return None | |
1397 | except TypeError: |
|
1397 | except TypeError: | |
1398 | pass |
|
1398 | pass | |
1399 |
|
1399 | |||
1400 | def lookup(self, id): |
|
1400 | def lookup(self, id): | |
1401 | """locate a node based on: |
|
1401 | """locate a node based on: | |
1402 | - revision number or str(revision number) |
|
1402 | - revision number or str(revision number) | |
1403 | - nodeid or subset of hex nodeid |
|
1403 | - nodeid or subset of hex nodeid | |
1404 | """ |
|
1404 | """ | |
1405 | n = self._match(id) |
|
1405 | n = self._match(id) | |
1406 | if n is not None: |
|
1406 | if n is not None: | |
1407 | return n |
|
1407 | return n | |
1408 | n = self._partialmatch(id) |
|
1408 | n = self._partialmatch(id) | |
1409 | if n: |
|
1409 | if n: | |
1410 | return n |
|
1410 | return n | |
1411 |
|
1411 | |||
1412 | raise error.LookupError(id, self.indexfile, _(b'no match found')) |
|
1412 | raise error.LookupError(id, self.indexfile, _(b'no match found')) | |
1413 |
|
1413 | |||
1414 | def shortest(self, node, minlength=1): |
|
1414 | def shortest(self, node, minlength=1): | |
1415 | """Find the shortest unambiguous prefix that matches node.""" |
|
1415 | """Find the shortest unambiguous prefix that matches node.""" | |
1416 |
|
1416 | |||
1417 | def isvalid(prefix): |
|
1417 | def isvalid(prefix): | |
1418 | try: |
|
1418 | try: | |
1419 | matchednode = self._partialmatch(prefix) |
|
1419 | matchednode = self._partialmatch(prefix) | |
1420 | except error.AmbiguousPrefixLookupError: |
|
1420 | except error.AmbiguousPrefixLookupError: | |
1421 | return False |
|
1421 | return False | |
1422 | except error.WdirUnsupported: |
|
1422 | except error.WdirUnsupported: | |
1423 | # single 'ff...' match |
|
1423 | # single 'ff...' match | |
1424 | return True |
|
1424 | return True | |
1425 | if matchednode is None: |
|
1425 | if matchednode is None: | |
1426 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1426 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1427 | return True |
|
1427 | return True | |
1428 |
|
1428 | |||
1429 | def maybewdir(prefix): |
|
1429 | def maybewdir(prefix): | |
1430 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) |
|
1430 | return all(c == b'f' for c in pycompat.iterbytestr(prefix)) | |
1431 |
|
1431 | |||
1432 | hexnode = hex(node) |
|
1432 | hexnode = hex(node) | |
1433 |
|
1433 | |||
1434 | def disambiguate(hexnode, minlength): |
|
1434 | def disambiguate(hexnode, minlength): | |
1435 | """Disambiguate against wdirid.""" |
|
1435 | """Disambiguate against wdirid.""" | |
1436 | for length in range(minlength, len(hexnode) + 1): |
|
1436 | for length in range(minlength, len(hexnode) + 1): | |
1437 | prefix = hexnode[:length] |
|
1437 | prefix = hexnode[:length] | |
1438 | if not maybewdir(prefix): |
|
1438 | if not maybewdir(prefix): | |
1439 | return prefix |
|
1439 | return prefix | |
1440 |
|
1440 | |||
1441 | if not getattr(self, 'filteredrevs', None): |
|
1441 | if not getattr(self, 'filteredrevs', None): | |
1442 | try: |
|
1442 | try: | |
1443 | length = max(self.index.shortest(node), minlength) |
|
1443 | length = max(self.index.shortest(node), minlength) | |
1444 | return disambiguate(hexnode, length) |
|
1444 | return disambiguate(hexnode, length) | |
1445 | except error.RevlogError: |
|
1445 | except error.RevlogError: | |
1446 | if node != self.nodeconstants.wdirid: |
|
1446 | if node != self.nodeconstants.wdirid: | |
1447 | raise error.LookupError(node, self.indexfile, _(b'no node')) |
|
1447 | raise error.LookupError(node, self.indexfile, _(b'no node')) | |
1448 | except AttributeError: |
|
1448 | except AttributeError: | |
1449 | # Fall through to pure code |
|
1449 | # Fall through to pure code | |
1450 | pass |
|
1450 | pass | |
1451 |
|
1451 | |||
1452 | if node == self.nodeconstants.wdirid: |
|
1452 | if node == self.nodeconstants.wdirid: | |
1453 | for length in range(minlength, len(hexnode) + 1): |
|
1453 | for length in range(minlength, len(hexnode) + 1): | |
1454 | prefix = hexnode[:length] |
|
1454 | prefix = hexnode[:length] | |
1455 | if isvalid(prefix): |
|
1455 | if isvalid(prefix): | |
1456 | return prefix |
|
1456 | return prefix | |
1457 |
|
1457 | |||
1458 | for length in range(minlength, len(hexnode) + 1): |
|
1458 | for length in range(minlength, len(hexnode) + 1): | |
1459 | prefix = hexnode[:length] |
|
1459 | prefix = hexnode[:length] | |
1460 | if isvalid(prefix): |
|
1460 | if isvalid(prefix): | |
1461 | return disambiguate(hexnode, length) |
|
1461 | return disambiguate(hexnode, length) | |
1462 |
|
1462 | |||
1463 | def cmp(self, node, text): |
|
1463 | def cmp(self, node, text): | |
1464 | """compare text with a given file revision |
|
1464 | """compare text with a given file revision | |
1465 |
|
1465 | |||
1466 | returns True if text is different than what is stored. |
|
1466 | returns True if text is different than what is stored. | |
1467 | """ |
|
1467 | """ | |
1468 | p1, p2 = self.parents(node) |
|
1468 | p1, p2 = self.parents(node) | |
1469 | return storageutil.hashrevisionsha1(text, p1, p2) != node |
|
1469 | return storageutil.hashrevisionsha1(text, p1, p2) != node | |
1470 |
|
1470 | |||
1471 | def _cachesegment(self, offset, data): |
|
1471 | def _cachesegment(self, offset, data): | |
1472 | """Add a segment to the revlog cache. |
|
1472 | """Add a segment to the revlog cache. | |
1473 |
|
1473 | |||
1474 | Accepts an absolute offset and the data that is at that location. |
|
1474 | Accepts an absolute offset and the data that is at that location. | |
1475 | """ |
|
1475 | """ | |
1476 | o, d = self._chunkcache |
|
1476 | o, d = self._chunkcache | |
1477 | # try to add to existing cache |
|
1477 | # try to add to existing cache | |
1478 | if o + len(d) == offset and len(d) + len(data) < _chunksize: |
|
1478 | if o + len(d) == offset and len(d) + len(data) < _chunksize: | |
1479 | self._chunkcache = o, d + data |
|
1479 | self._chunkcache = o, d + data | |
1480 | else: |
|
1480 | else: | |
1481 | self._chunkcache = offset, data |
|
1481 | self._chunkcache = offset, data | |
1482 |
|
1482 | |||
1483 | def _readsegment(self, offset, length, df=None): |
|
1483 | def _readsegment(self, offset, length, df=None): | |
1484 | """Load a segment of raw data from the revlog. |
|
1484 | """Load a segment of raw data from the revlog. | |
1485 |
|
1485 | |||
1486 | Accepts an absolute offset, length to read, and an optional existing |
|
1486 | Accepts an absolute offset, length to read, and an optional existing | |
1487 | file handle to read from. |
|
1487 | file handle to read from. | |
1488 |
|
1488 | |||
1489 | If an existing file handle is passed, it will be seeked and the |
|
1489 | If an existing file handle is passed, it will be seeked and the | |
1490 | original seek position will NOT be restored. |
|
1490 | original seek position will NOT be restored. | |
1491 |
|
1491 | |||
1492 | Returns a str or buffer of raw byte data. |
|
1492 | Returns a str or buffer of raw byte data. | |
1493 |
|
1493 | |||
1494 | Raises if the requested number of bytes could not be read. |
|
1494 | Raises if the requested number of bytes could not be read. | |
1495 | """ |
|
1495 | """ | |
1496 | # Cache data both forward and backward around the requested |
|
1496 | # Cache data both forward and backward around the requested | |
1497 | # data, in a fixed size window. This helps speed up operations |
|
1497 | # data, in a fixed size window. This helps speed up operations | |
1498 | # involving reading the revlog backwards. |
|
1498 | # involving reading the revlog backwards. | |
1499 | cachesize = self._chunkcachesize |
|
1499 | cachesize = self._chunkcachesize | |
1500 | realoffset = offset & ~(cachesize - 1) |
|
1500 | realoffset = offset & ~(cachesize - 1) | |
1501 | reallength = ( |
|
1501 | reallength = ( | |
1502 | (offset + length + cachesize) & ~(cachesize - 1) |
|
1502 | (offset + length + cachesize) & ~(cachesize - 1) | |
1503 | ) - realoffset |
|
1503 | ) - realoffset | |
1504 | with self._datareadfp(df) as df: |
|
1504 | with self._datareadfp(df) as df: | |
1505 | df.seek(realoffset) |
|
1505 | df.seek(realoffset) | |
1506 | d = df.read(reallength) |
|
1506 | d = df.read(reallength) | |
1507 |
|
1507 | |||
1508 | self._cachesegment(realoffset, d) |
|
1508 | self._cachesegment(realoffset, d) | |
1509 | if offset != realoffset or reallength != length: |
|
1509 | if offset != realoffset or reallength != length: | |
1510 | startoffset = offset - realoffset |
|
1510 | startoffset = offset - realoffset | |
1511 | if len(d) - startoffset < length: |
|
1511 | if len(d) - startoffset < length: | |
1512 | raise error.RevlogError( |
|
1512 | raise error.RevlogError( | |
1513 | _( |
|
1513 | _( | |
1514 | b'partial read of revlog %s; expected %d bytes from ' |
|
1514 | b'partial read of revlog %s; expected %d bytes from ' | |
1515 | b'offset %d, got %d' |
|
1515 | b'offset %d, got %d' | |
1516 | ) |
|
1516 | ) | |
1517 | % ( |
|
1517 | % ( | |
1518 | self.indexfile if self._inline else self.datafile, |
|
1518 | self.indexfile if self._inline else self.datafile, | |
1519 | length, |
|
1519 | length, | |
1520 | realoffset, |
|
1520 | realoffset, | |
1521 | len(d) - startoffset, |
|
1521 | len(d) - startoffset, | |
1522 | ) |
|
1522 | ) | |
1523 | ) |
|
1523 | ) | |
1524 |
|
1524 | |||
1525 | return util.buffer(d, startoffset, length) |
|
1525 | return util.buffer(d, startoffset, length) | |
1526 |
|
1526 | |||
1527 | if len(d) < length: |
|
1527 | if len(d) < length: | |
1528 | raise error.RevlogError( |
|
1528 | raise error.RevlogError( | |
1529 | _( |
|
1529 | _( | |
1530 | b'partial read of revlog %s; expected %d bytes from offset ' |
|
1530 | b'partial read of revlog %s; expected %d bytes from offset ' | |
1531 | b'%d, got %d' |
|
1531 | b'%d, got %d' | |
1532 | ) |
|
1532 | ) | |
1533 | % ( |
|
1533 | % ( | |
1534 | self.indexfile if self._inline else self.datafile, |
|
1534 | self.indexfile if self._inline else self.datafile, | |
1535 | length, |
|
1535 | length, | |
1536 | offset, |
|
1536 | offset, | |
1537 | len(d), |
|
1537 | len(d), | |
1538 | ) |
|
1538 | ) | |
1539 | ) |
|
1539 | ) | |
1540 |
|
1540 | |||
1541 | return d |
|
1541 | return d | |
1542 |
|
1542 | |||
1543 | def _getsegment(self, offset, length, df=None): |
|
1543 | def _getsegment(self, offset, length, df=None): | |
1544 | """Obtain a segment of raw data from the revlog. |
|
1544 | """Obtain a segment of raw data from the revlog. | |
1545 |
|
1545 | |||
1546 | Accepts an absolute offset, length of bytes to obtain, and an |
|
1546 | Accepts an absolute offset, length of bytes to obtain, and an | |
1547 | optional file handle to the already-opened revlog. If the file |
|
1547 | optional file handle to the already-opened revlog. If the file | |
1548 | handle is used, it's original seek position will not be preserved. |
|
1548 | handle is used, it's original seek position will not be preserved. | |
1549 |
|
1549 | |||
1550 | Requests for data may be returned from a cache. |
|
1550 | Requests for data may be returned from a cache. | |
1551 |
|
1551 | |||
1552 | Returns a str or a buffer instance of raw byte data. |
|
1552 | Returns a str or a buffer instance of raw byte data. | |
1553 | """ |
|
1553 | """ | |
1554 | o, d = self._chunkcache |
|
1554 | o, d = self._chunkcache | |
1555 | l = len(d) |
|
1555 | l = len(d) | |
1556 |
|
1556 | |||
1557 | # is it in the cache? |
|
1557 | # is it in the cache? | |
1558 | cachestart = offset - o |
|
1558 | cachestart = offset - o | |
1559 | cacheend = cachestart + length |
|
1559 | cacheend = cachestart + length | |
1560 | if cachestart >= 0 and cacheend <= l: |
|
1560 | if cachestart >= 0 and cacheend <= l: | |
1561 | if cachestart == 0 and cacheend == l: |
|
1561 | if cachestart == 0 and cacheend == l: | |
1562 | return d # avoid a copy |
|
1562 | return d # avoid a copy | |
1563 | return util.buffer(d, cachestart, cacheend - cachestart) |
|
1563 | return util.buffer(d, cachestart, cacheend - cachestart) | |
1564 |
|
1564 | |||
1565 | return self._readsegment(offset, length, df=df) |
|
1565 | return self._readsegment(offset, length, df=df) | |
1566 |
|
1566 | |||
1567 | def _getsegmentforrevs(self, startrev, endrev, df=None): |
|
1567 | def _getsegmentforrevs(self, startrev, endrev, df=None): | |
1568 | """Obtain a segment of raw data corresponding to a range of revisions. |
|
1568 | """Obtain a segment of raw data corresponding to a range of revisions. | |
1569 |
|
1569 | |||
1570 | Accepts the start and end revisions and an optional already-open |
|
1570 | Accepts the start and end revisions and an optional already-open | |
1571 | file handle to be used for reading. If the file handle is read, its |
|
1571 | file handle to be used for reading. If the file handle is read, its | |
1572 | seek position will not be preserved. |
|
1572 | seek position will not be preserved. | |
1573 |
|
1573 | |||
1574 | Requests for data may be satisfied by a cache. |
|
1574 | Requests for data may be satisfied by a cache. | |
1575 |
|
1575 | |||
1576 | Returns a 2-tuple of (offset, data) for the requested range of |
|
1576 | Returns a 2-tuple of (offset, data) for the requested range of | |
1577 | revisions. Offset is the integer offset from the beginning of the |
|
1577 | revisions. Offset is the integer offset from the beginning of the | |
1578 | revlog and data is a str or buffer of the raw byte data. |
|
1578 | revlog and data is a str or buffer of the raw byte data. | |
1579 |
|
1579 | |||
1580 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` |
|
1580 | Callers will need to call ``self.start(rev)`` and ``self.length(rev)`` | |
1581 | to determine where each revision's data begins and ends. |
|
1581 | to determine where each revision's data begins and ends. | |
1582 | """ |
|
1582 | """ | |
1583 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons |
|
1583 | # Inlined self.start(startrev) & self.end(endrev) for perf reasons | |
1584 | # (functions are expensive). |
|
1584 | # (functions are expensive). | |
1585 | index = self.index |
|
1585 | index = self.index | |
1586 | istart = index[startrev] |
|
1586 | istart = index[startrev] | |
1587 | start = int(istart[0] >> 16) |
|
1587 | start = int(istart[0] >> 16) | |
1588 | if startrev == endrev: |
|
1588 | if startrev == endrev: | |
1589 | end = start + istart[1] |
|
1589 | end = start + istart[1] | |
1590 | else: |
|
1590 | else: | |
1591 | iend = index[endrev] |
|
1591 | iend = index[endrev] | |
1592 | end = int(iend[0] >> 16) + iend[1] |
|
1592 | end = int(iend[0] >> 16) + iend[1] | |
1593 |
|
1593 | |||
1594 | if self._inline: |
|
1594 | if self._inline: | |
1595 | start += (startrev + 1) * self.index.entry_size |
|
1595 | start += (startrev + 1) * self.index.entry_size | |
1596 | end += (endrev + 1) * self.index.entry_size |
|
1596 | end += (endrev + 1) * self.index.entry_size | |
1597 | length = end - start |
|
1597 | length = end - start | |
1598 |
|
1598 | |||
1599 | return start, self._getsegment(start, length, df=df) |
|
1599 | return start, self._getsegment(start, length, df=df) | |
1600 |
|
1600 | |||
1601 | def _chunk(self, rev, df=None): |
|
1601 | def _chunk(self, rev, df=None): | |
1602 | """Obtain a single decompressed chunk for a revision. |
|
1602 | """Obtain a single decompressed chunk for a revision. | |
1603 |
|
1603 | |||
1604 | Accepts an integer revision and an optional already-open file handle |
|
1604 | Accepts an integer revision and an optional already-open file handle | |
1605 | to be used for reading. If used, the seek position of the file will not |
|
1605 | to be used for reading. If used, the seek position of the file will not | |
1606 | be preserved. |
|
1606 | be preserved. | |
1607 |
|
1607 | |||
1608 | Returns a str holding uncompressed data for the requested revision. |
|
1608 | Returns a str holding uncompressed data for the requested revision. | |
1609 | """ |
|
1609 | """ | |
1610 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) |
|
1610 | return self.decompress(self._getsegmentforrevs(rev, rev, df=df)[1]) | |
1611 |
|
1611 | |||
1612 | def _chunks(self, revs, df=None, targetsize=None): |
|
1612 | def _chunks(self, revs, df=None, targetsize=None): | |
1613 | """Obtain decompressed chunks for the specified revisions. |
|
1613 | """Obtain decompressed chunks for the specified revisions. | |
1614 |
|
1614 | |||
1615 | Accepts an iterable of numeric revisions that are assumed to be in |
|
1615 | Accepts an iterable of numeric revisions that are assumed to be in | |
1616 | ascending order. Also accepts an optional already-open file handle |
|
1616 | ascending order. Also accepts an optional already-open file handle | |
1617 | to be used for reading. If used, the seek position of the file will |
|
1617 | to be used for reading. If used, the seek position of the file will | |
1618 | not be preserved. |
|
1618 | not be preserved. | |
1619 |
|
1619 | |||
1620 | This function is similar to calling ``self._chunk()`` multiple times, |
|
1620 | This function is similar to calling ``self._chunk()`` multiple times, | |
1621 | but is faster. |
|
1621 | but is faster. | |
1622 |
|
1622 | |||
1623 | Returns a list with decompressed data for each requested revision. |
|
1623 | Returns a list with decompressed data for each requested revision. | |
1624 | """ |
|
1624 | """ | |
1625 | if not revs: |
|
1625 | if not revs: | |
1626 | return [] |
|
1626 | return [] | |
1627 | start = self.start |
|
1627 | start = self.start | |
1628 | length = self.length |
|
1628 | length = self.length | |
1629 | inline = self._inline |
|
1629 | inline = self._inline | |
1630 | iosize = self.index.entry_size |
|
1630 | iosize = self.index.entry_size | |
1631 | buffer = util.buffer |
|
1631 | buffer = util.buffer | |
1632 |
|
1632 | |||
1633 | l = [] |
|
1633 | l = [] | |
1634 | ladd = l.append |
|
1634 | ladd = l.append | |
1635 |
|
1635 | |||
1636 | if not self._withsparseread: |
|
1636 | if not self._withsparseread: | |
1637 | slicedchunks = (revs,) |
|
1637 | slicedchunks = (revs,) | |
1638 | else: |
|
1638 | else: | |
1639 | slicedchunks = deltautil.slicechunk( |
|
1639 | slicedchunks = deltautil.slicechunk( | |
1640 | self, revs, targetsize=targetsize |
|
1640 | self, revs, targetsize=targetsize | |
1641 | ) |
|
1641 | ) | |
1642 |
|
1642 | |||
1643 | for revschunk in slicedchunks: |
|
1643 | for revschunk in slicedchunks: | |
1644 | firstrev = revschunk[0] |
|
1644 | firstrev = revschunk[0] | |
1645 | # Skip trailing revisions with empty diff |
|
1645 | # Skip trailing revisions with empty diff | |
1646 | for lastrev in revschunk[::-1]: |
|
1646 | for lastrev in revschunk[::-1]: | |
1647 | if length(lastrev) != 0: |
|
1647 | if length(lastrev) != 0: | |
1648 | break |
|
1648 | break | |
1649 |
|
1649 | |||
1650 | try: |
|
1650 | try: | |
1651 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) |
|
1651 | offset, data = self._getsegmentforrevs(firstrev, lastrev, df=df) | |
1652 | except OverflowError: |
|
1652 | except OverflowError: | |
1653 | # issue4215 - we can't cache a run of chunks greater than |
|
1653 | # issue4215 - we can't cache a run of chunks greater than | |
1654 | # 2G on Windows |
|
1654 | # 2G on Windows | |
1655 | return [self._chunk(rev, df=df) for rev in revschunk] |
|
1655 | return [self._chunk(rev, df=df) for rev in revschunk] | |
1656 |
|
1656 | |||
1657 | decomp = self.decompress |
|
1657 | decomp = self.decompress | |
1658 | for rev in revschunk: |
|
1658 | for rev in revschunk: | |
1659 | chunkstart = start(rev) |
|
1659 | chunkstart = start(rev) | |
1660 | if inline: |
|
1660 | if inline: | |
1661 | chunkstart += (rev + 1) * iosize |
|
1661 | chunkstart += (rev + 1) * iosize | |
1662 | chunklength = length(rev) |
|
1662 | chunklength = length(rev) | |
1663 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) |
|
1663 | ladd(decomp(buffer(data, chunkstart - offset, chunklength))) | |
1664 |
|
1664 | |||
1665 | return l |
|
1665 | return l | |
1666 |
|
1666 | |||
1667 | def _chunkclear(self): |
|
1667 | def _chunkclear(self): | |
1668 | """Clear the raw chunk cache.""" |
|
1668 | """Clear the raw chunk cache.""" | |
1669 | self._chunkcache = (0, b'') |
|
1669 | self._chunkcache = (0, b'') | |
1670 |
|
1670 | |||
1671 | def deltaparent(self, rev): |
|
1671 | def deltaparent(self, rev): | |
1672 | """return deltaparent of the given revision""" |
|
1672 | """return deltaparent of the given revision""" | |
1673 | base = self.index[rev][3] |
|
1673 | base = self.index[rev][3] | |
1674 | if base == rev: |
|
1674 | if base == rev: | |
1675 | return nullrev |
|
1675 | return nullrev | |
1676 | elif self._generaldelta: |
|
1676 | elif self._generaldelta: | |
1677 | return base |
|
1677 | return base | |
1678 | else: |
|
1678 | else: | |
1679 | return rev - 1 |
|
1679 | return rev - 1 | |
1680 |
|
1680 | |||
1681 | def issnapshot(self, rev): |
|
1681 | def issnapshot(self, rev): | |
1682 | """tells whether rev is a snapshot""" |
|
1682 | """tells whether rev is a snapshot""" | |
1683 | if not self._sparserevlog: |
|
1683 | if not self._sparserevlog: | |
1684 | return self.deltaparent(rev) == nullrev |
|
1684 | return self.deltaparent(rev) == nullrev | |
1685 | elif util.safehasattr(self.index, b'issnapshot'): |
|
1685 | elif util.safehasattr(self.index, b'issnapshot'): | |
1686 | # directly assign the method to cache the testing and access |
|
1686 | # directly assign the method to cache the testing and access | |
1687 | self.issnapshot = self.index.issnapshot |
|
1687 | self.issnapshot = self.index.issnapshot | |
1688 | return self.issnapshot(rev) |
|
1688 | return self.issnapshot(rev) | |
1689 | if rev == nullrev: |
|
1689 | if rev == nullrev: | |
1690 | return True |
|
1690 | return True | |
1691 | entry = self.index[rev] |
|
1691 | entry = self.index[rev] | |
1692 | base = entry[3] |
|
1692 | base = entry[3] | |
1693 | if base == rev: |
|
1693 | if base == rev: | |
1694 | return True |
|
1694 | return True | |
1695 | if base == nullrev: |
|
1695 | if base == nullrev: | |
1696 | return True |
|
1696 | return True | |
1697 | p1 = entry[5] |
|
