##// END OF EJS Templates
diffutil: extract diff options code into a dedicated util-module...
Boris Feld -
r38581:c88d2c9b default
parent child Browse files
Show More
@@ -41,6 +41,7 b' from . import ('
41 )
41 )
42 from .utils import (
42 from .utils import (
43 dateutil,
43 dateutil,
44 diffutil,
44 procutil,
45 procutil,
45 stringutil,
46 stringutil,
46 )
47 )
@@ -2232,95 +2233,9 b' def changedfiles(ui, repo, patchpath, st'
2232 class GitDiffRequired(Exception):
2233 class GitDiffRequired(Exception):
2233 pass
2234 pass
2234
2235
2235 def diffallopts(ui, opts=None, untrusted=False, section='diff'):
2236 diffopts = diffutil.diffopts
2236 '''return diffopts with all features supported and parsed'''
2237 diffallopts = diffutil.diffallopts
2237 return difffeatureopts(ui, opts=opts, untrusted=untrusted, section=section,
2238 difffeatureopts = diffutil.difffeatureopts
2238 git=True, whitespace=True, formatchanging=True)
2239
2240 diffopts = diffallopts
2241
2242 def difffeatureopts(ui, opts=None, untrusted=False, section='diff', git=False,
2243 whitespace=False, formatchanging=False):
2244 '''return diffopts with only opted-in features parsed
2245
2246 Features:
2247 - git: git-style diffs
2248 - whitespace: whitespace options like ignoreblanklines and ignorews
2249 - formatchanging: options that will likely break or cause correctness issues
2250 with most diff parsers
2251 '''
2252 def get(key, name=None, getter=ui.configbool, forceplain=None):
2253 if opts:
2254 v = opts.get(key)
2255 # diffopts flags are either None-default (which is passed
2256 # through unchanged, so we can identify unset values), or
2257 # some other falsey default (eg --unified, which defaults
2258 # to an empty string). We only want to override the config
2259 # entries from hgrc with command line values if they
2260 # appear to have been set, which is any truthy value,
2261 # True, or False.
2262 if v or isinstance(v, bool):
2263 return v
2264 if forceplain is not None and ui.plain():
2265 return forceplain
2266 return getter(section, name or key, untrusted=untrusted)
2267
2268 # core options, expected to be understood by every diff parser
2269 buildopts = {
2270 'nodates': get('nodates'),
2271 'showfunc': get('show_function', 'showfunc'),
2272 'context': get('unified', getter=ui.config),
2273 }
2274 buildopts['worddiff'] = ui.configbool('experimental', 'worddiff')
2275 buildopts['xdiff'] = ui.configbool('experimental', 'xdiff')
2276
2277 if git:
2278 buildopts['git'] = get('git')
2279
2280 # since this is in the experimental section, we need to call
2281 # ui.configbool directory
2282 buildopts['showsimilarity'] = ui.configbool('experimental',
2283 'extendedheader.similarity')
2284
2285 # need to inspect the ui object instead of using get() since we want to
2286 # test for an int
2287 hconf = ui.config('experimental', 'extendedheader.index')
2288 if hconf is not None:
2289 hlen = None
2290 try:
2291 # the hash config could be an integer (for length of hash) or a
2292 # word (e.g. short, full, none)
2293 hlen = int(hconf)
2294 if hlen < 0 or hlen > 40:
2295 msg = _("invalid length for extendedheader.index: '%d'\n")
2296 ui.warn(msg % hlen)
2297 except ValueError:
2298 # default value
2299 if hconf == 'short' or hconf == '':
2300 hlen = 12
2301 elif hconf == 'full':
2302 hlen = 40
2303 elif hconf != 'none':
2304 msg = _("invalid value for extendedheader.index: '%s'\n")
2305 ui.warn(msg % hconf)
2306 finally:
2307 buildopts['index'] = hlen
2308
2309 if whitespace:
2310 buildopts['ignorews'] = get('ignore_all_space', 'ignorews')
2311 buildopts['ignorewsamount'] = get('ignore_space_change',
2312 'ignorewsamount')
2313 buildopts['ignoreblanklines'] = get('ignore_blank_lines',
2314 'ignoreblanklines')
2315 buildopts['ignorewseol'] = get('ignore_space_at_eol', 'ignorewseol')
2316 if formatchanging:
2317 buildopts['text'] = opts and opts.get('text')
2318 binary = None if opts is None else opts.get('binary')
2319 buildopts['nobinary'] = (not binary if binary is not None
2320 else get('nobinary', forceplain=False))
2321 buildopts['noprefix'] = get('noprefix', forceplain=False)
2322
2323 return mdiff.diffopts(**pycompat.strkwargs(buildopts))
2324
2239
2325 def diff(repo, node1=None, node2=None, match=None, changes=None,
2240 def diff(repo, node1=None, node2=None, match=None, changes=None,
2326 opts=None, losedatafn=None, prefix='', relroot='', copy=None,
2241 opts=None, losedatafn=None, prefix='', relroot='', copy=None,
This diff has been collapsed as it changes many lines, (2860 lines changed) Show them Hide them
@@ -1,2236 +1,20 b''
1 # patch.py - patch file parsing routines
1 # diffutil.py - utility functions related to diff and patch
2 #
2 #
3 # Copyright 2006 Brendan Cully <brendan@kublai.com>
3 # Copyright 2006 Brendan Cully <brendan@kublai.com>
4 # Copyright 2007 Chris Mason <chris.mason@oracle.com>
4 # Copyright 2007 Chris Mason <chris.mason@oracle.com>
5 # Copyright 2018 Octobus <octobus@octobus.net>
5 #
6 #
6 # This software may be used and distributed according to the terms of the
7 # This software may be used and distributed according to the terms of the
7 # GNU General Public License version 2 or any later version.
8 # GNU General Public License version 2 or any later version.
8
9
9 from __future__ import absolute_import, print_function
10 from __future__ import absolute_import
10
11 import collections
12 import contextlib
13 import copy
14 import email
15 import errno
16 import hashlib
17 import os
18 import posixpath
19 import re
20 import shutil
21 import zlib
22
23 from .i18n import _
24 from .node import (
25 hex,
26 short,
27 )
28 from . import (
29 copies,
30 diffhelper,
31 encoding,
32 error,
33 mail,
34 mdiff,
35 pathutil,
36 pycompat,
37 scmutil,
38 similar,
39 util,
40 vfs as vfsmod,
41 )
42 from .utils import (
43 dateutil,
44 procutil,
45 stringutil,
46 )
47
48 stringio = util.stringio
49
50 gitre = re.compile(br'diff --git a/(.*) b/(.*)')
51 tabsplitter = re.compile(br'(\t+|[^\t]+)')
52 wordsplitter = re.compile(br'(\t+| +|[a-zA-Z0-9_\x80-\xff]+|'
53 b'[^ \ta-zA-Z0-9_\x80-\xff])')
54
55 PatchError = error.PatchError
56
57 # public functions
58
59 def split(stream):
60 '''return an iterator of individual patches from a stream'''
61 def isheader(line, inheader):
62 if inheader and line.startswith((' ', '\t')):
63 # continuation
64 return True
65 if line.startswith((' ', '-', '+')):
66 # diff line - don't check for header pattern in there
67 return False
68 l = line.split(': ', 1)
69 return len(l) == 2 and ' ' not in l[0]
70
71 def chunk(lines):
72 return stringio(''.join(lines))
73
74 def hgsplit(stream, cur):
75 inheader = True
76
77 for line in stream:
78 if not line.strip():
79 inheader = False
80 if not inheader and line.startswith('# HG changeset patch'):
81 yield chunk(cur)
82 cur = []
83 inheader = True
84
85 cur.append(line)
86
87 if cur:
88 yield chunk(cur)
89
90 def mboxsplit(stream, cur):
91 for line in stream:
92 if line.startswith('From '):
93 for c in split(chunk(cur[1:])):
94 yield c
95 cur = []
96
97 cur.append(line)
98
99 if cur:
100 for c in split(chunk(cur[1:])):
101 yield c
102
103 def mimesplit(stream, cur):
104 def msgfp(m):
105 fp = stringio()
106 g = email.Generator.Generator(fp, mangle_from_=False)
107 g.flatten(m)
108 fp.seek(0)
109 return fp
110
111 for line in stream:
112 cur.append(line)
113 c = chunk(cur)
114
115 m = mail.parse(c)
116 if not m.is_multipart():
117 yield msgfp(m)
118 else:
119 ok_types = ('text/plain', 'text/x-diff', 'text/x-patch')
120 for part in m.walk():
121 ct = part.get_content_type()
122 if ct not in ok_types:
123 continue
124 yield msgfp(part)
125
126 def headersplit(stream, cur):
127 inheader = False
128
129 for line in stream:
130 if not inheader and isheader(line, inheader):
131 yield chunk(cur)
132 cur = []
133 inheader = True
134 if inheader and not isheader(line, inheader):
135 inheader = False
136
137 cur.append(line)
138
139 if cur:
140 yield chunk(cur)
141
142 def remainder(cur):
143 yield chunk(cur)
144
145 class fiter(object):
146 def __init__(self, fp):
147 self.fp = fp
148
149 def __iter__(self):
150 return self
151
152 def next(self):
153 l = self.fp.readline()
154 if not l:
155 raise StopIteration
156 return l
157
158 __next__ = next
159
160 inheader = False
161 cur = []
162
163 mimeheaders = ['content-type']
164
165 if not util.safehasattr(stream, 'next'):
166 # http responses, for example, have readline but not next
167 stream = fiter(stream)
168
169 for line in stream:
170 cur.append(line)
171 if line.startswith('# HG changeset patch'):
172 return hgsplit(stream, cur)
173 elif line.startswith('From '):
174 return mboxsplit(stream, cur)
175 elif isheader(line, inheader):
176 inheader = True
177 if line.split(':', 1)[0].lower() in mimeheaders:
178 # let email parser handle this
179 return mimesplit(stream, cur)
180 elif line.startswith('--- ') and inheader:
181 # No evil headers seen by diff start, split by hand
182 return headersplit(stream, cur)
183 # Not enough info, keep reading
184
185 # if we are here, we have a very plain patch
186 return remainder(cur)
187
188 ## Some facility for extensible patch parsing:
189 # list of pairs ("header to match", "data key")
190 patchheadermap = [('Date', 'date'),
191 ('Branch', 'branch'),
192 ('Node ID', 'nodeid'),
193 ]
194
195 @contextlib.contextmanager
196 def extract(ui, fileobj):
197 '''extract patch from data read from fileobj.
