Show More
@@ -1,1132 +1,1142 | |||||
1 | # copies.py - copy detection for Mercurial |
|
1 | # copies.py - copy detection for Mercurial | |
2 | # |
|
2 | # | |
3 | # Copyright 2008 Matt Mackall <mpm@selenic.com> |
|
3 | # Copyright 2008 Matt Mackall <mpm@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 collections |
|
10 | import collections | |
11 | import os |
|
11 | import os | |
12 |
|
12 | |||
13 | from .i18n import _ |
|
13 | from .i18n import _ | |
14 |
|
14 | |||
15 |
|
15 | |||
16 | from . import ( |
|
16 | from . import ( | |
17 | match as matchmod, |
|
17 | match as matchmod, | |
18 | node, |
|
18 | node, | |
19 | pathutil, |
|
19 | pathutil, | |
|
20 | policy, | |||
20 | pycompat, |
|
21 | pycompat, | |
21 | util, |
|
22 | util, | |
22 | ) |
|
23 | ) | |
23 |
|
24 | |||
24 |
|
25 | |||
25 | from .utils import stringutil |
|
26 | from .utils import stringutil | |
26 |
|
27 | |||
27 | from .revlogutils import flagutil |
|
28 | from .revlogutils import flagutil | |
28 |
|
29 | |||
|
30 | rustmod = policy.importrust("copy_tracing") | |||
|
31 | ||||
29 |
|
32 | |||
30 | def _filter(src, dst, t): |
|
33 | def _filter(src, dst, t): | |
31 | """filters out invalid copies after chaining""" |
|
34 | """filters out invalid copies after chaining""" | |
32 |
|
35 | |||
33 | # When _chain()'ing copies in 'a' (from 'src' via some other commit 'mid') |
|
36 | # When _chain()'ing copies in 'a' (from 'src' via some other commit 'mid') | |
34 | # with copies in 'b' (from 'mid' to 'dst'), we can get the different cases |
|
37 | # with copies in 'b' (from 'mid' to 'dst'), we can get the different cases | |
35 | # in the following table (not including trivial cases). For example, case 2 |
|
38 | # in the following table (not including trivial cases). For example, case 2 | |
36 | # is where a file existed in 'src' and remained under that name in 'mid' and |
|
39 | # is where a file existed in 'src' and remained under that name in 'mid' and | |
37 | # then was renamed between 'mid' and 'dst'. |
|
40 | # then was renamed between 'mid' and 'dst'. | |
38 | # |
|
41 | # | |
39 | # case src mid dst result |
|
42 | # case src mid dst result | |
40 | # 1 x y - - |
|
43 | # 1 x y - - | |
41 | # 2 x y y x->y |
|
44 | # 2 x y y x->y | |
42 | # 3 x y x - |
|
45 | # 3 x y x - | |
43 | # 4 x y z x->z |
|
46 | # 4 x y z x->z | |
44 | # 5 - x y - |
|
47 | # 5 - x y - | |
45 | # 6 x x y x->y |
|
48 | # 6 x x y x->y | |
46 | # |
|
49 | # | |
47 | # _chain() takes care of chaining the copies in 'a' and 'b', but it |
|
50 | # _chain() takes care of chaining the copies in 'a' and 'b', but it | |
48 | # cannot tell the difference between cases 1 and 2, between 3 and 4, or |
|
51 | # cannot tell the difference between cases 1 and 2, between 3 and 4, or | |
49 | # between 5 and 6, so it includes all cases in its result. |
|
52 | # between 5 and 6, so it includes all cases in its result. | |
50 | # Cases 1, 3, and 5 are then removed by _filter(). |
|
53 | # Cases 1, 3, and 5 are then removed by _filter(). | |
51 |
|
54 | |||
52 | for k, v in list(t.items()): |
|
55 | for k, v in list(t.items()): | |
53 | # remove copies from files that didn't exist |
|
56 | # remove copies from files that didn't exist | |
54 | if v not in src: |
|
57 | if v not in src: | |
55 | del t[k] |
|
58 | del t[k] | |
56 | # remove criss-crossed copies |
|
59 | # remove criss-crossed copies | |
57 | elif k in src and v in dst: |
|
60 | elif k in src and v in dst: | |
58 | del t[k] |
|
61 | del t[k] | |
59 | # remove copies to files that were then removed |
|
62 | # remove copies to files that were then removed | |
60 | elif k not in dst: |
|
63 | elif k not in dst: | |
61 | del t[k] |
|
64 | del t[k] | |
62 |
|
65 | |||
63 |
|
66 | |||
64 | def _chain(prefix, suffix): |
|
67 | def _chain(prefix, suffix): | |
65 | """chain two sets of copies 'prefix' and 'suffix'""" |
|
68 | """chain two sets of copies 'prefix' and 'suffix'""" | |
66 | result = prefix.copy() |
|
69 | result = prefix.copy() | |
67 | for key, value in pycompat.iteritems(suffix): |
|
70 | for key, value in pycompat.iteritems(suffix): | |
68 | result[key] = prefix.get(value, value) |
|
71 | result[key] = prefix.get(value, value) | |
69 | return result |
|
72 | return result | |
70 |
|
73 | |||
71 |
|
74 | |||
72 | def _tracefile(fctx, am, basemf): |
|
75 | def _tracefile(fctx, am, basemf): | |
73 | """return file context that is the ancestor of fctx present in ancestor |
|
76 | """return file context that is the ancestor of fctx present in ancestor | |
74 | manifest am |
|
77 | manifest am | |
75 |
|
78 | |||
76 | Note: we used to try and stop after a given limit, however checking if that |
|
79 | Note: we used to try and stop after a given limit, however checking if that | |
77 | limit is reached turned out to be very expensive. we are better off |
|
80 | limit is reached turned out to be very expensive. we are better off | |
78 | disabling that feature.""" |
|
81 | disabling that feature.""" | |
79 |
|
82 | |||
80 | for f in fctx.ancestors(): |
|
83 | for f in fctx.ancestors(): | |
81 | path = f.path() |
|
84 | path = f.path() | |
82 | if am.get(path, None) == f.filenode(): |
|
85 | if am.get(path, None) == f.filenode(): | |
83 | return path |
|
86 | return path | |
84 | if basemf and basemf.get(path, None) == f.filenode(): |
|
87 | if basemf and basemf.get(path, None) == f.filenode(): | |
85 | return path |
|
88 | return path | |
86 |
|
89 | |||
87 |
|
90 | |||
88 | def _dirstatecopies(repo, match=None): |
|
91 | def _dirstatecopies(repo, match=None): | |
89 | ds = repo.dirstate |
|
92 | ds = repo.dirstate | |
90 | c = ds.copies().copy() |
|
93 | c = ds.copies().copy() | |
91 | for k in list(c): |
|
94 | for k in list(c): | |
92 | if ds[k] not in b'anm' or (match and not match(k)): |
|
95 | if ds[k] not in b'anm' or (match and not match(k)): | |
93 | del c[k] |
|
96 | del c[k] | |
94 | return c |
|
97 | return c | |
95 |
|
98 | |||
96 |
|
99 | |||
97 | def _computeforwardmissing(a, b, match=None): |
|
100 | def _computeforwardmissing(a, b, match=None): | |
98 | """Computes which files are in b but not a. |
|
101 | """Computes which files are in b but not a. | |
99 | This is its own function so extensions can easily wrap this call to see what |
|
102 | This is its own function so extensions can easily wrap this call to see what | |
100 | files _forwardcopies is about to process. |
|
103 | files _forwardcopies is about to process. | |
101 | """ |
|
104 | """ | |
102 | ma = a.manifest() |
|
105 | ma = a.manifest() | |
103 | mb = b.manifest() |
|
106 | mb = b.manifest() | |
104 | return mb.filesnotin(ma, match=match) |
|
107 | return mb.filesnotin(ma, match=match) | |
105 |
|
108 | |||
106 |
|
109 | |||
107 | def usechangesetcentricalgo(repo): |
|
110 | def usechangesetcentricalgo(repo): | |
108 | """Checks if we should use changeset-centric copy algorithms""" |
|
111 | """Checks if we should use changeset-centric copy algorithms""" | |
109 | if repo.filecopiesmode == b'changeset-sidedata': |
|
112 | if repo.filecopiesmode == b'changeset-sidedata': | |
110 | return True |
|
113 | return True | |
111 | readfrom = repo.ui.config(b'experimental', b'copies.read-from') |
|
114 | readfrom = repo.ui.config(b'experimental', b'copies.read-from') | |
112 | changesetsource = (b'changeset-only', b'compatibility') |
|
115 | changesetsource = (b'changeset-only', b'compatibility') | |
113 | return readfrom in changesetsource |
|
116 | return readfrom in changesetsource | |
114 |
|
117 | |||
115 |
|
118 | |||
116 | def _committedforwardcopies(a, b, base, match): |
|
119 | def _committedforwardcopies(a, b, base, match): | |
117 | """Like _forwardcopies(), but b.rev() cannot be None (working copy)""" |
|
120 | """Like _forwardcopies(), but b.rev() cannot be None (working copy)""" | |
118 | # files might have to be traced back to the fctx parent of the last |
|
121 | # files might have to be traced back to the fctx parent of the last | |
119 | # one-side-only changeset, but not further back than that |
|
122 | # one-side-only changeset, but not further back than that | |
120 | repo = a._repo |
|
123 | repo = a._repo | |
121 |
|
124 | |||
122 | if usechangesetcentricalgo(repo): |
|
125 | if usechangesetcentricalgo(repo): | |
123 | return _changesetforwardcopies(a, b, match) |
|
126 | return _changesetforwardcopies(a, b, match) | |
124 |
|
127 | |||
125 | debug = repo.ui.debugflag and repo.ui.configbool(b'devel', b'debug.copies') |
|
128 | debug = repo.ui.debugflag and repo.ui.configbool(b'devel', b'debug.copies') | |
126 | dbg = repo.ui.debug |
|
129 | dbg = repo.ui.debug | |
127 | if debug: |
|
130 | if debug: | |
128 | dbg(b'debug.copies: looking into rename from %s to %s\n' % (a, b)) |
|
131 | dbg(b'debug.copies: looking into rename from %s to %s\n' % (a, b)) | |
129 | am = a.manifest() |
|
132 | am = a.manifest() | |
130 | basemf = None if base is None else base.manifest() |
|
133 | basemf = None if base is None else base.manifest() | |
131 |
|
134 | |||
132 | # find where new files came from |
|
135 | # find where new files came from | |
133 | # we currently don't try to find where old files went, too expensive |
|
136 | # we currently don't try to find where old files went, too expensive | |
134 | # this means we can miss a case like 'hg rm b; hg cp a b' |
|
137 | # this means we can miss a case like 'hg rm b; hg cp a b' | |
135 | cm = {} |
|
138 | cm = {} | |
136 |
|
139 | |||
137 | # Computing the forward missing is quite expensive on large manifests, since |
|
140 | # Computing the forward missing is quite expensive on large manifests, since | |
138 | # it compares the entire manifests. We can optimize it in the common use |
|
141 | # it compares the entire manifests. We can optimize it in the common use | |
139 | # case of computing what copies are in a commit versus its parent (like |
|
142 | # case of computing what copies are in a commit versus its parent (like | |
140 | # during a rebase or histedit). Note, we exclude merge commits from this |
|
143 | # during a rebase or histedit). Note, we exclude merge commits from this | |
141 | # optimization, since the ctx.files() for a merge commit is not correct for |
|
144 | # optimization, since the ctx.files() for a merge commit is not correct for | |
142 | # this comparison. |
|
145 | # this comparison. | |
143 | forwardmissingmatch = match |
|
146 | forwardmissingmatch = match | |
144 | if b.p1() == a and b.p2().node() == node.nullid: |
|
147 | if b.p1() == a and b.p2().node() == node.nullid: | |
145 | filesmatcher = matchmod.exact(b.files()) |
|
148 | filesmatcher = matchmod.exact(b.files()) | |