1697 | p1 = entry[5] | |
1698 | p2 = entry[6] |
|
1698 | p2 = entry[6] | |
1699 | if base == p1 or base == p2: |
|
1699 | if base == p1 or base == p2: | |
1700 | return False |
|
1700 | return False | |
1701 | return self.issnapshot(base) |
|
1701 | return self.issnapshot(base) | |
1702 |
|
1702 | |||
1703 | def snapshotdepth(self, rev): |
|
1703 | def snapshotdepth(self, rev): | |
1704 | """number of snapshot in the chain before this one""" |
|
1704 | """number of snapshot in the chain before this one""" | |
1705 | if not self.issnapshot(rev): |
|
1705 | if not self.issnapshot(rev): | |
1706 | raise error.ProgrammingError(b'revision %d not a snapshot') |
|
1706 | raise error.ProgrammingError(b'revision %d not a snapshot') | |
1707 | return len(self._deltachain(rev)[0]) - 1 |
|
1707 | return len(self._deltachain(rev)[0]) - 1 | |
1708 |
|
1708 | |||
1709 | def revdiff(self, rev1, rev2): |
|
1709 | def revdiff(self, rev1, rev2): | |
1710 | """return or calculate a delta between two revisions |
|
1710 | """return or calculate a delta between two revisions | |
1711 |
|
1711 | |||
1712 | The delta calculated is in binary form and is intended to be written to |
|
1712 | The delta calculated is in binary form and is intended to be written to | |
1713 | revlog data directly. So this function needs raw revision data. |
|
1713 | revlog data directly. So this function needs raw revision data. | |
1714 | """ |
|
1714 | """ | |
1715 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: |
|
1715 | if rev1 != nullrev and self.deltaparent(rev2) == rev1: | |
1716 | return bytes(self._chunk(rev2)) |
|
1716 | return bytes(self._chunk(rev2)) | |
1717 |
|
1717 | |||
1718 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) |
|
1718 | return mdiff.textdiff(self.rawdata(rev1), self.rawdata(rev2)) | |
1719 |
|
1719 | |||
1720 | def _processflags(self, text, flags, operation, raw=False): |
|
1720 | def _processflags(self, text, flags, operation, raw=False): | |
1721 | """deprecated entry point to access flag processors""" |
|
1721 | """deprecated entry point to access flag processors""" | |
1722 | msg = b'_processflag(...) use the specialized variant' |
|
1722 | msg = b'_processflag(...) use the specialized variant' | |
1723 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1723 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1724 | if raw: |
|
1724 | if raw: | |
1725 | return text, flagutil.processflagsraw(self, text, flags) |
|
1725 | return text, flagutil.processflagsraw(self, text, flags) | |
1726 | elif operation == b'read': |
|
1726 | elif operation == b'read': | |
1727 | return flagutil.processflagsread(self, text, flags) |
|
1727 | return flagutil.processflagsread(self, text, flags) | |
1728 | else: # write operation |
|
1728 | else: # write operation | |
1729 | return flagutil.processflagswrite(self, text, flags) |
|
1729 | return flagutil.processflagswrite(self, text, flags) | |
1730 |
|
1730 | |||
1731 | def revision(self, nodeorrev, _df=None, raw=False): |
|
1731 | def revision(self, nodeorrev, _df=None, raw=False): | |
1732 | """return an uncompressed revision of a given node or revision |
|
1732 | """return an uncompressed revision of a given node or revision | |
1733 | number. |
|
1733 | number. | |
1734 |
|
1734 | |||
1735 | _df - an existing file handle to read from. (internal-only) |
|
1735 | _df - an existing file handle to read from. (internal-only) | |
1736 | raw - an optional argument specifying if the revision data is to be |
|
1736 | raw - an optional argument specifying if the revision data is to be | |
1737 | treated as raw data when applying flag transforms. 'raw' should be set |
|
1737 | treated as raw data when applying flag transforms. 'raw' should be set | |
1738 | to True when generating changegroups or in debug commands. |
|
1738 | to True when generating changegroups or in debug commands. | |
1739 | """ |
|
1739 | """ | |
1740 | if raw: |
|
1740 | if raw: | |
1741 | msg = ( |
|
1741 | msg = ( | |
1742 | b'revlog.revision(..., raw=True) is deprecated, ' |
|
1742 | b'revlog.revision(..., raw=True) is deprecated, ' | |
1743 | b'use revlog.rawdata(...)' |
|
1743 | b'use revlog.rawdata(...)' | |
1744 | ) |
|
1744 | ) | |
1745 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) |
|
1745 | util.nouideprecwarn(msg, b'5.2', stacklevel=2) | |
1746 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] |
|
1746 | return self._revisiondata(nodeorrev, _df, raw=raw)[0] | |
1747 |
|
1747 | |||
1748 | def sidedata(self, nodeorrev, _df=None): |
|
1748 | def sidedata(self, nodeorrev, _df=None): | |
1749 | """a map of extra data related to the changeset but not part of the hash |
|
1749 | """a map of extra data related to the changeset but not part of the hash | |
1750 |
|
1750 | |||
1751 | This function currently return a dictionary. However, more advanced |
|
1751 | This function currently return a dictionary. However, more advanced | |
1752 | mapping object will likely be used in the future for a more |
|
1752 | mapping object will likely be used in the future for a more | |
1753 | efficient/lazy code. |
|
1753 | efficient/lazy code. | |
1754 | """ |
|
1754 | """ | |
1755 | return self._revisiondata(nodeorrev, _df)[1] |
|
1755 | return self._revisiondata(nodeorrev, _df)[1] | |
1756 |
|
1756 | |||
1757 | def _revisiondata(self, nodeorrev, _df=None, raw=False): |
|
1757 | def _revisiondata(self, nodeorrev, _df=None, raw=False): | |
1758 | # deal with <nodeorrev> argument type |
|
1758 | # deal with <nodeorrev> argument type | |
1759 | if isinstance(nodeorrev, int): |
|
1759 | if isinstance(nodeorrev, int): | |
1760 | rev = nodeorrev |
|
1760 | rev = nodeorrev | |
1761 | node = self.node(rev) |
|
1761 | node = self.node(rev) | |
1762 | else: |
|
1762 | else: | |
1763 | node = nodeorrev |
|
1763 | node = nodeorrev | |
1764 | rev = None |
|
1764 | rev = None | |
1765 |
|
1765 | |||
1766 | # fast path the special `nullid` rev |
|
1766 | # fast path the special `nullid` rev | |
1767 | if node == self.nullid: |
|
1767 | if node == self.nullid: | |
1768 | return b"", {} |
|
1768 | return b"", {} | |
1769 |
|
1769 | |||
1770 | # ``rawtext`` is the text as stored inside the revlog. Might be the |
|
1770 | # ``rawtext`` is the text as stored inside the revlog. Might be the | |
1771 | # revision or might need to be processed to retrieve the revision. |
|
1771 | # revision or might need to be processed to retrieve the revision. | |
1772 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) |
|
1772 | rev, rawtext, validated = self._rawtext(node, rev, _df=_df) | |
1773 |
|
1773 | |||
1774 | if self.version & 0xFFFF == REVLOGV2: |
|
1774 | if self.version & 0xFFFF == REVLOGV2: | |
1775 | if rev is None: |
|
1775 | if rev is None: | |
1776 | rev = self.rev(node) |
|
1776 | rev = self.rev(node) | |
1777 | sidedata = self._sidedata(rev) |
|
1777 | sidedata = self._sidedata(rev) | |
1778 | else: |
|
1778 | else: | |
1779 | sidedata = {} |
|
1779 | sidedata = {} | |
1780 |
|
1780 | |||
1781 | if raw and validated: |
|
1781 | if raw and validated: | |
1782 | # if we don't want to process the raw text and that raw |
|
1782 | # if we don't want to process the raw text and that raw | |
1783 | # text is cached, we can exit early. |
|
1783 | # text is cached, we can exit early. | |
1784 | return rawtext, sidedata |
|
1784 | return rawtext, sidedata | |
1785 | if rev is None: |
|
1785 | if rev is None: | |
1786 | rev = self.rev(node) |
|
1786 | rev = self.rev(node) | |
1787 | # the revlog's flag for this revision |
|
1787 | # the revlog's flag for this revision | |
1788 | # (usually alter its state or content) |
|
1788 | # (usually alter its state or content) | |
1789 | flags = self.flags(rev) |
|
1789 | flags = self.flags(rev) | |
1790 |
|
1790 | |||
1791 | if validated and flags == REVIDX_DEFAULT_FLAGS: |
|
1791 | if validated and flags == REVIDX_DEFAULT_FLAGS: | |
1792 | # no extra flags set, no flag processor runs, text = rawtext |
|
1792 | # no extra flags set, no flag processor runs, text = rawtext | |
1793 | return rawtext, sidedata |
|
1793 | return rawtext, sidedata | |
1794 |
|
1794 | |||
1795 | if raw: |
|
1795 | if raw: | |
1796 | validatehash = flagutil.processflagsraw(self, rawtext, flags) |
|
1796 | validatehash = flagutil.processflagsraw(self, rawtext, flags) | |
1797 | text = rawtext |
|
1797 | text = rawtext | |
1798 | else: |
|
1798 | else: | |
1799 | r = flagutil.processflagsread(self, rawtext, flags) |
|
1799 | r = flagutil.processflagsread(self, rawtext, flags) | |
1800 | text, validatehash = r |
|
1800 | text, validatehash = r | |
1801 | if validatehash: |
|
1801 | if validatehash: | |
1802 | self.checkhash(text, node, rev=rev) |
|
1802 | self.checkhash(text, node, rev=rev) | |
1803 | if not validated: |
|
1803 | if not validated: | |
1804 | self._revisioncache = (node, rev, rawtext) |
|
1804 | self._revisioncache = (node, rev, rawtext) | |
1805 |
|
1805 | |||
1806 | return text, sidedata |
|
1806 | return text, sidedata | |
1807 |
|
1807 | |||
1808 | def _rawtext(self, node, rev, _df=None): |
|
1808 | def _rawtext(self, node, rev, _df=None): | |
1809 | """return the possibly unvalidated rawtext for a revision |
|
1809 | """return the possibly unvalidated rawtext for a revision | |
1810 |
|
1810 | |||
1811 | returns (rev, rawtext, validated) |
|
1811 | returns (rev, rawtext, validated) | |
1812 | """ |
|
1812 | """ | |
1813 |
|
1813 | |||
1814 | # revision in the cache (could be useful to apply delta) |
|
1814 | # revision in the cache (could be useful to apply delta) | |
1815 | cachedrev = None |
|
1815 | cachedrev = None | |
1816 | # An intermediate text to apply deltas to |
|
1816 | # An intermediate text to apply deltas to | |
1817 | basetext = None |
|
1817 | basetext = None | |
1818 |
|
1818 | |||
1819 | # Check if we have the entry in cache |
|
1819 | # Check if we have the entry in cache | |
1820 | # The cache entry looks like (node, rev, rawtext) |
|
1820 | # The cache entry looks like (node, rev, rawtext) | |
1821 | if self._revisioncache: |
|
1821 | if self._revisioncache: | |
1822 | if self._revisioncache[0] == node: |
|
1822 | if self._revisioncache[0] == node: | |
1823 | return (rev, self._revisioncache[2], True) |
|
1823 | return (rev, self._revisioncache[2], True) | |
1824 | cachedrev = self._revisioncache[1] |
|
1824 | cachedrev = self._revisioncache[1] | |
1825 |
|
1825 | |||
1826 | if rev is None: |
|
1826 | if rev is None: | |
1827 | rev = self.rev(node) |
|
1827 | rev = self.rev(node) | |
1828 |
|
1828 | |||
1829 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) |
|
1829 | chain, stopped = self._deltachain(rev, stoprev=cachedrev) | |
1830 | if stopped: |
|
1830 | if stopped: | |
1831 | basetext = self._revisioncache[2] |
|
1831 | basetext = self._revisioncache[2] | |
1832 |
|
1832 | |||
1833 | # drop cache to save memory, the caller is expected to |
|
1833 | # drop cache to save memory, the caller is expected to | |
1834 | # update self._revisioncache after validating the text |
|
1834 | # update self._revisioncache after validating the text | |
1835 | self._revisioncache = None |
|
1835 | self._revisioncache = None | |
1836 |
|
1836 | |||
1837 | targetsize = None |
|
1837 | targetsize = None | |
1838 | rawsize = self.index[rev][2] |
|
1838 | rawsize = self.index[rev][2] | |
1839 | if 0 <= rawsize: |
|
1839 | if 0 <= rawsize: | |
1840 | targetsize = 4 * rawsize |
|
1840 | targetsize = 4 * rawsize | |
1841 |
|
1841 | |||
1842 | bins = self._chunks(chain, df=_df, targetsize=targetsize) |
|
1842 | bins = self._chunks(chain, df=_df, targetsize=targetsize) | |
1843 | if basetext is None: |
|
1843 | if basetext is None: | |
1844 | basetext = bytes(bins[0]) |
|
1844 | basetext = bytes(bins[0]) | |
1845 | bins = bins[1:] |
|
1845 | bins = bins[1:] | |
1846 |
|
1846 | |||
1847 | rawtext = mdiff.patches(basetext, bins) |
|
1847 | rawtext = mdiff.patches(basetext, bins) | |
1848 | del basetext # let us have a chance to free memory early |
|
1848 | del basetext # let us have a chance to free memory early | |
1849 | return (rev, rawtext, False) |
|
1849 | return (rev, rawtext, False) | |
1850 |
|
1850 | |||
1851 | def _sidedata(self, rev): |
|
1851 | def _sidedata(self, rev): | |
1852 | """Return the sidedata for a given revision number.""" |
|
1852 | """Return the sidedata for a given revision number.""" | |
1853 | index_entry = self.index[rev] |
|
1853 | index_entry = self.index[rev] | |
1854 | sidedata_offset = index_entry[8] |
|
1854 | sidedata_offset = index_entry[8] | |
1855 | sidedata_size = index_entry[9] |
|
1855 | sidedata_size = index_entry[9] | |
1856 |
|
1856 | |||
1857 | if self._inline: |
|
1857 | if self._inline: | |
1858 | sidedata_offset += self.index.entry_size * (1 + rev) |
|
1858 | sidedata_offset += self.index.entry_size * (1 + rev) | |
1859 | if sidedata_size == 0: |
|
1859 | if sidedata_size == 0: | |
1860 | return {} |
|
1860 | return {} | |
1861 |
|
1861 | |||
1862 | segment = self._getsegment(sidedata_offset, sidedata_size) |
|
1862 | segment = self._getsegment(sidedata_offset, sidedata_size) | |
1863 | sidedata = sidedatautil.deserialize_sidedata(segment) |
|
1863 | sidedata = sidedatautil.deserialize_sidedata(segment) | |
1864 | return sidedata |
|
1864 | return sidedata | |
1865 |
|
1865 | |||
1866 | def rawdata(self, nodeorrev, _df=None): |
|
1866 | def rawdata(self, nodeorrev, _df=None): | |
1867 | """return an uncompressed raw data of a given node or revision number. |
|
1867 | """return an uncompressed raw data of a given node or revision number. | |
1868 |
|
1868 | |||
1869 | _df - an existing file handle to read from. (internal-only) |
|
1869 | _df - an existing file handle to read from. (internal-only) | |
1870 | """ |
|
1870 | """ | |
1871 | return self._revisiondata(nodeorrev, _df, raw=True)[0] |
|
1871 | return self._revisiondata(nodeorrev, _df, raw=True)[0] | |
1872 |
|
1872 | |||
1873 | def hash(self, text, p1, p2): |
|
1873 | def hash(self, text, p1, p2): | |
1874 | """Compute a node hash. |
|
1874 | """Compute a node hash. | |
1875 |
|
1875 | |||
1876 | Available as a function so that subclasses can replace the hash |
|
1876 | Available as a function so that subclasses can replace the hash | |
1877 | as needed. |
|
1877 | as needed. | |
1878 | """ |
|
1878 | """ | |
1879 | return storageutil.hashrevisionsha1(text, p1, p2) |
|
1879 | return storageutil.hashrevisionsha1(text, p1, p2) | |
1880 |
|
1880 | |||
1881 | def checkhash(self, text, node, p1=None, p2=None, rev=None): |
|
1881 | def checkhash(self, text, node, p1=None, p2=None, rev=None): | |
1882 | """Check node hash integrity. |
|
1882 | """Check node hash integrity. | |
1883 |
|
1883 | |||
1884 | Available as a function so that subclasses can extend hash mismatch |
|
1884 | Available as a function so that subclasses can extend hash mismatch | |
1885 | behaviors as needed. |
|
1885 | behaviors as needed. | |
1886 | """ |
|
1886 | """ | |
1887 | try: |
|
1887 | try: | |
1888 | if p1 is None and p2 is None: |
|
1888 | if p1 is None and p2 is None: | |
1889 | p1, p2 = self.parents(node) |
|
1889 | p1, p2 = self.parents(node) | |
1890 | if node != self.hash(text, p1, p2): |
|
1890 | if node != self.hash(text, p1, p2): | |
1891 | # Clear the revision cache on hash failure. The revision cache |
|
1891 | # Clear the revision cache on hash failure. The revision cache | |
1892 | # only stores the raw revision and clearing the cache does have |
|
1892 | # only stores the raw revision and clearing the cache does have | |
1893 | # the side-effect that we won't have a cache hit when the raw |
|
1893 | # the side-effect that we won't have a cache hit when the raw | |
1894 | # revision data is accessed. But this case should be rare and |
|
1894 | # revision data is accessed. But this case should be rare and | |
1895 | # it is extra work to teach the cache about the hash |
|
1895 | # it is extra work to teach the cache about the hash | |
1896 | # verification state. |
|
1896 | # verification state. | |
1897 | if self._revisioncache and self._revisioncache[0] == node: |
|
1897 | if self._revisioncache and self._revisioncache[0] == node: | |
1898 | self._revisioncache = None |
|
1898 | self._revisioncache = None | |
1899 |
|
1899 | |||
1900 | revornode = rev |
|
1900 | revornode = rev | |
1901 | if revornode is None: |
|
1901 | if revornode is None: | |
1902 | revornode = templatefilters.short(hex(node)) |
|
1902 | revornode = templatefilters.short(hex(node)) | |
1903 | raise error.RevlogError( |
|
1903 | raise error.RevlogError( | |
1904 | _(b"integrity check failed on %s:%s") |
|
1904 | _(b"integrity check failed on %s:%s") | |
1905 | % (self.indexfile, pycompat.bytestr(revornode)) |
|
1905 | % (self.indexfile, pycompat.bytestr(revornode)) | |
1906 | ) |
|
1906 | ) | |
1907 | except error.RevlogError: |
|
1907 | except error.RevlogError: | |
1908 | if self._censorable and storageutil.iscensoredtext(text): |
|
1908 | if self._censorable and storageutil.iscensoredtext(text): | |
1909 | raise error.CensoredNodeError(self.indexfile, node, text) |
|
1909 | raise error.CensoredNodeError(self.indexfile, node, text) | |
1910 | raise |
|
1910 | raise | |
1911 |
|
1911 | |||
1912 | def _enforceinlinesize(self, tr, fp=None): |
|
1912 | def _enforceinlinesize(self, tr, fp=None): | |
1913 | """Check if the revlog is too big for inline and convert if so. |
|
1913 | """Check if the revlog is too big for inline and convert if so. | |
1914 |
|
1914 | |||
1915 | This should be called after revisions are added to the revlog. If the |
|
1915 | This should be called after revisions are added to the revlog. If the | |
1916 | revlog has grown too large to be an inline revlog, it will convert it |
|
1916 | revlog has grown too large to be an inline revlog, it will convert it | |
1917 | to use multiple index and data files. |
|
1917 | to use multiple index and data files. | |
1918 | """ |
|
1918 | """ | |
1919 | tiprev = len(self) - 1 |
|
1919 | tiprev = len(self) - 1 | |
1920 | if ( |
|
1920 | if ( | |
1921 | not self._inline |
|
1921 | not self._inline | |
1922 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline |
|
1922 | or (self.start(tiprev) + self.length(tiprev)) < _maxinline | |
1923 | ): |
|
1923 | ): | |
1924 | return |
|
1924 | return | |
1925 |
|
1925 | |||
1926 | troffset = tr.findoffset(self.indexfile) |
|
1926 | troffset = tr.findoffset(self.indexfile) | |
1927 | if troffset is None: |
|
1927 | if troffset is None: | |
1928 | raise error.RevlogError( |
|
1928 | raise error.RevlogError( | |
1929 | _(b"%s not found in the transaction") % self.indexfile |
|
1929 | _(b"%s not found in the transaction") % self.indexfile | |
1930 | ) |
|
1930 | ) | |
1931 | trindex = 0 |
|
1931 | trindex = 0 | |
1932 | tr.add(self.datafile, 0) |
|
1932 | tr.add(self.datafile, 0) | |
1933 |
|
1933 | |||
1934 | if fp: |
|
1934 | if fp: | |
1935 | fp.flush() |
|
1935 | fp.flush() | |
1936 | fp.close() |
|
1936 | fp.close() | |
1937 | # We can't use the cached file handle after close(). So prevent |
|
1937 | # We can't use the cached file handle after close(). So prevent | |
1938 | # its usage. |
|
1938 | # its usage. | |
1939 | self._writinghandles = None |
|
1939 | self._writinghandles = None | |
1940 |
|
1940 | |||
1941 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: |
|
1941 | with self._indexfp(b'r') as ifh, self._datafp(b'w') as dfh: | |
1942 | for r in self: |
|
1942 | for r in self: | |
1943 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) |
|
1943 | dfh.write(self._getsegmentforrevs(r, r, df=ifh)[1]) | |
1944 | if troffset <= self.start(r): |
|
1944 | if troffset <= self.start(r): | |
1945 | trindex = r |
|
1945 | trindex = r | |
1946 |
|
1946 | |||
1947 | with self._indexfp(b'w') as fp: |
|
1947 | with self._indexfp(b'w') as fp: | |
1948 | self.version &= ~FLAG_INLINE_DATA |
|
1948 | self.version &= ~FLAG_INLINE_DATA | |
1949 | self._inline = False |
|
1949 | self._inline = False | |
1950 | for i in self: |
|
1950 | for i in self: | |
1951 | e = self.index.entry_binary(i) |
|
1951 | e = self.index.entry_binary(i) | |
1952 | if i == 0: |
|
1952 | if i == 0: | |
1953 | header = self.index.pack_header(self.version) |
|
1953 | header = self.index.pack_header(self.version) | |
1954 | e = header + e |
|
1954 | e = header + e | |
1955 | fp.write(e) |
|
1955 | fp.write(e) | |
1956 |
|
1956 | |||
1957 | # the temp file replace the real index when we exit the context |
|
1957 | # the temp file replace the real index when we exit the context | |
1958 | # manager |
|
1958 | # manager | |
1959 |
|
1959 | |||
1960 | tr.replace(self.indexfile, trindex * self.index.entry_size) |
|
1960 | tr.replace(self.indexfile, trindex * self.index.entry_size) | |
1961 | nodemaputil.setup_persistent_nodemap(tr, self) |
|
1961 | nodemaputil.setup_persistent_nodemap(tr, self) | |
1962 | self._chunkclear() |
|
1962 | self._chunkclear() | |
1963 |
|
1963 | |||
1964 | def _nodeduplicatecallback(self, transaction, node): |
|
1964 | def _nodeduplicatecallback(self, transaction, node): | |
1965 | """called when trying to add a node already stored.""" |
|
1965 | """called when trying to add a node already stored.""" | |
1966 |
|
1966 | |||
1967 | def addrevision( |
|
1967 | def addrevision( | |
1968 | self, |
|
1968 | self, | |
1969 | text, |
|
1969 | text, | |
1970 | transaction, |
|
1970 | transaction, | |
1971 | link, |
|
1971 | link, | |
1972 | p1, |
|
1972 | p1, | |
1973 | p2, |
|
1973 | p2, | |
1974 | cachedelta=None, |
|
1974 | cachedelta=None, | |
1975 | node=None, |
|
1975 | node=None, | |
1976 | flags=REVIDX_DEFAULT_FLAGS, |
|
1976 | flags=REVIDX_DEFAULT_FLAGS, | |
1977 | deltacomputer=None, |
|
1977 | deltacomputer=None, | |
1978 | sidedata=None, |
|
1978 | sidedata=None, | |
1979 | ): |
|
1979 | ): | |
1980 | """add a revision to the log |
|
1980 | """add a revision to the log | |
1981 |
|
1981 | |||
1982 | text - the revision data to add |
|
1982 | text - the revision data to add | |
1983 | transaction - the transaction object used for rollback |
|
1983 | transaction - the transaction object used for rollback | |
