Show More
| @@ -1,277 +1,281 b'' | |||
|
|
1 | 1 | # mdiff.py - diff and patch routines for mercurial |
|
|
2 | 2 | # |
|
|
3 | 3 | # Copyright 2005, 2006 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 bdiff, mpatch, util |
|
|
10 | 10 | import re, struct |
|
|
11 | 11 | |
|
|
12 | 12 | def splitnewlines(text): |
|
|
13 | 13 | '''like str.splitlines, but only split on newlines.''' |
|
|
14 | 14 | lines = [l + '\n' for l in text.split('\n')] |
|
|
15 | 15 | if lines: |
|
|
16 | 16 | if lines[-1] == '\n': |
|
|
17 | 17 | lines.pop() |
|
|
18 | 18 | else: |
|
|
19 | 19 | lines[-1] = lines[-1][:-1] |
|
|
20 | 20 | return lines |
|
|
21 | 21 | |
|
|
22 | 22 | class diffopts(object): |
|
|
23 | 23 | '''context is the number of context lines |
|
|
24 | 24 | text treats all files as text |
|
|
25 | 25 | showfunc enables diff -p output |
|
|
26 | 26 | git enables the git extended patch format |
|
|
27 | 27 | nodates removes dates from diff headers |
|
|
28 | 28 | ignorews ignores all whitespace changes in the diff |
|
|
29 | 29 | ignorewsamount ignores changes in the amount of whitespace |
|
|
30 | 30 | ignoreblanklines ignores changes whose lines are all blank |
|
|
31 | 31 | upgrade generates git diffs to avoid data loss |
|
|
32 | 32 | ''' |
|
|
33 | 33 | |
|
|
34 | 34 | defaults = { |
|
|
35 | 35 | 'context': 3, |
|
|
36 | 36 | 'text': False, |
|
|
37 | 37 | 'showfunc': False, |
|
|
38 | 38 | 'git': False, |
|
|
39 | 39 | 'nodates': False, |
|
|
40 | 40 | 'ignorews': False, |
|
|
41 | 41 | 'ignorewsamount': False, |
|
|
42 | 42 | 'ignoreblanklines': False, |
|
|
43 | 43 | 'upgrade': False, |
|
|
44 | 44 | } |
|
|
45 | 45 | |
|
|
46 | 46 | __slots__ = defaults.keys() |
|
|
47 | 47 | |
|
|
48 | 48 | def __init__(self, **opts): |
|
|
49 | 49 | for k in self.__slots__: |
|
|
50 | 50 | v = opts.get(k) |
|
|
51 | 51 | if v is None: |
|
|
52 | 52 | v = self.defaults[k] |
|
|
53 | 53 | setattr(self, k, v) |
|
|
54 | 54 | |
|
|
55 | 55 | try: |
|
|
56 | 56 | self.context = int(self.context) |
|
|
57 | 57 | except ValueError: |
|
|
58 | 58 | raise util.Abort(_('diff context lines count must be ' |
|
|
59 | 59 | 'an integer, not %r') % self.context) |
|
|
60 | 60 | |
|
|
61 | 61 | def copy(self, **kwargs): |
|
|
62 | 62 | opts = dict((k, getattr(self, k)) for k in self.defaults) |
|
|
63 | 63 | opts.update(kwargs) |
|
|
64 | 64 | return diffopts(**opts) |
|
|
65 | 65 | |
|
|
66 | 66 | defaultopts = diffopts() |
|
|
67 | 67 | |
|
|
68 | 68 | def wsclean(opts, text, blank=True): |
|
|
69 | 69 | if opts.ignorews: |
|
|
70 | 70 | text = re.sub('[ \t\r]+', '', text) |
|
|
71 | 71 | elif opts.ignorewsamount: |
|
|
72 | 72 | text = re.sub('[ \t\r]+', ' ', text) |
|
|
73 | 73 | text = text.replace(' \n', '\n') |
|
|
74 | 74 | if blank and opts.ignoreblanklines: |
|
|
75 | 75 | text = re.sub('\n+', '', text) |
|
|
76 | 76 | return text |
|
|
77 | 77 | |
|
|
78 | 78 | def diffline(revs, a, b, opts): |
|
|