146 | forwardmissingmatch = matchmod.intersectmatchers(match, filesmatcher) |
|
149 | forwardmissingmatch = matchmod.intersectmatchers(match, filesmatcher) | |
147 | missing = _computeforwardmissing(a, b, match=forwardmissingmatch) |
|
150 | missing = _computeforwardmissing(a, b, match=forwardmissingmatch) | |
148 |
|
151 | |||
149 | ancestrycontext = a._repo.changelog.ancestors([b.rev()], inclusive=True) |
|
152 | ancestrycontext = a._repo.changelog.ancestors([b.rev()], inclusive=True) | |
150 |
|
153 | |||
151 | if debug: |
|
154 | if debug: | |
152 | dbg(b'debug.copies: missing files to search: %d\n' % len(missing)) |
|
155 | dbg(b'debug.copies: missing files to search: %d\n' % len(missing)) | |
153 |
|
156 | |||
154 | for f in sorted(missing): |
|
157 | for f in sorted(missing): | |
155 | if debug: |
|
158 | if debug: | |
156 | dbg(b'debug.copies: tracing file: %s\n' % f) |
|
159 | dbg(b'debug.copies: tracing file: %s\n' % f) | |
157 | fctx = b[f] |
|
160 | fctx = b[f] | |
158 | fctx._ancestrycontext = ancestrycontext |
|
161 | fctx._ancestrycontext = ancestrycontext | |
159 |
|
162 | |||
160 | if debug: |
|
163 | if debug: | |
161 | start = util.timer() |
|
164 | start = util.timer() | |
162 | opath = _tracefile(fctx, am, basemf) |
|
165 | opath = _tracefile(fctx, am, basemf) | |
163 | if opath: |
|
166 | if opath: | |
164 | if debug: |
|
167 | if debug: | |
165 | dbg(b'debug.copies: rename of: %s\n' % opath) |
|
168 | dbg(b'debug.copies: rename of: %s\n' % opath) | |
166 | cm[f] = opath |
|
169 | cm[f] = opath | |
167 | if debug: |
|
170 | if debug: | |
168 | dbg( |
|
171 | dbg( | |
169 | b'debug.copies: time: %f seconds\n' |
|
172 | b'debug.copies: time: %f seconds\n' | |
170 | % (util.timer() - start) |
|
173 | % (util.timer() - start) | |
171 | ) |
|
174 | ) | |
172 | return cm |
|
175 | return cm | |
173 |
|
176 | |||
174 |
|
177 | |||
175 | def _revinfo_getter(repo): |
|
178 | def _revinfo_getter(repo): | |
176 | """returns a function that returns the following data given a <rev>" |
|
179 | """returns a function that returns the following data given a <rev>" | |
177 |
|
180 | |||
178 | * p1: revision number of first parent |
|
181 | * p1: revision number of first parent | |
179 | * p2: revision number of first parent |
|
182 | * p2: revision number of first parent | |
180 | * changes: a ChangingFiles object |
|
183 | * changes: a ChangingFiles object | |
181 | """ |
|
184 | """ | |
182 | cl = repo.changelog |
|
185 | cl = repo.changelog | |
183 | parents = cl.parentrevs |
|
186 | parents = cl.parentrevs | |
184 | flags = cl.flags |
|
187 | flags = cl.flags | |
185 |
|
188 | |||
186 | HASCOPIESINFO = flagutil.REVIDX_HASCOPIESINFO |
|
189 | HASCOPIESINFO = flagutil.REVIDX_HASCOPIESINFO | |
187 |
|
190 | |||
188 | changelogrevision = cl.changelogrevision |
|
191 | changelogrevision = cl.changelogrevision | |
189 |
|
192 | |||
190 | # A small cache to avoid doing the work twice for merges |
|
193 | # A small cache to avoid doing the work twice for merges | |
191 | # |
|
194 | # | |
192 | # In the vast majority of cases, if we ask information for a revision |
|
195 | # In the vast majority of cases, if we ask information for a revision | |
193 | # about 1 parent, we'll later ask it for the other. So it make sense to |
|
196 | # about 1 parent, we'll later ask it for the other. So it make sense to | |
194 | # keep the information around when reaching the first parent of a merge |
|
197 | # keep the information around when reaching the first parent of a merge | |
195 | # and dropping it after it was provided for the second parents. |
|
198 | # and dropping it after it was provided for the second parents. | |
196 | # |
|
199 | # | |
197 | # It exists cases were only one parent of the merge will be walked. It |
|
200 | # It exists cases were only one parent of the merge will be walked. It | |
198 | # happens when the "destination" the copy tracing is descendant from a |
|
201 | # happens when the "destination" the copy tracing is descendant from a | |
199 | # new root, not common with the "source". In that case, we will only walk |
|
202 | # new root, not common with the "source". In that case, we will only walk | |
200 | # through merge parents that are descendant of changesets common |
|
203 | # through merge parents that are descendant of changesets common | |
201 | # between "source" and "destination". |
|
204 | # between "source" and "destination". | |
202 | # |
|
205 | # | |
203 | # With the current case implementation if such changesets have a copy |
|
206 | # With the current case implementation if such changesets have a copy | |
204 | # information, we'll keep them in memory until the end of |
|
207 | # information, we'll keep them in memory until the end of | |
205 | # _changesetforwardcopies. We don't expect the case to be frequent |
|
208 | # _changesetforwardcopies. We don't expect the case to be frequent | |
206 | # enough to matters. |
|
209 | # enough to matters. | |
207 | # |
|
210 | # | |
208 | # In addition, it would be possible to reach pathological case, were |
|
211 | # In addition, it would be possible to reach pathological case, were | |
209 | # many first parent are met before any second parent is reached. In |
|
212 | # many first parent are met before any second parent is reached. In | |
210 | # that case the cache could grow. If this even become an issue one can |
|
213 | # that case the cache could grow. If this even become an issue one can | |
211 | # safely introduce a maximum cache size. This would trade extra CPU/IO |
|
214 | # safely introduce a maximum cache size. This would trade extra CPU/IO | |
212 | # time to save memory. |
|
215 | # time to save memory. | |
213 | merge_caches = {} |
|
216 | merge_caches = {} | |
214 |
|
217 | |||
215 | def revinfo(rev): |
|
218 | def revinfo(rev): | |
216 | p1, p2 = parents(rev) |
|
219 | p1, p2 = parents(rev) | |
217 | value = None |
|
220 | value = None | |
218 | e = merge_caches.pop(rev, None) |
|
221 | e = merge_caches.pop(rev, None) | |
219 | if e is not None: |
|
222 | if e is not None: | |
220 | return e |
|
223 | return e | |
221 | changes = None |
|
224 | changes = None | |
222 | if flags(rev) & HASCOPIESINFO: |
|
225 | if flags(rev) & HASCOPIESINFO: | |
223 | changes = changelogrevision(rev).changes |
|
226 | changes = changelogrevision(rev).changes | |
224 | value = (p1, p2, changes) |
|
227 | value = (p1, p2, changes) | |
225 | if p1 != node.nullrev and p2 != node.nullrev: |
|
228 | if p1 != node.nullrev and p2 != node.nullrev: | |
226 | # XXX some case we over cache, IGNORE |
|
229 | # XXX some case we over cache, IGNORE | |
227 | merge_caches[rev] = value |
|
230 | merge_caches[rev] = value | |
228 | return value |
|
231 | return value | |
229 |
|
232 | |||
230 | return revinfo |
|
233 | return revinfo | |
231 |
|
234 | |||
232 |
|
235 | |||
233 | def cached_is_ancestor(is_ancestor): |
|
236 | def cached_is_ancestor(is_ancestor): | |
234 | """return a cached version of is_ancestor""" |
|
237 | """return a cached version of is_ancestor""" | |
235 | cache = {} |
|
238 | cache = {} | |
236 |
|
239 | |||
237 | def _is_ancestor(anc, desc): |
|
240 | def _is_ancestor(anc, desc): | |
238 | if anc > desc: |
|
241 | if anc > desc: | |
239 | return False |
|
242 | return False | |
240 | elif anc == desc: |
|
243 | elif anc == desc: | |
241 | return True |
|
244 | return True | |
242 | key = (anc, desc) |
|
245 | key = (anc, desc) | |
243 | ret = cache.get(key) |
|
246 | ret = cache.get(key) | |
244 | if ret is None: |
|
247 | if ret is None: | |
245 | ret = cache[key] = is_ancestor(anc, desc) |
|
248 | ret = cache[key] = is_ancestor(anc, desc) | |
246 | return ret |
|
249 | return ret | |
247 |
|
250 | |||
248 | return _is_ancestor |
|
251 | return _is_ancestor | |
249 |
|
252 | |||
250 |
|
253 | |||
251 | def _changesetforwardcopies(a, b, match): |
|
254 | def _changesetforwardcopies(a, b, match): | |
252 | if a.rev() in (node.nullrev, b.rev()): |
|
255 | if a.rev() in (node.nullrev, b.rev()): | |
253 | return {} |
|
256 | return {} | |
254 |
|
257 | |||
255 | repo = a.repo().unfiltered() |
|
258 | repo = a.repo().unfiltered() | |
256 | children = {} |
|
259 | children = {} | |
257 |
|
260 | |||
258 | cl = repo.changelog |
|
261 | cl = repo.changelog | |
259 | isancestor = cached_is_ancestor(cl.isancestorrev) |
|
262 | isancestor = cached_is_ancestor(cl.isancestorrev) | |
260 | missingrevs = cl.findmissingrevs(common=[a.rev()], heads=[b.rev()]) |
|
263 | missingrevs = cl.findmissingrevs(common=[a.rev()], heads=[b.rev()]) | |
261 | mrset = set(missingrevs) |
|
264 | mrset = set(missingrevs) | |
262 | roots = set() |
|
265 | roots = set() | |
263 | for r in missingrevs: |
|
266 | for r in missingrevs: | |
264 | for p in cl.parentrevs(r): |
|
267 | for p in cl.parentrevs(r): | |
265 | if p == node.nullrev: |
|
268 | if p == node.nullrev: | |
266 | continue |
|
269 | continue | |
267 | if p not in children: |
|
270 | if p not in children: | |
268 | children[p] = [r] |
|
271 | children[p] = [r] | |
269 | else: |
|
272 | else: | |
270 | children[p].append(r) |
|
273 | children[p].append(r) | |
271 | if p not in mrset: |
|
274 | if p not in mrset: | |
272 | roots.add(p) |
|
275 | roots.add(p) | |
273 | if not roots: |
|
276 | if not roots: | |
274 | # no common revision to track copies from |
|
277 | # no common revision to track copies from | |
275 | return {} |
|
278 | return {} | |
276 | min_root = min(roots) |
|
279 | min_root = min(roots) | |
277 |
|
280 | |||
278 | from_head = set( |
|
281 | from_head = set( | |
279 | cl.reachableroots(min_root, [b.rev()], list(roots), includepath=True) |
|
282 | cl.reachableroots(min_root, [b.rev()], list(roots), includepath=True) | |
280 | ) |
|
283 | ) | |
281 |
|
284 | |||
282 | iterrevs = set(from_head) |
|
285 | iterrevs = set(from_head) | |
283 | iterrevs &= mrset |
|
286 | iterrevs &= mrset | |
284 | iterrevs.update(roots) |
|
287 | iterrevs.update(roots) | |
285 | iterrevs.remove(b.rev()) |
|
288 | iterrevs.remove(b.rev()) | |
286 | revs = sorted(iterrevs) |
|
289 | revs = sorted(iterrevs) | |
287 |
|
290 | |||
288 | if repo.filecopiesmode == b'changeset-sidedata': |
|
291 | if repo.filecopiesmode == b'changeset-sidedata': | |
289 | revinfo = _revinfo_getter(repo) |
|
292 | revinfo = _revinfo_getter(repo) | |
290 | return _combine_changeset_copies( |
|
293 | return _combine_changeset_copies( | |
291 | revs, children, b.rev(), revinfo, match, isancestor |
|
294 | revs, children, b.rev(), revinfo, match, isancestor | |
292 | ) |
|
295 | ) | |
293 | else: |
|
296 | else: | |
294 | revinfo = _revinfo_getter_extra(repo) |
|
297 | revinfo = _revinfo_getter_extra(repo) | |
295 | return _combine_changeset_copies_extra( |