1984 | link - the linkrev data to add |
|
1984 | link - the linkrev data to add | |
1985 | p1, p2 - the parent nodeids of the revision |
|
1985 | p1, p2 - the parent nodeids of the revision | |
1986 | cachedelta - an optional precomputed delta |
|
1986 | cachedelta - an optional precomputed delta | |
1987 | node - nodeid of revision; typically node is not specified, and it is |
|
1987 | node - nodeid of revision; typically node is not specified, and it is | |
1988 | computed by default as hash(text, p1, p2), however subclasses might |
|
1988 | computed by default as hash(text, p1, p2), however subclasses might | |
1989 | use different hashing method (and override checkhash() in such case) |
|
1989 | use different hashing method (and override checkhash() in such case) | |
1990 | flags - the known flags to set on the revision |
|
1990 | flags - the known flags to set on the revision | |
1991 | deltacomputer - an optional deltacomputer instance shared between |
|
1991 | deltacomputer - an optional deltacomputer instance shared between | |
1992 | multiple calls |
|
1992 | multiple calls | |
1993 | """ |
|
1993 | """ | |
1994 | if link == nullrev: |
|
1994 | if link == nullrev: | |
1995 | raise error.RevlogError( |
|
1995 | raise error.RevlogError( | |
1996 | _(b"attempted to add linkrev -1 to %s") % self.indexfile |
|
1996 | _(b"attempted to add linkrev -1 to %s") % self.indexfile | |
1997 | ) |
|
1997 | ) | |
1998 |
|
1998 | |||
1999 | if sidedata is None: |
|
1999 | if sidedata is None: | |
2000 | sidedata = {} |
|
2000 | sidedata = {} | |
2001 | elif sidedata and not self.hassidedata: |
|
2001 | elif sidedata and not self.hassidedata: | |
2002 | raise error.ProgrammingError( |
|
2002 | raise error.ProgrammingError( | |
2003 | _(b"trying to add sidedata to a revlog who don't support them") |
|
2003 | _(b"trying to add sidedata to a revlog who don't support them") | |
2004 | ) |
|
2004 | ) | |
2005 |
|
2005 | |||
2006 | if flags: |
|
2006 | if flags: | |
2007 | node = node or self.hash(text, p1, p2) |
|
2007 | node = node or self.hash(text, p1, p2) | |
2008 |
|
2008 | |||
2009 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) |
|
2009 | rawtext, validatehash = flagutil.processflagswrite(self, text, flags) | |
2010 |
|
2010 | |||
2011 | # If the flag processor modifies the revision data, ignore any provided |
|
2011 | # If the flag processor modifies the revision data, ignore any provided | |
2012 | # cachedelta. |
|
2012 | # cachedelta. | |
2013 | if rawtext != text: |
|
2013 | if rawtext != text: | |
2014 | cachedelta = None |
|
2014 | cachedelta = None | |
2015 |
|
2015 | |||
2016 | if len(rawtext) > _maxentrysize: |
|
2016 | if len(rawtext) > _maxentrysize: | |
2017 | raise error.RevlogError( |
|
2017 | raise error.RevlogError( | |
2018 | _( |
|
2018 | _( | |
2019 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" |
|
2019 | b"%s: size of %d bytes exceeds maximum revlog storage of 2GiB" | |
2020 | ) |
|
2020 | ) | |
2021 | % (self.indexfile, len(rawtext)) |
|
2021 | % (self.indexfile, len(rawtext)) | |
2022 | ) |
|
2022 | ) | |
2023 |
|
2023 | |||
2024 | node = node or self.hash(rawtext, p1, p2) |
|
2024 | node = node or self.hash(rawtext, p1, p2) | |
2025 | rev = self.index.get_rev(node) |
|
2025 | rev = self.index.get_rev(node) | |
2026 | if rev is not None: |
|
2026 | if rev is not None: | |
2027 | return rev |
|
2027 | return rev | |
2028 |
|
2028 | |||
2029 | if validatehash: |
|
2029 | if validatehash: | |
2030 | self.checkhash(rawtext, node, p1=p1, p2=p2) |
|
2030 | self.checkhash(rawtext, node, p1=p1, p2=p2) | |
2031 |
|
2031 | |||
2032 | return self.addrawrevision( |
|
2032 | return self.addrawrevision( | |
2033 | rawtext, |
|
2033 | rawtext, | |
2034 | transaction, |
|
2034 | transaction, | |
2035 | link, |
|
2035 | link, | |
2036 | p1, |
|
2036 | p1, | |
2037 | p2, |
|
2037 | p2, | |
2038 | node, |
|
2038 | node, | |
2039 | flags, |
|
2039 | flags, | |
2040 | cachedelta=cachedelta, |
|
2040 | cachedelta=cachedelta, | |
2041 | deltacomputer=deltacomputer, |
|
2041 | deltacomputer=deltacomputer, | |
2042 | sidedata=sidedata, |
|
2042 | sidedata=sidedata, | |
2043 | ) |
|
2043 | ) | |
2044 |
|
2044 | |||
2045 | def addrawrevision( |
|
2045 | def addrawrevision( | |
2046 | self, |
|
2046 | self, | |
2047 | rawtext, |
|
2047 | rawtext, | |
2048 | transaction, |
|
2048 | transaction, | |
2049 | link, |
|
2049 | link, | |
2050 | p1, |
|
2050 | p1, | |
2051 | p2, |
|
2051 | p2, | |
2052 | node, |
|
2052 | node, | |
2053 | flags, |
|
2053 | flags, | |
2054 | cachedelta=None, |
|
2054 | cachedelta=None, | |
2055 | deltacomputer=None, |
|
2055 | deltacomputer=None, | |
2056 | sidedata=None, |
|
2056 | sidedata=None, | |
2057 | ): |
|
2057 | ): | |
2058 | """add a raw revision with known flags, node and parents |
|
2058 | """add a raw revision with known flags, node and parents | |
2059 | useful when reusing a revision not stored in this revlog (ex: received |
|
2059 | useful when reusing a revision not stored in this revlog (ex: received | |
2060 | over wire, or read from an external bundle). |
|
2060 | over wire, or read from an external bundle). | |
2061 | """ |
|
2061 | """ | |
2062 | dfh = None |
|
2062 | dfh = None | |
2063 | if not self._inline: |
|
2063 | if not self._inline: | |
2064 | dfh = self._datafp(b"a+") |
|
2064 | dfh = self._datafp(b"a+") | |
2065 | ifh = self._indexfp(b"a+") |
|
2065 | ifh = self._indexfp(b"a+") | |
2066 | try: |
|
2066 | try: | |
2067 | return self._addrevision( |
|
2067 | return self._addrevision( | |
2068 | node, |
|
2068 | node, | |
2069 | rawtext, |
|
2069 | rawtext, | |
2070 | transaction, |
|
2070 | transaction, | |
2071 | link, |
|
2071 | link, | |
2072 | p1, |
|
2072 | p1, | |
2073 | p2, |
|
2073 | p2, | |
2074 | flags, |
|
2074 | flags, | |
2075 | cachedelta, |
|
2075 | cachedelta, | |
2076 | ifh, |
|
2076 | ifh, | |
2077 | dfh, |
|
2077 | dfh, | |
2078 | deltacomputer=deltacomputer, |
|
2078 | deltacomputer=deltacomputer, | |
2079 | sidedata=sidedata, |
|
2079 | sidedata=sidedata, | |
2080 | ) |
|
2080 | ) | |
2081 | finally: |
|
2081 | finally: | |
2082 | if dfh: |
|
2082 | if dfh: | |
2083 | dfh.close() |
|
2083 | dfh.close() | |
2084 | ifh.close() |
|
2084 | ifh.close() | |
2085 |
|
2085 | |||
2086 | def compress(self, data): |
|
2086 | def compress(self, data): | |
2087 | """Generate a possibly-compressed representation of data.""" |
|
2087 | """Generate a possibly-compressed representation of data.""" | |
2088 | if not data: |
|
2088 | if not data: | |
2089 | return b'', data |
|
2089 | return b'', data | |
2090 |
|
2090 | |||
2091 | compressed = self._compressor.compress(data) |
|
2091 | compressed = self._compressor.compress(data) | |
2092 |
|
2092 | |||
2093 | if compressed: |
|
2093 | if compressed: | |
2094 | # The revlog compressor added the header in the returned data. |
|
2094 | # The revlog compressor added the header in the returned data. | |
2095 | return b'', compressed |
|
2095 | return b'', compressed | |
2096 |
|
2096 | |||
2097 | if data[0:1] == b'\0': |
|
2097 | if data[0:1] == b'\0': | |
2098 | return b'', data |
|
2098 | return b'', data | |
2099 | return b'u', data |
|
2099 | return b'u', data | |
2100 |
|
2100 | |||
2101 | def decompress(self, data): |
|
2101 | def decompress(self, data): | |
2102 | """Decompress a revlog chunk. |
|
2102 | """Decompress a revlog chunk. | |
2103 |
|
2103 | |||
2104 | The chunk is expected to begin with a header identifying the |
|
2104 | The chunk is expected to begin with a header identifying the | |
2105 | format type so it can be routed to an appropriate decompressor. |
|
2105 | format type so it can be routed to an appropriate decompressor. | |
2106 | """ |
|
2106 | """ | |
2107 | if not data: |
|
2107 | if not data: | |
2108 | return data |
|
2108 | return data | |
2109 |
|
2109 | |||
2110 | # Revlogs are read much more frequently than they are written and many |
|
2110 | # Revlogs are read much more frequently than they are written and many | |
2111 | # chunks only take microseconds to decompress, so performance is |
|
2111 | # chunks only take microseconds to decompress, so performance is | |
2112 | # important here. |
|
2112 | # important here. | |
2113 | # |
|
2113 | # | |
2114 | # We can make a few assumptions about revlogs: |
|
2114 | # We can make a few assumptions about revlogs: | |
2115 | # |
|
2115 | # | |
2116 | # 1) the majority of chunks will be compressed (as opposed to inline |
|
2116 | # 1) the majority of chunks will be compressed (as opposed to inline | |
2117 | # raw data). |
|
2117 | # raw data). | |
2118 | # 2) decompressing *any* data will likely by at least 10x slower than |
|
2118 | # 2) decompressing *any* data will likely by at least 10x slower than | |
2119 | # returning raw inline data. |
|
2119 | # returning raw inline data. | |
2120 | # 3) we want to prioritize common and officially supported compression |
|
2120 | # 3) we want to prioritize common and officially supported compression | |
2121 | # engines |
|
2121 | # engines | |
2122 | # |
|
2122 | # | |
2123 | # It follows that we want to optimize for "decompress compressed data |
|
2123 | # It follows that we want to optimize for "decompress compressed data | |
2124 | # when encoded with common and officially supported compression engines" |
|
2124 | # when encoded with common and officially supported compression engines" | |
2125 | # case over "raw data" and "data encoded by less common or non-official |
|
2125 | # case over "raw data" and "data encoded by less common or non-official | |
2126 | # compression engines." That is why we have the inline lookup first |
|
2126 | # compression engines." That is why we have the inline lookup first | |
2127 | # followed by the compengines lookup. |
|
2127 | # followed by the compengines lookup. | |
2128 | # |
|
2128 | # | |
2129 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib |
|
2129 | # According to `hg perfrevlogchunks`, this is ~0.5% faster for zlib | |
2130 | # compressed chunks. And this matters for changelog and manifest reads. |
|
2130 | # compressed chunks. And this matters for changelog and manifest reads. | |
2131 | t = data[0:1] |
|
2131 | t = data[0:1] | |
2132 |
|
2132 | |||
2133 | if t == b'x': |
|
2133 | if t == b'x': | |
2134 | try: |
|
2134 | try: | |
2135 | return _zlibdecompress(data) |
|
2135 | return _zlibdecompress(data) | |
2136 | except zlib.error as e: |
|
2136 | except zlib.error as e: | |
2137 | raise error.RevlogError( |
|
2137 | raise error.RevlogError( | |
2138 | _(b'revlog decompress error: %s') |
|
2138 | _(b'revlog decompress error: %s') | |
2139 | % stringutil.forcebytestr(e) |
|
2139 | % stringutil.forcebytestr(e) | |
2140 | ) |
|
2140 | ) | |
2141 | # '\0' is more common than 'u' so it goes first. |
|
2141 | # '\0' is more common than 'u' so it goes first. | |
2142 | elif t == b'\0': |
|
2142 | elif t == b'\0': | |
2143 | return data |
|
2143 | return data | |
2144 | elif t == b'u': |
|
2144 | elif t == b'u': | |
2145 | return util.buffer(data, 1) |
|
2145 | return util.buffer(data, 1) | |
2146 |
|
2146 | |||
2147 | try: |
|
2147 | try: | |
2148 | compressor = self._decompressors[t] |
|
2148 | compressor = self._decompressors[t] | |
2149 | except KeyError: |
|
2149 | except KeyError: | |
2150 | try: |
|
2150 | try: | |
2151 | engine = util.compengines.forrevlogheader(t) |
|
2151 | engine = util.compengines.forrevlogheader(t) | |
2152 | compressor = engine.revlogcompressor(self._compengineopts) |
|
2152 | compressor = engine.revlogcompressor(self._compengineopts) | |
2153 | self._decompressors[t] = compressor |
|
2153 | self._decompressors[t] = compressor | |
2154 | except KeyError: |
|
2154 | except KeyError: | |
2155 | raise error.RevlogError( |
|
2155 | raise error.RevlogError( | |
2156 | _(b'unknown compression type %s') % binascii.hexlify(t) |
|
2156 | _(b'unknown compression type %s') % binascii.hexlify(t) | |
2157 | ) |
|
2157 | ) | |
2158 |
|
2158 | |||
2159 | return compressor.decompress(data) |
|
2159 | return compressor.decompress(data) | |
2160 |
|
2160 | |||
2161 | def _addrevision( |
|
2161 | def _addrevision( | |
2162 | self, |
|
2162 | self, | |
2163 | node, |
|
2163 | node, | |
2164 | rawtext, |
|
2164 | rawtext, | |
2165 | transaction, |
|
2165 | transaction, | |
2166 | link, |
|
2166 | link, | |
2167 | p1, |
|
2167 | p1, | |
2168 | p2, |
|
2168 | p2, | |
2169 | flags, |
|
2169 | flags, | |
2170 | cachedelta, |
|
2170 | cachedelta, | |
2171 | ifh, |
|
2171 | ifh, | |
2172 | dfh, |
|
2172 | dfh, | |
2173 | alwayscache=False, |
|
2173 | alwayscache=False, | |
2174 | deltacomputer=None, |
|
2174 | deltacomputer=None, | |
2175 | sidedata=None, |
|
2175 | sidedata=None, | |
2176 | ): |
|
2176 | ): | |
2177 | """internal function to add revisions to the log |
|
2177 | """internal function to add revisions to the log | |
2178 |
|
2178 | |||
2179 | see addrevision for argument descriptions. |
|
2179 | see addrevision for argument descriptions. | |
2180 |
|
2180 | |||
2181 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. |
|
2181 | note: "addrevision" takes non-raw text, "_addrevision" takes raw text. | |
2182 |
|
2182 | |||
2183 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will |
|
2183 | if "deltacomputer" is not provided or None, a defaultdeltacomputer will | |
2184 | be used. |
|
2184 | be used. | |
2185 |
|
2185 | |||
2186 | invariants: |
|
2186 | invariants: | |
2187 | - rawtext is optional (can be None); if not set, cachedelta must be set. |
|
2187 | - rawtext is optional (can be None); if not set, cachedelta must be set. | |
2188 | if both are set, they must correspond to each other. |
|
2188 | if both are set, they must correspond to each other. | |
2189 | """ |
|
2189 | """ | |
2190 | if node == self.nullid: |
|
2190 | if node == self.nullid: | |
2191 | raise error.RevlogError( |
|
2191 | raise error.RevlogError( | |
2192 | _(b"%s: attempt to add null revision") % self.indexfile |
|
2192 | _(b"%s: attempt to add null revision") % self.indexfile | |
2193 | ) |
|
2193 | ) | |
2194 | if ( |
|
2194 | if ( | |
2195 | node == self.nodeconstants.wdirid |
|
2195 | node == self.nodeconstants.wdirid | |
2196 | or node in self.nodeconstants.wdirfilenodeids |
|
2196 | or node in self.nodeconstants.wdirfilenodeids | |
2197 | ): |
|
2197 | ): | |
2198 | raise error.RevlogError( |
|
2198 | raise error.RevlogError( | |
2199 | _(b"%s: attempt to add wdir revision") % self.indexfile |
|
2199 | _(b"%s: attempt to add wdir revision") % self.indexfile | |
2200 | ) |
|
2200 | ) | |
2201 |
|
2201 | |||
2202 | if self._inline: |
|
2202 | if self._inline: | |
2203 | fh = ifh |
|
2203 | fh = ifh | |
2204 | else: |
|
2204 | else: | |
2205 | fh = dfh |
|
2205 | fh = dfh | |
2206 |
|
2206 | |||
2207 | btext = [rawtext] |
|
2207 | btext = [rawtext] | |
2208 |
|
2208 | |||
2209 | curr = len(self) |
|
2209 | curr = len(self) | |
2210 | prev = curr - 1 |
|
2210 | prev = curr - 1 | |
2211 |
|
2211 | |||
2212 | offset = self._get_data_offset(prev) |
|
2212 | offset = self._get_data_offset(prev) | |
2213 |
|
2213 | |||
2214 | if self._concurrencychecker: |
|
2214 | if self._concurrencychecker: | |
2215 | if self._inline: |
|
2215 | if self._inline: | |
2216 | # offset is "as if" it were in the .d file, so we need to add on |
|
2216 | # offset is "as if" it were in the .d file, so we need to add on | |
2217 | # the size of the entry metadata. |
|
2217 | # the size of the entry metadata. | |
2218 | self._concurrencychecker( |
|
2218 | self._concurrencychecker( | |
2219 | ifh, self.indexfile, offset + curr * self.index.entry_size |
|
2219 | ifh, self.indexfile, offset + curr * self.index.entry_size | |
2220 | ) |
|
2220 | ) | |
2221 | else: |
|
2221 | else: | |
2222 | # Entries in the .i are a consistent size. |
|
2222 | # Entries in the .i are a consistent size. | |
2223 | self._concurrencychecker( |
|
2223 | self._concurrencychecker( | |
2224 | ifh, self.indexfile, curr * self.index.entry_size |
|
2224 | ifh, self.indexfile, curr * self.index.entry_size | |
2225 | ) |
|
2225 | ) | |
2226 | self._concurrencychecker(dfh, self.datafile, offset) |
|
2226 | self._concurrencychecker(dfh, self.datafile, offset) | |
2227 |
|
2227 | |||
2228 | p1r, p2r = self.rev(p1), self.rev(p2) |
|
2228 | p1r, p2r = self.rev(p1), self.rev(p2) | |
2229 |
|
2229 | |||
2230 | # full versions are inserted when the needed deltas |
|
2230 | # full versions are inserted when the needed deltas | |
2231 | # become comparable to the uncompressed text |
|
2231 | # become comparable to the uncompressed text | |
2232 | if rawtext is None: |
|
2232 | if rawtext is None: | |
2233 | # need rawtext size, before changed by flag processors, which is |
|
2233 | # need rawtext size, before changed by flag processors, which is | |
2234 | # the non-raw size. use revlog explicitly to avoid filelog's extra |
|
2234 | # the non-raw size. use revlog explicitly to avoid filelog's extra | |
2235 | # logic that might remove metadata size. |
|
2235 | # logic that might remove metadata size. | |
2236 | textlen = mdiff.patchedsize( |
|
2236 | textlen = mdiff.patchedsize( | |
2237 | revlog.size(self, cachedelta[0]), cachedelta[1] |
|
2237 | revlog.size(self, cachedelta[0]), cachedelta[1] | |
2238 | ) |
|
2238 | ) | |
2239 | else: |
|
2239 | else: | |
2240 | textlen = len(rawtext) |
|
2240 | textlen = len(rawtext) | |
2241 |
|
2241 | |||
2242 | if deltacomputer is None: |
|
2242 | if deltacomputer is None: | |
2243 | deltacomputer = deltautil.deltacomputer(self) |
|
2243 | deltacomputer = deltautil.deltacomputer(self) | |
2244 |
|
2244 | |||
2245 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) |
|
2245 | revinfo = _revisioninfo(node, p1, p2, btext, textlen, cachedelta, flags) | |
2246 |
|
2246 | |||
2247 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) |
|
2247 | deltainfo = deltacomputer.finddeltainfo(revinfo, fh) | |
2248 |
|
2248 | |||
2249 | if sidedata and self.version & 0xFFFF == REVLOGV2: |
|
2249 | if sidedata and self.version & 0xFFFF == REVLOGV2: | |
2250 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) |
|
2250 | serialized_sidedata = sidedatautil.serialize_sidedata(sidedata) | |
2251 | sidedata_offset = offset + deltainfo.deltalen |
|
2251 | sidedata_offset = offset + deltainfo.deltalen | |
2252 | else: |
|
2252 | else: | |
2253 | serialized_sidedata = b"" |
|
2253 | serialized_sidedata = b"" | |
2254 | # Don't store the offset if the sidedata is empty, that way |
|
2254 | # Don't store the offset if the sidedata is empty, that way | |
2255 | # we can easily detect empty sidedata and they will be no different |
|
2255 | # we can easily detect empty sidedata and they will be no different | |
2256 | # than ones we manually add. |
|
2256 | # than ones we manually add. | |
2257 | sidedata_offset = 0 |
|
2257 | sidedata_offset = 0 | |
2258 |
|
2258 | |||
2259 | e = ( |
|
2259 | e = ( | |
2260 | offset_type(offset, flags), |
|
2260 | offset_type(offset, flags), | |
2261 | deltainfo.deltalen, |
|
2261 | deltainfo.deltalen, | |
2262 | textlen, |
|
2262 | textlen, | |
2263 | deltainfo.base, |
|
2263 | deltainfo.base, | |
2264 | link, |
|
2264 | link, | |
2265 | p1r, |
|
2265 | p1r, | |
2266 | p2r, |
|
2266 | p2r, | |
2267 | node, |
|
2267 | node, | |
2268 | sidedata_offset, |
|
2268 | sidedata_offset, | |
2269 | len(serialized_sidedata), |
|
2269 | len(serialized_sidedata), | |
2270 | ) |
|
2270 | ) | |
2271 |
|
2271 | |||
2272 | if self.version & 0xFFFF != REVLOGV2: |
|
2272 | if self.version & 0xFFFF != REVLOGV2: | |
2273 | e = e[:8] |
|
2273 | e = e[:8] | |
2274 |
|
2274 | |||
2275 | self.index.append(e) |
|
2275 | self.index.append(e) | |
2276 | entry = self.index.entry_binary(curr) |
|
2276 | entry = self.index.entry_binary(curr) | |
2277 | if curr == 0: |
|
2277 | if curr == 0: | |
2278 | header = self.index.pack_header(self.version) |
|
2278 | header = self.index.pack_header(self.version) | |
2279 | entry = header + entry |
|
2279 | entry = header + entry | |
2280 | self._writeentry( |
|
2280 | self._writeentry( | |
2281 | transaction, |
|
2281 | transaction, | |
2282 | ifh, |
|
2282 | ifh, | |
2283 | dfh, |
|
2283 | dfh, | |
2284 | entry, |
|
2284 | entry, | |
2285 | deltainfo.data, |
|
2285 | deltainfo.data, | |
2286 | link, |
|
2286 | link, | |
2287 | offset, |
|
2287 | offset, | |
2288 | serialized_sidedata, |
|
2288 | serialized_sidedata, | |
2289 | ) |
|
2289 | ) | |
2290 |
|
2290 | |||
2291 | rawtext = btext[0] |
|
2291 | rawtext = btext[0] | |
2292 |
|
2292 | |||
2293 | if alwayscache and rawtext is None: |
|
2293 | if alwayscache and rawtext is None: | |
2294 | rawtext = deltacomputer.buildtext(revinfo, fh) |
|
2294 | rawtext = deltacomputer.buildtext(revinfo, fh) | |
2295 |
|
2295 | |||
2296 | if type(rawtext) == bytes: # only accept immutable objects |
|
2296 | if type(rawtext) == bytes: # only accept immutable objects | |
2297 | self._revisioncache = (node, curr, rawtext) |
|
2297 | self._revisioncache = (node, curr, rawtext) | |
2298 | self._chainbasecache[curr] = deltainfo.chainbase |
|
2298 | self._chainbasecache[curr] = deltainfo.chainbase | |
2299 | return curr |
|
2299 | return curr | |
2300 |
|
2300 | |||
2301 | def _get_data_offset(self, prev): |
|
2301 | def _get_data_offset(self, prev): | |
2302 | """Returns the current offset in the (in-transaction) data file. |
|
2302 | """Returns the current offset in the (in-transaction) data file. | |
2303 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket |
|
2303 | Versions < 2 of the revlog can get this 0(1), revlog v2 needs a docket | |
2304 | file to store that information: since sidedata can be rewritten to the |
|
2304 | file to store that information: since sidedata can be rewritten to the | |
2305 | end of the data file within a transaction, you can have cases where, for |
|
2305 | end of the data file within a transaction, you can have cases where, for | |
2306 | example, rev `n` does not have sidedata while rev `n - 1` does, leading |
|
2306 | example, rev `n` does not have sidedata while rev `n - 1` does, leading | |
2307 | to `n - 1`'s sidedata being written after `n`'s data. |
|