198
199 patch can be a normal patch or contained in an email message.
200
201 return a dictionary. Standard keys are:
202 - filename,
203 - message,
204 - user,
205 - date,
206 - branch,
207 - node,
208 - p1,
209 - p2.
210 Any item can be missing from the dictionary. If filename is missing,
211 fileobj did not contain a patch. Caller must unlink filename when done.'''
212
213 fd, tmpname = pycompat.mkstemp(prefix='hg-patch-')
214 tmpfp = os.fdopen(fd, r'wb')
215 try:
216 yield _extract(ui, fileobj, tmpname, tmpfp)
217 finally:
218 tmpfp.close()
219 os.unlink(tmpname)
220
221 def _extract(ui, fileobj, tmpname, tmpfp):
222
223 # attempt to detect the start of a patch
224 # (this heuristic is borrowed from quilt)
225 diffre = re.compile(br'^(?:Index:[ \t]|diff[ \t]-|RCS file: |'
226 br'retrieving revision [0-9]+(\.[0-9]+)*$|'
227 br'---[ \t].*?^\+\+\+[ \t]|'
228 br'\*\*\*[ \t].*?^---[ \t])',
229 re.MULTILINE | re.DOTALL)
230
231 data = {}
232
233 msg = mail.parse(fileobj)
234
235 subject = msg[r'Subject'] and mail.headdecode(msg[r'Subject'])
236 data['user'] = msg[r'From'] and mail.headdecode(msg[r'From'])
237 if not subject and not data['user']:
238 # Not an email, restore parsed headers if any
239 subject = '\n'.join(': '.join(map(encoding.strtolocal, h))
240 for h in msg.items()) + '\n'
241
242 # should try to parse msg['Date']
243 parents = []
244
245 if subject:
246 if subject.startswith('[PATCH'):
247 pend = subject.find(']')
248 if pend >= 0:
249 subject = subject[pend + 1:].lstrip()
250 subject = re.sub(br'\n[ \t]+', ' ', subject)
251 ui.debug('Subject: %s\n' % subject)
252 if data['user']:
253 ui.debug('From: %s\n' % data['user'])
254 diffs_seen = 0
255 ok_types = ('text/plain', 'text/x-diff', 'text/x-patch')
256 message = ''
257 for part in msg.walk():
258 content_type = pycompat.bytestr(part.get_content_type())
259 ui.debug('Content-Type: %s\n' % content_type)
260 if content_type not in ok_types:
261 continue
262 payload = part.get_payload(decode=True)
263 m = diffre.search(payload)
264 if m:
265 hgpatch = False
266 hgpatchheader = False
267 ignoretext = False
268
269 ui.debug('found patch at byte %d\n' % m.start(0))
270 diffs_seen += 1
271 cfp = stringio()
272 for line in payload[:m.start(0)].splitlines():
273 if line.startswith('# HG changeset patch') and not hgpatch:
274 ui.debug('patch generated by hg export\n')
275 hgpatch = True
276 hgpatchheader = True
277 # drop earlier commit message content
278 cfp.seek(0)
279 cfp.truncate()
280 subject = None
281 elif hgpatchheader:
282 if line.startswith('# User '):
283 data['user'] = line[7:]
284 ui.debug('From: %s\n' % data['user'])
285 elif line.startswith("# Parent "):
286 parents.append(line[9:].lstrip())
287 elif line.startswith("# "):
288 for header, key in patchheadermap:
289 prefix = '# %s ' % header
290 if line.startswith(prefix):
291 data[key] = line[len(prefix):]
292 else:
293 hgpatchheader = False
294 elif line == '---':
295 ignoretext = True
296 if not hgpatchheader and not ignoretext:
297 cfp.write(line)
298 cfp.write('\n')
299 message = cfp.getvalue()
300 if tmpfp:
301 tmpfp.write(payload)
302 if not payload.endswith('\n'):
303 tmpfp.write('\n')
304 elif not diffs_seen and message and content_type == 'text/plain':
305 message += '\n' + payload
306
307 if subject and not message.startswith(subject):
308 message = '%s\n%s' % (subject, message)
309 data['message'] = message
310 tmpfp.close()
311 if parents:
312 data['p1'] = parents.pop(0)
313 if parents:
314 data['p2'] = parents.pop(0)
315
316 if diffs_seen:
317 data['filename'] = tmpname
318
319 return data
320
321 class patchmeta(object):
322 """Patched file metadata
323
324 'op' is the performed operation within ADD, DELETE, RENAME, MODIFY
325 or COPY. 'path' is patched file path. 'oldpath' is set to the
326 origin file when 'op' is either COPY or RENAME, None otherwise. If
327 file mode is changed, 'mode' is a tuple (islink, isexec) where
328 'islink' is True if the file is a symlink and 'isexec' is True if
329 the file is executable. Otherwise, 'mode' is None.
330 """
331 def __init__(self, path):
332 self.path = path
333 self.oldpath = None
334 self.mode = None
335 self.op = 'MODIFY'
336 self.binary = False
337
338 def setmode(self, mode):
339 islink = mode & 0o20000
340 isexec = mode & 0o100
341 self.mode = (islink, isexec)
342
343 def copy(self):
344 other = patchmeta(self.path)
345 other.oldpath = self.oldpath
346 other.mode = self.mode
347 other.op = self.op
348 other.binary = self.binary
349 return other
350
351 def _ispatchinga(self, afile):
352 if afile == '/dev/null':
353 return self.op == 'ADD'
354 return afile == 'a/' + (self.oldpath or self.path)
355
356 def _ispatchingb(self, bfile):
357 if bfile == '/dev/null':
358 return self.op == 'DELETE'
359 return bfile == 'b/' + self.path
360
361 def ispatching(self, afile, bfile):
362 return self._ispatchinga(afile) and self._ispatchingb(bfile)
363
364 def __repr__(self):
365 return "<patchmeta %s %r>" % (self.op, self.path)
366
367 def readgitpatch(lr):
368 """extract git-style metadata about patches from <patchname>"""
369
370 # Filter patch for git information
371 gp = None
372 gitpatches = []
373 for line in lr:
374 line = line.rstrip(' \r\n')
375 if line.startswith('diff --git a/'):
376 m = gitre.match(line)
377 if m:
378 if gp:
379 gitpatches.append(gp)
380 dst = m.group(2)
381 gp = patchmeta(dst)
382 elif gp:
383 if line.startswith('--- '):
384 gitpatches.append(gp)
385 gp = None
386 continue
387 if line.startswith('rename from '):
388 gp.op = 'RENAME'
389 gp.oldpath = line[12:]
390 elif line.startswith('rename to '):
391 gp.path = line[10:]
392 elif line.startswith('copy from '):
393 gp.op = 'COPY'
394 gp.oldpath = line[10:]
395 elif line.startswith('copy to '):
396 gp.path = line[8:]
397 elif line.startswith('deleted file'):
398 gp.op = 'DELETE'
399 elif line.startswith('new file mode '):
400 gp.op = 'ADD'
401 gp.setmode(int(line[-6:], 8))
402 elif line.startswith('new mode '):
403 gp.setmode(int(line[-6:], 8))
404 elif line.startswith('GIT binary patch'):
405 gp.binary = True
406 if gp:
407 gitpatches.append(gp)
408
409 return gitpatches
410
411 class linereader(object):
412 # simple class to allow pushing lines back into the input stream
413 def __init__(self, fp):
414 self.fp = fp
415 self.buf = []
416
417 def push(self, line):
418 if line is not None:
419 self.buf.append(line)
420
421 def readline(self):
422 if self.buf:
423 l = self.buf[0]
424 del self.buf[0]
425 return l
426 return self.fp.readline()
427
428 def __iter__(self):
429 return iter(self.readline, '')
430
431 class abstractbackend(object):
432 def __init__(self, ui):
433 self.ui = ui
434
435 def getfile(self, fname):
436 """Return target file data and flags as a (data, (islink,
437 isexec)) tuple. Data is None if file is missing/deleted.
438 """
439 raise NotImplementedError
440
441 def setfile(self, fname, data, mode, copysource):
442 """Write data to target file fname and set its mode. mode is a
443 (islink, isexec) tuple. If data is None, the file content should
444 be left unchanged. If the file is modified after being copied,
445 copysource is set to the original file name.
446 """
447 raise NotImplementedError
448
449 def unlink(self, fname):
450 """Unlink target file."""
451 raise NotImplementedError
452
453 def writerej(self, fname, failed, total, lines):
454 """Write rejected lines for fname. total is the number of hunks
455 which failed to apply and total the total number of hunks for this
456 files.