|
298 | return _combine_changeset_copies_extra( | |
296 | revs, children, b.rev(), revinfo, match, isancestor |
|
299 | revs, children, b.rev(), revinfo, match, isancestor | |
297 | ) |
|
300 | ) | |
298 |
|
301 | |||
299 |
|
302 | |||
300 | def _combine_changeset_copies( |
|
303 | def _combine_changeset_copies( | |
301 | revs, children, targetrev, revinfo, match, isancestor |
|
304 | revs, children, targetrev, revinfo, match, isancestor | |
302 | ): |
|
305 | ): | |
303 | """combine the copies information for each item of iterrevs |
|
306 | """combine the copies information for each item of iterrevs | |
304 |
|
307 | |||
305 | revs: sorted iterable of revision to visit |
|
308 | revs: sorted iterable of revision to visit | |
306 | children: a {parent: [children]} mapping. |
|
309 | children: a {parent: [children]} mapping. | |
307 | targetrev: the final copies destination revision (not in iterrevs) |
|
310 | targetrev: the final copies destination revision (not in iterrevs) | |
308 | revinfo(rev): a function that return (p1, p2, p1copies, p2copies, removed) |
|
311 | revinfo(rev): a function that return (p1, p2, p1copies, p2copies, removed) | |
309 | match: a matcher |
|
312 | match: a matcher | |
310 |
|
313 | |||
311 | It returns the aggregated copies information for `targetrev`. |
|
314 | It returns the aggregated copies information for `targetrev`. | |
312 | """ |
|
315 | """ | |
|
316 | ||||
|
317 | alwaysmatch = match.always() | |||
|
318 | ||||
|
319 | if rustmod is not None and alwaysmatch: | |||
|
320 | return rustmod.combine_changeset_copies( | |||
|
321 | list(revs), children, targetrev, revinfo, isancestor | |||
|
322 | ) | |||
|
323 | ||||
313 | all_copies = {} |
|
324 | all_copies = {} | |
314 | alwaysmatch = match.always() |
|
|||
315 | for r in revs: |
|
325 | for r in revs: | |
316 | copies = all_copies.pop(r, None) |
|
326 | copies = all_copies.pop(r, None) | |
317 | if copies is None: |
|
327 | if copies is None: | |
318 | # this is a root |
|
328 | # this is a root | |
319 | copies = {} |
|
329 | copies = {} | |
320 | for i, c in enumerate(children[r]): |
|
330 | for i, c in enumerate(children[r]): | |
321 | p1, p2, changes = revinfo(c) |
|
331 | p1, p2, changes = revinfo(c) | |
322 | childcopies = {} |
|
332 | childcopies = {} | |
323 | if r == p1: |
|
333 | if r == p1: | |
324 | parent = 1 |
|
334 | parent = 1 | |
325 | if changes is not None: |
|
335 | if changes is not None: | |
326 | childcopies = changes.copied_from_p1 |
|
336 | childcopies = changes.copied_from_p1 | |
327 | else: |
|
337 | else: | |
328 | assert r == p2 |
|
338 | assert r == p2 | |
329 | parent = 2 |
|
339 | parent = 2 | |
330 | if changes is not None: |
|
340 | if changes is not None: | |
331 | childcopies = changes.copied_from_p2 |
|
341 | childcopies = changes.copied_from_p2 | |
332 | if not alwaysmatch: |
|
342 | if not alwaysmatch: | |
333 | childcopies = { |
|
343 | childcopies = { | |
334 | dst: src for dst, src in childcopies.items() if match(dst) |
|
344 | dst: src for dst, src in childcopies.items() if match(dst) | |
335 | } |
|
345 | } | |
336 | newcopies = copies |
|
346 | newcopies = copies | |
337 | if childcopies: |
|
347 | if childcopies: | |
338 | newcopies = copies.copy() |
|
348 | newcopies = copies.copy() | |
339 | for dest, source in pycompat.iteritems(childcopies): |
|
349 | for dest, source in pycompat.iteritems(childcopies): | |
340 | prev = copies.get(source) |
|
350 | prev = copies.get(source) | |
341 | if prev is not None and prev[1] is not None: |
|
351 | if prev is not None and prev[1] is not None: | |
342 | source = prev[1] |
|
352 | source = prev[1] | |
343 | newcopies[dest] = (c, source) |
|
353 | newcopies[dest] = (c, source) | |
344 | assert newcopies is not copies |
|
354 | assert newcopies is not copies | |
345 | if changes is not None: |
|
355 | if changes is not None: | |
346 | for f in changes.removed: |
|
356 | for f in changes.removed: | |
347 | if f in newcopies: |
|
357 | if f in newcopies: | |
348 | if newcopies is copies: |
|
358 | if newcopies is copies: | |
349 | # copy on write to avoid affecting potential other |
|
359 | # copy on write to avoid affecting potential other | |
350 | # branches. when there are no other branches, this |
|
360 | # branches. when there are no other branches, this | |
351 | # could be avoided. |
|
361 | # could be avoided. | |
352 | newcopies = copies.copy() |
|
362 | newcopies = copies.copy() | |
353 | newcopies[f] = (c, None) |
|
363 | newcopies[f] = (c, None) | |
354 | othercopies = all_copies.get(c) |
|
364 | othercopies = all_copies.get(c) | |
355 | if othercopies is None: |
|
365 | if othercopies is None: | |
356 | all_copies[c] = newcopies |
|
366 | all_copies[c] = newcopies | |
357 | else: |
|
367 | else: | |
358 | # we are the second parent to work on c, we need to merge our |
|
368 | # we are the second parent to work on c, we need to merge our | |
359 | # work with the other. |
|
369 | # work with the other. | |
360 | # |
|
370 | # | |
361 | # In case of conflict, parent 1 take precedence over parent 2. |
|
371 | # In case of conflict, parent 1 take precedence over parent 2. | |
362 | # This is an arbitrary choice made anew when implementing |
|
372 | # This is an arbitrary choice made anew when implementing | |
363 | # changeset based copies. It was made without regards with |
|
373 | # changeset based copies. It was made without regards with | |
364 | # potential filelog related behavior. |
|
374 | # potential filelog related behavior. | |
365 | if parent == 1: |
|
375 | if parent == 1: | |
366 | _merge_copies_dict( |
|
376 | _merge_copies_dict( | |
367 | othercopies, newcopies, isancestor, changes |
|
377 | othercopies, newcopies, isancestor, changes | |
368 | ) |
|
378 | ) | |
369 | else: |
|
379 | else: | |
370 | _merge_copies_dict( |
|
380 | _merge_copies_dict( | |
371 | newcopies, othercopies, isancestor, changes |
|
381 | newcopies, othercopies, isancestor, changes | |
372 | ) |
|
382 | ) | |
373 | all_copies[c] = newcopies |
|
383 | all_copies[c] = newcopies | |
374 |
|
384 | |||
375 | final_copies = {} |
|
385 | final_copies = {} | |
376 | for dest, (tt, source) in all_copies[targetrev].items(): |
|
386 | for dest, (tt, source) in all_copies[targetrev].items(): | |
377 | if source is not None: |
|
387 | if source is not None: | |
378 | final_copies[dest] = source |
|
388 | final_copies[dest] = source | |
379 | return final_copies |
|
389 | return final_copies | |
380 |
|
390 | |||
381 |
|
391 | |||
382 | def _merge_copies_dict(minor, major, isancestor, changes): |
|
392 | def _merge_copies_dict(minor, major, isancestor, changes): | |
383 | """merge two copies-mapping together, minor and major |
|
393 | """merge two copies-mapping together, minor and major | |
384 |
|
394 | |||
385 | In case of conflict, value from "major" will be picked. |
|
395 | In case of conflict, value from "major" will be picked. | |
386 |
|
396 | |||
387 | - `isancestors(low_rev, high_rev)`: callable return True if `low_rev` is an |
|
397 | - `isancestors(low_rev, high_rev)`: callable return True if `low_rev` is an | |
388 | ancestors of `high_rev`, |
|
398 | ancestors of `high_rev`, | |
389 |
|
399 | |||
390 | - `ismerged(path)`: callable return True if `path` have been merged in the |
|
400 | - `ismerged(path)`: callable return True if `path` have been merged in the | |
391 | current revision, |
|
401 | current revision, | |
392 | """ |
|
402 | """ | |
393 | for dest, value in major.items(): |
|
403 | for dest, value in major.items(): | |
394 | other = minor.get(dest) |
|
404 | other = minor.get(dest) | |
395 | if other is None: |
|
405 | if other is None: | |
396 | minor[dest] = value |
|
406 | minor[dest] = value | |
397 | else: |
|
407 | else: | |
398 | new_tt = value[0] |
|
408 | new_tt = value[0] | |
399 | other_tt = other[0] |
|
409 | other_tt = other[0] | |
400 | if value[1] == other[1]: |
|
410 | if value[1] == other[1]: | |
401 | continue |
|
411 | continue | |
402 | # content from "major" wins, unless it is older |
|
412 | # content from "major" wins, unless it is older | |
403 | # than the branch point or there is a merge |
|
413 | # than the branch point or there is a merge | |
404 | if new_tt == other_tt: |
|
414 | if new_tt == other_tt: | |
405 | minor[dest] = value |
|
415 | minor[dest] = value | |
406 | elif ( |
|
416 | elif ( | |
407 | changes is not None |
|
417 | changes is not None | |
408 | and value[1] is None |
|
418 | and value[1] is None | |
409 | and dest in changes.salvaged |
|
419 | and dest in changes.salvaged | |
410 | ): |
|
420 | ): | |
411 | pass |
|
421 | pass | |
412 | elif ( |
|
422 | elif ( | |
413 | changes is not None |
|
423 | changes is not None | |
414 | and other[1] is None |
|
424 | and other[1] is None | |
415 | and dest in changes.salvaged |
|
425 | and dest in changes.salvaged | |
416 | ): |
|
426 | ): | |
417 | minor[dest] = value |
|
427 | minor[dest] = value | |
418 | elif changes is not None and dest in changes.merged: |
|
428 | elif changes is not None and dest in changes.merged: | |
419 | minor[dest] = value |
|
429 | minor[dest] = value | |
420 | elif not isancestor(new_tt, other_tt): |
|
430 | elif not isancestor(new_tt, other_tt): | |
421 | if value[1] is not None: |
|
431 | if value[1] is not None: | |
422 | minor[dest] = value |
|
432 | minor[dest] = value | |
423 | elif isancestor(other_tt, new_tt): |
|
433 | elif isancestor(other_tt, new_tt): | |
424 | minor[dest] = value |
|
434 | minor[dest] = value | |
425 |
|
435 | |||
426 |
|
436 | |||
427 | def _revinfo_getter_extra(repo): |
|
437 | def _revinfo_getter_extra(repo): | |
428 | """return a function that return multiple data given a <rev>"i |
|
438 | """return a function that return multiple data given a <rev>"i | |
429 |
|
439 | |||
430 | * p1: revision number of first parent |
|
440 | * p1: revision number of first parent | |
431 | * p2: revision number of first parent |
|
441 | * p2: revision number of first parent | |
432 | * p1copies: mapping of copies from p1 |
|
442 | * p1copies: mapping of copies from p1 | |
433 | * p2copies: mapping of copies from p2 |
|
443 | * p2copies: mapping of copies from p2 | |
434 | * removed: a list of removed files |
|
444 | * removed: a list of removed files | |
435 | * ismerged: a callback to know if file was merged in that revision |
|
445 | * ismerged: a callback to know if file was merged in that revision | |
436 | """ |
|
446 | """ | |
437 | cl = repo.changelog |
|
447 | cl = repo.changelog | |
438 | parents = cl.parentrevs |
|
448 | parents = cl.parentrevs | |
439 |
|
449 | |||
440 | def get_ismerged(rev): |
|
450 | def get_ismerged(rev): | |