2307 | to `n - 1`'s sidedata being written after `n`'s data. | |
2308 |
|
2308 | |||
2309 | TODO cache this in a docket file before getting out of experimental.""" |
|
2309 | TODO cache this in a docket file before getting out of experimental.""" | |
2310 | if self.version & 0xFFFF != REVLOGV2: |
|
2310 | if self.version & 0xFFFF != REVLOGV2: | |
2311 | return self.end(prev) |
|
2311 | return self.end(prev) | |
2312 |
|
2312 | |||
2313 | offset = 0 |
|
2313 | offset = 0 | |
2314 | for rev, entry in enumerate(self.index): |
|
2314 | for rev, entry in enumerate(self.index): | |
2315 | sidedata_end = entry[8] + entry[9] |
|
2315 | sidedata_end = entry[8] + entry[9] | |
2316 | # Sidedata for a previous rev has potentially been written after |
|
2316 | # Sidedata for a previous rev has potentially been written after | |
2317 | # this rev's end, so take the max. |
|
2317 | # this rev's end, so take the max. | |
2318 | offset = max(self.end(rev), offset, sidedata_end) |
|
2318 | offset = max(self.end(rev), offset, sidedata_end) | |
2319 | return offset |
|
2319 | return offset | |
2320 |
|
2320 | |||
2321 | def _writeentry( |
|
2321 | def _writeentry( | |
2322 | self, transaction, ifh, dfh, entry, data, link, offset, sidedata |
|
2322 | self, transaction, ifh, dfh, entry, data, link, offset, sidedata | |
2323 | ): |
|
2323 | ): | |
2324 | # Files opened in a+ mode have inconsistent behavior on various |
|
2324 | # Files opened in a+ mode have inconsistent behavior on various | |
2325 | # platforms. Windows requires that a file positioning call be made |
|
2325 | # platforms. Windows requires that a file positioning call be made | |
2326 | # when the file handle transitions between reads and writes. See |
|
2326 | # when the file handle transitions between reads and writes. See | |
2327 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other |
|
2327 | # 3686fa2b8eee and the mixedfilemodewrapper in windows.py. On other | |
2328 | # platforms, Python or the platform itself can be buggy. Some versions |
|
2328 | # platforms, Python or the platform itself can be buggy. Some versions | |
2329 | # of Solaris have been observed to not append at the end of the file |
|
2329 | # of Solaris have been observed to not append at the end of the file | |
2330 | # if the file was seeked to before the end. See issue4943 for more. |
|
2330 | # if the file was seeked to before the end. See issue4943 for more. | |
2331 | # |
|
2331 | # | |
2332 | # We work around this issue by inserting a seek() before writing. |
|
2332 | # We work around this issue by inserting a seek() before writing. | |
2333 | # Note: This is likely not necessary on Python 3. However, because |
|
2333 | # Note: This is likely not necessary on Python 3. However, because | |
2334 | # the file handle is reused for reads and may be seeked there, we need |
|
2334 | # the file handle is reused for reads and may be seeked there, we need | |
2335 | # to be careful before changing this. |
|
2335 | # to be careful before changing this. | |
2336 | ifh.seek(0, os.SEEK_END) |
|
2336 | ifh.seek(0, os.SEEK_END) | |
2337 | if dfh: |
|
2337 | if dfh: | |
2338 | dfh.seek(0, os.SEEK_END) |
|
2338 | dfh.seek(0, os.SEEK_END) | |
2339 |
|
2339 | |||
2340 | curr = len(self) - 1 |
|
2340 | curr = len(self) - 1 | |
2341 | if not self._inline: |
|
2341 | if not self._inline: | |
2342 | transaction.add(self.datafile, offset) |
|
2342 | transaction.add(self.datafile, offset) | |
2343 | transaction.add(self.indexfile, curr * len(entry)) |
|
2343 | transaction.add(self.indexfile, curr * len(entry)) | |
2344 | if data[0]: |
|
2344 | if data[0]: | |
2345 | dfh.write(data[0]) |
|
2345 | dfh.write(data[0]) | |
2346 | dfh.write(data[1]) |
|
2346 | dfh.write(data[1]) | |
2347 | if sidedata: |
|
2347 | if sidedata: | |
2348 | dfh.write(sidedata) |
|
2348 | dfh.write(sidedata) | |
2349 | ifh.write(entry) |
|
2349 | ifh.write(entry) | |
2350 | else: |
|
2350 | else: | |
2351 | offset += curr * self.index.entry_size |
|
2351 | offset += curr * self.index.entry_size | |
2352 | transaction.add(self.indexfile, offset) |
|
2352 | transaction.add(self.indexfile, offset) | |
2353 | ifh.write(entry) |
|
2353 | ifh.write(entry) | |
2354 | ifh.write(data[0]) |
|
2354 | ifh.write(data[0]) | |
2355 | ifh.write(data[1]) |
|
2355 | ifh.write(data[1]) | |
2356 | if sidedata: |
|
2356 | if sidedata: | |
2357 | ifh.write(sidedata) |
|
2357 | ifh.write(sidedata) | |
2358 | self._enforceinlinesize(transaction, ifh) |
|
2358 | self._enforceinlinesize(transaction, ifh) | |
2359 | nodemaputil.setup_persistent_nodemap(transaction, self) |
|
2359 | nodemaputil.setup_persistent_nodemap(transaction, self) | |
2360 |
|
2360 | |||
2361 | def addgroup( |
|
2361 | def addgroup( | |
2362 | self, |
|
2362 | self, | |
2363 | deltas, |
|
2363 | deltas, | |
2364 | linkmapper, |
|
2364 | linkmapper, | |
2365 | transaction, |
|
2365 | transaction, | |
2366 | alwayscache=False, |
|
2366 | alwayscache=False, | |
2367 | addrevisioncb=None, |
|
2367 | addrevisioncb=None, | |
2368 | duplicaterevisioncb=None, |
|
2368 | duplicaterevisioncb=None, | |
2369 | ): |
|
2369 | ): | |
2370 | """ |
|
2370 | """ | |
2371 | add a delta group |
|
2371 | add a delta group | |
2372 |
|
2372 | |||
2373 | given a set of deltas, add them to the revision log. the |
|
2373 | given a set of deltas, add them to the revision log. the | |
2374 | first delta is against its parent, which should be in our |
|
2374 | first delta is against its parent, which should be in our | |
2375 | log, the rest are against the previous delta. |
|
2375 | log, the rest are against the previous delta. | |
2376 |
|
2376 | |||
2377 | If ``addrevisioncb`` is defined, it will be called with arguments of |
|
2377 | If ``addrevisioncb`` is defined, it will be called with arguments of | |
2378 | this revlog and the node that was added. |
|
2378 | this revlog and the node that was added. | |
2379 | """ |
|
2379 | """ | |
2380 |
|
2380 | |||
2381 | if self._writinghandles: |
|
2381 | if self._writinghandles: | |
2382 | raise error.ProgrammingError(b'cannot nest addgroup() calls') |
|
2382 | raise error.ProgrammingError(b'cannot nest addgroup() calls') | |
2383 |
|
2383 | |||
2384 | r = len(self) |
|
2384 | r = len(self) | |
2385 | end = 0 |
|
2385 | end = 0 | |
2386 | if r: |
|
2386 | if r: | |
2387 | end = self.end(r - 1) |
|
2387 | end = self.end(r - 1) | |
2388 | ifh = self._indexfp(b"a+") |
|
2388 | ifh = self._indexfp(b"a+") | |
2389 | isize = r * self.index.entry_size |
|
2389 | isize = r * self.index.entry_size | |
2390 | if self._inline: |
|
2390 | if self._inline: | |
2391 | transaction.add(self.indexfile, end + isize) |
|
2391 | transaction.add(self.indexfile, end + isize) | |
2392 | dfh = None |
|
2392 | dfh = None | |
2393 | else: |
|
2393 | else: | |
2394 | transaction.add(self.indexfile, isize) |
|
2394 | transaction.add(self.indexfile, isize) | |
2395 | transaction.add(self.datafile, end) |
|
2395 | transaction.add(self.datafile, end) | |
2396 | dfh = self._datafp(b"a+") |
|
2396 | dfh = self._datafp(b"a+") | |
2397 |
|
2397 | |||
2398 | def flush(): |
|
2398 | def flush(): | |
2399 | if dfh: |
|
2399 | if dfh: | |
2400 | dfh.flush() |
|
2400 | dfh.flush() | |
2401 | ifh.flush() |
|
2401 | ifh.flush() | |
2402 |
|
2402 | |||
2403 | self._writinghandles = (ifh, dfh) |
|
2403 | self._writinghandles = (ifh, dfh) | |
2404 | empty = True |
|
2404 | empty = True | |
2405 |
|
2405 | |||
2406 | try: |
|
2406 | try: | |
2407 | deltacomputer = deltautil.deltacomputer(self) |
|
2407 | deltacomputer = deltautil.deltacomputer(self) | |
2408 | # loop through our set of deltas |
|
2408 | # loop through our set of deltas | |
2409 | for data in deltas: |
|
2409 | for data in deltas: | |
2410 | node, p1, p2, linknode, deltabase, delta, flags, sidedata = data |
|
2410 | node, p1, p2, linknode, deltabase, delta, flags, sidedata = data | |
2411 | link = linkmapper(linknode) |
|
2411 | link = linkmapper(linknode) | |
2412 | flags = flags or REVIDX_DEFAULT_FLAGS |
|
2412 | flags = flags or REVIDX_DEFAULT_FLAGS | |
2413 |
|
2413 | |||
2414 | rev = self.index.get_rev(node) |
|
2414 | rev = self.index.get_rev(node) | |
2415 | if rev is not None: |
|
2415 | if rev is not None: | |
2416 | # this can happen if two branches make the same change |
|
2416 | # this can happen if two branches make the same change | |
2417 | self._nodeduplicatecallback(transaction, rev) |
|
2417 | self._nodeduplicatecallback(transaction, rev) | |
2418 | if duplicaterevisioncb: |
|
2418 | if duplicaterevisioncb: | |
2419 | duplicaterevisioncb(self, rev) |
|
2419 | duplicaterevisioncb(self, rev) | |
2420 | empty = False |
|
2420 | empty = False | |
2421 | continue |
|
2421 | continue | |
2422 |
|
2422 | |||
2423 | for p in (p1, p2): |
|
2423 | for p in (p1, p2): | |
2424 | if not self.index.has_node(p): |
|
2424 | if not self.index.has_node(p): | |
2425 | raise error.LookupError( |
|
2425 | raise error.LookupError( | |
2426 | p, self.indexfile, _(b'unknown parent') |
|
2426 | p, self.indexfile, _(b'unknown parent') | |
2427 | ) |
|
2427 | ) | |
2428 |
|
2428 | |||
2429 | if not self.index.has_node(deltabase): |
|
2429 | if not self.index.has_node(deltabase): | |
2430 | raise error.LookupError( |
|
2430 | raise error.LookupError( | |
2431 | deltabase, self.indexfile, _(b'unknown delta base') |
|
2431 | deltabase, self.indexfile, _(b'unknown delta base') | |
2432 | ) |
|
2432 | ) | |
2433 |
|
2433 | |||
2434 | baserev = self.rev(deltabase) |
|
2434 | baserev = self.rev(deltabase) | |
2435 |
|
2435 | |||
2436 | if baserev != nullrev and self.iscensored(baserev): |
|
2436 | if baserev != nullrev and self.iscensored(baserev): | |
2437 | # if base is censored, delta must be full replacement in a |
|
2437 | # if base is censored, delta must be full replacement in a | |
2438 | # single patch operation |
|
2438 | # single patch operation | |
2439 | hlen = struct.calcsize(b">lll") |
|
2439 | hlen = struct.calcsize(b">lll") | |
2440 | oldlen = self.rawsize(baserev) |
|
2440 | oldlen = self.rawsize(baserev) | |
2441 | newlen = len(delta) - hlen |
|
2441 | newlen = len(delta) - hlen | |
2442 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
2442 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
2443 | raise error.CensoredBaseError( |
|
2443 | raise error.CensoredBaseError( | |
2444 | self.indexfile, self.node(baserev) |
|
2444 | self.indexfile, self.node(baserev) | |
2445 | ) |
|
2445 | ) | |
2446 |
|
2446 | |||
2447 | if not flags and self._peek_iscensored(baserev, delta, flush): |
|
2447 | if not flags and self._peek_iscensored(baserev, delta, flush): | |
2448 | flags |= REVIDX_ISCENSORED |
|
2448 | flags |= REVIDX_ISCENSORED | |
2449 |
|
2449 | |||
2450 | # We assume consumers of addrevisioncb will want to retrieve |
|
2450 | # We assume consumers of addrevisioncb will want to retrieve | |
2451 | # the added revision, which will require a call to |
|
2451 | # the added revision, which will require a call to | |
2452 | # revision(). revision() will fast path if there is a cache |
|
2452 | # revision(). revision() will fast path if there is a cache | |
2453 | # hit. So, we tell _addrevision() to always cache in this case. |
|
2453 | # hit. So, we tell _addrevision() to always cache in this case. | |
2454 | # We're only using addgroup() in the context of changegroup |
|
2454 | # We're only using addgroup() in the context of changegroup | |
2455 | # generation so the revision data can always be handled as raw |
|
2455 | # generation so the revision data can always be handled as raw | |
2456 | # by the flagprocessor. |
|
2456 | # by the flagprocessor. | |
2457 | rev = self._addrevision( |
|
2457 | rev = self._addrevision( | |
2458 | node, |
|
2458 | node, | |
2459 | None, |
|
2459 | None, | |
2460 | transaction, |
|
2460 | transaction, | |
2461 | link, |
|
2461 | link, | |
2462 | p1, |
|
2462 | p1, | |
2463 | p2, |
|
2463 | p2, | |
2464 | flags, |
|
2464 | flags, | |
2465 | (baserev, delta), |
|
2465 | (baserev, delta), | |
2466 | ifh, |
|
2466 | ifh, | |
2467 | dfh, |
|
2467 | dfh, | |
2468 | alwayscache=alwayscache, |
|
2468 | alwayscache=alwayscache, | |
2469 | deltacomputer=deltacomputer, |
|
2469 | deltacomputer=deltacomputer, | |
2470 | sidedata=sidedata, |
|
2470 | sidedata=sidedata, | |
2471 | ) |
|
2471 | ) | |
2472 |
|
2472 | |||
2473 | if addrevisioncb: |
|
2473 | if addrevisioncb: | |
2474 | addrevisioncb(self, rev) |
|
2474 | addrevisioncb(self, rev) | |
2475 | empty = False |
|
2475 | empty = False | |
2476 |
|
2476 | |||
2477 | if not dfh and not self._inline: |
|
2477 | if not dfh and not self._inline: | |
2478 | # addrevision switched from inline to conventional |
|
2478 | # addrevision switched from inline to conventional | |
2479 | # reopen the index |
|
2479 | # reopen the index | |
2480 | ifh.close() |
|
2480 | ifh.close() | |
2481 | dfh = self._datafp(b"a+") |
|
2481 | dfh = self._datafp(b"a+") | |
2482 | ifh = self._indexfp(b"a+") |
|
2482 | ifh = self._indexfp(b"a+") | |
2483 | self._writinghandles = (ifh, dfh) |
|
2483 | self._writinghandles = (ifh, dfh) | |
2484 | finally: |
|
2484 | finally: | |
2485 | self._writinghandles = None |
|
2485 | self._writinghandles = None | |
2486 |
|
2486 | |||
2487 | if dfh: |
|
2487 | if dfh: | |
2488 | dfh.close() |
|
2488 | dfh.close() | |
2489 | ifh.close() |
|
2489 | ifh.close() | |
2490 | return not empty |
|
2490 | return not empty | |
2491 |
|
2491 | |||
2492 | def iscensored(self, rev): |
|
2492 | def iscensored(self, rev): | |
2493 | """Check if a file revision is censored.""" |
|
2493 | """Check if a file revision is censored.""" | |
2494 | if not self._censorable: |
|
2494 | if not self._censorable: | |
2495 | return False |
|
2495 | return False | |
2496 |
|
2496 | |||
2497 | return self.flags(rev) & REVIDX_ISCENSORED |
|
2497 | return self.flags(rev) & REVIDX_ISCENSORED | |
2498 |
|
2498 | |||
2499 | def _peek_iscensored(self, baserev, delta, flush): |
|
2499 | def _peek_iscensored(self, baserev, delta, flush): | |
2500 | """Quickly check if a delta produces a censored revision.""" |
|
2500 | """Quickly check if a delta produces a censored revision.""" | |
2501 | if not self._censorable: |
|
2501 | if not self._censorable: | |
2502 | return False |
|
2502 | return False | |
2503 |
|
2503 | |||
2504 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) |
|
2504 | return storageutil.deltaiscensored(delta, baserev, self.rawsize) | |
2505 |
|
2505 | |||
2506 | def getstrippoint(self, minlink): |
|
2506 | def getstrippoint(self, minlink): | |
2507 | """find the minimum rev that must be stripped to strip the linkrev |
|
2507 | """find the minimum rev that must be stripped to strip the linkrev | |
2508 |
|
2508 | |||
2509 | Returns a tuple containing the minimum rev and a set of all revs that |
|
2509 | Returns a tuple containing the minimum rev and a set of all revs that | |
2510 | have linkrevs that will be broken by this strip. |
|
2510 | have linkrevs that will be broken by this strip. | |
2511 | """ |
|
2511 | """ | |
2512 | return storageutil.resolvestripinfo( |
|
2512 | return storageutil.resolvestripinfo( | |
2513 | minlink, |
|
2513 | minlink, | |
2514 | len(self) - 1, |
|
2514 | len(self) - 1, | |
2515 | self.headrevs(), |
|
2515 | self.headrevs(), | |
2516 | self.linkrev, |
|
2516 | self.linkrev, | |
2517 | self.parentrevs, |
|
2517 | self.parentrevs, | |
2518 | ) |
|
2518 | ) | |
2519 |
|
2519 | |||
2520 | def strip(self, minlink, transaction): |
|
2520 | def strip(self, minlink, transaction): | |
2521 | """truncate the revlog on the first revision with a linkrev >= minlink |
|
2521 | """truncate the revlog on the first revision with a linkrev >= minlink | |
2522 |
|
2522 | |||
2523 | This function is called when we're stripping revision minlink and |
|
2523 | This function is called when we're stripping revision minlink and | |
2524 | its descendants from the repository. |
|
2524 | its descendants from the repository. | |
2525 |
|
2525 | |||
2526 | We have to remove all revisions with linkrev >= minlink, because |
|
2526 | We have to remove all revisions with linkrev >= minlink, because | |
2527 | the equivalent changelog revisions will be renumbered after the |
|
2527 | the equivalent changelog revisions will be renumbered after the | |
2528 | strip. |
|
2528 | strip. | |
2529 |
|
2529 | |||
2530 | So we truncate the revlog on the first of these revisions, and |
|
2530 | So we truncate the revlog on the first of these revisions, and | |
2531 | trust that the caller has saved the revisions that shouldn't be |
|
2531 | trust that the caller has saved the revisions that shouldn't be | |
2532 | removed and that it'll re-add them after this truncation. |
|
2532 | removed and that it'll re-add them after this truncation. | |
2533 | """ |
|
2533 | """ | |
2534 | if len(self) == 0: |
|
2534 | if len(self) == 0: | |
2535 | return |
|
2535 | return | |
2536 |
|
2536 | |||
2537 | rev, _ = self.getstrippoint(minlink) |
|
2537 | rev, _ = self.getstrippoint(minlink) | |
2538 | if rev == len(self): |
|
2538 | if rev == len(self): | |
2539 | return |
|
2539 | return | |
2540 |
|
2540 | |||
2541 | # first truncate the files on disk |
|
2541 | # first truncate the files on disk | |
2542 | end = self.start(rev) |
|
2542 | end = self.start(rev) | |
2543 | if not self._inline: |
|
2543 | if not self._inline: | |
2544 | transaction.add(self.datafile, end) |
|
2544 | transaction.add(self.datafile, end) | |
2545 | end = rev * self.index.entry_size |
|
2545 | end = rev * self.index.entry_size | |
2546 | else: |
|
2546 | else: | |
2547 | end += rev * self.index.entry_size |
|
2547 | end += rev * self.index.entry_size | |
2548 |
|
2548 | |||
2549 | transaction.add(self.indexfile, end) |
|
2549 | transaction.add(self.indexfile, end) | |
2550 |
|
2550 | |||
2551 | # then reset internal state in memory to forget those revisions |
|
2551 | # then reset internal state in memory to forget those revisions | |
2552 | self._revisioncache = None |
|
2552 | self._revisioncache = None | |
2553 | self._chaininfocache = util.lrucachedict(500) |
|
2553 | self._chaininfocache = util.lrucachedict(500) | |
2554 | self._chunkclear() |
|
2554 | self._chunkclear() | |
2555 |
|
2555 | |||
2556 | del self.index[rev:-1] |
|
2556 | del self.index[rev:-1] | |
2557 |
|
2557 | |||
2558 | def checksize(self): |
|
2558 | def checksize(self): | |
2559 | """Check size of index and data files |
|
2559 | """Check size of index and data files | |
2560 |
|
2560 | |||
2561 | return a (dd, di) tuple. |
|
2561 | return a (dd, di) tuple. | |
2562 | - dd: extra bytes for the "data" file |
|
2562 | - dd: extra bytes for the "data" file | |
2563 | - di: extra bytes for the "index" file |
|
2563 | - di: extra bytes for the "index" file | |
2564 |
|
2564 | |||
2565 | A healthy revlog will return (0, 0). |
|
2565 | A healthy revlog will return (0, 0). | |
2566 | """ |
|
2566 | """ | |
2567 | expected = 0 |
|
2567 | expected = 0 | |
2568 | if len(self): |
|
2568 | if len(self): | |
2569 | expected = max(0, self.end(len(self) - 1)) |
|
2569 | expected = max(0, self.end(len(self) - 1)) | |
2570 |
|
2570 | |||
2571 | try: |
|
2571 | try: | |
2572 | with self._datafp() as f: |
|
2572 | with self._datafp() as f: | |
2573 | f.seek(0, io.SEEK_END) |
|
2573 | f.seek(0, io.SEEK_END) | |
2574 | actual = f.tell() |
|
2574 | actual = f.tell() | |
2575 | dd = actual - expected |
|
2575 | dd = actual - expected | |
2576 | except IOError as inst: |
|
2576 | except IOError as inst: | |
2577 | if inst.errno != errno.ENOENT: |
|
2577 | if inst.errno != errno.ENOENT: | |
2578 | raise |
|
2578 | raise | |
2579 | dd = 0 |
|
2579 | dd = 0 | |
2580 |
|
2580 | |||
2581 | try: |
|
2581 | try: | |
2582 | f = self.opener(self.indexfile) |
|
2582 | f = self.opener(self.indexfile) | |
2583 | f.seek(0, io.SEEK_END) |
|
2583 | f.seek(0, io.SEEK_END) | |
2584 | actual = f.tell() |
|
2584 | actual = f.tell() | |
2585 | f.close() |
|
2585 | f.close() | |
2586 | s = self.index.entry_size |
|
2586 | s = self.index.entry_size | |
2587 | i = max(0, actual // s) |
|
2587 | i = max(0, actual // s) | |
2588 | di = actual - (i * s) |
|
2588 | di = actual - (i * s) | |
2589 | if self._inline: |
|
2589 | if self._inline: | |
2590 | databytes = 0 |
|
2590 | databytes = 0 | |
2591 | for r in self: |
|
2591 | for r in self: | |
2592 | databytes += max(0, self.length(r)) |
|
2592 | databytes += max(0, self.length(r)) | |
2593 | dd = 0 |
|
2593 | dd = 0 | |
2594 | di = actual - len(self) * s - databytes |
|
2594 | di = actual - len(self) * s - databytes | |
2595 | except IOError as inst: |
|
2595 | except IOError as inst: | |
2596 | if inst.errno != errno.ENOENT: |
|
2596 | if inst.errno != errno.ENOENT: | |
2597 | raise |
|
2597 | raise | |
2598 | di = 0 |
|
2598 | di = 0 | |
2599 |
|
2599 | |||
2600 | return (dd, di) |
|
2600 | return (dd, di) | |
2601 |
|
2601 | |||
2602 | def files(self): |
|
2602 | def files(self): | |
2603 | res = [self.indexfile] |
|
2603 | res = [self.indexfile] | |
2604 | if not self._inline: |
|
2604 | if not self._inline: | |
2605 | res.append(self.datafile) |
|
2605 | res.append(self.datafile) | |
2606 | return res |
|
2606 | return res | |
2607 |
|
2607 | |||
2608 | def emitrevisions( |
|
2608 | def emitrevisions( | |
2609 | self, |
|
2609 | self, | |
2610 | nodes, |
|
2610 | nodes, | |
2611 | nodesorder=None, |
|
2611 | nodesorder=None, | |
2612 | revisiondata=False, |
|
2612 | revisiondata=False, | |
2613 | assumehaveparentrevisions=False, |
|
2613 | assumehaveparentrevisions=False, | |
2614 | deltamode=repository.CG_DELTAMODE_STD, |
|
2614 | deltamode=repository.CG_DELTAMODE_STD, | |
2615 | sidedata_helpers=None, |
|
2615 | sidedata_helpers=None, | |
2616 | ): |
|
2616 | ): | |
2617 | if nodesorder not in (b'nodes', b'storage', b'linear', None): |
|
2617 | if nodesorder not in (b'nodes', b'storage', b'linear', None): | |
2618 | raise error.ProgrammingError( |
|
2618 | raise error.ProgrammingError( | |
2619 | b'unhandled value for nodesorder: %s' % nodesorder |
|
2619 | b'unhandled value for nodesorder: %s' % nodesorder | |
2620 | ) |
|