457 """
458
459 def exists(self, fname):
460 raise NotImplementedError
461
462 def close(self):
463 raise NotImplementedError
464
465 class fsbackend(abstractbackend):
466 def __init__(self, ui, basedir):
467 super(fsbackend, self).__init__(ui)
468 self.opener = vfsmod.vfs(basedir)
469
470 def getfile(self, fname):
471 if self.opener.islink(fname):
472 return (self.opener.readlink(fname), (True, False))
473
474 isexec = False
475 try:
476 isexec = self.opener.lstat(fname).st_mode & 0o100 != 0
477 except OSError as e:
478 if e.errno != errno.ENOENT:
479 raise
480 try:
481 return (self.opener.read(fname), (False, isexec))
482 except IOError as e:
483 if e.errno != errno.ENOENT:
484 raise
485 return None, None
486
487 def setfile(self, fname, data, mode, copysource):
488 islink, isexec = mode
489 if data is None:
490 self.opener.setflags(fname, islink, isexec)
491 return
492 if islink:
493 self.opener.symlink(data, fname)
494 else:
495 self.opener.write(fname, data)
496 if isexec:
497 self.opener.setflags(fname, False, True)
498
499 def unlink(self, fname):
500 rmdir = self.ui.configbool('experimental', 'removeemptydirs')
501 self.opener.unlinkpath(fname, ignoremissing=True, rmdir=rmdir)
502
503 def writerej(self, fname, failed, total, lines):
504 fname = fname + ".rej"
505 self.ui.warn(
506 _("%d out of %d hunks FAILED -- saving rejects to file %s\n") %
507 (failed, total, fname))
508 fp = self.opener(fname, 'w')
509 fp.writelines(lines)
510 fp.close()
511
512 def exists(self, fname):
513 return self.opener.lexists(fname)
514
515 class workingbackend(fsbackend):
516 def __init__(self, ui, repo, similarity):
517 super(workingbackend, self).__init__(ui, repo.root)
518 self.repo = repo
519 self.similarity = similarity
520 self.removed = set()
521 self.changed = set()
522 self.copied = []
523
524 def _checkknown(self, fname):
525 if self.repo.dirstate[fname] == '?' and self.exists(fname):
526 raise PatchError(_('cannot patch %s: file is not tracked') % fname)
527
528 def setfile(self, fname, data, mode, copysource):
529 self._checkknown(fname)
530 super(workingbackend, self).setfile(fname, data, mode, copysource)
531 if copysource is not None:
532 self.copied.append((copysource, fname))
533 self.changed.add(fname)
534
535 def unlink(self, fname):
536 self._checkknown(fname)
537 super(workingbackend, self).unlink(fname)
538 self.removed.add(fname)
539 self.changed.add(fname)
540
11
541 def close(self):
12 from ..i18n import _
542 wctx = self.repo[None]
543 changed = set(self.changed)
544 for src, dst in self.copied:
545 scmutil.dirstatecopy(self.ui, self.repo, wctx, src, dst)
546 if self.removed:
547 wctx.forget(sorted(self.removed))
548 for f in self.removed:
549 if f not in self.repo.dirstate:
550 # File was deleted and no longer belongs to the
551 # dirstate, it was probably marked added then
552 # deleted, and should not be considered by
553 # marktouched().
554 changed.discard(f)
555 if changed:
556 scmutil.marktouched(self.repo, changed, self.similarity)
557 return sorted(self.changed)
558
559 class filestore(object):
560 def __init__(self, maxsize=None):
561 self.opener = None
562 self.files = {}
563 self.created = 0
564 self.maxsize = maxsize
565 if self.maxsize is None:
566 self.maxsize = 4*(2**20)
567 self.size = 0
568 self.data = {}
569
570 def setfile(self, fname, data, mode, copied=None):
571 if self.maxsize < 0 or (len(data) + self.size) <= self.maxsize:
572 self.data[fname] = (data, mode, copied)
573 self.size += len(data)
574 else:
575 if self.opener is None:
576 root = pycompat.mkdtemp(prefix='hg-patch-')
577 self.opener = vfsmod.vfs(root)
578 # Avoid filename issues with these simple names
579 fn = '%d' % self.created
580 self.opener.write(fn, data)
581 self.created += 1
582 self.files[fname] = (fn, mode, copied)
583
584 def getfile(self, fname):
585 if fname in self.data:
586 return self.data[fname]
587 if not self.opener or fname not in self.files:
588 return None, None, None
589 fn, mode, copied = self.files[fname]
590 return self.opener.read(fn), mode, copied
591
592 def close(self):
593 if self.opener:
594 shutil.rmtree(self.opener.base)
595
596 class repobackend(abstractbackend):
597 def __init__(self, ui, repo, ctx, store):
598 super(repobackend, self).__init__(ui)
599 self.repo = repo
600 self.ctx = ctx
601 self.store = store
602 self.changed = set()
603 self.removed = set()
604 self.copied = {}
605
606 def _checkknown(self, fname):
607 if fname not in self.ctx:
608 raise PatchError(_('cannot patch %s: file is not tracked') % fname)
609
610 def getfile(self, fname):
611 try:
612 fctx = self.ctx[fname]
613 except error.LookupError:
614 return None, None
615 flags = fctx.flags()
616 return fctx.data(), ('l' in flags, 'x' in flags)
617
618 def setfile(self, fname, data, mode, copysource):
619 if copysource:
620 self._checkknown(copysource)
621 if data is None:
622 data = self.ctx[fname].data()
623 self.store.setfile(fname, data, mode, copysource)
624 self.changed.add(fname)
625 if copysource:
626 self.copied[fname] = copysource
627
628 def unlink(self, fname):
629 self._checkknown(fname)
630 self.removed.add(fname)
631
632 def exists(self, fname):
633 return fname in self.ctx
634
635 def close(self):
636 return self.changed | self.removed
637
638 # @@ -start,len +start,len @@ or @@ -start +start @@ if len is 1
639 unidesc = re.compile('@@ -(\d+)(?:,(\d+))? \+(\d+)(?:,(\d+))? @@')
640 contextdesc = re.compile('(?:---|\*\*\*) (\d+)(?:,(\d+))? (?:---|\*\*\*)')
641 eolmodes = ['strict', 'crlf', 'lf', 'auto']
642
643 class patchfile(object):
644 def __init__(self, ui, gp, backend, store, eolmode='strict'):
645 self.fname = gp.path
646 self.eolmode = eolmode
647 self.eol = None
648 self.backend = backend
649 self.ui = ui
650 self.lines = []
651 self.exists = False
652 self.missing = True
653 self.mode = gp.mode
654 self.copysource = gp.oldpath
655 self.create = gp.op in ('ADD', 'COPY', 'RENAME')
656 self.remove = gp.op == 'DELETE'
657 if self.copysource is None:
658 data, mode = backend.getfile(self.fname)
659 else:
660 data, mode = store.getfile(self.copysource)[:2]
661 if data is not None:
662 self.exists = self.copysource is None or backend.exists(self.fname)
663 self.missing = False
664 if data:
665 self.lines = mdiff.splitnewlines(data)
666 if self.mode is None:
667 self.mode = mode
668 if self.lines:
669 # Normalize line endings
670 if self.lines[0].endswith('\r\n'):
671 self.eol = '\r\n'
672 elif self.lines[0].endswith('\n'):
673 self.eol = '\n'
674 if eolmode != 'strict':
675 nlines = []
676 for l in self.lines:
677 if l.endswith('\r\n'):
678 l = l[:-2] + '\n'
679 nlines.append(l)
680 self.lines = nlines
681 else:
682 if self.create:
683 self.missing = False
684 if self.mode is None:
685 self.mode = (False, False)
686 if self.missing:
687 self.ui.warn(_("unable to find '%s' for patching\n") % self.fname)
688 self.ui.warn(_("(use '--prefix' to apply patch relative to the "
689 "current directory)\n"))
690
691 self.hash = {}
692 self.dirty = 0
693 self.offset = 0
694 self.skew = 0
695 self.rej = []
696 self.fileprinted = False
697 self.printfile(False)
698 self.hunks = 0
699
700 def writelines(self, fname, lines, mode):
701 if self.eolmode == 'auto':
702 eol = self.eol
703 elif self.eolmode == 'crlf':
704 eol = '\r\n'
705 else:
706 eol = '\n'
707
708 if self.eolmode != 'strict' and eol and eol != '\n':
709 rawlines = []
710 for l in lines:
711 if l and l.endswith('\n'):
712 l = l[:-1] + eol
713 rawlines.append(l)
714 lines = rawlines
715
716 self.backend.setfile(fname, ''.join(lines), mode, self.copysource)
717
718 def printfile(self, warn):
719 if self.fileprinted:
720 return
721 if warn or self.ui.verbose:
722 self.fileprinted = True
723 s = _("patching file %s\n") % self.fname
724 if warn:
725 self.ui.warn(s)
726 else:
727 self.ui.note(s)
728
729
730 def findlines(self, l, linenum):
731 # looks through the hash and finds candidate lines. The
732 # result is a list of line numbers sorted based on distance
733 # from linenum
734
735 cand = self.hash.get(l, [])
736 if len(cand) > 1:
737 # resort our list of potentials forward then back.
738 cand.sort(key=lambda x: abs(x - linenum))
739 return cand
740
741 def write_rej(self):
742 # our rejects are a little different from patch(1). This always
743 # creates rejects in the same form as the original patch. A file
744 # header is inserted so that you can run the reject through patch again
745 # without having to type the filename.
746 if not self.rej:
747 return
748 base = os.path.basename(self.fname)
749 lines = ["--- %s\n+++ %s\n" % (base, base)]
750 for x in self.rej:
751 for l in x.hunk:
752 lines.append(l)
753 if l[-1:] != '\n':
754 lines.append("\n\ No newline at end of file\n")
755 self.backend.writerej(self.fname, len(self.rej), self.hunks, lines)
756
757 def apply(self, h):
758 if not h.complete():
759 raise PatchError(_("bad hunk #%d %s (%d %d %d %d)") %
760 (h.number, h.desc, len(h.a), h.lena, len(h.b),
761 h.lenb))
762
763 self.hunks += 1
764
765 if self.missing:
766 self.rej.append(h)
767 return -1
768
769 if self.exists and self.create:
770 if self.copysource:
771 self.ui.warn(_("cannot create %s: destination already "
772 "exists\n") % self.fname)
773 else:
774 self.ui.warn(_("file %s already exists\n") % self.fname)
775 self.rej.append(h)
776 return -1
777
778 if isinstance(h, binhunk):
779 if self.remove:
780 self.backend.unlink(self.fname)
781 else:
782 l = h.new(self.lines)
783 self.lines[:] = l
784 self.offset += len(l)
785 self.dirty = True
786 return 0
787
788 horig = h
789 if (self.eolmode in ('crlf', 'lf')
790 or self.eolmode == 'auto' and self.eol):
791 # If new eols are going to be normalized, then normalize
792 # hunk data before patching. Otherwise, preserve input
793 # line-endings.