441 | ctx = repo[rev] |
|
451 | ctx = repo[rev] | |
442 |
|
452 | |||
443 | def ismerged(path): |
|
453 | def ismerged(path): | |
444 | if path not in ctx.files(): |
|
454 | if path not in ctx.files(): | |
445 | return False |
|
455 | return False | |
446 | fctx = ctx[path] |
|
456 | fctx = ctx[path] | |
447 | parents = fctx._filelog.parents(fctx._filenode) |
|
457 | parents = fctx._filelog.parents(fctx._filenode) | |
448 | nb_parents = 0 |
|
458 | nb_parents = 0 | |
449 | for n in parents: |
|
459 | for n in parents: | |
450 | if n != node.nullid: |
|
460 | if n != node.nullid: | |
451 | nb_parents += 1 |
|
461 | nb_parents += 1 | |
452 | return nb_parents >= 2 |
|
462 | return nb_parents >= 2 | |
453 |
|
463 | |||
454 | return ismerged |
|
464 | return ismerged | |
455 |
|
465 | |||
456 | def revinfo(rev): |
|
466 | def revinfo(rev): | |
457 | p1, p2 = parents(rev) |
|
467 | p1, p2 = parents(rev) | |
458 | ctx = repo[rev] |
|
468 | ctx = repo[rev] | |
459 | p1copies, p2copies = ctx._copies |
|
469 | p1copies, p2copies = ctx._copies | |
460 | removed = ctx.filesremoved() |
|
470 | removed = ctx.filesremoved() | |
461 | return p1, p2, p1copies, p2copies, removed, get_ismerged(rev) |
|
471 | return p1, p2, p1copies, p2copies, removed, get_ismerged(rev) | |
462 |
|
472 | |||
463 | return revinfo |
|
473 | return revinfo | |
464 |
|
474 | |||
465 |
|
475 | |||
466 | def _combine_changeset_copies_extra( |
|
476 | def _combine_changeset_copies_extra( | |
467 | revs, children, targetrev, revinfo, match, isancestor |
|
477 | revs, children, targetrev, revinfo, match, isancestor | |
468 | ): |
|
478 | ): | |
469 | """version of `_combine_changeset_copies` that works with the Google |
|
479 | """version of `_combine_changeset_copies` that works with the Google | |
470 | specific "extra" based storage for copy information""" |
|
480 | specific "extra" based storage for copy information""" | |
471 | all_copies = {} |
|
481 | all_copies = {} | |
472 | alwaysmatch = match.always() |
|
482 | alwaysmatch = match.always() | |
473 | for r in revs: |
|
483 | for r in revs: | |
474 | copies = all_copies.pop(r, None) |
|
484 | copies = all_copies.pop(r, None) | |
475 | if copies is None: |
|
485 | if copies is None: | |
476 | # this is a root |
|
486 | # this is a root | |
477 | copies = {} |
|
487 | copies = {} | |
478 | for i, c in enumerate(children[r]): |
|
488 | for i, c in enumerate(children[r]): | |
479 | p1, p2, p1copies, p2copies, removed, ismerged = revinfo(c) |
|
489 | p1, p2, p1copies, p2copies, removed, ismerged = revinfo(c) | |
480 | if r == p1: |
|
490 | if r == p1: | |
481 | parent = 1 |
|
491 | parent = 1 | |
482 | childcopies = p1copies |
|
492 | childcopies = p1copies | |
483 | else: |
|
493 | else: | |
484 | assert r == p2 |
|
494 | assert r == p2 | |
485 | parent = 2 |
|
495 | parent = 2 | |
486 | childcopies = p2copies |
|
496 | childcopies = p2copies | |
487 | if not alwaysmatch: |
|
497 | if not alwaysmatch: | |
488 | childcopies = { |
|
498 | childcopies = { | |
489 | dst: src for dst, src in childcopies.items() if match(dst) |
|
499 | dst: src for dst, src in childcopies.items() if match(dst) | |
490 | } |
|
500 | } | |
491 | newcopies = copies |
|
501 | newcopies = copies | |
492 | if childcopies: |
|
502 | if childcopies: | |
493 | newcopies = copies.copy() |
|
503 | newcopies = copies.copy() | |
494 | for dest, source in pycompat.iteritems(childcopies): |
|
504 | for dest, source in pycompat.iteritems(childcopies): | |
495 | prev = copies.get(source) |
|
505 | prev = copies.get(source) | |
496 | if prev is not None and prev[1] is not None: |
|
506 | if prev is not None and prev[1] is not None: | |
497 | source = prev[1] |
|
507 | source = prev[1] | |
498 | newcopies[dest] = (c, source) |
|
508 | newcopies[dest] = (c, source) | |
499 | assert newcopies is not copies |
|
509 | assert newcopies is not copies | |
500 | for f in removed: |
|
510 | for f in removed: | |
501 | if f in newcopies: |
|
511 | if f in newcopies: | |
502 | if newcopies is copies: |
|
512 | if newcopies is copies: | |
503 | # copy on write to avoid affecting potential other |
|
513 | # copy on write to avoid affecting potential other | |
504 | # branches. when there are no other branches, this |
|
514 | # branches. when there are no other branches, this | |
505 | # could be avoided. |
|
515 | # could be avoided. | |
506 | newcopies = copies.copy() |
|
516 | newcopies = copies.copy() | |
507 | newcopies[f] = (c, None) |
|
517 | newcopies[f] = (c, None) | |
508 | othercopies = all_copies.get(c) |
|
518 | othercopies = all_copies.get(c) | |
509 | if othercopies is None: |
|
519 | if othercopies is None: | |
510 | all_copies[c] = newcopies |
|
520 | all_copies[c] = newcopies | |
511 | else: |
|
521 | else: | |
512 | # we are the second parent to work on c, we need to merge our |
|
522 | # we are the second parent to work on c, we need to merge our | |
513 | # work with the other. |
|
523 | # work with the other. | |
514 | # |
|
524 | # | |
515 | # In case of conflict, parent 1 take precedence over parent 2. |
|
525 | # In case of conflict, parent 1 take precedence over parent 2. | |
516 | # This is an arbitrary choice made anew when implementing |
|
526 | # This is an arbitrary choice made anew when implementing | |
517 | # changeset based copies. It was made without regards with |
|
527 | # changeset based copies. It was made without regards with | |
518 | # potential filelog related behavior. |
|
528 | # potential filelog related behavior. | |
519 | if parent == 1: |
|
529 | if parent == 1: | |
520 | _merge_copies_dict_extra( |
|
530 | _merge_copies_dict_extra( | |
521 | othercopies, newcopies, isancestor, ismerged |
|
531 | othercopies, newcopies, isancestor, ismerged | |
522 | ) |
|
532 | ) | |
523 | else: |
|
533 | else: | |
524 | _merge_copies_dict_extra( |
|
534 | _merge_copies_dict_extra( | |
525 | newcopies, othercopies, isancestor, ismerged |
|
535 | newcopies, othercopies, isancestor, ismerged | |
526 | ) |
|
536 | ) | |
527 | all_copies[c] = newcopies |
|
537 | all_copies[c] = newcopies | |
528 |
|
538 | |||
529 | final_copies = {} |
|
539 | final_copies = {} | |
530 | for dest, (tt, source) in all_copies[targetrev].items(): |
|
540 | for dest, (tt, source) in all_copies[targetrev].items(): | |
531 | if source is not None: |
|
541 | if source is not None: | |
532 | final_copies[dest] = source |
|
542 | final_copies[dest] = source | |
533 | return final_copies |
|
543 | return final_copies | |
534 |
|
544 | |||
535 |
|
545 | |||
536 | def _merge_copies_dict_extra(minor, major, isancestor, ismerged): |
|
546 | def _merge_copies_dict_extra(minor, major, isancestor, ismerged): | |
537 | """version of `_merge_copies_dict` that works with the Google |
|
547 | """version of `_merge_copies_dict` that works with the Google | |
538 | specific "extra" based storage for copy information""" |
|
548 | specific "extra" based storage for copy information""" | |
539 | for dest, value in major.items(): |
|
549 | for dest, value in major.items(): | |
540 | other = minor.get(dest) |
|
550 | other = minor.get(dest) | |
541 | if other is None: |
|
551 | if other is None: | |
542 | minor[dest] = value |
|
552 | minor[dest] = value | |
543 | else: |
|
553 | else: | |
544 | new_tt = value[0] |
|
554 | new_tt = value[0] | |
545 | other_tt = other[0] |
|
555 | other_tt = other[0] | |
546 | if value[1] == other[1]: |
|
556 | if value[1] == other[1]: | |
547 | continue |
|
557 | continue | |
548 | # content from "major" wins, unless it is older |
|
558 | # content from "major" wins, unless it is older | |
549 | # than the branch point or there is a merge |
|
559 | # than the branch point or there is a merge | |
550 | if ( |
|
560 | if ( | |
551 | new_tt == other_tt |
|
561 | new_tt == other_tt | |
552 | or not isancestor(new_tt, other_tt) |
|
562 | or not isancestor(new_tt, other_tt) | |
553 | or ismerged(dest) |
|
563 | or ismerged(dest) | |
554 | ): |
|
564 | ): | |
555 | minor[dest] = value |
|
565 | minor[dest] = value | |
556 |
|
566 | |||
557 |
|
567 | |||
558 | def _forwardcopies(a, b, base=None, match=None): |
|
568 | def _forwardcopies(a, b, base=None, match=None): | |
559 | """find {dst@b: src@a} copy mapping where a is an ancestor of b""" |
|
569 | """find {dst@b: src@a} copy mapping where a is an ancestor of b""" | |
560 |
|
570 | |||
561 | if base is None: |
|
571 | if base is None: | |
562 | base = a |
|
572 | base = a | |
563 | match = a.repo().narrowmatch(match) |
|
573 | match = a.repo().narrowmatch(match) | |
564 | # check for working copy |
|
574 | # check for working copy | |
565 | if b.rev() is None: |
|
575 | if b.rev() is None: | |
566 | cm = _committedforwardcopies(a, b.p1(), base, match) |
|
576 | cm = _committedforwardcopies(a, b.p1(), base, match) | |
567 | # combine copies from dirstate if necessary |
|
577 | # combine copies from dirstate if necessary | |
568 | copies = _chain(cm, _dirstatecopies(b._repo, match)) |
|
578 | copies = _chain(cm, _dirstatecopies(b._repo, match)) | |
569 | else: |
|
579 | else: | |
570 | copies = _committedforwardcopies(a, b, base, match) |
|
580 | copies = _committedforwardcopies(a, b, base, match) | |
571 | return copies |
|
581 | return copies | |
572 |
|
582 | |||
573 |
|
583 | |||
574 | def _backwardrenames(a, b, match): |
|
584 | def _backwardrenames(a, b, match): | |
575 | if a._repo.ui.config(b'experimental', b'copytrace') == b'off': |
|
585 | if a._repo.ui.config(b'experimental', b'copytrace') == b'off': | |
576 | return {} |
|
586 | return {} | |
577 |
|
587 | |||
578 | # Even though we're not taking copies into account, 1:n rename situations |
|
588 | # Even though we're not taking copies into account, 1:n rename situations | |
579 | # can still exist (e.g. hg cp a b; hg mv a c). In those cases we |
|
589 | # can still exist (e.g. hg cp a b; hg mv a c). In those cases we | |
580 | # arbitrarily pick one of the renames. |
|
590 | # arbitrarily pick one of the renames. | |
581 | # We don't want to pass in "match" here, since that would filter |
|
591 | # We don't want to pass in "match" here, since that would filter | |
582 | # the destination by it. Since we're reversing the copies, we want |
|
592 | # the destination by it. Since we're reversing the copies, we want | |
583 | # to filter the source instead. |
|
593 | # to filter the source instead. | |
584 | f = _forwardcopies(b, a) |
|
594 | f = _forwardcopies(b, a) | |
585 | r = {} |
|
595 | r = {} | |
586 | for k, v in sorted(pycompat.iteritems(f)): |
|
596 | for k, v in sorted(pycompat.iteritems(f)): | |
587 | if match and not match(v): |
|
597 | if match and not match(v): | |
588 | continue |