2620 | ) | |
2621 |
|
2621 | |||
2622 | if nodesorder is None and not self._generaldelta: |
|
2622 | if nodesorder is None and not self._generaldelta: | |
2623 | nodesorder = b'storage' |
|
2623 | nodesorder = b'storage' | |
2624 |
|
2624 | |||
2625 | if ( |
|
2625 | if ( | |
2626 | not self._storedeltachains |
|
2626 | not self._storedeltachains | |
2627 | and deltamode != repository.CG_DELTAMODE_PREV |
|
2627 | and deltamode != repository.CG_DELTAMODE_PREV | |
2628 | ): |
|
2628 | ): | |
2629 | deltamode = repository.CG_DELTAMODE_FULL |
|
2629 | deltamode = repository.CG_DELTAMODE_FULL | |
2630 |
|
2630 | |||
2631 | return storageutil.emitrevisions( |
|
2631 | return storageutil.emitrevisions( | |
2632 | self, |
|
2632 | self, | |
2633 | nodes, |
|
2633 | nodes, | |
2634 | nodesorder, |
|
2634 | nodesorder, | |
2635 | revlogrevisiondelta, |
|
2635 | revlogrevisiondelta, | |
2636 | deltaparentfn=self.deltaparent, |
|
2636 | deltaparentfn=self.deltaparent, | |
2637 | candeltafn=self.candelta, |
|
2637 | candeltafn=self.candelta, | |
2638 | rawsizefn=self.rawsize, |
|
2638 | rawsizefn=self.rawsize, | |
2639 | revdifffn=self.revdiff, |
|
2639 | revdifffn=self.revdiff, | |
2640 | flagsfn=self.flags, |
|
2640 | flagsfn=self.flags, | |
2641 | deltamode=deltamode, |
|
2641 | deltamode=deltamode, | |
2642 | revisiondata=revisiondata, |
|
2642 | revisiondata=revisiondata, | |
2643 | assumehaveparentrevisions=assumehaveparentrevisions, |
|
2643 | assumehaveparentrevisions=assumehaveparentrevisions, | |
2644 | sidedata_helpers=sidedata_helpers, |
|
2644 | sidedata_helpers=sidedata_helpers, | |
2645 | ) |
|
2645 | ) | |
2646 |
|
2646 | |||
2647 | DELTAREUSEALWAYS = b'always' |
|
2647 | DELTAREUSEALWAYS = b'always' | |
2648 | DELTAREUSESAMEREVS = b'samerevs' |
|
2648 | DELTAREUSESAMEREVS = b'samerevs' | |
2649 | DELTAREUSENEVER = b'never' |
|
2649 | DELTAREUSENEVER = b'never' | |
2650 |
|
2650 | |||
2651 | DELTAREUSEFULLADD = b'fulladd' |
|
2651 | DELTAREUSEFULLADD = b'fulladd' | |
2652 |
|
2652 | |||
2653 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} |
|
2653 | DELTAREUSEALL = {b'always', b'samerevs', b'never', b'fulladd'} | |
2654 |
|
2654 | |||
2655 | def clone( |
|
2655 | def clone( | |
2656 | self, |
|
2656 | self, | |
2657 | tr, |
|
2657 | tr, | |
2658 | destrevlog, |
|
2658 | destrevlog, | |
2659 | addrevisioncb=None, |
|
2659 | addrevisioncb=None, | |
2660 | deltareuse=DELTAREUSESAMEREVS, |
|
2660 | deltareuse=DELTAREUSESAMEREVS, | |
2661 | forcedeltabothparents=None, |
|
2661 | forcedeltabothparents=None, | |
2662 | sidedata_helpers=None, |
|
2662 | sidedata_helpers=None, | |
2663 | ): |
|
2663 | ): | |
2664 | """Copy this revlog to another, possibly with format changes. |
|
2664 | """Copy this revlog to another, possibly with format changes. | |
2665 |
|
2665 | |||
2666 | The destination revlog will contain the same revisions and nodes. |
|
2666 | The destination revlog will contain the same revisions and nodes. | |
2667 | However, it may not be bit-for-bit identical due to e.g. delta encoding |
|
2667 | However, it may not be bit-for-bit identical due to e.g. delta encoding | |
2668 | differences. |
|
2668 | differences. | |
2669 |
|
2669 | |||
2670 | The ``deltareuse`` argument control how deltas from the existing revlog |
|
2670 | The ``deltareuse`` argument control how deltas from the existing revlog | |
2671 | are preserved in the destination revlog. The argument can have the |
|
2671 | are preserved in the destination revlog. The argument can have the | |
2672 | following values: |
|
2672 | following values: | |
2673 |
|
2673 | |||
2674 | DELTAREUSEALWAYS |
|
2674 | DELTAREUSEALWAYS | |
2675 | Deltas will always be reused (if possible), even if the destination |
|
2675 | Deltas will always be reused (if possible), even if the destination | |
2676 | revlog would not select the same revisions for the delta. This is the |
|
2676 | revlog would not select the same revisions for the delta. This is the | |
2677 | fastest mode of operation. |
|
2677 | fastest mode of operation. | |
2678 | DELTAREUSESAMEREVS |
|
2678 | DELTAREUSESAMEREVS | |
2679 | Deltas will be reused if the destination revlog would pick the same |
|
2679 | Deltas will be reused if the destination revlog would pick the same | |
2680 | revisions for the delta. This mode strikes a balance between speed |
|
2680 | revisions for the delta. This mode strikes a balance between speed | |
2681 | and optimization. |
|
2681 | and optimization. | |
2682 | DELTAREUSENEVER |
|
2682 | DELTAREUSENEVER | |
2683 | Deltas will never be reused. This is the slowest mode of execution. |
|
2683 | Deltas will never be reused. This is the slowest mode of execution. | |
2684 | This mode can be used to recompute deltas (e.g. if the diff/delta |
|
2684 | This mode can be used to recompute deltas (e.g. if the diff/delta | |
2685 | algorithm changes). |
|
2685 | algorithm changes). | |
2686 | DELTAREUSEFULLADD |
|
2686 | DELTAREUSEFULLADD | |
2687 | Revision will be re-added as if their were new content. This is |
|
2687 | Revision will be re-added as if their were new content. This is | |
2688 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. |
|
2688 | slower than DELTAREUSEALWAYS but allow more mechanism to kicks in. | |
2689 | eg: large file detection and handling. |
|
2689 | eg: large file detection and handling. | |
2690 |
|
2690 | |||
2691 | Delta computation can be slow, so the choice of delta reuse policy can |
|
2691 | Delta computation can be slow, so the choice of delta reuse policy can | |
2692 | significantly affect run time. |
|
2692 | significantly affect run time. | |
2693 |
|
2693 | |||
2694 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between |
|
2694 | The default policy (``DELTAREUSESAMEREVS``) strikes a balance between | |
2695 | two extremes. Deltas will be reused if they are appropriate. But if the |
|
2695 | two extremes. Deltas will be reused if they are appropriate. But if the | |
2696 | delta could choose a better revision, it will do so. This means if you |
|
2696 | delta could choose a better revision, it will do so. This means if you | |
2697 | are converting a non-generaldelta revlog to a generaldelta revlog, |
|
2697 | are converting a non-generaldelta revlog to a generaldelta revlog, | |
2698 | deltas will be recomputed if the delta's parent isn't a parent of the |
|
2698 | deltas will be recomputed if the delta's parent isn't a parent of the | |
2699 | revision. |
|
2699 | revision. | |
2700 |
|
2700 | |||
2701 | In addition to the delta policy, the ``forcedeltabothparents`` |
|
2701 | In addition to the delta policy, the ``forcedeltabothparents`` | |
2702 | argument controls whether to force compute deltas against both parents |
|
2702 | argument controls whether to force compute deltas against both parents | |
2703 | for merges. By default, the current default is used. |
|
2703 | for merges. By default, the current default is used. | |
2704 |
|
2704 | |||
2705 | See `storageutil.emitrevisions` for the doc on `sidedata_helpers`. |
|
2705 | See `revlogutil.sidedata.get_sidedata_helpers` for the doc on | |
|
2706 | `sidedata_helpers`. | |||
2706 | """ |
|
2707 | """ | |
2707 | if deltareuse not in self.DELTAREUSEALL: |
|
2708 | if deltareuse not in self.DELTAREUSEALL: | |
2708 | raise ValueError( |
|
2709 | raise ValueError( | |
2709 | _(b'value for deltareuse invalid: %s') % deltareuse |
|
2710 | _(b'value for deltareuse invalid: %s') % deltareuse | |
2710 | ) |
|
2711 | ) | |
2711 |
|
2712 | |||
2712 | if len(destrevlog): |
|
2713 | if len(destrevlog): | |
2713 | raise ValueError(_(b'destination revlog is not empty')) |
|
2714 | raise ValueError(_(b'destination revlog is not empty')) | |
2714 |
|
2715 | |||
2715 | if getattr(self, 'filteredrevs', None): |
|
2716 | if getattr(self, 'filteredrevs', None): | |
2716 | raise ValueError(_(b'source revlog has filtered revisions')) |
|
2717 | raise ValueError(_(b'source revlog has filtered revisions')) | |
2717 | if getattr(destrevlog, 'filteredrevs', None): |
|
2718 | if getattr(destrevlog, 'filteredrevs', None): | |
2718 | raise ValueError(_(b'destination revlog has filtered revisions')) |
|
2719 | raise ValueError(_(b'destination revlog has filtered revisions')) | |
2719 |
|
2720 | |||
2720 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, |
|
2721 | # lazydelta and lazydeltabase controls whether to reuse a cached delta, | |
2721 | # if possible. |
|
2722 | # if possible. | |
2722 | oldlazydelta = destrevlog._lazydelta |
|
2723 | oldlazydelta = destrevlog._lazydelta | |
2723 | oldlazydeltabase = destrevlog._lazydeltabase |
|
2724 | oldlazydeltabase = destrevlog._lazydeltabase | |
2724 | oldamd = destrevlog._deltabothparents |
|
2725 | oldamd = destrevlog._deltabothparents | |
2725 |
|
2726 | |||
2726 | try: |
|
2727 | try: | |
2727 | if deltareuse == self.DELTAREUSEALWAYS: |
|
2728 | if deltareuse == self.DELTAREUSEALWAYS: | |
2728 | destrevlog._lazydeltabase = True |
|
2729 | destrevlog._lazydeltabase = True | |
2729 | destrevlog._lazydelta = True |
|
2730 | destrevlog._lazydelta = True | |
2730 | elif deltareuse == self.DELTAREUSESAMEREVS: |
|
2731 | elif deltareuse == self.DELTAREUSESAMEREVS: | |
2731 | destrevlog._lazydeltabase = False |
|
2732 | destrevlog._lazydeltabase = False | |
2732 | destrevlog._lazydelta = True |
|
2733 | destrevlog._lazydelta = True | |
2733 | elif deltareuse == self.DELTAREUSENEVER: |
|
2734 | elif deltareuse == self.DELTAREUSENEVER: | |
2734 | destrevlog._lazydeltabase = False |
|
2735 | destrevlog._lazydeltabase = False | |
2735 | destrevlog._lazydelta = False |
|
2736 | destrevlog._lazydelta = False | |
2736 |
|
2737 | |||
2737 | destrevlog._deltabothparents = forcedeltabothparents or oldamd |
|
2738 | destrevlog._deltabothparents = forcedeltabothparents or oldamd | |
2738 |
|
2739 | |||
2739 | self._clone( |
|
2740 | self._clone( | |
2740 | tr, |
|
2741 | tr, | |
2741 | destrevlog, |
|
2742 | destrevlog, | |
2742 | addrevisioncb, |
|
2743 | addrevisioncb, | |
2743 | deltareuse, |
|
2744 | deltareuse, | |
2744 | forcedeltabothparents, |
|
2745 | forcedeltabothparents, | |
2745 | sidedata_helpers, |
|
2746 | sidedata_helpers, | |
2746 | ) |
|
2747 | ) | |
2747 |
|
2748 | |||
2748 | finally: |
|
2749 | finally: | |
2749 | destrevlog._lazydelta = oldlazydelta |
|
2750 | destrevlog._lazydelta = oldlazydelta | |
2750 | destrevlog._lazydeltabase = oldlazydeltabase |
|
2751 | destrevlog._lazydeltabase = oldlazydeltabase | |
2751 | destrevlog._deltabothparents = oldamd |
|
2752 | destrevlog._deltabothparents = oldamd | |
2752 |
|
2753 | |||
2753 | def _clone( |
|
2754 | def _clone( | |
2754 | self, |
|
2755 | self, | |
2755 | tr, |
|
2756 | tr, | |
2756 | destrevlog, |
|
2757 | destrevlog, | |
2757 | addrevisioncb, |
|
2758 | addrevisioncb, | |
2758 | deltareuse, |
|
2759 | deltareuse, | |
2759 | forcedeltabothparents, |
|
2760 | forcedeltabothparents, | |
2760 | sidedata_helpers, |
|
2761 | sidedata_helpers, | |
2761 | ): |
|
2762 | ): | |
2762 | """perform the core duty of `revlog.clone` after parameter processing""" |
|
2763 | """perform the core duty of `revlog.clone` after parameter processing""" | |
2763 | deltacomputer = deltautil.deltacomputer(destrevlog) |
|
2764 | deltacomputer = deltautil.deltacomputer(destrevlog) | |
2764 | index = self.index |
|
2765 | index = self.index | |
2765 | for rev in self: |
|
2766 | for rev in self: | |
2766 | entry = index[rev] |
|
2767 | entry = index[rev] | |
2767 |
|
2768 | |||
2768 | # Some classes override linkrev to take filtered revs into |
|
2769 | # Some classes override linkrev to take filtered revs into | |
2769 | # account. Use raw entry from index. |
|
2770 | # account. Use raw entry from index. | |
2770 | flags = entry[0] & 0xFFFF |
|
2771 | flags = entry[0] & 0xFFFF | |
2771 | linkrev = entry[4] |
|
2772 | linkrev = entry[4] | |
2772 | p1 = index[entry[5]][7] |
|
2773 | p1 = index[entry[5]][7] | |
2773 | p2 = index[entry[6]][7] |
|
2774 | p2 = index[entry[6]][7] | |
2774 | node = entry[7] |
|
2775 | node = entry[7] | |
2775 |
|
2776 | |||
2776 | # (Possibly) reuse the delta from the revlog if allowed and |
|
2777 | # (Possibly) reuse the delta from the revlog if allowed and | |
2777 | # the revlog chunk is a delta. |
|
2778 | # the revlog chunk is a delta. | |
2778 | cachedelta = None |
|
2779 | cachedelta = None | |
2779 | rawtext = None |
|
2780 | rawtext = None | |
2780 | if deltareuse == self.DELTAREUSEFULLADD: |
|
2781 | if deltareuse == self.DELTAREUSEFULLADD: | |
2781 | text, sidedata = self._revisiondata(rev) |
|
2782 | text, sidedata = self._revisiondata(rev) | |
2782 |
|
2783 | |||
2783 | if sidedata_helpers is not None: |
|
2784 | if sidedata_helpers is not None: | |
2784 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
2785 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
2785 | self, sidedata_helpers, sidedata, rev |
|
2786 | self, sidedata_helpers, sidedata, rev | |
2786 | ) |
|
2787 | ) | |
2787 | flags = flags | new_flags[0] & ~new_flags[1] |
|
2788 | flags = flags | new_flags[0] & ~new_flags[1] | |
2788 |
|
2789 | |||
2789 | destrevlog.addrevision( |
|
2790 | destrevlog.addrevision( | |
2790 | text, |
|
2791 | text, | |
2791 | tr, |
|
2792 | tr, | |
2792 | linkrev, |
|
2793 | linkrev, | |
2793 | p1, |
|
2794 | p1, | |
2794 | p2, |
|
2795 | p2, | |
2795 | cachedelta=cachedelta, |
|
2796 | cachedelta=cachedelta, | |
2796 | node=node, |
|
2797 | node=node, | |
2797 | flags=flags, |
|
2798 | flags=flags, | |
2798 | deltacomputer=deltacomputer, |
|
2799 | deltacomputer=deltacomputer, | |
2799 | sidedata=sidedata, |
|
2800 | sidedata=sidedata, | |
2800 | ) |
|
2801 | ) | |
2801 | else: |
|
2802 | else: | |
2802 | if destrevlog._lazydelta: |
|
2803 | if destrevlog._lazydelta: | |
2803 | dp = self.deltaparent(rev) |
|
2804 | dp = self.deltaparent(rev) | |
2804 | if dp != nullrev: |
|
2805 | if dp != nullrev: | |
2805 | cachedelta = (dp, bytes(self._chunk(rev))) |
|
2806 | cachedelta = (dp, bytes(self._chunk(rev))) | |
2806 |
|
2807 | |||
2807 | sidedata = None |
|
2808 | sidedata = None | |
2808 | if not cachedelta: |
|
2809 | if not cachedelta: | |
2809 | rawtext, sidedata = self._revisiondata(rev) |
|
2810 | rawtext, sidedata = self._revisiondata(rev) | |
2810 | if sidedata is None: |
|
2811 | if sidedata is None: | |
2811 | sidedata = self.sidedata(rev) |
|
2812 | sidedata = self.sidedata(rev) | |
2812 |
|
2813 | |||
2813 | if sidedata_helpers is not None: |
|
2814 | if sidedata_helpers is not None: | |
2814 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( |
|
2815 | (sidedata, new_flags) = sidedatautil.run_sidedata_helpers( | |
2815 | self, sidedata_helpers, sidedata, rev |
|
2816 | self, sidedata_helpers, sidedata, rev | |
2816 | ) |
|
2817 | ) | |
2817 | flags = flags | new_flags[0] & ~new_flags[1] |
|
2818 | flags = flags | new_flags[0] & ~new_flags[1] | |
2818 |
|
2819 | |||
2819 | ifh = destrevlog.opener( |
|
2820 | ifh = destrevlog.opener( | |
2820 | destrevlog.indexfile, b'a+', checkambig=False |
|
2821 | destrevlog.indexfile, b'a+', checkambig=False | |
2821 | ) |
|
2822 | ) | |
2822 | dfh = None |
|
2823 | dfh = None | |
2823 | if not destrevlog._inline: |
|
2824 | if not destrevlog._inline: | |
2824 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') |
|
2825 | dfh = destrevlog.opener(destrevlog.datafile, b'a+') | |
2825 | try: |
|
2826 | try: | |
2826 | destrevlog._addrevision( |
|
2827 | destrevlog._addrevision( | |
2827 | node, |
|
2828 | node, | |
2828 | rawtext, |
|
2829 | rawtext, | |
2829 | tr, |
|
2830 | tr, | |
2830 | linkrev, |
|
2831 | linkrev, | |
2831 | p1, |
|
2832 | p1, | |
2832 | p2, |
|
2833 | p2, | |
2833 | flags, |
|
2834 | flags, | |
2834 | cachedelta, |
|
2835 | cachedelta, | |
2835 | ifh, |
|
2836 | ifh, | |
2836 | dfh, |
|
2837 | dfh, | |
2837 | deltacomputer=deltacomputer, |
|
2838 | deltacomputer=deltacomputer, | |
2838 | sidedata=sidedata, |
|
2839 | sidedata=sidedata, | |
2839 | ) |
|
2840 | ) | |
2840 | finally: |
|
2841 | finally: | |
2841 | if dfh: |
|
2842 | if dfh: | |
2842 | dfh.close() |
|
2843 | dfh.close() | |
2843 | ifh.close() |
|
2844 | ifh.close() | |
2844 |
|
2845 | |||
2845 | if addrevisioncb: |
|
2846 | if addrevisioncb: | |
2846 | addrevisioncb(self, rev, node) |
|
2847 | addrevisioncb(self, rev, node) | |
2847 |
|
2848 | |||
2848 | def censorrevision(self, tr, censornode, tombstone=b''): |
|
2849 | def censorrevision(self, tr, censornode, tombstone=b''): | |
2849 | if (self.version & 0xFFFF) == REVLOGV0: |
|
2850 | if (self.version & 0xFFFF) == REVLOGV0: | |
2850 | raise error.RevlogError( |
|
2851 | raise error.RevlogError( | |
2851 | _(b'cannot censor with version %d revlogs') % self.version |
|
2852 | _(b'cannot censor with version %d revlogs') % self.version | |
2852 | ) |
|
2853 | ) | |
2853 |
|
2854 | |||
2854 | censorrev = self.rev(censornode) |
|
2855 | censorrev = self.rev(censornode) | |
2855 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') |
|
2856 | tombstone = storageutil.packmeta({b'censored': tombstone}, b'') | |
2856 |
|
2857 | |||
2857 | if len(tombstone) > self.rawsize(censorrev): |
|
2858 | if len(tombstone) > self.rawsize(censorrev): | |
2858 | raise error.Abort( |
|
2859 | raise error.Abort( | |
2859 | _(b'censor tombstone must be no longer than censored data') |
|
2860 | _(b'censor tombstone must be no longer than censored data') | |
2860 | ) |
|
2861 | ) | |
2861 |
|
2862 | |||
2862 | # Rewriting the revlog in place is hard. Our strategy for censoring is |
|
2863 | # Rewriting the revlog in place is hard. Our strategy for censoring is | |
2863 | # to create a new revlog, copy all revisions to it, then replace the |
|
2864 | # to create a new revlog, copy all revisions to it, then replace the | |
2864 | # revlogs on transaction close. |
|
2865 | # revlogs on transaction close. | |
2865 |
|
2866 | |||
2866 | newindexfile = self.indexfile + b'.tmpcensored' |
|
2867 | newindexfile = self.indexfile + b'.tmpcensored' | |
2867 | newdatafile = self.datafile + b'.tmpcensored' |
|
2868 | newdatafile = self.datafile + b'.tmpcensored' | |
2868 |
|
2869 | |||
2869 | # This is a bit dangerous. We could easily have a mismatch of state. |
|
2870 | # This is a bit dangerous. We could easily have a mismatch of state. | |
2870 | newrl = revlog( |
|
2871 | newrl = revlog( | |
2871 | self.opener, |
|
2872 | self.opener, | |
2872 | target=self.target, |
|
2873 | target=self.target, | |
2873 | indexfile=newindexfile, |
|
2874 | indexfile=newindexfile, | |
2874 | datafile=newdatafile, |
|
2875 | datafile=newdatafile, | |
2875 | censorable=True, |
|
2876 | censorable=True, | |
2876 | ) |
|
2877 | ) | |
2877 | newrl.version = self.version |
|
2878 | newrl.version = self.version | |
2878 | newrl._generaldelta = self._generaldelta |
|
2879 | newrl._generaldelta = self._generaldelta | |
2879 | newrl._parse_index = self._parse_index |
|
2880 | newrl._parse_index = self._parse_index | |
2880 |
|
2881 | |||
2881 | for rev in self.revs(): |
|
2882 | for rev in self.revs(): | |
2882 | node = self.node(rev) |
|
2883 | node = self.node(rev) | |
2883 | p1, p2 = self.parents(node) |
|
2884 | p1, p2 = self.parents(node) | |
2884 |
|
2885 | |||
2885 | if rev == censorrev: |
|
2886 | if rev == censorrev: | |
2886 | newrl.addrawrevision( |
|
2887 | newrl.addrawrevision( | |
2887 | tombstone, |
|
2888 | tombstone, | |
2888 | tr, |
|
2889 | tr, | |
2889 | self.linkrev(censorrev), |
|
2890 | self.linkrev(censorrev), | |
2890 | p1, |
|
2891 | p1, | |
2891 | p2, |
|
2892 | p2, | |
2892 | censornode, |
|
2893 | censornode, | |
2893 | REVIDX_ISCENSORED, |
|
2894 | REVIDX_ISCENSORED, | |
2894 | ) |
|
2895 | ) | |
2895 |
|
2896 | |||
2896 | if newrl.deltaparent(rev) != nullrev: |
|
2897 | if newrl.deltaparent(rev) != nullrev: | |
2897 | raise error.Abort( |
|
2898 | raise error.Abort( | |
2898 | _( |
|
2899 | _( | |
2899 | b'censored revision stored as delta; ' |
|
2900 | b'censored revision stored as delta; ' | |
2900 | b'cannot censor' |
|
2901 | b'cannot censor' | |
2901 | ), |
|
2902 | ), | |
2902 | hint=_( |
|
2903 | hint=_( | |
2903 | b'censoring of revlogs is not ' |
|
2904 | b'censoring of revlogs is not ' | |
2904 | b'fully implemented; please report ' |
|
2905 | b'fully implemented; please report ' | |
2905 | b'this bug' |
|
2906 | b'this bug' | |
2906 | ), |
|
2907 | ), | |
2907 | ) |
|
2908 | ) | |
2908 | continue |
|
2909 | continue | |
2909 |
|
2910 | |||
2910 | if self.iscensored(rev): |
|
2911 | if self.iscensored(rev): | |
2911 | if self.deltaparent(rev) != nullrev: |
|
2912 | if self.deltaparent(rev) != nullrev: | |
2912 | raise error.Abort( |
|
2913 | raise error.Abort( | |
2913 | _( |
|
2914 | _( | |
2914 | b'cannot censor due to censored ' |
|
2915 | b'cannot censor due to censored ' | |
2915 | b'revision having delta stored' |
|
2916 | b'revision having delta stored' | |
2916 | ) |
|
2917 | ) | |
2917 | ) |
|
2918 | ) | |
2918 | rawtext = self._chunk(rev) |
|
2919 | rawtext = self._chunk(rev) | |
2919 | else: |
|
2920 | else: | |
2920 | rawtext = self.rawdata(rev) |
|
2921 | rawtext = self.rawdata(rev) | |
2921 |
|
2922 | |||
2922 | newrl.addrawrevision( |
|
2923 | newrl.addrawrevision( | |
2923 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) |
|
2924 | rawtext, tr, self.linkrev(rev), p1, p2, node, self.flags(rev) | |
2924 | ) |
|
2925 | ) | |
2925 |
|
2926 | |||
2926 | tr.addbackup(self.indexfile, location=b'store') |
|
2927 | tr.addbackup(self.indexfile, location=b'store') | |
2927 | if not self._inline: |
|
2928 | if not self._inline: | |
2928 | tr.addbackup(self.datafile, location=b'store') |
|
2929 | tr.addbackup(self.datafile, location=b'store') | |
2929 |
|
2930 | |||
2930 | self.opener.rename(newrl.indexfile, self.indexfile) |
|
2931 | self.opener.rename(newrl.indexfile, self.indexfile) | |
2931 | if not self._inline: |
|
2932 | if not self._inline: | |
2932 | self.opener.rename(newrl.datafile, self.datafile) |
|
2933 | self.opener.rename(newrl.datafile, self.datafile) | |
2933 |
|
2934 | |||
2934 | self.clearcaches() |
|
2935 | self.clearcaches() | |
2935 | self._loadindex() |