794 h = h.getnormalized()
795
796 # fast case first, no offsets, no fuzz
797 old, oldstart, new, newstart = h.fuzzit(0, False)
798 oldstart += self.offset
799 orig_start = oldstart
800 # if there's skew we want to emit the "(offset %d lines)" even
801 # when the hunk cleanly applies at start + skew, so skip the
802 # fast case code
803 if self.skew == 0 and diffhelper.testhunk(old, self.lines, oldstart):
804 if self.remove:
805 self.backend.unlink(self.fname)
806 else:
807 self.lines[oldstart:oldstart + len(old)] = new
808 self.offset += len(new) - len(old)
809 self.dirty = True
810 return 0
811
812 # ok, we couldn't match the hunk. Lets look for offsets and fuzz it
813 self.hash = {}
814 for x, s in enumerate(self.lines):
815 self.hash.setdefault(s, []).append(x)
816
817 for fuzzlen in xrange(self.ui.configint("patch", "fuzz") + 1):
818 for toponly in [True, False]:
819 old, oldstart, new, newstart = h.fuzzit(fuzzlen, toponly)
820 oldstart = oldstart + self.offset + self.skew
821 oldstart = min(oldstart, len(self.lines))
822 if old:
823 cand = self.findlines(old[0][1:], oldstart)
824 else:
825 # Only adding lines with no or fuzzed context, just
826 # take the skew in account
827 cand = [oldstart]
828
829 for l in cand:
830 if not old or diffhelper.testhunk(old, self.lines, l):
831 self.lines[l : l + len(old)] = new
832 self.offset += len(new) - len(old)
833 self.skew = l - orig_start
834 self.dirty = True
835 offset = l - orig_start - fuzzlen
836 if fuzzlen:
837 msg = _("Hunk #%d succeeded at %d "
838 "with fuzz %d "
839 "(offset %d lines).\n")
840 self.printfile(True)
841 self.ui.warn(msg %
842 (h.number, l + 1, fuzzlen, offset))
843 else:
844 msg = _("Hunk #%d succeeded at %d "
845 "(offset %d lines).\n")
846 self.ui.note(msg % (h.number, l + 1, offset))
847 return fuzzlen
848 self.printfile(True)
849 self.ui.warn(_("Hunk #%d FAILED at %d\n") % (h.number, orig_start))
850 self.rej.append(horig)
851 return -1
852
853 def close(self):
854 if self.dirty:
855 self.writelines(self.fname, self.lines, self.mode)
856 self.write_rej()
857 return len(self.rej)
858
859 class header(object):
860 """patch header
861 """
862 diffgit_re = re.compile('diff --git a/(.*) b/(.*)$')
863 diff_re = re.compile('diff -r .* (.*)$')
864 allhunks_re = re.compile('(?:index|deleted file) ')
865 pretty_re = re.compile('(?:new file|deleted file) ')
866 special_re = re.compile('(?:index|deleted|copy|rename) ')
867 newfile_re = re.compile('(?:new file)')
868
869 def __init__(self, header):
870 self.header = header
871 self.hunks = []
872
873 def binary(self):
874 return any(h.startswith('index ') for h in self.header)
875
876 def pretty(self, fp):
877 for h in self.header:
878 if h.startswith('index '):
879 fp.write(_('this modifies a binary file (all or nothing)\n'))
880 break
881 if self.pretty_re.match(h):
882 fp.write(h)
883 if self.binary():
884 fp.write(_('this is a binary file\n'))
885 break
886 if h.startswith('---'):
887 fp.write(_('%d hunks, %d lines changed\n') %
888 (len(self.hunks),
889 sum([max(h.added, h.removed) for h in self.hunks])))
890 break
891 fp.write(h)
892
893 def write(self, fp):
894 fp.write(''.join(self.header))
895
896 def allhunks(self):
897 return any(self.allhunks_re.match(h) for h in self.header)
898
899 def files(self):
900 match = self.diffgit_re.match(self.header[0])
901 if match:
902 fromfile, tofile = match.groups()
903 if fromfile == tofile:
904 return [fromfile]
905 return [fromfile, tofile]
906 else:
907 return self.diff_re.match(self.header[0]).groups()
908
909 def filename(self):
910 return self.files()[-1]
911
912 def __repr__(self):
913 return '<header %s>' % (' '.join(map(repr, self.files())))
914
915 def isnewfile(self):
916 return any(self.newfile_re.match(h) for h in self.header)
917
918 def special(self):
919 # Special files are shown only at the header level and not at the hunk
920 # level for example a file that has been deleted is a special file.
921 # The user cannot change the content of the operation, in the case of
922 # the deleted file he has to take the deletion or not take it, he
923 # cannot take some of it.
924 # Newly added files are special if they are empty, they are not special
925 # if they have some content as we want to be able to change it
926 nocontent = len(self.header) == 2
927 emptynewfile = self.isnewfile() and nocontent
928 return emptynewfile or \
929 any(self.special_re.match(h) for h in self.header)
930
931 class recordhunk(object):
932 """patch hunk
933
934 XXX shouldn't we merge this with the other hunk class?
935 """
936
937 def __init__(self, header, fromline, toline, proc, before, hunk, after,
938 maxcontext=None):
939 def trimcontext(lines, reverse=False):
940 if maxcontext is not None:
941 delta = len(lines) - maxcontext
942 if delta > 0:
943 if reverse:
944 return delta, lines[delta:]
945 else:
946 return delta, lines[:maxcontext]
947 return 0, lines
948
949 self.header = header
950 trimedbefore, self.before = trimcontext(before, True)
951 self.fromline = fromline + trimedbefore
952 self.toline = toline + trimedbefore
953 _trimedafter, self.after = trimcontext(after, False)
954 self.proc = proc
955 self.hunk = hunk
956 self.added, self.removed = self.countchanges(self.hunk)
957
958 def __eq__(self, v):
959 if not isinstance(v, recordhunk):
960 return False
961
962 return ((v.hunk == self.hunk) and
963 (v.proc == self.proc) and
964 (self.fromline == v.fromline) and
965 (self.header.files() == v.header.files()))
966
967 def __hash__(self):
968 return hash((tuple(self.hunk),
969 tuple(self.header.files()),
970 self.fromline,
971 self.proc))
972
973 def countchanges(self, hunk):
974 """hunk -> (n+,n-)"""
975 add = len([h for h in hunk if h.startswith('+')])
976 rem = len([h for h in hunk if h.startswith('-')])
977 return add, rem
978
979 def reversehunk(self):
980 """return another recordhunk which is the reverse of the hunk
981
982 If this hunk is diff(A, B), the returned hunk is diff(B, A). To do
983 that, swap fromline/toline and +/- signs while keep other things
984 unchanged.
985 """
986 m = {'+': '-', '-': '+', '\\': '\\'}
987 hunk = ['%s%s' % (m[l[0:1]], l[1:]) for l in self.hunk]
988 return recordhunk(self.header, self.toline, self.fromline, self.proc,
989 self.before, hunk, self.after)
990
991 def write(self, fp):
992 delta = len(self.before) + len(self.after)
993 if self.after and self.after[-1] == '\\ No newline at end of file\n':
994 delta -= 1
995 fromlen = delta + self.removed
996 tolen = delta + self.added
997 fp.write('@@ -%d,%d +%d,%d @@%s\n' %
998 (self.fromline, fromlen, self.toline, tolen,
999 self.proc and (' ' + self.proc)))
1000 fp.write(''.join(self.before + self.hunk + self.after))
1001
1002 pretty = write
1003
1004 def filename(self):
1005 return self.header.filename()
1006
1007 def __repr__(self):
1008 return '<hunk %r@%d>' % (self.filename(), self.fromline)
1009
1010 def getmessages():
1011 return {
1012 'multiple': {
1013 'apply': _("apply change %d/%d to '%s'?"),
1014 'discard': _("discard change %d/%d to '%s'?"),
1015 'record': _("record change %d/%d to '%s'?"),
1016 },
1017 'single': {
1018 'apply': _("apply this change to '%s'?"),
1019 'discard': _("discard this change to '%s'?"),
1020 'record': _("record this change to '%s'?"),
1021 },
1022 'help': {
1023 'apply': _('[Ynesfdaq?]'
1024 '$$ &Yes, apply this change'
1025 '$$ &No, skip this change'
1026 '$$ &Edit this change manually'
1027 '$$ &Skip remaining changes to this file'
1028 '$$ Apply remaining changes to this &file'
1029 '$$ &Done, skip remaining changes and files'
1030 '$$ Apply &all changes to all remaining files'
1031 '$$ &Quit, applying no changes'
1032 '$$ &? (display help)'),
1033 'discard': _('[Ynesfdaq?]'
1034 '$$ &Yes, discard this change'
1035 '$$ &No, skip this change'
1036 '$$ &Edit this change manually'
1037 '$$ &Skip remaining changes to this file'
1038 '$$ Discard remaining changes to this &file'
1039 '$$ &Done, skip remaining changes and files'
1040 '$$ Discard &all changes to all remaining files'
1041 '$$ &Quit, discarding no changes'
1042 '$$ &? (display help)'),
1043 'record': _('[Ynesfdaq?]'
1044 '$$ &Yes, record this change'
1045 '$$ &No, skip this change'
1046 '$$ &Edit this change manually'
1047 '$$ &Skip remaining changes to this file'
1048 '$$ Record remaining changes to this &file'
1049 '$$ &Done, skip remaining changes and files'
1050 '$$ Record &all changes to all remaining files'
1051 '$$ &Quit, recording no changes'
1052 '$$ &? (display help)'),
1053 }
1054 }
1055
1056 def filterpatch(ui, headers, operation=None):
1057 """Interactively filter patch chunks into applied-only chunks"""
1058 messages = getmessages()
1059
1060 if operation is None:
1061 operation = 'record'
1062
1063 def prompt(skipfile, skipall, query, chunk):
1064 """prompt query, and process base inputs
1065
1066 - y/n for the rest of file
1067 - y/n for the rest
1068 - ? (help)
1069 - q (quit)
1070
1071 Return True/False and possibly updated skipfile and skipall.
1072 """
1073 newpatches = None
1074 if skipall is not None:
1075 return skipall, skipfile, skipall, newpatches
1076 if skipfile is not None:
1077 return skipfile, skipfile, skipall, newpatches
1078 while True:
1079 resps = messages['help'][operation]
1080 r = ui.promptchoice("%s %s" % (query, resps))
1081 ui.write("\n")
1082 if r == 8: # ?
1083 for c, t in ui.extractchoices(resps)[1]:
1084 ui.write('%s - %s\n' % (c, encoding.lower(t)))
1085 continue
1086 elif r == 0: # yes
1087 ret = True
1088 elif r == 1: # no
1089 ret = False
1090 elif r == 2: # Edit patch
1091 if chunk is None:
1092 ui.write(_('cannot edit patch for whole file'))
1093 ui.write("\n")
1094 continue
1095 if chunk.header.binary():
1096 ui.write(_('cannot edit patch for binary file'))
1097 ui.write("\n")
1098 continue
1099 # Patch comment based on the Git one (based on comment at end of
1100 # https://mercurial-scm.org/wiki/RecordExtension)
1101 phelp = '---' + _("""
1102 To remove '-' lines, make them ' ' lines (context).
1103 To remove '+' lines, delete them.
1104 Lines starting with # will be removed from the patch.
1105
13
1106 If the patch applies cleanly, the edited hunk will immediately be
14 from .. import (
1107 added to the record list. If it does not apply cleanly, a rejects
15 mdiff,
1108 file will be generated: you can use that when you try again. If
16 pycompat,
1109 all lines of the hunk are removed, then the edit is aborted and
17 )
1110 the hunk is left unchanged.