79 | 79 | parts = ['diff'] |
|
|
80 | 80 | if opts.git: |
|
|
81 | 81 | parts.append('--git') |
|
|
82 | 82 | if revs and not opts.git: |
|
|
83 | 83 | parts.append(' '.join(["-r %s" % rev for rev in revs])) |
|
|
84 | 84 | if opts.git: |
|
|
85 | 85 | parts.append('a/%s' % a) |
|
|
86 | 86 | parts.append('b/%s' % b) |
|
|
87 | 87 | else: |
|
|
88 | 88 | parts.append(a) |
|
|
89 | 89 | return ' '.join(parts) + '\n' |
|
|
90 | 90 | |
|
|
91 | 91 | def unidiff(a, ad, b, bd, fn1, fn2, r=None, opts=defaultopts): |
|
|
92 | 92 | def datetag(date, addtab=True): |
|
|
93 | 93 | if not opts.git and not opts.nodates: |
|
|
94 | 94 | return '\t%s\n' % date |
|
|
95 | 95 | if addtab and ' ' in fn1: |
|
|
96 | 96 | return '\t\n' |
|
|
97 | 97 | return '\n' |
|
|
98 | 98 | |
|
|
99 | 99 | if not a and not b: |
|
|
100 | 100 | return "" |
|
|
101 | 101 | epoch = util.datestr((0, 0)) |
|
|
102 | 102 | |
|
|
103 | 103 | if not opts.text and (util.binary(a) or util.binary(b)): |
|
|
104 | 104 | if a and b and len(a) == len(b) and a == b: |
|
|
105 | 105 | return "" |
|
|
106 | 106 | l = ['Binary file %s has changed\n' % fn1] |
|
|
107 | 107 | elif not a: |
|
|
108 | 108 | b = splitnewlines(b) |
|
|
109 | 109 | if a is None: |
|
|
110 | 110 | l1 = '--- /dev/null%s' % datetag(epoch, False) |
|
|
111 | 111 | else: |
|
|
112 | 112 | l1 = "--- %s%s" % ("a/" + fn1, datetag(ad)) |
|
|
113 | 113 | l2 = "+++ %s%s" % ("b/" + fn2, datetag(bd)) |
|
|
114 | 114 | l3 = "@@ -0,0 +1,%d @@\n" % len(b) |
|
|
115 | 115 | l = [l1, l2, l3] + ["+" + e for e in b] |
|
|
116 | 116 | elif not b: |
|
|
117 | 117 | a = splitnewlines(a) |
|
|
118 | 118 | l1 = "--- %s%s" % ("a/" + fn1, datetag(ad)) |
|
|
119 | 119 | if b is None: |
|
|
120 | 120 | l2 = '+++ /dev/null%s' % datetag(epoch, False) |
|
|
121 | 121 | else: |
|
|
122 | 122 | l2 = "+++ %s%s" % ("b/" + fn2, datetag(bd)) |
|
|
123 | 123 | l3 = "@@ -1,%d +0,0 @@\n" % len(a) |
|
|
124 | 124 | l = [l1, l2, l3] + ["-" + e for e in a] |
|
|
125 | 125 | else: |
|
|
126 | 126 | al = splitnewlines(a) |
|
|
127 | 127 | bl = splitnewlines(b) |
|
|
128 | 128 | l = list(_unidiff(a, b, al, bl, opts=opts)) |
|
|
129 | 129 | if not l: |
|
|
130 | 130 | return "" |
|
|
131 | 131 | |
|
|
132 | 132 | l.insert(0, "--- a/%s%s" % (fn1, datetag(ad))) |
|
|
133 | 133 | l.insert(1, "+++ b/%s%s" % (fn2, datetag(bd))) |
|
|
134 | 134 | |
|
|
135 | 135 | for ln in xrange(len(l)): |
|
|
136 | 136 | if l[ln][-1] != '\n': |
|
|
137 | 137 | l[ln] += "\n\ No newline at end of file\n" |
|
|
138 | 138 | |
|
|
139 | 139 | if r: |
|
|
140 | 140 | l.insert(0, diffline(r, fn1, fn2, opts)) |
|
|
141 | 141 | |
|
|
142 | 142 | return "".join(l) |
|
|
143 | 143 | |
|
|
144 | 144 | # creates a headerless unified diff |
|
|
145 | 145 | # t1 and t2 are the text to be diffed |
|
|
146 | 146 | # l1 and l2 are the text broken up into lines |
|
|
147 | 147 | def _unidiff(t1, t2, l1, l2, opts=defaultopts): |
|
|
148 | 148 | def contextend(l, len): |