|
2936 | self._loadindex() | |
2936 |
|
2937 | |||
2937 | def verifyintegrity(self, state): |
|
2938 | def verifyintegrity(self, state): | |
2938 | """Verifies the integrity of the revlog. |
|
2939 | """Verifies the integrity of the revlog. | |
2939 |
|
2940 | |||
2940 | Yields ``revlogproblem`` instances describing problems that are |
|
2941 | Yields ``revlogproblem`` instances describing problems that are | |
2941 | found. |
|
2942 | found. | |
2942 | """ |
|
2943 | """ | |
2943 | dd, di = self.checksize() |
|
2944 | dd, di = self.checksize() | |
2944 | if dd: |
|
2945 | if dd: | |
2945 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) |
|
2946 | yield revlogproblem(error=_(b'data length off by %d bytes') % dd) | |
2946 | if di: |
|
2947 | if di: | |
2947 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) |
|
2948 | yield revlogproblem(error=_(b'index contains %d extra bytes') % di) | |
2948 |
|
2949 | |||
2949 | version = self.version & 0xFFFF |
|
2950 | version = self.version & 0xFFFF | |
2950 |
|
2951 | |||
2951 | # The verifier tells us what version revlog we should be. |
|
2952 | # The verifier tells us what version revlog we should be. | |
2952 | if version != state[b'expectedversion']: |
|
2953 | if version != state[b'expectedversion']: | |
2953 | yield revlogproblem( |
|
2954 | yield revlogproblem( | |
2954 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") |
|
2955 | warning=_(b"warning: '%s' uses revlog format %d; expected %d") | |
2955 | % (self.indexfile, version, state[b'expectedversion']) |
|
2956 | % (self.indexfile, version, state[b'expectedversion']) | |
2956 | ) |
|
2957 | ) | |
2957 |
|
2958 | |||
2958 | state[b'skipread'] = set() |
|
2959 | state[b'skipread'] = set() | |
2959 | state[b'safe_renamed'] = set() |
|
2960 | state[b'safe_renamed'] = set() | |
2960 |
|
2961 | |||
2961 | for rev in self: |
|
2962 | for rev in self: | |
2962 | node = self.node(rev) |
|
2963 | node = self.node(rev) | |
2963 |
|
2964 | |||
2964 | # Verify contents. 4 cases to care about: |
|
2965 | # Verify contents. 4 cases to care about: | |
2965 | # |
|
2966 | # | |
2966 | # common: the most common case |
|
2967 | # common: the most common case | |
2967 | # rename: with a rename |
|
2968 | # rename: with a rename | |
2968 | # meta: file content starts with b'\1\n', the metadata |
|
2969 | # meta: file content starts with b'\1\n', the metadata | |
2969 | # header defined in filelog.py, but without a rename |
|
2970 | # header defined in filelog.py, but without a rename | |
2970 | # ext: content stored externally |
|
2971 | # ext: content stored externally | |
2971 | # |
|
2972 | # | |
2972 | # More formally, their differences are shown below: |
|
2973 | # More formally, their differences are shown below: | |
2973 | # |
|
2974 | # | |
2974 | # | common | rename | meta | ext |
|
2975 | # | common | rename | meta | ext | |
2975 | # ------------------------------------------------------- |
|
2976 | # ------------------------------------------------------- | |
2976 | # flags() | 0 | 0 | 0 | not 0 |
|
2977 | # flags() | 0 | 0 | 0 | not 0 | |
2977 | # renamed() | False | True | False | ? |
|
2978 | # renamed() | False | True | False | ? | |
2978 | # rawtext[0:2]=='\1\n'| False | True | True | ? |
|
2979 | # rawtext[0:2]=='\1\n'| False | True | True | ? | |
2979 | # |
|
2980 | # | |
2980 | # "rawtext" means the raw text stored in revlog data, which |
|
2981 | # "rawtext" means the raw text stored in revlog data, which | |
2981 | # could be retrieved by "rawdata(rev)". "text" |
|
2982 | # could be retrieved by "rawdata(rev)". "text" | |
2982 | # mentioned below is "revision(rev)". |
|
2983 | # mentioned below is "revision(rev)". | |
2983 | # |
|
2984 | # | |
2984 | # There are 3 different lengths stored physically: |
|
2985 | # There are 3 different lengths stored physically: | |
2985 | # 1. L1: rawsize, stored in revlog index |
|
2986 | # 1. L1: rawsize, stored in revlog index | |
2986 | # 2. L2: len(rawtext), stored in revlog data |
|
2987 | # 2. L2: len(rawtext), stored in revlog data | |
2987 | # 3. L3: len(text), stored in revlog data if flags==0, or |
|
2988 | # 3. L3: len(text), stored in revlog data if flags==0, or | |
2988 | # possibly somewhere else if flags!=0 |
|
2989 | # possibly somewhere else if flags!=0 | |
2989 | # |
|
2990 | # | |
2990 | # L1 should be equal to L2. L3 could be different from them. |
|
2991 | # L1 should be equal to L2. L3 could be different from them. | |
2991 | # "text" may or may not affect commit hash depending on flag |
|
2992 | # "text" may or may not affect commit hash depending on flag | |
2992 | # processors (see flagutil.addflagprocessor). |
|
2993 | # processors (see flagutil.addflagprocessor). | |
2993 | # |
|
2994 | # | |
2994 | # | common | rename | meta | ext |
|
2995 | # | common | rename | meta | ext | |
2995 | # ------------------------------------------------- |
|
2996 | # ------------------------------------------------- | |
2996 | # rawsize() | L1 | L1 | L1 | L1 |
|
2997 | # rawsize() | L1 | L1 | L1 | L1 | |
2997 | # size() | L1 | L2-LM | L1(*) | L1 (?) |
|
2998 | # size() | L1 | L2-LM | L1(*) | L1 (?) | |
2998 | # len(rawtext) | L2 | L2 | L2 | L2 |
|
2999 | # len(rawtext) | L2 | L2 | L2 | L2 | |
2999 | # len(text) | L2 | L2 | L2 | L3 |
|
3000 | # len(text) | L2 | L2 | L2 | L3 | |
3000 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) |
|
3001 | # len(read()) | L2 | L2-LM | L2-LM | L3 (?) | |
3001 | # |
|
3002 | # | |
3002 | # LM: length of metadata, depending on rawtext |
|
3003 | # LM: length of metadata, depending on rawtext | |
3003 | # (*): not ideal, see comment in filelog.size |
|
3004 | # (*): not ideal, see comment in filelog.size | |
3004 | # (?): could be "- len(meta)" if the resolved content has |
|
3005 | # (?): could be "- len(meta)" if the resolved content has | |
3005 | # rename metadata |
|
3006 | # rename metadata | |
3006 | # |
|
3007 | # | |
3007 | # Checks needed to be done: |
|
3008 | # Checks needed to be done: | |
3008 | # 1. length check: L1 == L2, in all cases. |
|
3009 | # 1. length check: L1 == L2, in all cases. | |
3009 | # 2. hash check: depending on flag processor, we may need to |
|
3010 | # 2. hash check: depending on flag processor, we may need to | |
3010 | # use either "text" (external), or "rawtext" (in revlog). |
|
3011 | # use either "text" (external), or "rawtext" (in revlog). | |
3011 |
|
3012 | |||
3012 | try: |
|
3013 | try: | |
3013 | skipflags = state.get(b'skipflags', 0) |
|
3014 | skipflags = state.get(b'skipflags', 0) | |
3014 | if skipflags: |
|
3015 | if skipflags: | |
3015 | skipflags &= self.flags(rev) |
|
3016 | skipflags &= self.flags(rev) | |
3016 |
|
3017 | |||
3017 | _verify_revision(self, skipflags, state, node) |
|
3018 | _verify_revision(self, skipflags, state, node) | |
3018 |
|
3019 | |||
3019 | l1 = self.rawsize(rev) |
|
3020 | l1 = self.rawsize(rev) | |
3020 | l2 = len(self.rawdata(node)) |
|
3021 | l2 = len(self.rawdata(node)) | |
3021 |
|
3022 | |||
3022 | if l1 != l2: |
|
3023 | if l1 != l2: | |
3023 | yield revlogproblem( |
|
3024 | yield revlogproblem( | |
3024 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), |
|
3025 | error=_(b'unpacked size is %d, %d expected') % (l2, l1), | |
3025 | node=node, |
|
3026 | node=node, | |
3026 | ) |
|
3027 | ) | |
3027 |
|
3028 | |||
3028 | except error.CensoredNodeError: |
|
3029 | except error.CensoredNodeError: | |
3029 | if state[b'erroroncensored']: |
|
3030 | if state[b'erroroncensored']: | |
3030 | yield revlogproblem( |
|
3031 | yield revlogproblem( | |
3031 | error=_(b'censored file data'), node=node |
|
3032 | error=_(b'censored file data'), node=node | |
3032 | ) |
|
3033 | ) | |
3033 | state[b'skipread'].add(node) |
|
3034 | state[b'skipread'].add(node) | |
3034 | except Exception as e: |
|
3035 | except Exception as e: | |
3035 | yield revlogproblem( |
|
3036 | yield revlogproblem( | |
3036 | error=_(b'unpacking %s: %s') |
|
3037 | error=_(b'unpacking %s: %s') | |
3037 | % (short(node), stringutil.forcebytestr(e)), |
|
3038 | % (short(node), stringutil.forcebytestr(e)), | |
3038 | node=node, |
|
3039 | node=node, | |
3039 | ) |
|
3040 | ) | |
3040 | state[b'skipread'].add(node) |
|
3041 | state[b'skipread'].add(node) | |
3041 |
|
3042 | |||
3042 | def storageinfo( |
|
3043 | def storageinfo( | |
3043 | self, |
|
3044 | self, | |
3044 | exclusivefiles=False, |
|
3045 | exclusivefiles=False, | |
3045 | sharedfiles=False, |
|
3046 | sharedfiles=False, | |
3046 | revisionscount=False, |
|
3047 | revisionscount=False, | |
3047 | trackedsize=False, |
|
3048 | trackedsize=False, | |
3048 | storedsize=False, |
|
3049 | storedsize=False, | |
3049 | ): |
|
3050 | ): | |
3050 | d = {} |
|
3051 | d = {} | |
3051 |
|
3052 | |||
3052 | if exclusivefiles: |
|
3053 | if exclusivefiles: | |
3053 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] |
|
3054 | d[b'exclusivefiles'] = [(self.opener, self.indexfile)] | |
3054 | if not self._inline: |
|
3055 | if not self._inline: | |
3055 | d[b'exclusivefiles'].append((self.opener, self.datafile)) |
|
3056 | d[b'exclusivefiles'].append((self.opener, self.datafile)) | |
3056 |
|
3057 | |||
3057 | if sharedfiles: |
|
3058 | if sharedfiles: | |
3058 | d[b'sharedfiles'] = [] |
|
3059 | d[b'sharedfiles'] = [] | |
3059 |
|
3060 | |||
3060 | if revisionscount: |
|
3061 | if revisionscount: | |
3061 | d[b'revisionscount'] = len(self) |
|
3062 | d[b'revisionscount'] = len(self) | |
3062 |
|
3063 | |||
3063 | if trackedsize: |
|
3064 | if trackedsize: | |
3064 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) |
|
3065 | d[b'trackedsize'] = sum(map(self.rawsize, iter(self))) | |
3065 |
|
3066 | |||
3066 | if storedsize: |
|
3067 | if storedsize: | |
3067 | d[b'storedsize'] = sum( |
|
3068 | d[b'storedsize'] = sum( | |
3068 | self.opener.stat(path).st_size for path in self.files() |
|
3069 | self.opener.stat(path).st_size for path in self.files() | |
3069 | ) |
|
3070 | ) | |
3070 |
|
3071 | |||
3071 | return d |
|
3072 | return d | |
3072 |
|
3073 | |||
3073 | def rewrite_sidedata(self, helpers, startrev, endrev): |
|
3074 | def rewrite_sidedata(self, helpers, startrev, endrev): | |
3074 | if self.version & 0xFFFF != REVLOGV2: |
|
3075 | if self.version & 0xFFFF != REVLOGV2: | |
3075 | return |
|
3076 | return | |
3076 | # inline are not yet supported because they suffer from an issue when |
|
3077 | # inline are not yet supported because they suffer from an issue when | |
3077 | # rewriting them (since it's not an append-only operation). |
|
3078 | # rewriting them (since it's not an append-only operation). | |
3078 | # See issue6485. |
|
3079 | # See issue6485. | |
3079 | assert not self._inline |
|
3080 | assert not self._inline | |
3080 | if not helpers[1] and not helpers[2]: |
|
3081 | if not helpers[1] and not helpers[2]: | |
3081 | # Nothing to generate or remove |
|
3082 | # Nothing to generate or remove | |
3082 | return |
|
3083 | return | |
3083 |
|
3084 | |||
3084 | new_entries = [] |
|
3085 | new_entries = [] | |
3085 | # append the new sidedata |
|
3086 | # append the new sidedata | |
3086 | with self._datafp(b'a+') as fp: |
|
3087 | with self._datafp(b'a+') as fp: | |
3087 | # Maybe this bug still exists, see revlog._writeentry |
|
3088 | # Maybe this bug still exists, see revlog._writeentry | |
3088 | fp.seek(0, os.SEEK_END) |
|
3089 | fp.seek(0, os.SEEK_END) | |
3089 | current_offset = fp.tell() |
|
3090 | current_offset = fp.tell() | |
3090 | for rev in range(startrev, endrev + 1): |
|
3091 | for rev in range(startrev, endrev + 1): | |
3091 | entry = self.index[rev] |
|
3092 | entry = self.index[rev] | |
3092 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( |
|
3093 | new_sidedata, flags = sidedatautil.run_sidedata_helpers( | |
3093 | store=self, |
|
3094 | store=self, | |
3094 | sidedata_helpers=helpers, |
|
3095 | sidedata_helpers=helpers, | |
3095 | sidedata={}, |
|
3096 | sidedata={}, | |
3096 | rev=rev, |
|
3097 | rev=rev, | |
3097 | ) |
|
3098 | ) | |
3098 |
|
3099 | |||
3099 | serialized_sidedata = sidedatautil.serialize_sidedata( |
|
3100 | serialized_sidedata = sidedatautil.serialize_sidedata( | |
3100 | new_sidedata |
|
3101 | new_sidedata | |
3101 | ) |
|
3102 | ) | |
3102 | if entry[8] != 0 or entry[9] != 0: |
|
3103 | if entry[8] != 0 or entry[9] != 0: | |
3103 | # rewriting entries that already have sidedata is not |
|
3104 | # rewriting entries that already have sidedata is not | |
3104 | # supported yet, because it introduces garbage data in the |
|
3105 | # supported yet, because it introduces garbage data in the | |
3105 | # revlog. |
|
3106 | # revlog. | |
3106 | msg = b"Rewriting existing sidedata is not supported yet" |
|
3107 | msg = b"Rewriting existing sidedata is not supported yet" | |
3107 | raise error.Abort(msg) |
|
3108 | raise error.Abort(msg) | |
3108 |
|
3109 | |||
3109 | # Apply (potential) flags to add and to remove after running |
|
3110 | # Apply (potential) flags to add and to remove after running | |
3110 | # the sidedata helpers |
|
3111 | # the sidedata helpers | |
3111 | new_offset_flags = entry[0] | flags[0] & ~flags[1] |
|
3112 | new_offset_flags = entry[0] | flags[0] & ~flags[1] | |
3112 | entry = (new_offset_flags,) + entry[1:8] |
|
3113 | entry = (new_offset_flags,) + entry[1:8] | |
3113 | entry += (current_offset, len(serialized_sidedata)) |
|
3114 | entry += (current_offset, len(serialized_sidedata)) | |
3114 |
|
3115 | |||
3115 | fp.write(serialized_sidedata) |
|
3116 | fp.write(serialized_sidedata) | |
3116 | new_entries.append(entry) |
|
3117 | new_entries.append(entry) | |
3117 | current_offset += len(serialized_sidedata) |
|
3118 | current_offset += len(serialized_sidedata) | |
3118 |
|
3119 | |||
3119 | # rewrite the new index entries |
|
3120 | # rewrite the new index entries | |
3120 | with self._indexfp(b'w+') as fp: |
|
3121 | with self._indexfp(b'w+') as fp: | |
3121 | fp.seek(startrev * self.index.entry_size) |
|
3122 | fp.seek(startrev * self.index.entry_size) | |
3122 | for i, e in enumerate(new_entries): |
|
3123 | for i, e in enumerate(new_entries): | |
3123 | rev = startrev + i |
|
3124 | rev = startrev + i | |
3124 | self.index.replace_sidedata_info(rev, e[8], e[9], e[0]) |
|
3125 | self.index.replace_sidedata_info(rev, e[8], e[9], e[0]) | |
3125 | packed = self.index.entry_binary(rev) |
|
3126 | packed = self.index.entry_binary(rev) | |
3126 | if rev == 0: |
|
3127 | if rev == 0: | |
3127 | header = self.index.pack_header(self.version) |
|
3128 | header = self.index.pack_header(self.version) | |
3128 | packed = header + packed |
|
3129 | packed = header + packed | |
3129 | fp.write(packed) |
|
3130 | fp.write(packed) |
@@ -1,155 +1,175 b'' | |||||
1 | # sidedata.py - Logic around store extra data alongside revlog revisions |
|
1 | # sidedata.py - Logic around store extra data alongside revlog revisions | |
2 | # |
|
2 | # | |
3 | # Copyright 2019 Pierre-Yves David <pierre-yves.david@octobus.net) |
|
3 | # Copyright 2019 Pierre-Yves David <pierre-yves.david@octobus.net) | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 | """core code for "sidedata" support |
|
7 | """core code for "sidedata" support | |
8 |
|
8 | |||
9 | The "sidedata" are stored alongside the revision without actually being part of |
|
9 | The "sidedata" are stored alongside the revision without actually being part of | |
10 | its content and not affecting its hash. It's main use cases is to cache |
|
10 | its content and not affecting its hash. It's main use cases is to cache | |
11 | important information related to a changesets. |
|
11 | important information related to a changesets. | |
12 |
|
12 | |||
13 | The current implementation is experimental and subject to changes. Do not rely |
|
13 | The current implementation is experimental and subject to changes. Do not rely | |
14 | on it in production. |
|
14 | on it in production. | |
15 |
|
15 | |||
16 | Sidedata are stored in the revlog itself, thanks to a new version of the |
|
16 | Sidedata are stored in the revlog itself, thanks to a new version of the | |
17 | revlog. The following format is currently used:: |
|
17 | revlog. The following format is currently used:: | |
18 |
|
18 | |||
19 | initial header: |
|
19 | initial header: | |
20 | <number of sidedata; 2 bytes> |
|
20 | <number of sidedata; 2 bytes> | |
21 | sidedata (repeated N times): |
|
21 | sidedata (repeated N times): | |
22 | <sidedata-key; 2 bytes> |
|
22 | <sidedata-key; 2 bytes> | |
23 | <sidedata-entry-length: 4 bytes> |
|
23 | <sidedata-entry-length: 4 bytes> | |
24 | <sidedata-content-sha1-digest: 20 bytes> |
|
24 | <sidedata-content-sha1-digest: 20 bytes> | |
25 | <sidedata-content; X bytes> |
|
25 | <sidedata-content; X bytes> | |
26 | normal raw text: |
|
26 | normal raw text: | |
27 | <all bytes remaining in the rawtext> |
|
27 | <all bytes remaining in the rawtext> | |
28 |
|
28 | |||
29 | This is a simple and effective format. It should be enough to experiment with |
|
29 | This is a simple and effective format. It should be enough to experiment with | |
30 | the concept. |
|
30 | the concept. | |
31 | """ |
|
31 | """ | |
32 |
|
32 | |||
33 | from __future__ import absolute_import |
|
33 | from __future__ import absolute_import | |
34 |
|
34 | |||
35 | import collections |
|
35 | import collections | |
36 | import struct |
|
36 | import struct | |
37 |
|
37 | |||
38 | from .. import error, requirements as requirementsmod |
|
38 | from .. import error, requirements as requirementsmod | |
39 | from ..revlogutils import constants, flagutil |
|
39 | from ..revlogutils import constants, flagutil | |
40 | from ..utils import hashutil |
|
40 | from ..utils import hashutil | |
41 |
|
41 | |||
42 | ## sidedata type constant |
|
42 | ## sidedata type constant | |
43 | # reserve a block for testing purposes. |
|
43 | # reserve a block for testing purposes. | |
44 | SD_TEST1 = 1 |
|
44 | SD_TEST1 = 1 | |
45 | SD_TEST2 = 2 |
|
45 | SD_TEST2 = 2 | |
46 | SD_TEST3 = 3 |
|
46 | SD_TEST3 = 3 | |
47 | SD_TEST4 = 4 |
|
47 | SD_TEST4 = 4 | |
48 | SD_TEST5 = 5 |
|
48 | SD_TEST5 = 5 | |
49 | SD_TEST6 = 6 |
|
49 | SD_TEST6 = 6 | |
50 | SD_TEST7 = 7 |
|
50 | SD_TEST7 = 7 | |
51 |
|
51 | |||
52 | # key to store copies related information |
|
52 | # key to store copies related information | |
53 | SD_P1COPIES = 8 |
|
53 | SD_P1COPIES = 8 | |
54 | SD_P2COPIES = 9 |
|
54 | SD_P2COPIES = 9 | |
55 | SD_FILESADDED = 10 |
|
55 | SD_FILESADDED = 10 | |
56 | SD_FILESREMOVED = 11 |
|
56 | SD_FILESREMOVED = 11 | |
57 | SD_FILES = 12 |
|
57 | SD_FILES = 12 | |
58 |
|
58 | |||
59 | # internal format constant |
|
59 | # internal format constant | |
60 | SIDEDATA_HEADER = struct.Struct('>H') |
|
60 | SIDEDATA_HEADER = struct.Struct('>H') | |
61 | SIDEDATA_ENTRY = struct.Struct('>HL20s') |
|
61 | SIDEDATA_ENTRY = struct.Struct('>HL20s') | |
62 |
|
62 | |||
63 |
|
63 | |||
64 | def serialize_sidedata(sidedata): |
|
64 | def serialize_sidedata(sidedata): | |
65 | sidedata = list(sidedata.items()) |
|
65 | sidedata = list(sidedata.items()) | |
66 | sidedata.sort() |
|
66 | sidedata.sort() | |
67 | buf = [SIDEDATA_HEADER.pack(len(sidedata))] |
|
67 | buf = [SIDEDATA_HEADER.pack(len(sidedata))] | |
68 | for key, value in sidedata: |
|
68 | for key, value in sidedata: | |
69 | digest = hashutil.sha1(value).digest() |
|
69 | digest = hashutil.sha1(value).digest() | |
70 | buf.append(SIDEDATA_ENTRY.pack(key, len(value), digest)) |
|
70 | buf.append(SIDEDATA_ENTRY.pack(key, len(value), digest)) | |
71 | for key, value in sidedata: |
|
71 | for key, value in sidedata: | |
72 | buf.append(value) |
|
72 | buf.append(value) | |
73 | buf = b''.join(buf) |
|
73 | buf = b''.join(buf) | |
74 | return buf |
|
74 | return buf | |
75 |
|
75 | |||
76 |
|
76 | |||
77 | def deserialize_sidedata(blob): |
|
77 | def deserialize_sidedata(blob): | |
78 | sidedata = {} |
|
78 | sidedata = {} | |
79 | offset = 0 |
|
79 | offset = 0 | |
80 | (nbentry,) = SIDEDATA_HEADER.unpack(blob[: SIDEDATA_HEADER.size]) |
|
80 | (nbentry,) = SIDEDATA_HEADER.unpack(blob[: SIDEDATA_HEADER.size]) | |
81 | offset += SIDEDATA_HEADER.size |
|
81 | offset += SIDEDATA_HEADER.size | |
82 | dataoffset = SIDEDATA_HEADER.size + (SIDEDATA_ENTRY.size * nbentry) |
|
82 | dataoffset = SIDEDATA_HEADER.size + (SIDEDATA_ENTRY.size * nbentry) | |
83 | for i in range(nbentry): |
|
83 | for i in range(nbentry): | |
84 | nextoffset = offset + SIDEDATA_ENTRY.size |
|
84 | nextoffset = offset + SIDEDATA_ENTRY.size | |
85 | key, size, storeddigest = SIDEDATA_ENTRY.unpack(blob[offset:nextoffset]) |
|
85 | key, size, storeddigest = SIDEDATA_ENTRY.unpack(blob[offset:nextoffset]) | |
86 | offset = nextoffset |
|
86 | offset = nextoffset | |
87 | # read the data associated with that entry |
|
87 | # read the data associated with that entry | |
88 | nextdataoffset = dataoffset + size |
|
88 | nextdataoffset = dataoffset + size | |
89 | entrytext = bytes(blob[dataoffset:nextdataoffset]) |
|
89 | entrytext = bytes(blob[dataoffset:nextdataoffset]) | |
90 | readdigest = hashutil.sha1(entrytext).digest() |
|
90 | readdigest = hashutil.sha1(entrytext).digest() | |
91 | if storeddigest != readdigest: |
|
91 | if storeddigest != readdigest: | |
92 | raise error.SidedataHashError(key, storeddigest, readdigest) |
|
92 | raise error.SidedataHashError(key, storeddigest, readdigest) | |
93 | sidedata[key] = entrytext |
|
93 | sidedata[key] = entrytext | |
94 | dataoffset = nextdataoffset |
|
94 | dataoffset = nextdataoffset | |
95 | return sidedata |
|
95 | return sidedata | |
96 |
|
96 | |||
97 |
|
97 | |||
98 | def get_sidedata_helpers(repo, remote_sd_categories, pull=False): |
|
98 | def get_sidedata_helpers(repo, remote_sd_categories, pull=False): | |
|
99 | """ | |||
|
100 | Returns a dictionary mapping revlog types to tuples of | |||
|
101 | `(repo, computers, removers)`: | |||
|
102 | * `repo` is used as an argument for computers | |||
|
103 | * `computers` is a list of `(category, (keys, computer, flags)` that | |||
|
104 | compute the missing sidedata categories that were asked: | |||
|