|
598 | continue | |
589 | # remove copies |
|
599 | # remove copies | |
590 | if v in a: |
|
600 | if v in a: | |
591 | continue |
|
601 | continue | |
592 | r[v] = k |
|
602 | r[v] = k | |
593 | return r |
|
603 | return r | |
594 |
|
604 | |||
595 |
|
605 | |||
596 | def pathcopies(x, y, match=None): |
|
606 | def pathcopies(x, y, match=None): | |
597 | """find {dst@y: src@x} copy mapping for directed compare""" |
|
607 | """find {dst@y: src@x} copy mapping for directed compare""" | |
598 | repo = x._repo |
|
608 | repo = x._repo | |
599 | debug = repo.ui.debugflag and repo.ui.configbool(b'devel', b'debug.copies') |
|
609 | debug = repo.ui.debugflag and repo.ui.configbool(b'devel', b'debug.copies') | |
600 | if debug: |
|
610 | if debug: | |
601 | repo.ui.debug( |
|
611 | repo.ui.debug( | |
602 | b'debug.copies: searching copies from %s to %s\n' % (x, y) |
|
612 | b'debug.copies: searching copies from %s to %s\n' % (x, y) | |
603 | ) |
|
613 | ) | |
604 | if x == y or not x or not y: |
|
614 | if x == y or not x or not y: | |
605 | return {} |
|
615 | return {} | |
606 | if y.rev() is None and x == y.p1(): |
|
616 | if y.rev() is None and x == y.p1(): | |
607 | if debug: |
|
617 | if debug: | |
608 | repo.ui.debug(b'debug.copies: search mode: dirstate\n') |
|
618 | repo.ui.debug(b'debug.copies: search mode: dirstate\n') | |
609 | # short-circuit to avoid issues with merge states |
|
619 | # short-circuit to avoid issues with merge states | |
610 | return _dirstatecopies(repo, match) |
|
620 | return _dirstatecopies(repo, match) | |
611 | a = y.ancestor(x) |
|
621 | a = y.ancestor(x) | |
612 | if a == x: |
|
622 | if a == x: | |
613 | if debug: |
|
623 | if debug: | |
614 | repo.ui.debug(b'debug.copies: search mode: forward\n') |
|
624 | repo.ui.debug(b'debug.copies: search mode: forward\n') | |
615 | copies = _forwardcopies(x, y, match=match) |
|
625 | copies = _forwardcopies(x, y, match=match) | |
616 | elif a == y: |
|
626 | elif a == y: | |
617 | if debug: |
|
627 | if debug: | |
618 | repo.ui.debug(b'debug.copies: search mode: backward\n') |
|
628 | repo.ui.debug(b'debug.copies: search mode: backward\n') | |
619 | copies = _backwardrenames(x, y, match=match) |
|
629 | copies = _backwardrenames(x, y, match=match) | |
620 | else: |
|
630 | else: | |
621 | if debug: |
|
631 | if debug: | |
622 | repo.ui.debug(b'debug.copies: search mode: combined\n') |
|
632 | repo.ui.debug(b'debug.copies: search mode: combined\n') | |
623 | base = None |
|
633 | base = None | |
624 | if a.rev() != node.nullrev: |
|
634 | if a.rev() != node.nullrev: | |
625 | base = x |
|
635 | base = x | |
626 | copies = _chain( |
|
636 | copies = _chain( | |
627 | _backwardrenames(x, a, match=match), |
|
637 | _backwardrenames(x, a, match=match), | |
628 | _forwardcopies(a, y, base, match=match), |
|
638 | _forwardcopies(a, y, base, match=match), | |
629 | ) |
|
639 | ) | |
630 | _filter(x, y, copies) |
|
640 | _filter(x, y, copies) | |
631 | return copies |
|
641 | return copies | |
632 |
|
642 | |||
633 |
|
643 | |||
634 | def mergecopies(repo, c1, c2, base): |
|
644 | def mergecopies(repo, c1, c2, base): | |
635 | """ |
|
645 | """ | |
636 | Finds moves and copies between context c1 and c2 that are relevant for |
|
646 | Finds moves and copies between context c1 and c2 that are relevant for | |
637 | merging. 'base' will be used as the merge base. |
|
647 | merging. 'base' will be used as the merge base. | |
638 |
|
648 | |||
639 | Copytracing is used in commands like rebase, merge, unshelve, etc to merge |
|
649 | Copytracing is used in commands like rebase, merge, unshelve, etc to merge | |
640 | files that were moved/ copied in one merge parent and modified in another. |
|
650 | files that were moved/ copied in one merge parent and modified in another. | |
641 | For example: |
|
651 | For example: | |
642 |
|
652 | |||
643 | o ---> 4 another commit |
|
653 | o ---> 4 another commit | |
644 | | |
|
654 | | | |
645 | | o ---> 3 commit that modifies a.txt |
|
655 | | o ---> 3 commit that modifies a.txt | |
646 | | / |
|
656 | | / | |
647 | o / ---> 2 commit that moves a.txt to b.txt |
|
657 | o / ---> 2 commit that moves a.txt to b.txt | |
648 | |/ |
|
658 | |/ | |
649 | o ---> 1 merge base |
|
659 | o ---> 1 merge base | |
650 |
|
660 | |||
651 | If we try to rebase revision 3 on revision 4, since there is no a.txt in |
|
661 | If we try to rebase revision 3 on revision 4, since there is no a.txt in | |
652 | revision 4, and if user have copytrace disabled, we prints the following |
|
662 | revision 4, and if user have copytrace disabled, we prints the following | |
653 | message: |
|
663 | message: | |
654 |
|
664 | |||
655 | ```other changed <file> which local deleted``` |
|
665 | ```other changed <file> which local deleted``` | |
656 |
|
666 | |||
657 | Returns a tuple where: |
|
667 | Returns a tuple where: | |
658 |
|
668 | |||
659 | "branch_copies" an instance of branch_copies. |
|
669 | "branch_copies" an instance of branch_copies. | |
660 |
|
670 | |||
661 | "diverge" is a mapping of source name -> list of destination names |
|
671 | "diverge" is a mapping of source name -> list of destination names | |
662 | for divergent renames. |
|
672 | for divergent renames. | |
663 |
|
673 | |||
664 | This function calls different copytracing algorithms based on config. |
|
674 | This function calls different copytracing algorithms based on config. | |
665 | """ |
|
675 | """ | |
666 | # avoid silly behavior for update from empty dir |
|
676 | # avoid silly behavior for update from empty dir | |
667 | if not c1 or not c2 or c1 == c2: |
|
677 | if not c1 or not c2 or c1 == c2: | |
668 | return branch_copies(), branch_copies(), {} |
|
678 | return branch_copies(), branch_copies(), {} | |
669 |
|
679 | |||
670 | narrowmatch = c1.repo().narrowmatch() |
|
680 | narrowmatch = c1.repo().narrowmatch() | |
671 |
|
681 | |||
672 | # avoid silly behavior for parent -> working dir |
|
682 | # avoid silly behavior for parent -> working dir | |
673 | if c2.node() is None and c1.node() == repo.dirstate.p1(): |
|
683 | if c2.node() is None and c1.node() == repo.dirstate.p1(): | |
674 | return ( |
|
684 | return ( | |
675 | branch_copies(_dirstatecopies(repo, narrowmatch)), |
|
685 | branch_copies(_dirstatecopies(repo, narrowmatch)), | |
676 | branch_copies(), |
|
686 | branch_copies(), | |
677 | {}, |
|
687 | {}, | |
678 | ) |
|
688 | ) | |
679 |
|
689 | |||
680 | copytracing = repo.ui.config(b'experimental', b'copytrace') |
|
690 | copytracing = repo.ui.config(b'experimental', b'copytrace') | |
681 | if stringutil.parsebool(copytracing) is False: |
|
691 | if stringutil.parsebool(copytracing) is False: | |
682 | # stringutil.parsebool() returns None when it is unable to parse the |
|
692 | # stringutil.parsebool() returns None when it is unable to parse the | |
683 | # value, so we should rely on making sure copytracing is on such cases |
|
693 | # value, so we should rely on making sure copytracing is on such cases | |
684 | return branch_copies(), branch_copies(), {} |
|
694 | return branch_copies(), branch_copies(), {} | |
685 |
|
695 | |||
686 | if usechangesetcentricalgo(repo): |
|
696 | if usechangesetcentricalgo(repo): | |
687 | # The heuristics don't make sense when we need changeset-centric algos |
|
697 | # The heuristics don't make sense when we need changeset-centric algos | |
688 | return _fullcopytracing(repo, c1, c2, base) |
|
698 | return _fullcopytracing(repo, c1, c2, base) | |
689 |
|
699 | |||
690 | # Copy trace disabling is explicitly below the node == p1 logic above |
|
700 | # Copy trace disabling is explicitly below the node == p1 logic above | |
691 | # because the logic above is required for a simple copy to be kept across a |
|
701 | # because the logic above is required for a simple copy to be kept across a | |
692 | # rebase. |
|
702 | # rebase. | |
693 | if copytracing == b'heuristics': |
|
703 | if copytracing == b'heuristics': | |
694 | # Do full copytracing if only non-public revisions are involved as |
|
704 | # Do full copytracing if only non-public revisions are involved as | |
695 | # that will be fast enough and will also cover the copies which could |
|
705 | # that will be fast enough and will also cover the copies which could | |
696 | # be missed by heuristics |
|
706 | # be missed by heuristics | |
697 | if _isfullcopytraceable(repo, c1, base): |
|
707 | if _isfullcopytraceable(repo, c1, base): | |
698 | return _fullcopytracing(repo, c1, c2, base) |
|
708 | return _fullcopytracing(repo, c1, c2, base) | |
699 | return _heuristicscopytracing(repo, c1, c2, base) |
|
709 | return _heuristicscopytracing(repo, c1, c2, base) | |
700 | else: |
|
710 | else: | |
701 | return _fullcopytracing(repo, c1, c2, base) |
|
711 | return _fullcopytracing(repo, c1, c2, base) | |
702 |
|
712 | |||
703 |
|
713 | |||
704 | def _isfullcopytraceable(repo, c1, base): |
|
714 | def _isfullcopytraceable(repo, c1, base): | |
705 | """Checks that if base, source and destination are all no-public branches, |
|
715 | """Checks that if base, source and destination are all no-public branches, | |
706 | if yes let's use the full copytrace algorithm for increased capabilities |
|
716 | if yes let's use the full copytrace algorithm for increased capabilities | |
707 | since it will be fast enough. |
|
717 | since it will be fast enough. | |
708 |
|
718 | |||
709 | `experimental.copytrace.sourcecommitlimit` can be used to set a limit for |
|
719 | `experimental.copytrace.sourcecommitlimit` can be used to set a limit for | |
710 | number of changesets from c1 to base such that if number of changesets are |
|
720 | number of changesets from c1 to base such that if number of changesets are | |
711 | more than the limit, full copytracing algorithm won't be used. |
|
721 | more than the limit, full copytracing algorithm won't be used. | |
712 | """ |
|
722 | """ | |
713 | if c1.rev() is None: |
|
723 | if c1.rev() is None: | |
714 | c1 = c1.p1() |
|
724 | c1 = c1.p1() | |
715 | if c1.mutable() and base.mutable(): |
|
725 | if c1.mutable() and base.mutable(): | |
716 | sourcecommitlimit = repo.ui.configint( |
|
726 | sourcecommitlimit = repo.ui.configint( | |
717 | b'experimental', b'copytrace.sourcecommitlimit' |
|
727 | b'experimental', b'copytrace.sourcecommitlimit' | |
718 | ) |
|
728 | ) | |
719 | commits = len(repo.revs(b'%d::%d', base.rev(), c1.rev())) |
|
729 | commits = len(repo.revs(b'%d::%d', base.rev(), c1.rev())) | |
720 | return commits < sourcecommitlimit |
|
730 | return commits < sourcecommitlimit | |
721 | return False |
|
731 | return False | |
722 |
|
732 | |||
723 |
|
733 | |||
724 | def _checksinglesidecopies( |