1111 """)
1112 (patchfd, patchfn) = pycompat.mkstemp(prefix="hg-editor-",
1113 suffix=".diff")
1114 ncpatchfp = None
1115 try:
1116 # Write the initial patch
1117 f = util.nativeeolwriter(os.fdopen(patchfd, r'wb'))
1118 chunk.header.write(f)
1119 chunk.write(f)
1120 f.write('\n'.join(['# ' + i for i in phelp.splitlines()]))
1121 f.close()
1122 # Start the editor and wait for it to complete
1123 editor = ui.geteditor()
1124 ret = ui.system("%s \"%s\"" % (editor, patchfn),
1125 environ={'HGUSER': ui.username()},
1126 blockedtag='filterpatch')
1127 if ret != 0:
1128 ui.warn(_("editor exited with exit code %d\n") % ret)
1129 continue
1130 # Remove comment lines
1131 patchfp = open(patchfn, r'rb')
1132 ncpatchfp = stringio()
1133 for line in util.iterfile(patchfp):
1134 line = util.fromnativeeol(line)
1135 if not line.startswith('#'):
1136 ncpatchfp.write(line)
1137 patchfp.close()
1138 ncpatchfp.seek(0)
1139 newpatches = parsepatch(ncpatchfp)
1140 finally:
1141 os.unlink(patchfn)
1142 del ncpatchfp
1143 # Signal that the chunk shouldn't be applied as-is, but
1144 # provide the new patch to be used instead.
1145 ret = False
1146 elif r == 3: # Skip
1147 ret = skipfile = False
1148 elif r == 4: # file (Record remaining)
1149 ret = skipfile = True
1150 elif r == 5: # done, skip remaining
1151 ret = skipall = False
1152 elif r == 6: # all
1153 ret = skipall = True
1154 elif r == 7: # quit
1155 raise error.Abort(_('user quit'))
1156 return ret, skipfile, skipall, newpatches
1157
1158 seen = set()
1159 applied = {} # 'filename' -> [] of chunks
1160 skipfile, skipall = None, None
1161 pos, total = 1, sum(len(h.hunks) for h in headers)
1162 for h in headers:
1163 pos += len(h.hunks)
1164 skipfile = None
1165 fixoffset = 0
1166 hdr = ''.join(h.header)
1167 if hdr in seen:
1168 continue
1169 seen.add(hdr)
1170 if skipall is None:
1171 h.pretty(ui)
1172 msg = (_('examine changes to %s?') %
1173 _(' and ').join("'%s'" % f for f in h.files()))
1174 r, skipfile, skipall, np = prompt(skipfile, skipall, msg, None)
1175 if not r:
1176 continue
1177 applied[h.filename()] = [h]
1178 if h.allhunks():
1179 applied[h.filename()] += h.hunks
1180 continue
1181 for i, chunk in enumerate(h.hunks):
1182 if skipfile is None and skipall is None:
1183 chunk.pretty(ui)
1184 if total == 1:
1185 msg = messages['single'][operation] % chunk.filename()
1186 else:
1187 idx = pos - len(h.hunks) + i
1188 msg = messages['multiple'][operation] % (idx, total,
1189 chunk.filename())
1190 r, skipfile, skipall, newpatches = prompt(skipfile,
1191 skipall, msg, chunk)
1192 if r:
1193 if fixoffset:
1194 chunk = copy.copy(chunk)
1195 chunk.toline += fixoffset
1196 applied[chunk.filename()].append(chunk)
1197 elif newpatches is not None:
1198 for newpatch in newpatches:
1199 for newhunk in newpatch.hunks:
1200 if fixoffset:
1201 newhunk.toline += fixoffset
1202 applied[newhunk.filename()].append(newhunk)
1203 else:
1204 fixoffset += chunk.removed - chunk.added
1205 return (sum([h for h in applied.itervalues()
1206 if h[0].special() or len(h) > 1], []), {})
1207 class hunk(object):
1208 def __init__(self, desc, num, lr, context):
1209 self.number = num
1210 self.desc = desc
1211 self.hunk = [desc]
1212 self.a = []
1213 self.b = []
1214 self.starta = self.lena = None
1215 self.startb = self.lenb = None
1216 if lr is not None:
1217 if context:
1218 self.read_context_hunk(lr)
1219 else:
1220 self.read_unified_hunk(lr)
1221
1222 def getnormalized(self):
1223 """Return a copy with line endings normalized to LF."""
1224
1225 def normalize(lines):
1226 nlines = []
1227 for line in lines:
1228 if line.endswith('\r\n'):
1229 line = line[:-2] + '\n'
1230 nlines.append(line)
1231 return nlines
1232
1233 # Dummy object, it is rebuilt manually
1234 nh = hunk(self.desc, self.number, None, None)
1235 nh.number = self.number
1236 nh.desc = self.desc
1237 nh.hunk = self.hunk
1238 nh.a = normalize(self.a)
1239 nh.b = normalize(self.b)
1240 nh.starta = self.starta
1241 nh.startb = self.startb
1242 nh.lena = self.lena
1243 nh.lenb = self.lenb
1244 return nh
1245
1246 def read_unified_hunk(self, lr):
1247 m = unidesc.match(self.desc)
1248 if not m:
1249 raise PatchError(_("bad hunk #%d") % self.number)
1250 self.starta, self.lena, self.startb, self.lenb = m.groups()
1251 if self.lena is None:
1252 self.lena = 1
1253 else:
1254 self.lena = int(self.lena)
1255 if self.lenb is None:
1256 self.lenb = 1
1257 else:
1258 self.lenb = int(self.lenb)
1259 self.starta = int(self.starta)
1260 self.startb = int(self.startb)
1261 try:
1262 diffhelper.addlines(lr, self.hunk, self.lena, self.lenb,
1263 self.a, self.b)
1264 except error.ParseError as e:
1265 raise PatchError(_("bad hunk #%d: %s") % (self.number, e))
1266 # if we hit eof before finishing out the hunk, the last line will
1267 # be zero length. Lets try to fix it up.
1268 while len(self.hunk[-1]) == 0:
1269 del self.hunk[-1]
1270 del self.a[-1]
1271 del self.b[-1]
1272 self.lena -= 1
1273 self.lenb -= 1
1274 self._fixnewline(lr)
1275
1276 def read_context_hunk(self, lr):
1277 self.desc = lr.readline()
1278 m = contextdesc.match(self.desc)
1279 if not m:
1280 raise PatchError(_("bad hunk #%d") % self.number)
1281 self.starta, aend = m.groups()
1282 self.starta = int(self.starta)
1283 if aend is None:
1284 aend = self.starta
1285 self.lena = int(aend) - self.starta
1286 if self.starta:
1287 self.lena += 1
1288 for x in xrange(self.lena):
1289 l = lr.readline()
1290 if l.startswith('---'):
1291 # lines addition, old block is empty
1292 lr.push(l)
1293 break
1294 s = l[2:]
1295 if l.startswith('- ') or l.startswith('! '):
1296 u = '-' + s
1297 elif l.startswith(' '):
1298 u = ' ' + s
1299 else:
1300 raise PatchError(_("bad hunk #%d old text line %d") %
1301 (self.number, x))
1302 self.a.append(u)
1303 self.hunk.append(u)
1304
1305 l = lr.readline()
1306 if l.startswith('\ '):
1307 s = self.a[-1][:-1]
1308 self.a[-1] = s
1309 self.hunk[-1] = s
1310 l = lr.readline()
1311 m = contextdesc.match(l)
1312 if not m:
1313 raise PatchError(_("bad hunk #%d") % self.number)
1314 self.startb, bend = m.groups()
1315 self.startb = int(self.startb)
1316 if bend is None:
1317 bend = self.startb
1318 self.lenb = int(bend) - self.startb
1319 if self.startb:
1320 self.lenb += 1
1321 hunki = 1
1322 for x in xrange(self.lenb):
1323 l = lr.readline()
1324 if l.startswith('\ '):
1325 # XXX: the only way to hit this is with an invalid line range.
1326 # The no-eol marker is not counted in the line range, but I
1327 # guess there are diff(1) out there which behave differently.
1328 s = self.b[-1][:-1]
1329 self.b[-1] = s
1330 self.hunk[hunki - 1] = s
1331 continue
1332 if not l:
1333 # line deletions, new block is empty and we hit EOF
1334 lr.push(l)
1335 break
1336 s = l[2:]
1337 if l.startswith('+ ') or l.startswith('! '):
1338 u = '+' + s
1339 elif l.startswith(' '):
1340 u = ' ' + s
1341 elif len(self.b) == 0:
1342 # line deletions, new block is empty
1343 lr.push(l)
1344 break
1345 else:
1346 raise PatchError(_("bad hunk #%d old text line %d") %
1347 (self.number, x))
1348 self.b.append(s)
1349 while True:
1350 if hunki >= len(self.hunk):
1351 h = ""
1352 else:
1353 h = self.hunk[hunki]
1354 hunki += 1
1355 if h == u:
1356 break
1357 elif h.startswith('-'):
1358 continue
1359 else:
1360 self.hunk.insert(hunki - 1, u)
1361 break
1362
1363 if not self.a:
1364 # this happens when lines were only added to the hunk
1365 for x in self.hunk:
1366 if x.startswith('-') or x.startswith(' '):
1367 self.a.append(x)
1368 if not self.b:
1369 # this happens when lines were only deleted from the hunk
1370 for x in self.hunk:
1371 if x.startswith('+') or x.startswith(' '):
1372 self.b.append(x[1:])
1373 # @@ -start,len +start,len @@
1374 self.desc = "@@ -%d,%d +%d,%d @@\n" % (self.starta, self.lena,
1375 self.startb, self.lenb)
1376 self.hunk[0] = self.desc
1377 self._fixnewline(lr)
1378
1379 def _fixnewline(self, lr):
1380 l = lr.readline()
1381 if l.startswith('\ '):
1382 diffhelper.fixnewline(self.hunk, self.a, self.b)
1383 else:
1384 lr.push(l)
1385
1386 def complete(self):
1387 return len(self.a) == self.lena and len(self.b) == self.lenb
1388
1389 def _fuzzit(self, old, new, fuzz, toponly):
1390 # this removes context lines from the top and bottom of list 'l'. It
1391 # checks the hunk to make sure only context lines are removed, and then
1392 # returns a new shortened list of lines.