105 | * `category` is the sidedata category | |||
|
106 | * `keys` are the sidedata keys to be affected | |||
|
107 | * `flags` is a bitmask (an integer) of flags to remove when | |||
|
108 | removing the category. | |||
|
109 | * `computer` is the function `(repo, store, rev, sidedata)` that | |||
|
110 | returns a tuple of | |||
|
111 | `(new sidedata dict, (flags to add, flags to remove))`. | |||
|
112 | For example, it will return `({}, (0, 1 << 15))` to return no | |||
|
113 | sidedata, with no flags to add and one flag to remove. | |||
|
114 | * `removers` will remove the keys corresponding to the categories | |||
|
115 | that are present, but not needed. | |||
|
116 | If both `computers` and `removers` are empty, sidedata will simply not | |||
|
117 | be transformed. | |||
|
118 | """ | |||
99 | # Computers for computing sidedata on-the-fly |
|
119 | # Computers for computing sidedata on-the-fly | |
100 | sd_computers = collections.defaultdict(list) |
|
120 | sd_computers = collections.defaultdict(list) | |
101 | # Computers for categories to remove from sidedata |
|
121 | # Computers for categories to remove from sidedata | |
102 | sd_removers = collections.defaultdict(list) |
|
122 | sd_removers = collections.defaultdict(list) | |
103 | to_generate = remote_sd_categories - repo._wanted_sidedata |
|
123 | to_generate = remote_sd_categories - repo._wanted_sidedata | |
104 | to_remove = repo._wanted_sidedata - remote_sd_categories |
|
124 | to_remove = repo._wanted_sidedata - remote_sd_categories | |
105 | if pull: |
|
125 | if pull: | |
106 | to_generate, to_remove = to_remove, to_generate |
|
126 | to_generate, to_remove = to_remove, to_generate | |
107 |
|
127 | |||
108 | for revlog_kind, computers in repo._sidedata_computers.items(): |
|
128 | for revlog_kind, computers in repo._sidedata_computers.items(): | |
109 | for category, computer in computers.items(): |
|
129 | for category, computer in computers.items(): | |
110 | if category in to_generate: |
|
130 | if category in to_generate: | |
111 | sd_computers[revlog_kind].append(computer) |
|
131 | sd_computers[revlog_kind].append(computer) | |
112 | if category in to_remove: |
|
132 | if category in to_remove: | |
113 | sd_removers[revlog_kind].append(computer) |
|
133 | sd_removers[revlog_kind].append(computer) | |
114 |
|
134 | |||
115 | sidedata_helpers = (repo, sd_computers, sd_removers) |
|
135 | sidedata_helpers = (repo, sd_computers, sd_removers) | |
116 | return sidedata_helpers |
|
136 | return sidedata_helpers | |
117 |
|
137 | |||
118 |
|
138 | |||
119 | def run_sidedata_helpers(store, sidedata_helpers, sidedata, rev): |
|
139 | def run_sidedata_helpers(store, sidedata_helpers, sidedata, rev): | |
120 | """Returns the sidedata for the given revision after running through |
|
140 | """Returns the sidedata for the given revision after running through | |
121 | the given helpers. |
|
141 | the given helpers. | |
122 | - `store`: the revlog this applies to (changelog, manifest, or filelog |
|
142 | - `store`: the revlog this applies to (changelog, manifest, or filelog | |
123 | instance) |
|
143 | instance) | |
124 |
- `sidedata_helpers`: see ` |
|
144 | - `sidedata_helpers`: see `get_sidedata_helpers` | |
125 | - `sidedata`: previous sidedata at the given rev, if any |
|
145 | - `sidedata`: previous sidedata at the given rev, if any | |
126 | - `rev`: affected rev of `store` |
|
146 | - `rev`: affected rev of `store` | |
127 | """ |
|
147 | """ | |
128 | repo, sd_computers, sd_removers = sidedata_helpers |
|
148 | repo, sd_computers, sd_removers = sidedata_helpers | |
129 | kind = store.revlog_kind |
|
149 | kind = store.revlog_kind | |
130 | flags_to_add = 0 |
|
150 | flags_to_add = 0 | |
131 | flags_to_remove = 0 |
|
151 | flags_to_remove = 0 | |
132 | for _keys, sd_computer, _flags in sd_computers.get(kind, []): |
|
152 | for _keys, sd_computer, _flags in sd_computers.get(kind, []): | |
133 | sidedata, flags = sd_computer(repo, store, rev, sidedata) |
|
153 | sidedata, flags = sd_computer(repo, store, rev, sidedata) | |
134 | flags_to_add |= flags[0] |
|
154 | flags_to_add |= flags[0] | |
135 | flags_to_remove |= flags[1] |
|
155 | flags_to_remove |= flags[1] | |
136 | for keys, _computer, flags in sd_removers.get(kind, []): |
|
156 | for keys, _computer, flags in sd_removers.get(kind, []): | |
137 | for key in keys: |
|
157 | for key in keys: | |
138 | sidedata.pop(key, None) |
|
158 | sidedata.pop(key, None) | |
139 | flags_to_remove |= flags |
|
159 | flags_to_remove |= flags | |
140 | return sidedata, (flags_to_add, flags_to_remove) |
|
160 | return sidedata, (flags_to_add, flags_to_remove) | |
141 |
|
161 | |||
142 |
|
162 | |||
143 | def set_sidedata_spec_for_repo(repo): |
|
163 | def set_sidedata_spec_for_repo(repo): | |
144 | # prevent cycle metadata -> revlogutils.sidedata -> metadata |
|
164 | # prevent cycle metadata -> revlogutils.sidedata -> metadata | |
145 | from .. import metadata |
|
165 | from .. import metadata | |
146 |
|
166 | |||
147 | if requirementsmod.COPIESSDC_REQUIREMENT in repo.requirements: |
|
167 | if requirementsmod.COPIESSDC_REQUIREMENT in repo.requirements: | |
148 | repo.register_wanted_sidedata(SD_FILES) |
|
168 | repo.register_wanted_sidedata(SD_FILES) | |
149 | repo.register_sidedata_computer( |
|
169 | repo.register_sidedata_computer( | |
150 | constants.KIND_CHANGELOG, |
|
170 | constants.KIND_CHANGELOG, | |
151 | SD_FILES, |
|
171 | SD_FILES, | |
152 | (SD_FILES,), |
|
172 | (SD_FILES,), | |
153 | metadata.copies_sidedata_computer, |
|
173 | metadata.copies_sidedata_computer, | |
154 | flagutil.REVIDX_HASCOPIESINFO, |
|
174 | flagutil.REVIDX_HASCOPIESINFO, | |
155 | ) |
|
175 | ) |
@@ -1,561 +1,544 b'' | |||||
1 | # storageutil.py - Storage functionality agnostic of backend implementation. |
|
1 | # storageutil.py - Storage functionality agnostic of backend implementation. | |
2 | # |
|
2 | # | |
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> |
|
3 | # Copyright 2018 Gregory Szorc <gregory.szorc@gmail.com> | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import re |
|
10 | import re | |
11 | import struct |
|
11 | import struct | |
12 |
|
12 | |||
13 | from ..i18n import _ |
|
13 | from ..i18n import _ | |
14 | from ..node import ( |
|
14 | from ..node import ( | |
15 | bin, |
|
15 | bin, | |
16 | nullrev, |
|
16 | nullrev, | |
17 | sha1nodeconstants, |
|
17 | sha1nodeconstants, | |
18 | ) |
|
18 | ) | |
19 | from .. import ( |
|
19 | from .. import ( | |
20 | dagop, |
|
20 | dagop, | |
21 | error, |
|
21 | error, | |
22 | mdiff, |
|
22 | mdiff, | |
23 | pycompat, |
|
23 | pycompat, | |
24 | ) |
|
24 | ) | |
25 | from ..interfaces import repository |
|
25 | from ..interfaces import repository | |
26 | from ..revlogutils import sidedata as sidedatamod |
|
26 | from ..revlogutils import sidedata as sidedatamod | |
27 | from ..utils import hashutil |
|
27 | from ..utils import hashutil | |
28 |
|
28 | |||
29 | _nullhash = hashutil.sha1(sha1nodeconstants.nullid) |
|
29 | _nullhash = hashutil.sha1(sha1nodeconstants.nullid) | |
30 |
|
30 | |||
31 | # revision data contains extra metadata not part of the official digest |
|
31 | # revision data contains extra metadata not part of the official digest | |
32 | # Only used in changegroup >= v4. |
|
32 | # Only used in changegroup >= v4. | |
33 | CG_FLAG_SIDEDATA = 1 |
|
33 | CG_FLAG_SIDEDATA = 1 | |
34 |
|
34 | |||
35 |
|
35 | |||
36 | def hashrevisionsha1(text, p1, p2): |
|
36 | def hashrevisionsha1(text, p1, p2): | |
37 | """Compute the SHA-1 for revision data and its parents. |
|
37 | """Compute the SHA-1 for revision data and its parents. | |
38 |
|
38 | |||
39 | This hash combines both the current file contents and its history |
|
39 | This hash combines both the current file contents and its history | |
40 | in a manner that makes it easy to distinguish nodes with the same |
|
40 | in a manner that makes it easy to distinguish nodes with the same | |
41 | content in the revision graph. |
|
41 | content in the revision graph. | |
42 | """ |
|
42 | """ | |
43 | # As of now, if one of the parent node is null, p2 is null |
|
43 | # As of now, if one of the parent node is null, p2 is null | |
44 | if p2 == sha1nodeconstants.nullid: |
|
44 | if p2 == sha1nodeconstants.nullid: | |
45 | # deep copy of a hash is faster than creating one |
|
45 | # deep copy of a hash is faster than creating one | |
46 | s = _nullhash.copy() |
|
46 | s = _nullhash.copy() | |
47 | s.update(p1) |
|
47 | s.update(p1) | |
48 | else: |
|
48 | else: | |
49 | # none of the parent nodes are nullid |
|
49 | # none of the parent nodes are nullid | |
50 | if p1 < p2: |
|
50 | if p1 < p2: | |
51 | a = p1 |
|
51 | a = p1 | |
52 | b = p2 |
|
52 | b = p2 | |
53 | else: |
|
53 | else: | |
54 | a = p2 |
|
54 | a = p2 | |
55 | b = p1 |
|
55 | b = p1 | |
56 | s = hashutil.sha1(a) |
|
56 | s = hashutil.sha1(a) | |
57 | s.update(b) |
|
57 | s.update(b) | |
58 | s.update(text) |
|
58 | s.update(text) | |
59 | return s.digest() |
|
59 | return s.digest() | |
60 |
|
60 | |||
61 |
|
61 | |||
62 | METADATA_RE = re.compile(b'\x01\n') |
|
62 | METADATA_RE = re.compile(b'\x01\n') | |
63 |
|
63 | |||
64 |
|
64 | |||
65 | def parsemeta(text): |
|
65 | def parsemeta(text): | |
66 | """Parse metadata header from revision data. |
|
66 | """Parse metadata header from revision data. | |
67 |
|
67 | |||
68 | Returns a 2-tuple of (metadata, offset), where both can be None if there |
|
68 | Returns a 2-tuple of (metadata, offset), where both can be None if there | |
69 | is no metadata. |
|
69 | is no metadata. | |
70 | """ |
|
70 | """ | |
71 | # text can be buffer, so we can't use .startswith or .index |
|
71 | # text can be buffer, so we can't use .startswith or .index | |
72 | if text[:2] != b'\x01\n': |
|
72 | if text[:2] != b'\x01\n': | |
73 | return None, None |
|
73 | return None, None | |
74 | s = METADATA_RE.search(text, 2).start() |
|
74 | s = METADATA_RE.search(text, 2).start() | |
75 | mtext = text[2:s] |
|
75 | mtext = text[2:s] | |
76 | meta = {} |
|
76 | meta = {} | |
77 | for l in mtext.splitlines(): |
|
77 | for l in mtext.splitlines(): | |
78 | k, v = l.split(b': ', 1) |
|
78 | k, v = l.split(b': ', 1) | |
79 | meta[k] = v |
|
79 | meta[k] = v | |
80 | return meta, s + 2 |
|
80 | return meta, s + 2 | |
81 |
|
81 | |||
82 |
|
82 | |||
83 | def packmeta(meta, text): |
|
83 | def packmeta(meta, text): | |
84 | """Add metadata to fulltext to produce revision text.""" |
|
84 | """Add metadata to fulltext to produce revision text.""" | |
85 | keys = sorted(meta) |
|
85 | keys = sorted(meta) | |
86 | metatext = b''.join(b'%s: %s\n' % (k, meta[k]) for k in keys) |
|
86 | metatext = b''.join(b'%s: %s\n' % (k, meta[k]) for k in keys) | |
87 | return b'\x01\n%s\x01\n%s' % (metatext, text) |
|
87 | return b'\x01\n%s\x01\n%s' % (metatext, text) | |
88 |
|
88 | |||
89 |
|
89 | |||
90 | def iscensoredtext(text): |
|
90 | def iscensoredtext(text): | |
91 | meta = parsemeta(text)[0] |
|
91 | meta = parsemeta(text)[0] | |
92 | return meta and b'censored' in meta |
|
92 | return meta and b'censored' in meta | |
93 |
|
93 | |||
94 |
|
94 | |||
95 | def filtermetadata(text): |
|
95 | def filtermetadata(text): | |
96 | """Extract just the revision data from source text. |
|
96 | """Extract just the revision data from source text. | |
97 |
|
97 | |||
98 | Returns ``text`` unless it has a metadata header, in which case we return |
|
98 | Returns ``text`` unless it has a metadata header, in which case we return | |
99 | a new buffer without hte metadata. |
|
99 | a new buffer without hte metadata. | |
100 | """ |
|
100 | """ | |
101 | if not text.startswith(b'\x01\n'): |
|
101 | if not text.startswith(b'\x01\n'): | |
102 | return text |
|
102 | return text | |
103 |
|
103 | |||
104 | offset = text.index(b'\x01\n', 2) |
|
104 | offset = text.index(b'\x01\n', 2) | |
105 | return text[offset + 2 :] |
|
105 | return text[offset + 2 :] | |
106 |
|
106 | |||
107 |
|
107 | |||
108 | def filerevisioncopied(store, node): |
|
108 | def filerevisioncopied(store, node): | |
109 | """Resolve file revision copy metadata. |
|
109 | """Resolve file revision copy metadata. | |
110 |
|
110 | |||
111 | Returns ``False`` if the file has no copy metadata. Otherwise a |
|
111 | Returns ``False`` if the file has no copy metadata. Otherwise a | |
112 | 2-tuple of the source filename and node. |
|
112 | 2-tuple of the source filename and node. | |
113 | """ |
|
113 | """ | |
114 | if store.parents(node)[0] != sha1nodeconstants.nullid: |
|
114 | if store.parents(node)[0] != sha1nodeconstants.nullid: | |
115 | return False |
|
115 | return False | |
116 |
|
116 | |||
117 | meta = parsemeta(store.revision(node))[0] |
|
117 | meta = parsemeta(store.revision(node))[0] | |
118 |
|
118 | |||
119 | # copy and copyrev occur in pairs. In rare cases due to old bugs, |
|
119 | # copy and copyrev occur in pairs. In rare cases due to old bugs, | |
120 | # one can occur without the other. So ensure both are present to flag |
|
120 | # one can occur without the other. So ensure both are present to flag | |
121 | # as a copy. |
|
121 | # as a copy. | |
122 | if meta and b'copy' in meta and b'copyrev' in meta: |
|
122 | if meta and b'copy' in meta and b'copyrev' in meta: | |
123 | return meta[b'copy'], bin(meta[b'copyrev']) |
|
123 | return meta[b'copy'], bin(meta[b'copyrev']) | |
124 |
|
124 | |||
125 | return False |
|
125 | return False | |
126 |
|
126 | |||
127 |
|
127 | |||
128 | def filedataequivalent(store, node, filedata): |
|
128 | def filedataequivalent(store, node, filedata): | |
129 | """Determines whether file data is equivalent to a stored node. |
|
129 | """Determines whether file data is equivalent to a stored node. | |
130 |
|
130 | |||
131 | Returns True if the passed file data would hash to the same value |
|
131 | Returns True if the passed file data would hash to the same value | |
132 | as a stored revision and False otherwise. |
|
132 | as a stored revision and False otherwise. | |
133 |
|
133 | |||
134 | When a stored revision is censored, filedata must be empty to have |
|
134 | When a stored revision is censored, filedata must be empty to have | |
135 | equivalence. |
|
135 | equivalence. | |
136 |
|
136 | |||
137 | When a stored revision has copy metadata, it is ignored as part |
|
137 | When a stored revision has copy metadata, it is ignored as part | |
138 | of the compare. |
|
138 | of the compare. | |
139 | """ |
|
139 | """ | |
140 |
|
140 | |||
141 | if filedata.startswith(b'\x01\n'): |
|
141 | if filedata.startswith(b'\x01\n'): | |
142 | revisiontext = b'\x01\n\x01\n' + filedata |
|
142 | revisiontext = b'\x01\n\x01\n' + filedata | |
143 | else: |
|
143 | else: | |
144 | revisiontext = filedata |
|
144 | revisiontext = filedata | |
145 |
|
145 | |||
146 | p1, p2 = store.parents(node) |
|
146 | p1, p2 = store.parents(node) | |
147 |
|
147 | |||
148 | computednode = hashrevisionsha1(revisiontext, p1, p2) |
|
148 | computednode = hashrevisionsha1(revisiontext, p1, p2) | |
149 |
|
149 | |||
150 | if computednode == node: |
|
150 | if computednode == node: | |
151 | return True |
|
151 | return True | |
152 |
|
152 | |||
153 | # Censored files compare against the empty file. |
|
153 | # Censored files compare against the empty file. | |
154 | if store.iscensored(store.rev(node)): |
|
154 | if store.iscensored(store.rev(node)): | |
155 | return filedata == b'' |
|
155 | return filedata == b'' | |
156 |
|
156 | |||
157 | # Renaming a file produces a different hash, even if the data |
|
157 | # Renaming a file produces a different hash, even if the data | |
158 | # remains unchanged. Check if that's the case. |
|
158 | # remains unchanged. Check if that's the case. | |
159 | if store.renamed(node): |
|
159 | if store.renamed(node): | |
160 | return store.read(node) == filedata |
|
160 | return store.read(node) == filedata | |
161 |
|
161 | |||
162 | return False |
|
162 | return False | |
163 |
|
163 | |||
164 |
|
164 | |||
165 | def iterrevs(storelen, start=0, stop=None): |
|
165 | def iterrevs(storelen, start=0, stop=None): | |
166 | """Iterate over revision numbers in a store.""" |
|
166 | """Iterate over revision numbers in a store.""" | |
167 | step = 1 |
|
167 | step = 1 | |
168 |
|
168 | |||
169 | if stop is not None: |
|
169 | if stop is not None: | |
170 | if start > stop: |
|
170 | if start > stop: | |
171 | step = -1 |
|
171 | step = -1 | |
172 | stop += step |
|
172 | stop += step | |
173 | if stop > storelen: |
|
173 | if stop > storelen: | |
174 | stop = storelen |
|
174 | stop = storelen | |
175 | else: |
|
175 | else: | |
176 | stop = storelen |
|
176 | stop = storelen | |
177 |
|
177 | |||
178 | return pycompat.xrange(start, stop, step) |
|
178 | return pycompat.xrange(start, stop, step) | |
179 |
|
179 | |||
180 |
|
180 | |||
181 | def fileidlookup(store, fileid, identifier): |
|
181 | def fileidlookup(store, fileid, identifier): | |
182 | """Resolve the file node for a value. |
|
182 | """Resolve the file node for a value. | |
183 |
|
183 | |||
184 | ``store`` is an object implementing the ``ifileindex`` interface. |
|
184 | ``store`` is an object implementing the ``ifileindex`` interface. | |
185 |
|
185 | |||
186 | ``fileid`` can be: |
|
186 | ``fileid`` can be: | |
187 |
|
187 | |||
188 | * A 20 or 32 byte binary node. |
|
188 | * A 20 or 32 byte binary node. | |
189 | * An integer revision number |
|
189 | * An integer revision number | |
190 | * A 40 or 64 byte hex node. |
|
190 | * A 40 or 64 byte hex node. | |
191 | * A bytes that can be parsed as an integer representing a revision number. |
|
191 | * A bytes that can be parsed as an integer representing a revision number. | |
192 |
|
192 | |||
193 | ``identifier`` is used to populate ``error.LookupError`` with an identifier |
|
193 | ``identifier`` is used to populate ``error.LookupError`` with an identifier | |
194 | for the store. |
|
194 | for the store. | |
195 |
|
195 | |||
196 | Raises ``error.LookupError`` on failure. |
|
196 | Raises ``error.LookupError`` on failure. | |
197 | """ |
|
197 | """ | |
198 | if isinstance(fileid, int): |
|
198 | if isinstance(fileid, int): | |
199 | try: |
|
199 | try: | |
200 | return store.node(fileid) |
|
200 | return store.node(fileid) | |
201 | except IndexError: |
|
201 | except IndexError: | |
202 | raise error.LookupError( |
|
202 | raise error.LookupError( | |
203 | b'%d' % fileid, identifier, _(b'no match found') |
|
203 | b'%d' % fileid, identifier, _(b'no match found') | |
204 | ) |
|
204 | ) | |
205 |
|
205 | |||
206 | if len(fileid) in (20, 32): |
|
206 | if len(fileid) in (20, 32): | |
207 | try: |
|
207 | try: | |
208 | store.rev(fileid) |
|
208 | store.rev(fileid) | |
209 | return fileid |
|
209 | return fileid | |
210 | except error.LookupError: |
|
210 | except error.LookupError: | |
211 | pass |
|
211 | pass | |
212 |
|
212 | |||
213 | if len(fileid) in (40, 64): |
|
213 | if len(fileid) in (40, 64): | |
214 | try: |
|
214 | try: | |
215 | rawnode = bin(fileid) |
|
215 | rawnode = bin(fileid) | |
216 | store.rev(rawnode) |
|
216 | store.rev(rawnode) | |
217 | return rawnode |
|
217 | return rawnode | |
218 | except TypeError: |
|
218 | except TypeError: | |
219 | pass |
|
219 | pass | |
220 |
|
220 | |||
221 | try: |
|
221 | try: | |
222 | rev = int(fileid) |
|
222 | rev = int(fileid) | |
223 |
|
223 | |||
224 | if b'%d' % rev != fileid: |
|
224 | if b'%d' % rev != fileid: | |
225 | raise ValueError |
|
225 | raise ValueError | |
226 |
|
226 | |||
227 | try: |
|
227 | try: | |
228 | return store.node(rev) |
|
228 | return store.node(rev) | |
229 | except (IndexError, TypeError): |
|
229 | except (IndexError, TypeError): | |
230 | pass |
|
230 | pass | |
231 | except (ValueError, OverflowError): |
|
231 | except (ValueError, OverflowError): | |
232 | pass |
|
232 | pass | |
233 |
|
233 | |||
234 | raise error.LookupError(fileid, identifier, _(b'no match found')) |
|
234 | raise error.LookupError(fileid, identifier, _(b'no match found')) | |
235 |
|
235 | |||
236 |
|
236 | |||
237 | def resolvestripinfo(minlinkrev, tiprev, headrevs, linkrevfn, parentrevsfn): |
|
237 | def resolvestripinfo(minlinkrev, tiprev, headrevs, linkrevfn, parentrevsfn): | |
238 | """Resolve information needed to strip revisions. |
|
238 | """Resolve information needed to strip revisions. | |
239 |
|
239 | |||
240 | Finds the minimum revision number that must be stripped in order to |
|
240 | Finds the minimum revision number that must be stripped in order to | |
241 | strip ``minlinkrev``. |
|
241 | strip ``minlinkrev``. | |
242 |
|
242 | |||
243 | Returns a 2-tuple of the minimum revision number to do that and a set |
|
243 | Returns a 2-tuple of the minimum revision number to do that and a set | |
244 | of all revision numbers that have linkrevs that would be broken |
|
244 | of all revision numbers that have linkrevs that would be broken | |
245 | by that strip. |
|
245 | by that strip. | |
246 |
|
246 | |||
247 | ``tiprev`` is the current tip-most revision. It is ``len(store) - 1``. |
|
247 | ``tiprev`` is the current tip-most revision. It is ``len(store) - 1``. | |
248 | ``headrevs`` is an iterable of head revisions. |
|
248 | ``headrevs`` is an iterable of head revisions. | |
249 | ``linkrevfn`` is a callable that receives a revision and returns a linked |
|
249 | ``linkrevfn`` is a callable that receives a revision and returns a linked | |
250 | revision. |
|
250 | revision. | |
251 | ``parentrevsfn`` is a callable that receives a revision number and returns |
|
251 | ``parentrevsfn`` is a callable that receives a revision number and returns | |
252 | an iterable of its parent revision numbers. |
|
252 | an iterable of its parent revision numbers. | |
253 | """ |
|
253 | """ | |
254 | brokenrevs = set() |
|
254 | brokenrevs = set() | |
255 | strippoint = tiprev + 1 |
|
255 | strippoint = tiprev + 1 | |
256 |
|
256 | |||
257 | heads = {} |
|
257 | heads = {} | |
258 | futurelargelinkrevs = set() |
|
258 | futurelargelinkrevs = set() | |
259 | for head in headrevs: |
|
259 | for head in headrevs: | |
