##// END OF EJS Templates
manifest: move addlistdelta to module-level...
Augie Fackler -
r22409:8f09b785 default
parent child Browse files
Show More
@@ -1,233 +1,234
1 1 # manifest.py - manifest revision class for mercurial
2 2 #
3 3 # Copyright 2005-2007 Matt Mackall <mpm@selenic.com>
4 4 #
5 5 # This software may be used and distributed according to the terms of the
6 6 # GNU General Public License version 2 or any later version.
7 7
8 8 from i18n import _
9 9 import mdiff, parsers, error, revlog, util, dicthelpers
10 10 import array, struct
11 11
12 12 class manifestdict(dict):
13 13 def __init__(self, mapping=None, flags=None):
14 14 if mapping is None:
15 15 mapping = {}
16 16 if flags is None:
17 17 flags = {}
18 18 dict.__init__(self, mapping)
19 19 self._flags = flags
20 20 def flags(self, f):
21 21 return self._flags.get(f, "")
22 22 def withflags(self):
23 23 return set(self._flags.keys())
24 24 def set(self, f, flags):
25 25 self._flags[f] = flags
26 26 def copy(self):
27 27 return manifestdict(self, dict.copy(self._flags))
28 28 def intersectfiles(self, files):
29 29 '''make a new manifestdict with the intersection of self with files
30 30
31 31 The algorithm assumes that files is much smaller than self.'''
32 32 ret = manifestdict()
33 33 for fn in files:
34 34 if fn in self:
35 35 ret[fn] = self[fn]
36 36 flags = self._flags.get(fn, None)
37 37 if flags:
38 38 ret._flags[fn] = flags
39 39 return ret
40 40 def flagsdiff(self, d2):
41 41 return dicthelpers.diff(self._flags, d2._flags, "")
42 42
43 43
44 44 def checkforbidden(l):
45 45 """Check filenames for illegal characters."""
46 46 for f in l:
47 47 if '\n' in f or '\r' in f:
48 48 raise error.RevlogError(
49 49 _("'\\n' and '\\r' disallowed in filenames: %r") % f)
50 50
51 51
52 # apply the changes collected during the bisect loop to our addlist
53 # return a delta suitable for addrevision
54 def addlistdelta(addlist, x):
55 # for large addlist arrays, building a new array is cheaper
56 # than repeatedly modifying the existing one
57 currentposition = 0
58 newaddlist = array.array('c')
59
60 for start, end, content in x:
61 newaddlist += addlist[currentposition:start]
62 if content:
63 newaddlist += array.array('c', content)
64
65 currentposition = end
66
67 newaddlist += addlist[currentposition:]
68
69 deltatext = "".join(struct.pack(">lll", start, end, len(content))
70 + content for start, end, content in x)
71 return deltatext, newaddlist
72
73
52 74 class manifest(revlog.revlog):
53 75 def __init__(self, opener):
54 76 # we expect to deal with not more than four revs at a time,
55 77 # during a commit --amend
56 78 self._mancache = util.lrucachedict(4)
57 79 revlog.revlog.__init__(self, opener, "00manifest.i")
58 80
59 81 def parse(self, lines):
60 82 mfdict = manifestdict()
61 83 parsers.parse_manifest(mfdict, mfdict._flags, lines)
62 84 return mfdict
63 85
64 86 def readdelta(self, node):
65 87 r = self.rev(node)
66 88 return self.parse(mdiff.patchtext(self.revdiff(self.deltaparent(r), r)))
67 89
68 90 def readfast(self, node):
69 91 '''use the faster of readdelta or read'''
70 92 r = self.rev(node)
71 93 deltaparent = self.deltaparent(r)
72 94 if deltaparent != revlog.nullrev and deltaparent in self.parentrevs(r):
73 95 return self.readdelta(node)
74 96 return self.read(node)
75 97
76 98 def read(self, node):
77 99 if node == revlog.nullid:
78 100 return manifestdict() # don't upset local cache
79 101 if node in self._mancache:
80 102 return self._mancache[node][0]
81 103 text = self.revision(node)
82 104 arraytext = array.array('c', text)
83 105 mapping = self.parse(text)
84 106 self._mancache[node] = (mapping, arraytext)
85 107 return mapping
86 108
87 109 def _search(self, m, s, lo=0, hi=None):
88 110 '''return a tuple (start, end) that says where to find s within m.
89 111
90 112 If the string is found m[start:end] are the line containing
91 113 that string. If start == end the string was not found and
92 114 they indicate the proper sorted insertion point.