|
734 | def _checksinglesidecopies( | |
725 | src, dsts1, m1, m2, mb, c2, base, copy, renamedelete |
|
735 | src, dsts1, m1, m2, mb, c2, base, copy, renamedelete | |
726 | ): |
|
736 | ): | |
727 | if src not in m2: |
|
737 | if src not in m2: | |
728 | # deleted on side 2 |
|
738 | # deleted on side 2 | |
729 | if src not in m1: |
|
739 | if src not in m1: | |
730 | # renamed on side 1, deleted on side 2 |
|
740 | # renamed on side 1, deleted on side 2 | |
731 | renamedelete[src] = dsts1 |
|
741 | renamedelete[src] = dsts1 | |
732 | elif src not in mb: |
|
742 | elif src not in mb: | |
733 | # Work around the "short-circuit to avoid issues with merge states" |
|
743 | # Work around the "short-circuit to avoid issues with merge states" | |
734 | # thing in pathcopies(): pathcopies(x, y) can return a copy where the |
|
744 | # thing in pathcopies(): pathcopies(x, y) can return a copy where the | |
735 | # destination doesn't exist in y. |
|
745 | # destination doesn't exist in y. | |
736 | pass |
|
746 | pass | |
737 | elif mb[src] != m2[src] and not _related(c2[src], base[src]): |
|
747 | elif mb[src] != m2[src] and not _related(c2[src], base[src]): | |
738 | return |
|
748 | return | |
739 | elif mb[src] != m2[src] or mb.flags(src) != m2.flags(src): |
|
749 | elif mb[src] != m2[src] or mb.flags(src) != m2.flags(src): | |
740 | # modified on side 2 |
|
750 | # modified on side 2 | |
741 | for dst in dsts1: |
|
751 | for dst in dsts1: | |
742 | copy[dst] = src |
|
752 | copy[dst] = src | |
743 |
|
753 | |||
744 |
|
754 | |||
745 | class branch_copies(object): |
|
755 | class branch_copies(object): | |
746 | """Information about copies made on one side of a merge/graft. |
|
756 | """Information about copies made on one side of a merge/graft. | |
747 |
|
757 | |||
748 | "copy" is a mapping from destination name -> source name, |
|
758 | "copy" is a mapping from destination name -> source name, | |
749 | where source is in c1 and destination is in c2 or vice-versa. |
|
759 | where source is in c1 and destination is in c2 or vice-versa. | |
750 |
|
760 | |||
751 | "movewithdir" is a mapping from source name -> destination name, |
|
761 | "movewithdir" is a mapping from source name -> destination name, | |
752 | where the file at source present in one context but not the other |
|
762 | where the file at source present in one context but not the other | |
753 | needs to be moved to destination by the merge process, because the |
|
763 | needs to be moved to destination by the merge process, because the | |
754 | other context moved the directory it is in. |
|
764 | other context moved the directory it is in. | |
755 |
|
765 | |||
756 | "renamedelete" is a mapping of source name -> list of destination |
|
766 | "renamedelete" is a mapping of source name -> list of destination | |
757 | names for files deleted in c1 that were renamed in c2 or vice-versa. |
|
767 | names for files deleted in c1 that were renamed in c2 or vice-versa. | |
758 |
|
768 | |||
759 | "dirmove" is a mapping of detected source dir -> destination dir renames. |
|
769 | "dirmove" is a mapping of detected source dir -> destination dir renames. | |
760 | This is needed for handling changes to new files previously grafted into |
|
770 | This is needed for handling changes to new files previously grafted into | |
761 | renamed directories. |
|
771 | renamed directories. | |
762 | """ |
|
772 | """ | |
763 |
|
773 | |||
764 | def __init__( |
|
774 | def __init__( | |
765 | self, copy=None, renamedelete=None, dirmove=None, movewithdir=None |
|
775 | self, copy=None, renamedelete=None, dirmove=None, movewithdir=None | |
766 | ): |
|
776 | ): | |
767 | self.copy = {} if copy is None else copy |
|
777 | self.copy = {} if copy is None else copy | |
768 | self.renamedelete = {} if renamedelete is None else renamedelete |
|
778 | self.renamedelete = {} if renamedelete is None else renamedelete | |
769 | self.dirmove = {} if dirmove is None else dirmove |
|
779 | self.dirmove = {} if dirmove is None else dirmove | |
770 | self.movewithdir = {} if movewithdir is None else movewithdir |
|
780 | self.movewithdir = {} if movewithdir is None else movewithdir | |
771 |
|
781 | |||
772 | def __repr__(self): |
|
782 | def __repr__(self): | |
773 | return '<branch_copies\n copy=%r\n renamedelete=%r\n dirmove=%r\n movewithdir=%r\n>' % ( |
|
783 | return '<branch_copies\n copy=%r\n renamedelete=%r\n dirmove=%r\n movewithdir=%r\n>' % ( | |
774 | self.copy, |
|
784 | self.copy, | |
775 | self.renamedelete, |
|
785 | self.renamedelete, | |
776 | self.dirmove, |
|
786 | self.dirmove, | |
777 | self.movewithdir, |
|
787 | self.movewithdir, | |
778 | ) |
|
788 | ) | |
779 |
|
789 | |||
780 |
|
790 | |||
781 | def _fullcopytracing(repo, c1, c2, base): |
|
791 | def _fullcopytracing(repo, c1, c2, base): | |
782 | """The full copytracing algorithm which finds all the new files that were |
|
792 | """The full copytracing algorithm which finds all the new files that were | |
783 | added from merge base up to the top commit and for each file it checks if |
|
793 | added from merge base up to the top commit and for each file it checks if | |
784 | this file was copied from another file. |
|
794 | this file was copied from another file. | |
785 |
|
795 | |||
786 | This is pretty slow when a lot of changesets are involved but will track all |
|
796 | This is pretty slow when a lot of changesets are involved but will track all | |
787 | the copies. |
|
797 | the copies. | |
788 | """ |
|
798 | """ | |
789 | m1 = c1.manifest() |
|
799 | m1 = c1.manifest() | |
790 | m2 = c2.manifest() |
|
800 | m2 = c2.manifest() | |
791 | mb = base.manifest() |
|
801 | mb = base.manifest() | |
792 |
|
802 | |||
793 | copies1 = pathcopies(base, c1) |
|
803 | copies1 = pathcopies(base, c1) | |
794 | copies2 = pathcopies(base, c2) |
|
804 | copies2 = pathcopies(base, c2) | |
795 |
|
805 | |||
796 | if not (copies1 or copies2): |
|
806 | if not (copies1 or copies2): | |
797 | return branch_copies(), branch_copies(), {} |
|
807 | return branch_copies(), branch_copies(), {} | |
798 |
|
808 | |||
799 | inversecopies1 = {} |
|
809 | inversecopies1 = {} | |
800 | inversecopies2 = {} |
|
810 | inversecopies2 = {} | |
801 | for dst, src in copies1.items(): |
|
811 | for dst, src in copies1.items(): | |
802 | inversecopies1.setdefault(src, []).append(dst) |
|
812 | inversecopies1.setdefault(src, []).append(dst) | |
803 | for dst, src in copies2.items(): |
|
813 | for dst, src in copies2.items(): | |
804 | inversecopies2.setdefault(src, []).append(dst) |
|
814 | inversecopies2.setdefault(src, []).append(dst) | |
805 |
|
815 | |||
806 | copy1 = {} |
|
816 | copy1 = {} | |
807 | copy2 = {} |
|
817 | copy2 = {} | |
808 | diverge = {} |
|
818 | diverge = {} | |
809 | renamedelete1 = {} |
|
819 | renamedelete1 = {} | |
810 | renamedelete2 = {} |
|
820 | renamedelete2 = {} | |
811 | allsources = set(inversecopies1) | set(inversecopies2) |
|
821 | allsources = set(inversecopies1) | set(inversecopies2) | |
812 | for src in allsources: |
|
822 | for src in allsources: | |
813 | dsts1 = inversecopies1.get(src) |
|
823 | dsts1 = inversecopies1.get(src) | |
814 | dsts2 = inversecopies2.get(src) |
|
824 | dsts2 = inversecopies2.get(src) | |
815 | if dsts1 and dsts2: |
|
825 | if dsts1 and dsts2: | |
816 | # copied/renamed on both sides |
|
826 | # copied/renamed on both sides | |
817 | if src not in m1 and src not in m2: |
|
827 | if src not in m1 and src not in m2: | |
818 | # renamed on both sides |
|
828 | # renamed on both sides | |
819 | dsts1 = set(dsts1) |
|
829 | dsts1 = set(dsts1) | |
820 | dsts2 = set(dsts2) |
|
830 | dsts2 = set(dsts2) | |
821 | # If there's some overlap in the rename destinations, we |
|
831 | # If there's some overlap in the rename destinations, we | |
822 | # consider it not divergent. For example, if side 1 copies 'a' |
|
832 | # consider it not divergent. For example, if side 1 copies 'a' | |
823 | # to 'b' and 'c' and deletes 'a', and side 2 copies 'a' to 'c' |
|
833 | # to 'b' and 'c' and deletes 'a', and side 2 copies 'a' to 'c' | |
824 | # and 'd' and deletes 'a'. |
|
834 | # and 'd' and deletes 'a'. | |
825 | if dsts1 & dsts2: |
|
835 | if dsts1 & dsts2: | |
826 | for dst in dsts1 & dsts2: |
|
836 | for dst in dsts1 & dsts2: | |
827 | copy1[dst] = src |
|
837 | copy1[dst] = src | |
828 | copy2[dst] = src |
|
838 | copy2[dst] = src | |
829 | else: |
|
839 | else: | |
830 | diverge[src] = sorted(dsts1 | dsts2) |
|
840 | diverge[src] = sorted(dsts1 | dsts2) | |
831 | elif src in m1 and src in m2: |
|
841 | elif src in m1 and src in m2: | |
832 | # copied on both sides |
|
842 | # copied on both sides | |
833 | dsts1 = set(dsts1) |
|
843 | dsts1 = set(dsts1) | |
834 | dsts2 = set(dsts2) |
|
844 | dsts2 = set(dsts2) | |
835 | for dst in dsts1 & dsts2: |
|
845 | for dst in dsts1 & dsts2: | |
836 | copy1[dst] = src |
|
846 | copy1[dst] = src | |
837 | copy2[dst] = src |
|
847 | copy2[dst] = src | |
838 | # TODO: Handle cases where it was renamed on one side and copied |
|
848 | # TODO: Handle cases where it was renamed on one side and copied | |
839 | # on the other side |
|
849 | # on the other side | |
840 | elif dsts1: |
|
850 | elif dsts1: | |
841 | # copied/renamed only on side 1 |
|
851 | # copied/renamed only on side 1 | |
842 | _checksinglesidecopies( |
|
852 | _checksinglesidecopies( | |
843 | src, dsts1, m1, m2, mb, c2, base, copy1, renamedelete1 |
|
853 | src, dsts1, m1, m2, mb, c2, base, copy1, renamedelete1 | |
844 | ) |
|
854 | ) | |
845 | elif dsts2: |
|
855 | elif dsts2: | |
846 | # copied/renamed only on side 2 |
|
856 | # copied/renamed only on side 2 | |
847 | _checksinglesidecopies( |
|
857 | _checksinglesidecopies( | |
848 | src, dsts2, m2, m1, mb, c1, base, copy2, renamedelete2 |
|
858 | src, dsts2, m2, m1, mb, c1, base, copy2, renamedelete2 | |
849 | ) |
|
859 | ) | |
850 |
|
860 | |||
851 | # find interesting file sets from manifests |
|
861 | # find interesting file sets from manifests | |
852 | addedinm1 = m1.filesnotin(mb, repo.narrowmatch()) |
|
862 | addedinm1 = m1.filesnotin(mb, repo.narrowmatch()) | |
853 | addedinm2 = m2.filesnotin(mb, repo.narrowmatch()) |
|
863 | addedinm2 = m2.filesnotin(mb, repo.narrowmatch()) | |
854 | u1 = sorted(addedinm1 - addedinm2) |
|
864 | u1 = sorted(addedinm1 - addedinm2) | |
855 | u2 = sorted(addedinm2 - addedinm1) |
|
865 | u2 = sorted(addedinm2 - addedinm1) | |
856 |
|
866 | |||
857 | header = b" unmatched files in %s" |
|
867 | header = b" unmatched files in %s" | |
858 | if u1: |
|
868 | if u1: | |
859 | repo.ui.debug(b"%s:\n %s\n" % (header % b'local', b"\n ".join(u1))) |
|