260 | headlinkrev = linkrevfn(head) |
|
260 | headlinkrev = linkrevfn(head) | |
261 | heads[head] = headlinkrev |
|
261 | heads[head] = headlinkrev | |
262 | if headlinkrev >= minlinkrev: |
|
262 | if headlinkrev >= minlinkrev: | |
263 | futurelargelinkrevs.add(headlinkrev) |
|
263 | futurelargelinkrevs.add(headlinkrev) | |
264 |
|
264 | |||
265 | # This algorithm involves walking down the rev graph, starting at the |
|
265 | # This algorithm involves walking down the rev graph, starting at the | |
266 | # heads. Since the revs are topologically sorted according to linkrev, |
|
266 | # heads. Since the revs are topologically sorted according to linkrev, | |
267 | # once all head linkrevs are below the minlink, we know there are |
|
267 | # once all head linkrevs are below the minlink, we know there are | |
268 | # no more revs that could have a linkrev greater than minlink. |
|
268 | # no more revs that could have a linkrev greater than minlink. | |
269 | # So we can stop walking. |
|
269 | # So we can stop walking. | |
270 | while futurelargelinkrevs: |
|
270 | while futurelargelinkrevs: | |
271 | strippoint -= 1 |
|
271 | strippoint -= 1 | |
272 | linkrev = heads.pop(strippoint) |
|
272 | linkrev = heads.pop(strippoint) | |
273 |
|
273 | |||
274 | if linkrev < minlinkrev: |
|
274 | if linkrev < minlinkrev: | |
275 | brokenrevs.add(strippoint) |
|
275 | brokenrevs.add(strippoint) | |
276 | else: |
|
276 | else: | |
277 | futurelargelinkrevs.remove(linkrev) |
|
277 | futurelargelinkrevs.remove(linkrev) | |
278 |
|
278 | |||
279 | for p in parentrevsfn(strippoint): |
|
279 | for p in parentrevsfn(strippoint): | |
280 | if p != nullrev: |
|
280 | if p != nullrev: | |
281 | plinkrev = linkrevfn(p) |
|
281 | plinkrev = linkrevfn(p) | |
282 | heads[p] = plinkrev |
|
282 | heads[p] = plinkrev | |
283 | if plinkrev >= minlinkrev: |
|
283 | if plinkrev >= minlinkrev: | |
284 | futurelargelinkrevs.add(plinkrev) |
|
284 | futurelargelinkrevs.add(plinkrev) | |
285 |
|
285 | |||
286 | return strippoint, brokenrevs |
|
286 | return strippoint, brokenrevs | |
287 |
|
287 | |||
288 |
|
288 | |||
289 | def emitrevisions( |
|
289 | def emitrevisions( | |
290 | store, |
|
290 | store, | |
291 | nodes, |
|
291 | nodes, | |
292 | nodesorder, |
|
292 | nodesorder, | |
293 | resultcls, |
|
293 | resultcls, | |
294 | deltaparentfn=None, |
|
294 | deltaparentfn=None, | |
295 | candeltafn=None, |
|
295 | candeltafn=None, | |
296 | rawsizefn=None, |
|
296 | rawsizefn=None, | |
297 | revdifffn=None, |
|
297 | revdifffn=None, | |
298 | flagsfn=None, |
|
298 | flagsfn=None, | |
299 | deltamode=repository.CG_DELTAMODE_STD, |
|
299 | deltamode=repository.CG_DELTAMODE_STD, | |
300 | revisiondata=False, |
|
300 | revisiondata=False, | |
301 | assumehaveparentrevisions=False, |
|
301 | assumehaveparentrevisions=False, | |
302 | sidedata_helpers=None, |
|
302 | sidedata_helpers=None, | |
303 | ): |
|
303 | ): | |
304 | """Generic implementation of ifiledata.emitrevisions(). |
|
304 | """Generic implementation of ifiledata.emitrevisions(). | |
305 |
|
305 | |||
306 | Emitting revision data is subtly complex. This function attempts to |
|
306 | Emitting revision data is subtly complex. This function attempts to | |
307 | encapsulate all the logic for doing so in a backend-agnostic way. |
|
307 | encapsulate all the logic for doing so in a backend-agnostic way. | |
308 |
|
308 | |||
309 | ``store`` |
|
309 | ``store`` | |
310 | Object conforming to ``ifilestorage`` interface. |
|
310 | Object conforming to ``ifilestorage`` interface. | |
311 |
|
311 | |||
312 | ``nodes`` |
|
312 | ``nodes`` | |
313 | List of revision nodes whose data to emit. |
|
313 | List of revision nodes whose data to emit. | |
314 |
|
314 | |||
315 | ``resultcls`` |
|
315 | ``resultcls`` | |
316 | A type implementing the ``irevisiondelta`` interface that will be |
|
316 | A type implementing the ``irevisiondelta`` interface that will be | |
317 | constructed and returned. |
|
317 | constructed and returned. | |
318 |
|
318 | |||
319 | ``deltaparentfn`` (optional) |
|
319 | ``deltaparentfn`` (optional) | |
320 | Callable receiving a revision number and returning the revision number |
|
320 | Callable receiving a revision number and returning the revision number | |
321 | of a revision that the internal delta is stored against. This delta |
|
321 | of a revision that the internal delta is stored against. This delta | |
322 | will be preferred over computing a new arbitrary delta. |
|
322 | will be preferred over computing a new arbitrary delta. | |
323 |
|
323 | |||
324 | If not defined, a delta will always be computed from raw revision |
|
324 | If not defined, a delta will always be computed from raw revision | |
325 | data. |
|
325 | data. | |
326 |
|
326 | |||
327 | ``candeltafn`` (optional) |
|
327 | ``candeltafn`` (optional) | |
328 | Callable receiving a pair of revision numbers that returns a bool |
|
328 | Callable receiving a pair of revision numbers that returns a bool | |
329 | indicating whether a delta between them can be produced. |
|
329 | indicating whether a delta between them can be produced. | |
330 |
|
330 | |||
331 | If not defined, it is assumed that any two revisions can delta with |
|
331 | If not defined, it is assumed that any two revisions can delta with | |
332 | each other. |
|
332 | each other. | |
333 |
|
333 | |||
334 | ``rawsizefn`` (optional) |
|
334 | ``rawsizefn`` (optional) | |
335 | Callable receiving a revision number and returning the length of the |
|
335 | Callable receiving a revision number and returning the length of the | |
336 | ``store.rawdata(rev)``. |
|
336 | ``store.rawdata(rev)``. | |
337 |
|
337 | |||
338 | If not defined, ``len(store.rawdata(rev))`` will be called. |
|
338 | If not defined, ``len(store.rawdata(rev))`` will be called. | |
339 |
|
339 | |||
340 | ``revdifffn`` (optional) |
|
340 | ``revdifffn`` (optional) | |
341 | Callable receiving a pair of revision numbers that returns a delta |
|
341 | Callable receiving a pair of revision numbers that returns a delta | |
342 | between them. |
|
342 | between them. | |
343 |
|
343 | |||
344 | If not defined, a delta will be computed by invoking mdiff code |
|
344 | If not defined, a delta will be computed by invoking mdiff code | |
345 | on ``store.revision()`` results. |
|
345 | on ``store.revision()`` results. | |
346 |
|
346 | |||
347 | Defining this function allows a precomputed or stored delta to be |
|
347 | Defining this function allows a precomputed or stored delta to be | |
348 | used without having to compute on. |
|
348 | used without having to compute on. | |
349 |
|
349 | |||
350 | ``flagsfn`` (optional) |
|
350 | ``flagsfn`` (optional) | |
351 | Callable receiving a revision number and returns the integer flags |
|
351 | Callable receiving a revision number and returns the integer flags | |
352 | value for it. If not defined, flags value will be 0. |
|
352 | value for it. If not defined, flags value will be 0. | |
353 |
|
353 | |||
354 | ``deltamode`` |
|
354 | ``deltamode`` | |
355 | constaint on delta to be sent: |
|
355 | constaint on delta to be sent: | |
356 | * CG_DELTAMODE_STD - normal mode, try to reuse storage deltas, |
|
356 | * CG_DELTAMODE_STD - normal mode, try to reuse storage deltas, | |
357 | * CG_DELTAMODE_PREV - only delta against "prev", |
|
357 | * CG_DELTAMODE_PREV - only delta against "prev", | |
358 | * CG_DELTAMODE_FULL - only issue full snapshot. |
|
358 | * CG_DELTAMODE_FULL - only issue full snapshot. | |
359 |
|
359 | |||
360 | Whether to send fulltext revisions instead of deltas, if allowed. |
|
360 | Whether to send fulltext revisions instead of deltas, if allowed. | |
361 |
|
361 | |||
362 | ``nodesorder`` |
|
362 | ``nodesorder`` | |
363 | ``revisiondata`` |
|
363 | ``revisiondata`` | |
364 | ``assumehaveparentrevisions`` |
|
364 | ``assumehaveparentrevisions`` | |
365 | ``sidedata_helpers`` (optional) |
|
365 | ``sidedata_helpers`` (optional) | |
366 | If not None, means that sidedata should be included. |
|
366 | If not None, means that sidedata should be included. | |
367 | A dictionary of revlog type to tuples of `(repo, computers, removers)`: |
|
367 | See `revlogutil.sidedata.get_sidedata_helpers`. | |
368 | * `repo` is used as an argument for computers |
|
|||
369 | * `computers` is a list of `(category, (keys, computer, flags)` that |
|
|||
370 | compute the missing sidedata categories that were asked: |
|
|||
371 | * `category` is the sidedata category |
|
|||
372 | * `keys` are the sidedata keys to be affected |
|
|||
373 | * `flags` is a bitmask (an integer) of flags to remove when |
|
|||
374 | removing the category. |
|
|||
375 | * `computer` is the function `(repo, store, rev, sidedata)` that |
|
|||
376 | returns a tuple of |
|
|||
377 | `(new sidedata dict, (flags to add, flags to remove))`. |
|
|||
378 | For example, it will return `({}, (0, 1 << 15))` to return no |
|
|||
379 | sidedata, with no flags to add and one flag to remove. |
|
|||
380 | * `removers` will remove the keys corresponding to the categories |
|
|||
381 | that are present, but not needed. |
|
|||
382 | If both `computers` and `removers` are empty, sidedata are simply not |
|
|||
383 | transformed. |
|
|||
384 | Revlog types are `changelog`, `manifest` or `filelog`. |
|
|||
385 | """ |
|
368 | """ | |
386 |
|
369 | |||
387 | fnode = store.node |
|
370 | fnode = store.node | |
388 | frev = store.rev |
|
371 | frev = store.rev | |
389 |
|
372 | |||
390 | if nodesorder == b'nodes': |
|
373 | if nodesorder == b'nodes': | |
391 | revs = [frev(n) for n in nodes] |
|
374 | revs = [frev(n) for n in nodes] | |
392 | elif nodesorder == b'linear': |
|
375 | elif nodesorder == b'linear': | |
393 | revs = {frev(n) for n in nodes} |
|
376 | revs = {frev(n) for n in nodes} | |
394 | revs = dagop.linearize(revs, store.parentrevs) |
|
377 | revs = dagop.linearize(revs, store.parentrevs) | |
395 | else: # storage and default |
|
378 | else: # storage and default | |
396 | revs = sorted(frev(n) for n in nodes) |
|
379 | revs = sorted(frev(n) for n in nodes) | |
397 |
|
380 | |||
398 | prevrev = None |
|
381 | prevrev = None | |
399 |
|
382 | |||
400 | if deltamode == repository.CG_DELTAMODE_PREV or assumehaveparentrevisions: |
|
383 | if deltamode == repository.CG_DELTAMODE_PREV or assumehaveparentrevisions: | |
401 | prevrev = store.parentrevs(revs[0])[0] |
|
384 | prevrev = store.parentrevs(revs[0])[0] | |
402 |
|
385 | |||
403 | # Set of revs available to delta against. |
|
386 | # Set of revs available to delta against. | |
404 | available = set() |
|
387 | available = set() | |
405 |
|
388 | |||
406 | for rev in revs: |
|
389 | for rev in revs: | |
407 | if rev == nullrev: |
|
390 | if rev == nullrev: | |
408 | continue |
|
391 | continue | |
409 |
|
392 | |||
410 | node = fnode(rev) |
|
393 | node = fnode(rev) | |
411 | p1rev, p2rev = store.parentrevs(rev) |
|
394 | p1rev, p2rev = store.parentrevs(rev) | |
412 |
|
395 | |||
413 | if deltaparentfn: |
|
396 | if deltaparentfn: | |
414 | deltaparentrev = deltaparentfn(rev) |
|
397 | deltaparentrev = deltaparentfn(rev) | |
415 | else: |
|
398 | else: | |
416 | deltaparentrev = nullrev |
|
399 | deltaparentrev = nullrev | |
417 |
|
400 | |||
418 | # Forced delta against previous mode. |
|
401 | # Forced delta against previous mode. | |
419 | if deltamode == repository.CG_DELTAMODE_PREV: |
|
402 | if deltamode == repository.CG_DELTAMODE_PREV: | |
420 | baserev = prevrev |
|
403 | baserev = prevrev | |
421 |
|
404 | |||
422 | # We're instructed to send fulltext. Honor that. |
|
405 | # We're instructed to send fulltext. Honor that. | |
423 | elif deltamode == repository.CG_DELTAMODE_FULL: |
|
406 | elif deltamode == repository.CG_DELTAMODE_FULL: | |
424 | baserev = nullrev |
|
407 | baserev = nullrev | |
425 | # We're instructed to use p1. Honor that |
|
408 | # We're instructed to use p1. Honor that | |
426 | elif deltamode == repository.CG_DELTAMODE_P1: |
|
409 | elif deltamode == repository.CG_DELTAMODE_P1: | |
427 | baserev = p1rev |
|
410 | baserev = p1rev | |
428 |
|
411 | |||
429 | # There is a delta in storage. We try to use that because it |
|
412 | # There is a delta in storage. We try to use that because it | |
430 | # amounts to effectively copying data from storage and is |
|
413 | # amounts to effectively copying data from storage and is | |
431 | # therefore the fastest. |
|
414 | # therefore the fastest. | |
432 | elif deltaparentrev != nullrev: |
|
415 | elif deltaparentrev != nullrev: | |
433 | # Base revision was already emitted in this group. We can |
|
416 | # Base revision was already emitted in this group. We can | |
434 | # always safely use the delta. |
|
417 | # always safely use the delta. | |
435 | if deltaparentrev in available: |
|
418 | if deltaparentrev in available: | |
436 | baserev = deltaparentrev |
|
419 | baserev = deltaparentrev | |
437 |
|
420 | |||
438 | # Base revision is a parent that hasn't been emitted already. |
|
421 | # Base revision is a parent that hasn't been emitted already. | |
439 | # Use it if we can assume the receiver has the parent revision. |
|
422 | # Use it if we can assume the receiver has the parent revision. | |
440 | elif assumehaveparentrevisions and deltaparentrev in (p1rev, p2rev): |
|
423 | elif assumehaveparentrevisions and deltaparentrev in (p1rev, p2rev): | |
441 | baserev = deltaparentrev |
|
424 | baserev = deltaparentrev | |
442 |
|
425 | |||
443 | # No guarantee the receiver has the delta parent. Send delta |
|
426 | # No guarantee the receiver has the delta parent. Send delta | |
444 | # against last revision (if possible), which in the common case |
|
427 | # against last revision (if possible), which in the common case | |
445 | # should be similar enough to this revision that the delta is |
|
428 | # should be similar enough to this revision that the delta is | |
446 | # reasonable. |
|
429 | # reasonable. | |
447 | elif prevrev is not None: |
|
430 | elif prevrev is not None: | |
448 | baserev = prevrev |
|
431 | baserev = prevrev | |
449 | else: |
|
432 | else: | |
450 | baserev = nullrev |
|
433 | baserev = nullrev | |
451 |
|
434 | |||
452 | # Storage has a fulltext revision. |
|
435 | # Storage has a fulltext revision. | |
453 |
|
436 | |||
454 | # Let's use the previous revision, which is as good a guess as any. |
|
437 | # Let's use the previous revision, which is as good a guess as any. | |
455 | # There is definitely room to improve this logic. |
|
438 | # There is definitely room to improve this logic. | |
456 | elif prevrev is not None: |
|
439 | elif prevrev is not None: | |
457 | baserev = prevrev |
|
440 | baserev = prevrev | |
458 | else: |
|
441 | else: | |
459 | baserev = nullrev |
|
442 | baserev = nullrev | |
460 |
|
443 | |||
461 | # But we can't actually use our chosen delta base for whatever |
|
444 | # But we can't actually use our chosen delta base for whatever | |
462 | # reason. Reset to fulltext. |
|
445 | # reason. Reset to fulltext. | |
463 | if baserev != nullrev and (candeltafn and not candeltafn(baserev, rev)): |
|
446 | if baserev != nullrev and (candeltafn and not candeltafn(baserev, rev)): | |
464 | baserev = nullrev |
|
447 | baserev = nullrev | |
465 |
|
448 | |||
466 | revision = None |
|
449 | revision = None | |
467 | delta = None |
|
450 | delta = None | |
468 | baserevisionsize = None |
|
451 | baserevisionsize = None | |
469 |
|
452 | |||
470 | if revisiondata: |
|
453 | if revisiondata: | |
471 | if store.iscensored(baserev) or store.iscensored(rev): |
|
454 | if store.iscensored(baserev) or store.iscensored(rev): | |
472 | try: |
|
455 | try: | |
473 | revision = store.rawdata(node) |
|
456 | revision = store.rawdata(node) | |
474 | except error.CensoredNodeError as e: |
|
457 | except error.CensoredNodeError as e: | |
475 | revision = e.tombstone |
|
458 | revision = e.tombstone | |
476 |
|
459 | |||
477 | if baserev != nullrev: |
|
460 | if baserev != nullrev: | |
478 | if rawsizefn: |
|
461 | if rawsizefn: | |
479 | baserevisionsize = rawsizefn(baserev) |
|
462 | baserevisionsize = rawsizefn(baserev) | |
480 | else: |
|
463 | else: | |
481 | baserevisionsize = len(store.rawdata(baserev)) |
|
464 | baserevisionsize = len(store.rawdata(baserev)) | |
482 |
|
465 | |||
483 | elif ( |
|
466 | elif ( | |
484 | baserev == nullrev and deltamode != repository.CG_DELTAMODE_PREV |
|
467 | baserev == nullrev and deltamode != repository.CG_DELTAMODE_PREV | |
485 | ): |
|
468 | ): | |
486 | revision = store.rawdata(node) |
|
469 | revision = store.rawdata(node) | |
487 | available.add(rev) |
|
470 | available.add(rev) | |
488 | else: |
|
471 | else: | |
489 | if revdifffn: |
|
472 | if revdifffn: | |
490 | delta = revdifffn(baserev, rev) |
|
473 | delta = revdifffn(baserev, rev) | |
491 | else: |
|
474 | else: | |
492 | delta = mdiff.textdiff( |
|
475 | delta = mdiff.textdiff( | |
493 | store.rawdata(baserev), store.rawdata(rev) |
|
476 | store.rawdata(baserev), store.rawdata(rev) | |
494 | ) |
|
477 | ) | |
495 |
|
478 | |||
496 | available.add(rev) |
|
479 | available.add(rev) | |
497 |
|
480 | |||
498 | serialized_sidedata = None |
|
481 | serialized_sidedata = None | |
499 | sidedata_flags = (0, 0) |
|
482 | sidedata_flags = (0, 0) | |
500 | if sidedata_helpers: |
|
483 | if sidedata_helpers: | |
501 | old_sidedata = store.sidedata(rev) |
|
484 | old_sidedata = store.sidedata(rev) | |
502 | sidedata, sidedata_flags = sidedatamod.run_sidedata_helpers( |
|
485 | sidedata, sidedata_flags = sidedatamod.run_sidedata_helpers( | |
503 | store=store, |
|
486 | store=store, | |
504 | sidedata_helpers=sidedata_helpers, |
|
487 | sidedata_helpers=sidedata_helpers, | |
505 | sidedata=old_sidedata, |
|
488 | sidedata=old_sidedata, | |
506 | rev=rev, |
|
489 | rev=rev, | |
507 | ) |
|
490 | ) | |
508 | if sidedata: |
|
491 | if sidedata: | |
509 | serialized_sidedata = sidedatamod.serialize_sidedata(sidedata) |
|
492 | serialized_sidedata = sidedatamod.serialize_sidedata(sidedata) | |
510 |
|
493 | |||
511 | flags = flagsfn(rev) if flagsfn else 0 |
|
494 | flags = flagsfn(rev) if flagsfn else 0 | |
512 | protocol_flags = 0 |
|
495 | protocol_flags = 0 | |
513 | if serialized_sidedata: |
|
496 | if serialized_sidedata: | |
514 | # Advertise that sidedata exists to the other side |
|
497 | # Advertise that sidedata exists to the other side | |
515 | protocol_flags |= CG_FLAG_SIDEDATA |
|
498 | protocol_flags |= CG_FLAG_SIDEDATA | |
516 | # Computers and removers can return flags to add and/or remove |
|
499 | # Computers and removers can return flags to add and/or remove | |
517 | flags = flags | sidedata_flags[0] & ~sidedata_flags[1] |
|
500 | flags = flags | sidedata_flags[0] & ~sidedata_flags[1] | |
518 |
|
501 | |||
519 | yield resultcls( |
|
502 | yield resultcls( | |
520 | node=node, |
|
503 | node=node, | |
521 | p1node=fnode(p1rev), |
|
504 | p1node=fnode(p1rev), | |
522 | p2node=fnode(p2rev), |
|
505 | p2node=fnode(p2rev), | |
523 | basenode=fnode(baserev), |
|
506 | basenode=fnode(baserev), | |
524 | flags=flags, |
|
507 | flags=flags, | |
525 | baserevisionsize=baserevisionsize, |
|
508 | baserevisionsize=baserevisionsize, | |
526 | revision=revision, |
|
509 | revision=revision, | |
527 | delta=delta, |
|
510 | delta=delta, | |
528 | sidedata=serialized_sidedata, |
|
511 | sidedata=serialized_sidedata, | |
529 | protocol_flags=protocol_flags, |
|
512 | protocol_flags=protocol_flags, | |
530 | ) |
|
513 | ) | |
531 |
|
514 | |||
532 | prevrev = rev |
|
515 | prevrev = rev | |
533 |
|
516 | |||
534 |
|
517 | |||
535 | def deltaiscensored(delta, baserev, baselenfn): |
|
518 | def deltaiscensored(delta, baserev, baselenfn): | |
536 | """Determine if a delta represents censored revision data. |
|
519 | """Determine if a delta represents censored revision data. | |
537 |
|
520 | |||
538 | ``baserev`` is the base revision this delta is encoded against. |
|
521 | ``baserev`` is the base revision this delta is encoded against. | |
539 | ``baselenfn`` is a callable receiving a revision number that resolves the |
|
522 | ``baselenfn`` is a callable receiving a revision number that resolves the | |
540 | length of the revision fulltext. |
|
523 | length of the revision fulltext. | |
541 |
|
524 | |||
542 | Returns a bool indicating if the result of the delta represents a censored |
|
525 | Returns a bool indicating if the result of the delta represents a censored | |
543 | revision. |
|
526 | revision. | |
544 | """ |
|
527 | """ | |
545 | # Fragile heuristic: unless new file meta keys are added alphabetically |
|
528 | # Fragile heuristic: unless new file meta keys are added alphabetically | |
546 | # preceding "censored", all censored revisions are prefixed by |
|
529 | # preceding "censored", all censored revisions are prefixed by | |
547 | # "\1\ncensored:". A delta producing such a censored revision must be a |
|
530 | # "\1\ncensored:". A delta producing such a censored revision must be a | |
548 | # full-replacement delta, so we inspect the first and only patch in the |
|
531 | # full-replacement delta, so we inspect the first and only patch in the | |
549 | # delta for this prefix. |
|
532 | # delta for this prefix. | |
550 | hlen = struct.calcsize(b">lll") |
|
533 | hlen = struct.calcsize(b">lll") | |
551 | if len(delta) <= hlen: |
|
534 | if len(delta) <= hlen: | |
552 | return False |
|
535 | return False | |
553 |
|
536 | |||
554 | oldlen = baselenfn(baserev) |
|
537 | oldlen = baselenfn(baserev) | |
555 | newlen = len(delta) - hlen |
|
538 | newlen = len(delta) - hlen | |
556 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): |
|
539 | if delta[:hlen] != mdiff.replacediffheader(oldlen, newlen): | |
557 | return False |
|
540 | return False | |
558 |
|
541 | |||
559 | add = b"\1\ncensored:" |
|
542 | add = b"\1\ncensored:" | |
560 | addlen = len(add) |
|
543 | addlen = len(add) | |
561 | return newlen >= addlen and delta[hlen : hlen + addlen] == add |
|
544 | return newlen >= addlen and delta[hlen : hlen + addlen] == add |
General Comments 0
You need to be logged in to leave comments.
Login now