1393 fuzz = min(fuzz, len(old))
1394 if fuzz:
1395 top = 0
1396 bot = 0
1397 hlen = len(self.hunk)
1398 for x in xrange(hlen - 1):
1399 # the hunk starts with the @@ line, so use x+1
1400 if self.hunk[x + 1].startswith(' '):
1401 top += 1
1402 else:
1403 break
1404 if not toponly:
1405 for x in xrange(hlen - 1):
1406 if self.hunk[hlen - bot - 1].startswith(' '):
1407 bot += 1
1408 else:
1409 break
1410
1411 bot = min(fuzz, bot)
1412 top = min(fuzz, top)
1413 return old[top:len(old) - bot], new[top:len(new) - bot], top
1414 return old, new, 0
1415
1416 def fuzzit(self, fuzz, toponly):
1417 old, new, top = self._fuzzit(self.a, self.b, fuzz, toponly)
1418 oldstart = self.starta + top
1419 newstart = self.startb + top
1420 # zero length hunk ranges already have their start decremented
1421 if self.lena and oldstart > 0:
1422 oldstart -= 1
1423 if self.lenb and newstart > 0:
1424 newstart -= 1
1425 return old, oldstart, new, newstart
1426
1427 class binhunk(object):
1428 'A binary patch file.'
1429 def __init__(self, lr, fname):
1430 self.text = None
1431 self.delta = False
1432 self.hunk = ['GIT binary patch\n']
1433 self._fname = fname
1434 self._read(lr)
1435
1436 def complete(self):
1437 return self.text is not None
1438
1439 def new(self, lines):
1440 if self.delta:
1441 return [applybindelta(self.text, ''.join(lines))]
1442 return [self.text]
1443
1444 def _read(self, lr):
1445 def getline(lr, hunk):
1446 l = lr.readline()
1447 hunk.append(l)
1448 return l.rstrip('\r\n')
1449
1450 size = 0
1451 while True:
1452 line = getline(lr, self.hunk)
1453 if not line:
1454 raise PatchError(_('could not extract "%s" binary data')
1455 % self._fname)
1456 if line.startswith('literal '):
1457 size = int(line[8:].rstrip())
1458 break
1459 if line.startswith('delta '):
1460 size = int(line[6:].rstrip())
1461 self.delta = True
1462 break
1463 dec = []
1464 line = getline(lr, self.hunk)
1465 while len(line) > 1:
1466 l = line[0:1]
1467 if l <= 'Z' and l >= 'A':
1468 l = ord(l) - ord('A') + 1
1469 else:
1470 l = ord(l) - ord('a') + 27
1471 try:
1472 dec.append(util.b85decode(line[1:])[:l])
1473 except ValueError as e:
1474 raise PatchError(_('could not decode "%s" binary patch: %s')
1475 % (self._fname, stringutil.forcebytestr(e)))
1476 line = getline(lr, self.hunk)
1477 text = zlib.decompress(''.join(dec))
1478 if len(text) != size:
1479 raise PatchError(_('"%s" length is %d bytes, should be %d')
1480 % (self._fname, len(text), size))
1481 self.text = text
1482
1483 def parsefilename(str):
1484 # --- filename \t|space stuff
1485 s = str[4:].rstrip('\r\n')
1486 i = s.find('\t')
1487 if i < 0:
1488 i = s.find(' ')
1489 if i < 0:
1490 return s
1491 return s[:i]
1492
1493 def reversehunks(hunks):
1494 '''reverse the signs in the hunks given as argument
1495
1496 This function operates on hunks coming out of patch.filterpatch, that is
1497 a list of the form: [header1, hunk1, hunk2, header2...]. Example usage:
1498
1499 >>> rawpatch = b"""diff --git a/folder1/g b/folder1/g
1500 ... --- a/folder1/g
1501 ... +++ b/folder1/g
1502 ... @@ -1,7 +1,7 @@
1503 ... +firstline
1504 ... c
1505 ... 1
1506 ... 2
1507 ... + 3
1508 ... -4
1509 ... 5
1510 ... d
1511 ... +lastline"""
1512 >>> hunks = parsepatch([rawpatch])
1513 >>> hunkscomingfromfilterpatch = []
1514 >>> for h in hunks:
1515 ... hunkscomingfromfilterpatch.append(h)
1516 ... hunkscomingfromfilterpatch.extend(h.hunks)
1517
1518 >>> reversedhunks = reversehunks(hunkscomingfromfilterpatch)
1519 >>> from . import util
1520 >>> fp = util.stringio()
1521 >>> for c in reversedhunks:
1522 ... c.write(fp)
1523 >>> fp.seek(0) or None
1524 >>> reversedpatch = fp.read()
1525 >>> print(pycompat.sysstr(reversedpatch))
1526 diff --git a/folder1/g b/folder1/g
1527 --- a/folder1/g
1528 +++ b/folder1/g
1529 @@ -1,4 +1,3 @@
1530 -firstline
1531 c
1532 1
1533 2
1534 @@ -2,6 +1,6 @@
1535 c
1536 1
1537 2
1538 - 3
1539 +4
1540 5
1541 d
1542 @@ -6,3 +5,2 @@
1543 5
1544 d
1545 -lastline
1546
1547 '''
1548
1549 newhunks = []
1550 for c in hunks:
1551 if util.safehasattr(c, 'reversehunk'):
1552 c = c.reversehunk()
1553 newhunks.append(c)
1554 return newhunks
1555
1556 def parsepatch(originalchunks, maxcontext=None):
1557 """patch -> [] of headers -> [] of hunks
1558
1559 If maxcontext is not None, trim context lines if necessary.
1560
1561 >>> rawpatch = b'''diff --git a/folder1/g b/folder1/g
1562 ... --- a/folder1/g
1563 ... +++ b/folder1/g
1564 ... @@ -1,8 +1,10 @@
1565 ... 1
1566 ... 2
1567 ... -3
1568 ... 4
1569 ... 5
1570 ... 6
1571 ... +6.1
1572 ... +6.2
1573 ... 7
1574 ... 8
1575 ... +9'''
1576 >>> out = util.stringio()
1577 >>> headers = parsepatch([rawpatch], maxcontext=1)
1578 >>> for header in headers:
1579 ... header.write(out)
1580 ... for hunk in header.hunks:
1581 ... hunk.write(out)
1582 >>> print(pycompat.sysstr(out.getvalue()))
1583 diff --git a/folder1/g b/folder1/g
1584 --- a/folder1/g
1585 +++ b/folder1/g
1586 @@ -2,3 +2,2 @@
1587 2
1588 -3
1589 4
1590 @@ -6,2 +5,4 @@
1591 6
1592 +6.1
1593 +6.2
1594 7
1595 @@ -8,1 +9,2 @@
1596 8
1597 +9
1598 """
1599 class parser(object):
1600 """patch parsing state machine"""
1601 def __init__(self):
1602 self.fromline = 0
1603 self.toline = 0
1604 self.proc = ''
1605 self.header = None
1606 self.context = []
1607 self.before = []
1608 self.hunk = []
1609 self.headers = []
1610
1611 def addrange(self, limits):
1612 fromstart, fromend, tostart, toend, proc = limits
1613 self.fromline = int(fromstart)
1614 self.toline = int(tostart)
1615 self.proc = proc
1616
1617 def addcontext(self, context):
1618 if self.hunk:
1619 h = recordhunk(self.header, self.fromline, self.toline,
1620 self.proc, self.before, self.hunk, context, maxcontext)
1621 self.header.hunks.append(h)
1622 self.fromline += len(self.before) + h.removed
1623 self.toline += len(self.before) + h.added
1624 self.before = []
1625 self.hunk = []
1626 self.context = context
1627
1628 def addhunk(self, hunk):
1629 if self.context:
1630 self.before = self.context
1631 self.context = []
1632 self.hunk = hunk
1633
1634 def newfile(self, hdr):
1635 self.addcontext([])
1636 h = header(hdr)
1637 self.headers.append(h)
1638 self.header = h
1639
1640 def addother(self, line):
1641 pass # 'other' lines are ignored
1642
1643 def finished(self):
1644 self.addcontext([])
1645 return self.headers
1646
1647 transitions = {
1648 'file': {'context': addcontext,
1649 'file': newfile,
1650 'hunk': addhunk,
1651 'range': addrange},
1652 'context': {'file': newfile,
1653 'hunk': addhunk,
1654 'range': addrange,
1655 'other': addother},
1656 'hunk': {'context': addcontext,
1657 'file': newfile,
1658 'range': addrange},
1659 'range': {'context': addcontext,
1660 'hunk': addhunk},
1661 'other': {'other': addother},
1662 }
1663
1664 p = parser()
1665 fp = stringio()
1666 fp.write(''.join(originalchunks))
1667 fp.seek(0)
1668
1669 state = 'context'
1670 for newstate, data in scanpatch(fp):
1671 try:
1672 p.transitions[state][newstate](p, data)
1673 except KeyError:
1674 raise PatchError('unhandled transition: %s -> %s' %
1675 (state, newstate))
1676 state = newstate
1677 del fp
1678 return p.finished()
1679
1680 def pathtransform(path, strip, prefix):
1681 '''turn a path from a patch into a path suitable for the repository
1682
1683 prefix, if not empty, is expected to be normalized with a / at the end.