869 | repo.ui.debug(b"%s:\n %s\n" % (header % b'local', b"\n ".join(u1))) | |
860 | if u2: |
|
870 | if u2: | |
861 | repo.ui.debug(b"%s:\n %s\n" % (header % b'other', b"\n ".join(u2))) |
|
871 | repo.ui.debug(b"%s:\n %s\n" % (header % b'other', b"\n ".join(u2))) | |
862 |
|
872 | |||
863 | if repo.ui.debugflag: |
|
873 | if repo.ui.debugflag: | |
864 | renamedeleteset = set() |
|
874 | renamedeleteset = set() | |
865 | divergeset = set() |
|
875 | divergeset = set() | |
866 | for dsts in diverge.values(): |
|
876 | for dsts in diverge.values(): | |
867 | divergeset.update(dsts) |
|
877 | divergeset.update(dsts) | |
868 | for dsts in renamedelete1.values(): |
|
878 | for dsts in renamedelete1.values(): | |
869 | renamedeleteset.update(dsts) |
|
879 | renamedeleteset.update(dsts) | |
870 | for dsts in renamedelete2.values(): |
|
880 | for dsts in renamedelete2.values(): | |
871 | renamedeleteset.update(dsts) |
|
881 | renamedeleteset.update(dsts) | |
872 |
|
882 | |||
873 | repo.ui.debug( |
|
883 | repo.ui.debug( | |
874 | b" all copies found (* = to merge, ! = divergent, " |
|
884 | b" all copies found (* = to merge, ! = divergent, " | |
875 | b"% = renamed and deleted):\n" |
|
885 | b"% = renamed and deleted):\n" | |
876 | ) |
|
886 | ) | |
877 | for side, copies in ((b"local", copies1), (b"remote", copies2)): |
|
887 | for side, copies in ((b"local", copies1), (b"remote", copies2)): | |
878 | if not copies: |
|
888 | if not copies: | |
879 | continue |
|
889 | continue | |
880 | repo.ui.debug(b" on %s side:\n" % side) |
|
890 | repo.ui.debug(b" on %s side:\n" % side) | |
881 | for f in sorted(copies): |
|
891 | for f in sorted(copies): | |
882 | note = b"" |
|
892 | note = b"" | |
883 | if f in copy1 or f in copy2: |
|
893 | if f in copy1 or f in copy2: | |
884 | note += b"*" |
|
894 | note += b"*" | |
885 | if f in divergeset: |
|
895 | if f in divergeset: | |
886 | note += b"!" |
|
896 | note += b"!" | |
887 | if f in renamedeleteset: |
|
897 | if f in renamedeleteset: | |
888 | note += b"%" |
|
898 | note += b"%" | |
889 | repo.ui.debug( |
|
899 | repo.ui.debug( | |
890 | b" src: '%s' -> dst: '%s' %s\n" % (copies[f], f, note) |
|
900 | b" src: '%s' -> dst: '%s' %s\n" % (copies[f], f, note) | |
891 | ) |
|
901 | ) | |
892 | del renamedeleteset |
|
902 | del renamedeleteset | |
893 | del divergeset |
|
903 | del divergeset | |
894 |
|
904 | |||
895 | repo.ui.debug(b" checking for directory renames\n") |
|
905 | repo.ui.debug(b" checking for directory renames\n") | |
896 |
|
906 | |||
897 | dirmove1, movewithdir2 = _dir_renames(repo, c1, copy1, copies1, u2) |
|
907 | dirmove1, movewithdir2 = _dir_renames(repo, c1, copy1, copies1, u2) | |
898 | dirmove2, movewithdir1 = _dir_renames(repo, c2, copy2, copies2, u1) |
|
908 | dirmove2, movewithdir1 = _dir_renames(repo, c2, copy2, copies2, u1) | |
899 |
|
909 | |||
900 | branch_copies1 = branch_copies(copy1, renamedelete1, dirmove1, movewithdir1) |
|
910 | branch_copies1 = branch_copies(copy1, renamedelete1, dirmove1, movewithdir1) | |
901 | branch_copies2 = branch_copies(copy2, renamedelete2, dirmove2, movewithdir2) |
|
911 | branch_copies2 = branch_copies(copy2, renamedelete2, dirmove2, movewithdir2) | |
902 |
|
912 | |||
903 | return branch_copies1, branch_copies2, diverge |
|
913 | return branch_copies1, branch_copies2, diverge | |
904 |
|
914 | |||
905 |
|
915 | |||
906 | def _dir_renames(repo, ctx, copy, fullcopy, addedfiles): |
|
916 | def _dir_renames(repo, ctx, copy, fullcopy, addedfiles): | |
907 | """Finds moved directories and files that should move with them. |
|
917 | """Finds moved directories and files that should move with them. | |
908 |
|
918 | |||
909 | ctx: the context for one of the sides |
|
919 | ctx: the context for one of the sides | |
910 | copy: files copied on the same side (as ctx) |
|
920 | copy: files copied on the same side (as ctx) | |
911 | fullcopy: files copied on the same side (as ctx), including those that |
|
921 | fullcopy: files copied on the same side (as ctx), including those that | |
912 | merge.manifestmerge() won't care about |
|
922 | merge.manifestmerge() won't care about | |
913 | addedfiles: added files on the other side (compared to ctx) |
|
923 | addedfiles: added files on the other side (compared to ctx) | |
914 | """ |
|
924 | """ | |
915 | # generate a directory move map |
|
925 | # generate a directory move map | |
916 | d = ctx.dirs() |
|
926 | d = ctx.dirs() | |
917 | invalid = set() |
|
927 | invalid = set() | |
918 | dirmove = {} |
|
928 | dirmove = {} | |
919 |
|
929 | |||
920 | # examine each file copy for a potential directory move, which is |
|
930 | # examine each file copy for a potential directory move, which is | |
921 | # when all the files in a directory are moved to a new directory |
|
931 | # when all the files in a directory are moved to a new directory | |
922 | for dst, src in pycompat.iteritems(fullcopy): |
|
932 | for dst, src in pycompat.iteritems(fullcopy): | |
923 | dsrc, ddst = pathutil.dirname(src), pathutil.dirname(dst) |
|
933 | dsrc, ddst = pathutil.dirname(src), pathutil.dirname(dst) | |
924 | if dsrc in invalid: |
|
934 | if dsrc in invalid: | |
925 | # already seen to be uninteresting |
|
935 | # already seen to be uninteresting | |
926 | continue |
|
936 | continue | |
927 | elif dsrc in d and ddst in d: |
|
937 | elif dsrc in d and ddst in d: | |
928 | # directory wasn't entirely moved locally |
|
938 | # directory wasn't entirely moved locally | |
929 | invalid.add(dsrc) |
|
939 | invalid.add(dsrc) | |
930 | elif dsrc in dirmove and dirmove[dsrc] != ddst: |
|
940 | elif dsrc in dirmove and dirmove[dsrc] != ddst: | |
931 | # files from the same directory moved to two different places |
|
941 | # files from the same directory moved to two different places | |
932 | invalid.add(dsrc) |
|
942 | invalid.add(dsrc) | |
933 | else: |
|
943 | else: | |
934 | # looks good so far |
|
944 | # looks good so far | |
935 | dirmove[dsrc] = ddst |
|
945 | dirmove[dsrc] = ddst | |
936 |
|
946 | |||
937 | for i in invalid: |
|
947 | for i in invalid: | |
938 | if i in dirmove: |
|
948 | if i in dirmove: | |
939 | del dirmove[i] |
|
949 | del dirmove[i] | |
940 | del d, invalid |
|
950 | del d, invalid | |
941 |
|
951 | |||
942 | if not dirmove: |
|
952 | if not dirmove: | |
943 | return {}, {} |
|
953 | return {}, {} | |
944 |
|
954 | |||
945 | dirmove = {k + b"/": v + b"/" for k, v in pycompat.iteritems(dirmove)} |
|
955 | dirmove = {k + b"/": v + b"/" for k, v in pycompat.iteritems(dirmove)} | |
946 |
|
956 | |||
947 | for d in dirmove: |
|
957 | for d in dirmove: | |
948 | repo.ui.debug( |
|
958 | repo.ui.debug( | |
949 | b" discovered dir src: '%s' -> dst: '%s'\n" % (d, dirmove[d]) |
|
959 | b" discovered dir src: '%s' -> dst: '%s'\n" % (d, dirmove[d]) | |
950 | ) |
|
960 | ) | |
951 |
|
961 | |||
952 | movewithdir = {} |
|
962 | movewithdir = {} | |
953 | # check unaccounted nonoverlapping files against directory moves |
|
963 | # check unaccounted nonoverlapping files against directory moves | |
954 | for f in addedfiles: |
|
964 | for f in addedfiles: | |
955 | if f not in fullcopy: |
|
965 | if f not in fullcopy: | |
956 | for d in dirmove: |
|
966 | for d in dirmove: | |
957 | if f.startswith(d): |
|
967 | if f.startswith(d): | |
958 | # new file added in a directory that was moved, move it |
|
968 | # new file added in a directory that was moved, move it | |
959 | df = dirmove[d] + f[len(d) :] |
|
969 | df = dirmove[d] + f[len(d) :] | |
960 | if df not in copy: |
|
970 | if df not in copy: | |
961 | movewithdir[f] = df |
|
971 | movewithdir[f] = df | |
962 | repo.ui.debug( |
|
972 | repo.ui.debug( | |
963 | b" pending file src: '%s' -> dst: '%s'\n" |
|
973 | b" pending file src: '%s' -> dst: '%s'\n" | |
964 | % (f, df) |
|
974 | % (f, df) | |
965 | ) |
|
975 | ) | |
966 | break |
|
976 | break | |
967 |
|
977 | |||
968 | return dirmove, movewithdir |
|
978 | return dirmove, movewithdir | |
969 |
|
979 | |||
970 |
|
980 | |||
971 | def _heuristicscopytracing(repo, c1, c2, base): |
|
981 | def _heuristicscopytracing(repo, c1, c2, base): | |
972 | """Fast copytracing using filename heuristics |
|
982 | """Fast copytracing using filename heuristics | |
973 |
|
983 | |||
974 | Assumes that moves or renames are of following two types: |
|
984 | Assumes that moves or renames are of following two types: | |
975 |
|
985 | |||
976 | 1) Inside a directory only (same directory name but different filenames) |
|
986 | 1) Inside a directory only (same directory name but different filenames) | |
977 | 2) Move from one directory to another |
|
987 | 2) Move from one directory to another | |
978 | (same filenames but different directory names) |
|
988 | (same filenames but different directory names) | |
979 |
|
989 | |||
980 | Works only when there are no merge commits in the "source branch". |
|
990 | Works only when there are no merge commits in the "source branch". | |
981 | Source branch is commits from base up to c2 not including base. |
|
991 | Source branch is commits from base up to c2 not including base. | |
982 |
|
992 | |||
983 | If merge is involved it fallbacks to _fullcopytracing(). |
|
993 | If merge is involved it fallbacks to _fullcopytracing(). | |
984 |
|
994 | |||
985 | Can be used by setting the following config: |
|
995 | Can be used by setting the following config: | |
986 |
|
996 | |||
987 | [experimental] |
|
997 | [experimental] | |
988 | copytrace = heuristics |
|
998 | copytrace = heuristics | |
989 |
|
999 | |||
990 | In some cases the copy/move candidates found by heuristics can be very large |
|
1000 | In some cases the copy/move candidates found by heuristics can be very large | |
991 | in number and that will make the algorithm slow. The number of possible |
|
1001 | in number and that will make the algorithm slow. The number of possible | |
992 | candidates to check can be limited by using the config |
|
1002 | candidates to check can be limited by using the config | |
993 | `experimental.copytrace.movecandidateslimit` which defaults to 100. |
|
1003 | `experimental.copytrace.movecandidateslimit` which defaults to 100. | |
994 | """ |
|
1004 | """ | |
995 |
|
1005 | |||
996 | if c1.rev() is None: |
|
1006 | if c1.rev() is None: | |
997 | c1 = c1.p1() |
|
1007 | c1 = c1.p1() | |
998 | if c2.rev() is None: |
|
1008 | if c2.rev() is None: | |
999 | c2 = c2.p1() |
|
1009 | c2 = c2.p1() | |
1000 |
|
1010 | |||
1001 | changedfiles = set() |
|
1011 | changedfiles = set() | |
1002 | m1 = c1.manifest() |
|
1012 | m1 = c1.manifest() | |
1003 | if not repo.revs(b'%d::%d', base.rev(), c2.rev()): |
|
1013 | if not repo.revs(b'%d::%d', base.rev(), c2.rev()): | |