93 115
94 116 m should be a buffer or a string
95 117 s is a string'''
96 118 def advance(i, c):
97 119 while i < lenm and m[i] != c:
98 120 i += 1
99 121 return i
100 122 if not s:
101 123 return (lo, lo)
102 124 lenm = len(m)
103 125 if not hi:
104 126 hi = lenm
105 127 while lo < hi:
106 128 mid = (lo + hi) // 2
107 129 start = mid
108 130 while start > 0 and m[start - 1] != '\n':
109 131 start -= 1
110 132 end = advance(start, '\0')
111 133 if m[start:end] < s:
112 134 # we know that after the null there are 40 bytes of sha1
113 135 # this translates to the bisect lo = mid + 1
114 136 lo = advance(end + 40, '\n') + 1
115 137 else:
116 138 # this translates to the bisect hi = mid
117 139 hi = start
118 140 end = advance(lo, '\0')
119 141 found = m[lo:end]
120 142 if s == found:
121 143 # we know that after the null there are 40 bytes of sha1
122 144 end = advance(end + 40, '\n')
123 145 return (lo, end + 1)
124 146 else:
125 147 return (lo, lo)
126 148
127 149 def find(self, node, f):
128 150 '''look up entry for a single file efficiently.
129 151 return (node, flags) pair if found, (None, None) if not.'''
130 152 if node in self._mancache:
131 153 mapping = self._mancache[node][0]
132 154 return mapping.get(f), mapping.flags(f)
133 155 text = self.revision(node)
134 156 start, end = self._search(text, f)
135 157 if start == end:
136 158 return None, None
137 159 l = text[start:end]
138 160 f, n = l.split('\0')
139 161 return revlog.bin(n[:40]), n[40:-1]
140 162
141 163 def add(self, map, transaction, link, p1=None, p2=None,
142 164 changed=None):
143 # apply the changes collected during the bisect loop to our addlist
144 # return a delta suitable for addrevision
145 def addlistdelta(addlist, x):
146 # for large addlist arrays, building a new array is cheaper
147 # than repeatedly modifying the existing one
148 currentposition = 0
149 newaddlist = array.array('c')
150
151 for start, end, content in x:
152 newaddlist += addlist[currentposition:start]
153 if content:
154 newaddlist += array.array('c', content)
155
156 currentposition = end
157
158 newaddlist += addlist[currentposition:]
159
160 deltatext = "".join(struct.pack(">lll", start, end, len(content))
161 + content for start, end, content in x)
162 return deltatext, newaddlist
163
164 165 # if we're using the cache, make sure it is valid and
165 166 # parented by the same node we're diffing against
166 167 if not (changed and p1 and (p1 in self._mancache)):
167 168 files = sorted(map)
168 169 checkforbidden(files)
169 170
170 171 # if this is changed to support newlines in filenames,
171 172 # be sure to check the templates/ dir again (especially *-raw.tmpl)
172 173 hex, flags = revlog.hex, map.flags
173 174 text = ''.join("%s\0%s%s\n" % (f, hex(map[f]), flags(f))
174 175 for f in files)
175 176 arraytext = array.array('c', text)
176 177 cachedelta = None
177 178 else:
178 179 added, removed = changed
179 180 addlist = self._mancache[p1][1]
180 181
181 182 checkforbidden(added)
182 183 # combine the changed lists into one list for sorting
183 184 work = [(x, False) for x in added]
184 185 work.extend((x, True) for x in removed)
185 186 # this could use heapq.merge() (from Python 2.6+) or equivalent
186 187 # since the lists are already sorted
187 188 work.sort()
188 189
189 190 delta = []
190 191 dstart = None
191 192 dend = None
192 193 dline = [""]
193 194 start = 0
194 195 # zero copy representation of addlist as a buffer
195 196 addbuf = util.buffer(addlist)
196 197
197 198 # start with a readonly loop that finds the offset of
198 199 # each line and creates the deltas
199 200 for f, todelete in work:
200 201 # bs will either be the index of the item or the insert point
201 202 start, end = self._search(addbuf, f, start)
202 203 if not todelete:
203 204 l = "%s\0%s%s\n" % (f, revlog.hex(map[f]), map.flags(f))
204 205 else:
205 206 if start == end:
206 207 # item we want to delete was not found, error out
207 208 raise AssertionError(
208 209 _("failed to remove %s from manifest") % f)
209 210 l = ""
210 211 if dstart is not None and dstart <= start and dend >= start:
211 212 if dend < end:
212 213 dend = end
213 214 if l:
214 215 dline.append(l)
215 216 else:
216 217 if dstart is not None:
217 218 delta.append([dstart, dend, "".join(dline)])
218 219 dstart = start
219 220 dend = end
220 221 dline = [l]
221 222
222 223 if dstart is not None:
223 224 delta.append([dstart, dend, "".join(dline)])
224 225 # apply the delta to the addlist, and get a delta for addrevision
225 226 deltatext, addlist = addlistdelta(addlist, delta)
226 227 cachedelta = (self.rev(p1), deltatext)
227 228 arraytext = addlist
228 229 text = util.buffer(arraytext)
229 230
230 231 n = self.addrevision(text, transaction, link, p1, p2, cachedelta)
231 232 self._mancache[n] = (map, arraytext)
232 233
233 234 return n
General Comments 0
You need to be logged in to leave comments. Login now