1684
1685 Returns (stripped components, path in repository).
1686
1687 >>> pathtransform(b'a/b/c', 0, b'')
1688 ('', 'a/b/c')
1689 >>> pathtransform(b' a/b/c ', 0, b'')
1690 ('', ' a/b/c')
1691 >>> pathtransform(b' a/b/c ', 2, b'')
1692 ('a/b/', 'c')
1693 >>> pathtransform(b'a/b/c', 0, b'd/e/')
1694 ('', 'd/e/a/b/c')
1695 >>> pathtransform(b' a//b/c ', 2, b'd/e/')
1696 ('a//b/', 'd/e/c')
1697 >>> pathtransform(b'a/b/c', 3, b'')
1698 Traceback (most recent call last):
1699 PatchError: unable to strip away 1 of 3 dirs from a/b/c
1700 '''
1701 pathlen = len(path)
1702 i = 0
1703 if strip == 0:
1704 return '', prefix + path.rstrip()
1705 count = strip
1706 while count > 0:
1707 i = path.find('/', i)
1708 if i == -1:
1709 raise PatchError(_("unable to strip away %d of %d dirs from %s") %
1710 (count, strip, path))
1711 i += 1
1712 # consume '//' in the path
1713 while i < pathlen - 1 and path[i:i + 1] == '/':
1714 i += 1
1715 count -= 1
1716 return path[:i].lstrip(), prefix + path[i:].rstrip()
1717
1718 def makepatchmeta(backend, afile_orig, bfile_orig, hunk, strip, prefix):
1719 nulla = afile_orig == "/dev/null"
1720 nullb = bfile_orig == "/dev/null"
1721 create = nulla and hunk.starta == 0 and hunk.lena == 0
1722 remove = nullb and hunk.startb == 0 and hunk.lenb == 0
1723 abase, afile = pathtransform(afile_orig, strip, prefix)
1724 gooda = not nulla and backend.exists(afile)
1725 bbase, bfile = pathtransform(bfile_orig, strip, prefix)
1726 if afile == bfile:
1727 goodb = gooda
1728 else:
1729 goodb = not nullb and backend.exists(bfile)
1730 missing = not goodb and not gooda and not create
1731
1732 # some diff programs apparently produce patches where the afile is
1733 # not /dev/null, but afile starts with bfile
1734 abasedir = afile[:afile.rfind('/') + 1]
1735 bbasedir = bfile[:bfile.rfind('/') + 1]
1736 if (missing and abasedir == bbasedir and afile.startswith(bfile)
1737 and hunk.starta == 0 and hunk.lena == 0):
1738 create = True
1739 missing = False
1740
1741 # If afile is "a/b/foo" and bfile is "a/b/foo.orig" we assume the
1742 # diff is between a file and its backup. In this case, the original
1743 # file should be patched (see original mpatch code).
1744 isbackup = (abase == bbase and bfile.startswith(afile))
1745 fname = None
1746 if not missing:
1747 if gooda and goodb:
1748 if isbackup:
1749 fname = afile
1750 else:
1751 fname = bfile
1752 elif gooda:
1753 fname = afile
1754
1755 if not fname:
1756 if not nullb:
1757 if isbackup:
1758 fname = afile
1759 else:
1760 fname = bfile
1761 elif not nulla:
1762 fname = afile
1763 else:
1764 raise PatchError(_("undefined source and destination files"))
1765
1766 gp = patchmeta(fname)
1767 if create:
1768 gp.op = 'ADD'
1769 elif remove:
1770 gp.op = 'DELETE'
1771 return gp
1772
1773 def scanpatch(fp):
1774 """like patch.iterhunks, but yield different events
1775
1776 - ('file', [header_lines + fromfile + tofile])
1777 - ('context', [context_lines])
1778 - ('hunk', [hunk_lines])
1779 - ('range', (-start,len, +start,len, proc))
1780 """
1781 lines_re = re.compile(br'@@ -(\d+),(\d+) \+(\d+),(\d+) @@\s*(.*)')
1782 lr = linereader(fp)
1783
1784 def scanwhile(first, p):
1785 """scan lr while predicate holds"""
1786 lines = [first]
1787 for line in iter(lr.readline, ''):
1788 if p(line):
1789 lines.append(line)
1790 else:
1791 lr.push(line)
1792 break
1793 return lines
1794
1795 for line in iter(lr.readline, ''):
1796 if line.startswith('diff --git a/') or line.startswith('diff -r '):
1797 def notheader(line):
1798 s = line.split(None, 1)
1799 return not s or s[0] not in ('---', 'diff')
1800 header = scanwhile(line, notheader)
1801 fromfile = lr.readline()
1802 if fromfile.startswith('---'):
1803 tofile = lr.readline()
1804 header += [fromfile, tofile]
1805 else:
1806 lr.push(fromfile)
1807 yield 'file', header
1808 elif line.startswith(' '):
1809 cs = (' ', '\\')
1810 yield 'context', scanwhile(line, lambda l: l.startswith(cs))
1811 elif line.startswith(('-', '+')):
1812 cs = ('-', '+', '\\')
1813 yield 'hunk', scanwhile(line, lambda l: l.startswith(cs))
1814 else:
1815 m = lines_re.match(line)
1816 if m:
1817 yield 'range', m.groups()
1818 else:
1819 yield 'other', line
1820
1821 def scangitpatch(lr, firstline):
1822 """
1823 Git patches can emit:
1824 - rename a to b
1825 - change b
1826 - copy a to c
1827 - change c
1828
1829 We cannot apply this sequence as-is, the renamed 'a' could not be
1830 found for it would have been renamed already. And we cannot copy
1831 from 'b' instead because 'b' would have been changed already. So
1832 we scan the git patch for copy and rename commands so we can
1833 perform the copies ahead of time.
1834 """
1835 pos = 0
1836 try:
1837 pos = lr.fp.tell()
1838 fp = lr.fp
1839 except IOError:
1840 fp = stringio(lr.fp.read())
1841 gitlr = linereader(fp)
1842 gitlr.push(firstline)
1843 gitpatches = readgitpatch(gitlr)
1844 fp.seek(pos)
1845 return gitpatches
1846
1847 def iterhunks(fp):
1848 """Read a patch and yield the following events:
1849 - ("file", afile, bfile, firsthunk): select a new target file.
1850 - ("hunk", hunk): a new hunk is ready to be applied, follows a
1851 "file" event.
1852 - ("git", gitchanges): current diff is in git format, gitchanges
1853 maps filenames to gitpatch records. Unique event.
1854 """
1855 afile = ""
1856 bfile = ""
1857 state = None
1858 hunknum = 0
1859 emitfile = newfile = False
1860 gitpatches = None
1861
1862 # our states
1863 BFILE = 1
1864 context = None
1865 lr = linereader(fp)
1866
1867 for x in iter(lr.readline, ''):
1868 if state == BFILE and (
1869 (not context and x.startswith('@'))
1870 or (context is not False and x.startswith('***************'))
1871 or x.startswith('GIT binary patch')):
1872 gp = None
1873 if (gitpatches and
1874 gitpatches[-1].ispatching(afile, bfile)):
1875 gp = gitpatches.pop()
1876 if x.startswith('GIT binary patch'):
1877 h = binhunk(lr, gp.path)
1878 else:
1879 if context is None and x.startswith('***************'):
1880 context = True
1881 h = hunk(x, hunknum + 1, lr, context)
1882 hunknum += 1
1883 if emitfile:
1884 emitfile = False
1885 yield 'file', (afile, bfile, h, gp and gp.copy() or None)
1886 yield 'hunk', h
1887 elif x.startswith('diff --git a/'):
1888 m = gitre.match(x.rstrip(' \r\n'))
1889 if not m:
1890 continue
1891 if gitpatches is None:
1892 # scan whole input for git metadata
1893 gitpatches = scangitpatch(lr, x)
1894 yield 'git', [g.copy() for g in gitpatches
1895 if g.op in ('COPY', 'RENAME')]
1896 gitpatches.reverse()
1897 afile = 'a/' + m.group(1)
1898 bfile = 'b/' + m.group(2)
1899 while gitpatches and not gitpatches[-1].ispatching(afile, bfile):
1900 gp = gitpatches.pop()
1901 yield 'file', ('a/' + gp.path, 'b/' + gp.path, None, gp.copy())
1902 if not gitpatches:
1903 raise PatchError(_('failed to synchronize metadata for "%s"')
1904 % afile[2:])
1905 gp = gitpatches[-1]
1906 newfile = True
1907 elif x.startswith('---'):
1908 # check for a unified diff
1909 l2 = lr.readline()
1910 if not l2.startswith('+++'):
1911 lr.push(l2)
1912 continue
1913 newfile = True
1914 context = False
1915 afile = parsefilename(x)
1916 bfile = parsefilename(l2)
1917 elif x.startswith('***'):
1918 # check for a context diff
1919 l2 = lr.readline()
1920 if not l2.startswith('---'):
1921 lr.push(l2)
1922 continue
1923 l3 = lr.readline()
1924 lr.push(l3)
1925 if not l3.startswith("***************"):
1926 lr.push(l2)
1927 continue
1928 newfile = True
1929 context = True
1930 afile = parsefilename(x)
1931 bfile = parsefilename(l2)
1932
1933 if newfile:
1934 newfile = False
1935 emitfile = True
1936 state = BFILE
1937 hunknum = 0
1938
1939 while gitpatches:
1940 gp = gitpatches.pop()
1941 yield 'file', ('a/' + gp.path, 'b/' + gp.path, None, gp.copy())
1942
1943 def applybindelta(binchunk, data):
1944 """Apply a binary delta hunk
1945 The algorithm used is the algorithm from git's patch-delta.c
1946 """
1947 def deltahead(binchunk):
1948 i = 0
1949 for c in pycompat.bytestr(binchunk):
1950 i += 1
1951 if not (ord(c) & 0x80):
1952 return i
1953 return i
1954 out = ""
1955 s = deltahead(binchunk)
1956 binchunk = binchunk[s:]
1957 s = deltahead(binchunk)
1958 binchunk = binchunk[s:]
1959 i = 0
1960 while i < len(binchunk):
1961 cmd = ord(binchunk[i:i + 1])
1962 i += 1
1963 if (cmd & 0x80):
1964 offset = 0
1965 size = 0
1966 if (cmd & 0x01):
1967 offset = ord(binchunk[i:i + 1])
1968 i += 1
1969 if (cmd & 0x02):
1970 offset |= ord(binchunk[i:i + 1]) << 8
1971 i += 1
1972 if (cmd & 0x04):
1973 offset |= ord(binchunk[i:i + 1]) << 16
1974 i += 1
1975 if (cmd & 0x08):
1976 offset |= ord(binchunk[i:i + 1]) << 24
1977 i += 1
1978 if (cmd & 0x10):
1979 size = ord(binchunk[i:i + 1])
1980 i += 1
1981 if (cmd & 0x20):
1982 size |= ord(binchunk[i:i + 1]) << 8
1983 i += 1
1984 if (cmd & 0x40):
1985 size |= ord(binchunk[i:i + 1]) << 16
1986 i += 1
1987 if size == 0:
1988 size = 0x10000
1989 offset_end = offset + size
1990 out += data[offset:offset_end]
1991 elif cmd != 0:
1992 offset_end = i + cmd
1993 out += binchunk[i:offset_end]
1994 i += cmd
1995 else:
1996 raise PatchError(_('unexpected delta opcode 0'))
1997 return out
1998
1999 def applydiff(ui, fp, backend, store, strip=1, prefix='', eolmode='strict'):
2000 """Reads a patch from fp and tries to apply it.