1004 | # If base is not in c2 branch, we switch to fullcopytracing |
|
1014 | # If base is not in c2 branch, we switch to fullcopytracing | |
1005 | repo.ui.debug( |
|
1015 | repo.ui.debug( | |
1006 | b"switching to full copytracing as base is not " |
|
1016 | b"switching to full copytracing as base is not " | |
1007 | b"an ancestor of c2\n" |
|
1017 | b"an ancestor of c2\n" | |
1008 | ) |
|
1018 | ) | |
1009 | return _fullcopytracing(repo, c1, c2, base) |
|
1019 | return _fullcopytracing(repo, c1, c2, base) | |
1010 |
|
1020 | |||
1011 | ctx = c2 |
|
1021 | ctx = c2 | |
1012 | while ctx != base: |
|
1022 | while ctx != base: | |
1013 | if len(ctx.parents()) == 2: |
|
1023 | if len(ctx.parents()) == 2: | |
1014 | # To keep things simple let's not handle merges |
|
1024 | # To keep things simple let's not handle merges | |
1015 | repo.ui.debug(b"switching to full copytracing because of merges\n") |
|
1025 | repo.ui.debug(b"switching to full copytracing because of merges\n") | |
1016 | return _fullcopytracing(repo, c1, c2, base) |
|
1026 | return _fullcopytracing(repo, c1, c2, base) | |
1017 | changedfiles.update(ctx.files()) |
|
1027 | changedfiles.update(ctx.files()) | |
1018 | ctx = ctx.p1() |
|
1028 | ctx = ctx.p1() | |
1019 |
|
1029 | |||
1020 | copies2 = {} |
|
1030 | copies2 = {} | |
1021 | cp = _forwardcopies(base, c2) |
|
1031 | cp = _forwardcopies(base, c2) | |
1022 | for dst, src in pycompat.iteritems(cp): |
|
1032 | for dst, src in pycompat.iteritems(cp): | |
1023 | if src in m1: |
|
1033 | if src in m1: | |
1024 | copies2[dst] = src |
|
1034 | copies2[dst] = src | |
1025 |
|
1035 | |||
1026 | # file is missing if it isn't present in the destination, but is present in |
|
1036 | # file is missing if it isn't present in the destination, but is present in | |
1027 | # the base and present in the source. |
|
1037 | # the base and present in the source. | |
1028 | # Presence in the base is important to exclude added files, presence in the |
|
1038 | # Presence in the base is important to exclude added files, presence in the | |
1029 | # source is important to exclude removed files. |
|
1039 | # source is important to exclude removed files. | |
1030 | filt = lambda f: f not in m1 and f in base and f in c2 |
|
1040 | filt = lambda f: f not in m1 and f in base and f in c2 | |
1031 | missingfiles = [f for f in changedfiles if filt(f)] |
|
1041 | missingfiles = [f for f in changedfiles if filt(f)] | |
1032 |
|
1042 | |||
1033 | copies1 = {} |
|
1043 | copies1 = {} | |
1034 | if missingfiles: |
|
1044 | if missingfiles: | |
1035 | basenametofilename = collections.defaultdict(list) |
|
1045 | basenametofilename = collections.defaultdict(list) | |
1036 | dirnametofilename = collections.defaultdict(list) |
|
1046 | dirnametofilename = collections.defaultdict(list) | |
1037 |
|
1047 | |||
1038 | for f in m1.filesnotin(base.manifest()): |
|
1048 | for f in m1.filesnotin(base.manifest()): | |
1039 | basename = os.path.basename(f) |
|
1049 | basename = os.path.basename(f) | |
1040 | dirname = os.path.dirname(f) |
|
1050 | dirname = os.path.dirname(f) | |
1041 | basenametofilename[basename].append(f) |
|
1051 | basenametofilename[basename].append(f) | |
1042 | dirnametofilename[dirname].append(f) |
|
1052 | dirnametofilename[dirname].append(f) | |
1043 |
|
1053 | |||
1044 | for f in missingfiles: |
|
1054 | for f in missingfiles: | |
1045 | basename = os.path.basename(f) |
|
1055 | basename = os.path.basename(f) | |
1046 | dirname = os.path.dirname(f) |
|
1056 | dirname = os.path.dirname(f) | |
1047 | samebasename = basenametofilename[basename] |
|
1057 | samebasename = basenametofilename[basename] | |
1048 | samedirname = dirnametofilename[dirname] |
|
1058 | samedirname = dirnametofilename[dirname] | |
1049 | movecandidates = samebasename + samedirname |
|
1059 | movecandidates = samebasename + samedirname | |
1050 | # f is guaranteed to be present in c2, that's why |
|
1060 | # f is guaranteed to be present in c2, that's why | |
1051 | # c2.filectx(f) won't fail |
|
1061 | # c2.filectx(f) won't fail | |
1052 | f2 = c2.filectx(f) |
|
1062 | f2 = c2.filectx(f) | |
1053 | # we can have a lot of candidates which can slow down the heuristics |
|
1063 | # we can have a lot of candidates which can slow down the heuristics | |
1054 | # config value to limit the number of candidates moves to check |
|
1064 | # config value to limit the number of candidates moves to check | |
1055 | maxcandidates = repo.ui.configint( |
|
1065 | maxcandidates = repo.ui.configint( | |
1056 | b'experimental', b'copytrace.movecandidateslimit' |
|
1066 | b'experimental', b'copytrace.movecandidateslimit' | |
1057 | ) |
|
1067 | ) | |
1058 |
|
1068 | |||
1059 | if len(movecandidates) > maxcandidates: |
|
1069 | if len(movecandidates) > maxcandidates: | |
1060 | repo.ui.status( |
|
1070 | repo.ui.status( | |
1061 | _( |
|
1071 | _( | |
1062 | b"skipping copytracing for '%s', more " |
|
1072 | b"skipping copytracing for '%s', more " | |
1063 | b"candidates than the limit: %d\n" |
|
1073 | b"candidates than the limit: %d\n" | |
1064 | ) |
|
1074 | ) | |
1065 | % (f, len(movecandidates)) |
|
1075 | % (f, len(movecandidates)) | |
1066 | ) |
|
1076 | ) | |
1067 | continue |
|
1077 | continue | |
1068 |
|
1078 | |||
1069 | for candidate in movecandidates: |
|
1079 | for candidate in movecandidates: | |
1070 | f1 = c1.filectx(candidate) |
|
1080 | f1 = c1.filectx(candidate) | |
1071 | if _related(f1, f2): |
|
1081 | if _related(f1, f2): | |
1072 | # if there are a few related copies then we'll merge |
|
1082 | # if there are a few related copies then we'll merge | |
1073 | # changes into all of them. This matches the behaviour |
|
1083 | # changes into all of them. This matches the behaviour | |
1074 | # of upstream copytracing |
|
1084 | # of upstream copytracing | |
1075 | copies1[candidate] = f |
|
1085 | copies1[candidate] = f | |
1076 |
|
1086 | |||
1077 | return branch_copies(copies1), branch_copies(copies2), {} |
|
1087 | return branch_copies(copies1), branch_copies(copies2), {} | |
1078 |
|
1088 | |||
1079 |
|
1089 | |||
1080 | def _related(f1, f2): |
|
1090 | def _related(f1, f2): | |
1081 | """return True if f1 and f2 filectx have a common ancestor |
|
1091 | """return True if f1 and f2 filectx have a common ancestor | |
1082 |
|
1092 | |||
1083 | Walk back to common ancestor to see if the two files originate |
|
1093 | Walk back to common ancestor to see if the two files originate | |
1084 | from the same file. Since workingfilectx's rev() is None it messes |
|
1094 | from the same file. Since workingfilectx's rev() is None it messes | |
1085 | up the integer comparison logic, hence the pre-step check for |
|
1095 | up the integer comparison logic, hence the pre-step check for | |
1086 | None (f1 and f2 can only be workingfilectx's initially). |
|
1096 | None (f1 and f2 can only be workingfilectx's initially). | |
1087 | """ |
|
1097 | """ | |
1088 |
|
1098 | |||
1089 | if f1 == f2: |
|
1099 | if f1 == f2: | |
1090 | return True # a match |
|
1100 | return True # a match | |
1091 |
|
1101 | |||
1092 | g1, g2 = f1.ancestors(), f2.ancestors() |
|
1102 | g1, g2 = f1.ancestors(), f2.ancestors() | |
1093 | try: |
|
1103 | try: | |
1094 | f1r, f2r = f1.linkrev(), f2.linkrev() |
|
1104 | f1r, f2r = f1.linkrev(), f2.linkrev() | |
1095 |
|
1105 | |||
1096 | if f1r is None: |
|
1106 | if f1r is None: | |
1097 | f1 = next(g1) |
|
1107 | f1 = next(g1) | |
1098 | if f2r is None: |
|
1108 | if f2r is None: | |
1099 | f2 = next(g2) |
|
1109 | f2 = next(g2) | |
1100 |
|
1110 | |||
1101 | while True: |
|
1111 | while True: | |
1102 | f1r, f2r = f1.linkrev(), f2.linkrev() |
|
1112 | f1r, f2r = f1.linkrev(), f2.linkrev() | |
1103 | if f1r > f2r: |
|
1113 | if f1r > f2r: | |
1104 | f1 = next(g1) |
|
1114 | f1 = next(g1) | |
1105 | elif f2r > f1r: |
|
1115 | elif f2r > f1r: | |
1106 | f2 = next(g2) |
|
1116 | f2 = next(g2) | |
1107 | else: # f1 and f2 point to files in the same linkrev |
|
1117 | else: # f1 and f2 point to files in the same linkrev | |
1108 | return f1 == f2 # true if they point to the same file |
|
1118 | return f1 == f2 # true if they point to the same file | |
1109 | except StopIteration: |
|
1119 | except StopIteration: | |
1110 | return False |
|
1120 | return False | |
1111 |
|
1121 | |||
1112 |
|
1122 | |||
1113 | def graftcopies(wctx, ctx, base): |
|
1123 | def graftcopies(wctx, ctx, base): | |
1114 | """reproduce copies between base and ctx in the wctx |
|
1124 | """reproduce copies between base and ctx in the wctx | |
1115 |
|
1125 | |||
1116 | Unlike mergecopies(), this function will only consider copies between base |
|
1126 | Unlike mergecopies(), this function will only consider copies between base | |
1117 | and ctx; it will ignore copies between base and wctx. Also unlike |
|
1127 | and ctx; it will ignore copies between base and wctx. Also unlike | |
1118 | mergecopies(), this function will apply copies to the working copy (instead |
|
1128 | mergecopies(), this function will apply copies to the working copy (instead | |
1119 | of just returning information about the copies). That makes it cheaper |
|
1129 | of just returning information about the copies). That makes it cheaper | |
1120 | (especially in the common case of base==ctx.p1()) and useful also when |
|
1130 | (especially in the common case of base==ctx.p1()) and useful also when | |
1121 | experimental.copytrace=off. |
|
1131 | experimental.copytrace=off. | |
1122 |
|
1132 | |||
1123 | merge.update() will have already marked most copies, but it will only |
|
1133 | merge.update() will have already marked most copies, but it will only | |
1124 | mark copies if it thinks the source files are related (see |
|
1134 | mark copies if it thinks the source files are related (see | |
1125 | merge._related()). It will also not mark copies if the file wasn't modified |
|
1135 | merge._related()). It will also not mark copies if the file wasn't modified | |
1126 | on the local side. This function adds the copies that were "missed" |
|
1136 | on the local side. This function adds the copies that were "missed" | |
1127 | by merge.update(). |
|
1137 | by merge.update(). | |
1128 | """ |
|
1138 | """ | |
1129 | new_copies = pathcopies(base, ctx) |
|
1139 | new_copies = pathcopies(base, ctx) | |
1130 | _filter(wctx.p1(), wctx, new_copies) |
|
1140 | _filter(wctx.p1(), wctx, new_copies) | |
1131 | for dst, src in pycompat.iteritems(new_copies): |
|
1141 | for dst, src in pycompat.iteritems(new_copies): | |
1132 | wctx[dst].markcopied(src) |
|
1142 | wctx[dst].markcopied(src) |
General Comments 0
You need to be logged in to leave comments.
Login now