|
|
149 | 149 | ret = l + opts.context |
|
|
150 | 150 | if ret > len: |
|
|
151 | 151 | ret = len |
|
|
152 | 152 | return ret |
|
|
153 | 153 | |
|
|
154 | 154 | def contextstart(l): |
|
|
155 | 155 | ret = l - opts.context |
|
|
156 | 156 | if ret < 0: |
|
|
157 | 157 | return 0 |
|
|
158 | 158 | return ret |
|
|
159 | 159 | |
|
|
160 | lastfunc = [0, ''] | |
|
|
160 | 161 | def yieldhunk(hunk): |
|
|
161 | 162 | (astart, a2, bstart, b2, delta) = hunk |
|
|
162 | 163 | aend = contextend(a2, len(l1)) |
|
|
163 | 164 | alen = aend - astart |
|
|
164 | 165 | blen = b2 - bstart + aend - a2 |
|
|
165 | 166 | |
|
|
166 | 167 | func = "" |
|
|
167 | 168 | if opts.showfunc: |
|
|
168 | # walk backwards from the start of the context | |
|
|
169 | # to find a line starting with an alphanumeric char. | |
|
|
170 | for x in xrange(astart - 1, -1, -1): | |
|
|
171 | t = l1[x].rstrip() | |
|
|
172 | if funcre.match(t): | |
|
|
173 | func = ' ' + t[:40] | |
|
|
169 | lastpos, func = lastfunc | |
|
|
170 | # walk backwards from the start of the context up to the start of | |
|
|
171 | # the previous hunk context until we find a line starting with an | |
|
|
172 | # alphanumeric char. | |
|
|
173 | for i in xrange(astart - 1, lastpos - 1, -1): | |
|
|
174 | if l1[i][0].isalnum(): | |
|
|
175 | func = ' ' + l1[i].rstrip()[:40] | |
|
|
176 | lastfunc[1] = func | |
|
|
174 | 177 | break |
|
|
178 | # by recording this hunk's starting point as the next place to | |
|
|
179 | # start looking for function lines, we avoid reading any line in | |
|
|
180 | # the file more than once. | |
|
|
181 | lastfunc[0] = astart | |
|
|
175 | 182 | |
|
|
176 | 183 | yield "@@ -%d,%d +%d,%d @@%s\n" % (astart + 1, alen, |
|
|
177 | 184 | bstart + 1, blen, func) |
|
|
178 | 185 | for x in delta: |
|
|
179 | 186 | yield x |
|
|
180 | 187 | for x in xrange(a2, aend): |
|
|
181 | 188 | yield ' ' + l1[x] |
|
|
182 | 189 | |
|
|
183 | if opts.showfunc: | |
|
|
184 | funcre = re.compile('\w') | |
|
|
185 | ||
|
|
186 | 190 | # bdiff.blocks gives us the matching sequences in the files. The loop |
|
|
187 | 191 | # below finds the spaces between those matching sequences and translates |
|
|
188 | 192 | # them into diff output. |
|
|
189 | 193 | # |
|
|
190 | 194 | if opts.ignorews or opts.ignorewsamount: |
|
|
191 | 195 | t1 = wsclean(opts, t1, False) |
|
|
192 | 196 | t2 = wsclean(opts, t2, False) |
|
|
193 | 197 | |
|
|
194 | 198 | diff = bdiff.blocks(t1, t2) |
|
|
195 | 199 | hunk = None |
|
|
196 | 200 | for i, s1 in enumerate(diff): |
|
|
197 | 201 | # The first match is special. |
|
|
198 | 202 | # we've either found a match starting at line 0 or a match later |
|
|
199 | 203 | # in the file. If it starts later, old and new below will both be |
|
|
200 | 204 | # empty and we'll continue to the next match. |
|
|
201 | 205 | if i > 0: |
|
|
202 | 206 | s = diff[i - 1] |
|
|
203 | 207 | else: |
|
|
204 | 208 | s = [0, 0, 0, 0] |
|
|
205 | 209 | delta = [] |
|
|
206 | 210 | a1 = s[1] |
|
|
207 | 211 | a2 = s1[0] |
|
|
208 | 212 | b1 = s[3] |
|
|
209 | 213 | b2 = s1[2] |
|
|
210 | 214 | |
|
|
211 | 215 | old = l1[a1:a2] |
|
|
212 | 216 | new = l2[b1:b2] |
|
|
213 | 217 | |
|
|
214 | 218 | # bdiff sometimes gives huge matches past eof, this check eats them, |
|
|
215 | 219 | # and deals with the special first match case described above |
|
|
216 | 220 | if not old and not new: |
|
|
217 | 221 | continue |
|
|
218 | 222 | |
|
|
219 | 223 | if opts.ignoreblanklines: |
|
|
220 | 224 | if wsclean(opts, "".join(old)) == wsclean(opts, "".join(new)): |
|
|
221 | 225 | continue |
|
|
222 | 226 | |
|
|
223 | 227 | astart = contextstart(a1) |
|
|
224 | 228 | bstart = contextstart(b1) |
|
|
225 | 229 | prev = None |
|
|
226 | 230 | if hunk: |
|
|
227 | 231 | # join with the previous hunk if it falls inside the context |
|
|
228 | 232 | if astart < hunk[1] + opts.context + 1: |
|
|
229 | 233 | prev = hunk |
|
|
230 | 234 | astart = hunk[1] |
|
|
231 | 235 | bstart = hunk[3] |
|
|
232 | 236 | else: |
|
|
233 | 237 | for x in yieldhunk(hunk): |
|
|
234 | 238 | yield x |
|
|
235 | 239 | if prev: |
|
|
236 | 240 | # we've joined the previous hunk, record the new ending points. |
|
|
237 | 241 | hunk[1] = a2 |
|
|
238 | 242 | hunk[3] = b2 |
|
|
239 | 243 | delta = hunk[4] |
|
|
240 | 244 | else: |
|
|
241 | 245 | # create a new hunk |
|
|
242 | 246 | hunk = [astart, a2, bstart, b2, delta] |
|
|
243 | 247 | |
|
|
244 | 248 | delta[len(delta):] = [' ' + x for x in l1[astart:a1]] |
|
|
245 | 249 | delta[len(delta):] = ['-' + x for x in old] |
|
|
246 | 250 | delta[len(delta):] = ['+' + x for x in new] |
|
|
247 | 251 | |
|
|
248 | 252 | if hunk: |
|
|
249 | 253 | for x in yieldhunk(hunk): |
|
|
250 | 254 | yield x |
|
|
251 | 255 | |
|
|
252 | 256 | def patchtext(bin): |
|
|
253 | 257 | pos = 0 |
|
|
254 | 258 | t = [] |
|
|
255 | 259 | while pos < len(bin): |
|
|
256 | 260 | p1, p2, l = struct.unpack(">lll", bin[pos:pos + 12]) |
|
|
257 | 261 | pos += 12 |
|
|
258 | 262 | t.append(bin[pos:pos + l]) |
|
|
259 | 263 | pos += l |
|
|
260 | 264 | return "".join(t) |
|
|
261 | 265 | |
|
|
262 | 266 | def patch(a, bin): |
|
|
263 | 267 | if len(a) == 0: |
|
|
264 | 268 | # skip over trivial delta header |
|
|
265 | 269 | return buffer(bin, 12) |
|
|
266 | 270 | return mpatch.patches(a, [bin]) |
|
|
267 | 271 | |
|
|
268 | 272 | # similar to difflib.SequenceMatcher.get_matching_blocks |
|
|
269 | 273 | def get_matching_blocks(a, b): |
|
|
270 | 274 | return [(d[0], d[2], d[1] - d[0]) for d in bdiff.blocks(a, b)] |
|
|
271 | 275 | |
|
|
272 | 276 | def trivialdiffheader(length): |
|
|
273 | 277 | return struct.pack(">lll", 0, 0, length) |
|
|
274 | 278 | |
|
|
275 | 279 | patches = mpatch.patches |
|
|
276 | 280 | patchedsize = mpatch.patchedsize |
|
|
277 | 281 | textdiff = bdiff.bdiff |
General Comments 0
You need to be logged in to leave comments.
Login now