2001
2002 Returns 0 for a clean patch, -1 if any rejects were found and 1 if
2003 there was any fuzz.
2004
2005 If 'eolmode' is 'strict', the patch content and patched file are
2006 read in binary mode. Otherwise, line endings are ignored when
2007 patching then normalized according to 'eolmode'.
2008 """
2009 return _applydiff(ui, fp, patchfile, backend, store, strip=strip,
2010 prefix=prefix, eolmode=eolmode)
2011
2012 def _canonprefix(repo, prefix):
2013 if prefix:
2014 prefix = pathutil.canonpath(repo.root, repo.getcwd(), prefix)
2015 if prefix != '':
2016 prefix += '/'
2017 return prefix
2018
2019 def _applydiff(ui, fp, patcher, backend, store, strip=1, prefix='',
2020 eolmode='strict'):
2021 prefix = _canonprefix(backend.repo, prefix)
2022 def pstrip(p):
2023 return pathtransform(p, strip - 1, prefix)[1]
2024
2025 rejects = 0
2026 err = 0
2027 current_file = None
2028
2029 for state, values in iterhunks(fp):
2030 if state == 'hunk':
2031 if not current_file:
2032 continue
2033 ret = current_file.apply(values)
2034 if ret > 0:
2035 err = 1
2036 elif state == 'file':
2037 if current_file:
2038 rejects += current_file.close()
2039 current_file = None
2040 afile, bfile, first_hunk, gp = values
2041 if gp:
2042 gp.path = pstrip(gp.path)
2043 if gp.oldpath:
2044 gp.oldpath = pstrip(gp.oldpath)
2045 else:
2046 gp = makepatchmeta(backend, afile, bfile, first_hunk, strip,
2047 prefix)
2048 if gp.op == 'RENAME':
2049 backend.unlink(gp.oldpath)
2050 if not first_hunk:
2051 if gp.op == 'DELETE':
2052 backend.unlink(gp.path)
2053 continue
2054 data, mode = None, None
2055 if gp.op in ('RENAME', 'COPY'):
2056 data, mode = store.getfile(gp.oldpath)[:2]
2057 if data is None:
2058 # This means that the old path does not exist
2059 raise PatchError(_("source file '%s' does not exist")
2060 % gp.oldpath)
2061 if gp.mode:
2062 mode = gp.mode
2063 if gp.op == 'ADD':
2064 # Added files without content have no hunk and
2065 # must be created
2066 data = ''
2067 if data or mode:
2068 if (gp.op in ('ADD', 'RENAME', 'COPY')
2069 and backend.exists(gp.path)):
2070 raise PatchError(_("cannot create %s: destination "
2071 "already exists") % gp.path)
2072 backend.setfile(gp.path, data, mode, gp.oldpath)
2073 continue
2074 try:
2075 current_file = patcher(ui, gp, backend, store,
2076 eolmode=eolmode)
2077 except PatchError as inst:
2078 ui.warn(str(inst) + '\n')
2079 current_file = None
2080 rejects += 1
2081 continue
2082 elif state == 'git':
2083 for gp in values:
2084 path = pstrip(gp.oldpath)
2085 data, mode = backend.getfile(path)
2086 if data is None:
2087 # The error ignored here will trigger a getfile()
2088 # error in a place more appropriate for error
2089 # handling, and will not interrupt the patching
2090 # process.
2091 pass
2092 else:
2093 store.setfile(path, data, mode)
2094 else:
2095 raise error.Abort(_('unsupported parser state: %s') % state)
2096
2097 if current_file:
2098 rejects += current_file.close()
2099
2100 if rejects:
2101 return -1
2102 return err
2103
2104 def _externalpatch(ui, repo, patcher, patchname, strip, files,
2105 similarity):
2106 """use <patcher> to apply <patchname> to the working directory.
2107 returns whether patch was applied with fuzz factor."""
2108
2109 fuzz = False
2110 args = []
2111 cwd = repo.root
2112 if cwd:
2113 args.append('-d %s' % procutil.shellquote(cwd))
2114 cmd = ('%s %s -p%d < %s'
2115 % (patcher, ' '.join(args), strip, procutil.shellquote(patchname)))
2116 ui.debug('Using external patch tool: %s\n' % cmd)
2117 fp = procutil.popen(cmd, 'rb')
2118 try:
2119 for line in util.iterfile(fp):
2120 line = line.rstrip()
2121 ui.note(line + '\n')
2122 if line.startswith('patching file '):
2123 pf = util.parsepatchoutput(line)
2124 printed_file = False
2125 files.add(pf)
2126 elif line.find('with fuzz') >= 0:
2127 fuzz = True
2128 if not printed_file:
2129 ui.warn(pf + '\n')
2130 printed_file = True
2131 ui.warn(line + '\n')
2132 elif line.find('saving rejects to file') >= 0:
2133 ui.warn(line + '\n')
2134 elif line.find('FAILED') >= 0:
2135 if not printed_file:
2136 ui.warn(pf + '\n')
2137 printed_file = True
2138 ui.warn(line + '\n')
2139 finally:
2140 if files:
2141 scmutil.marktouched(repo, files, similarity)
2142 code = fp.close()
2143 if code:
2144 raise PatchError(_("patch command failed: %s") %
2145 procutil.explainexit(code))
2146 return fuzz
2147
2148 def patchbackend(ui, backend, patchobj, strip, prefix, files=None,
2149 eolmode='strict'):
2150 if files is None:
2151 files = set()
2152 if eolmode is None:
2153 eolmode = ui.config('patch', 'eol')
2154 if eolmode.lower() not in eolmodes:
2155 raise error.Abort(_('unsupported line endings type: %s') % eolmode)
2156 eolmode = eolmode.lower()
2157
2158 store = filestore()
2159 try:
2160 fp = open(patchobj, 'rb')
2161 except TypeError:
2162 fp = patchobj
2163 try:
2164 ret = applydiff(ui, fp, backend, store, strip=strip, prefix=prefix,
2165 eolmode=eolmode)
2166 finally:
2167 if fp != patchobj:
2168 fp.close()
2169 files.update(backend.close())
2170 store.close()
2171 if ret < 0:
2172 raise PatchError(_('patch failed to apply'))
2173 return ret > 0
2174
2175 def internalpatch(ui, repo, patchobj, strip, prefix='', files=None,
2176 eolmode='strict', similarity=0):
2177 """use builtin patch to apply <patchobj> to the working directory.
2178 returns whether patch was applied with fuzz factor."""
2179 backend = workingbackend(ui, repo, similarity)
2180 return patchbackend(ui, backend, patchobj, strip, prefix, files, eolmode)
2181
2182 def patchrepo(ui, repo, ctx, store, patchobj, strip, prefix, files=None,
2183 eolmode='strict'):
2184 backend = repobackend(ui, repo, ctx, store)
2185 return patchbackend(ui, backend, patchobj, strip, prefix, files, eolmode)
2186
2187 def patch(ui, repo, patchname, strip=1, prefix='', files=None, eolmode='strict',
2188 similarity=0):
2189 """Apply <patchname> to the working directory.
2190
2191 'eolmode' specifies how end of lines should be handled. It can be:
2192 - 'strict': inputs are read in binary mode, EOLs are preserved
2193 - 'crlf': EOLs are ignored when patching and reset to CRLF
2194 - 'lf': EOLs are ignored when patching and reset to LF
2195 - None: get it from user settings, default to 'strict'
2196 'eolmode' is ignored when using an external patcher program.
2197
2198 Returns whether patch was applied with fuzz factor.
2199 """
2200 patcher = ui.config('ui', 'patch')
2201 if files is None:
2202 files = set()
2203 if patcher:
2204 return _externalpatch(ui, repo, patcher, patchname, strip,
2205 files, similarity)
2206 return internalpatch(ui, repo, patchname, strip, prefix, files, eolmode,
2207 similarity)
2208
2209 def changedfiles(ui, repo, patchpath, strip=1, prefix=''):
2210 backend = fsbackend(ui, repo.root)
2211 prefix = _canonprefix(repo, prefix)
2212 with open(patchpath, 'rb') as fp:
2213 changed = set()
2214 for state, values in iterhunks(fp):
2215 if state == 'file':
2216 afile, bfile, first_hunk, gp = values
2217 if gp:
2218 gp.path = pathtransform(gp.path, strip - 1, prefix)[1]
2219 if gp.oldpath:
2220 gp.oldpath = pathtransform(gp.oldpath, strip - 1,
2221 prefix)[1]
2222 else:
2223 gp = makepatchmeta(backend, afile, bfile, first_hunk, strip,
2224 prefix)
2225 changed.add(gp.path)
2226 if gp.op == 'RENAME':
2227 changed.add(gp.oldpath)
2228 elif state not in ('hunk', 'git'):
2229 raise error.Abort(_('unsupported parser state: %s') % state)
2230 return changed
2231
2232 class GitDiffRequired(Exception):
2233 pass
2234
18
2235 def diffallopts(ui, opts=None, untrusted=False, section='diff'):
19 def diffallopts(ui, opts=None, untrusted=False, section='diff'):
2236 '''return diffopts with all features supported and parsed'''
20 '''return diffopts with all features supported and parsed'''
@@ -2321,631 +105,3 b' def difffeatureopts(ui, opts=None, untru'
2321 buildopts['noprefix'] = get('noprefix', forceplain=False)
105 buildopts['noprefix'] = get('noprefix', forceplain=False)
2322
106
2323 return mdiff.diffopts(**pycompat.strkwargs(buildopts))
107 return mdiff.diffopts(**pycompat.strkwargs(buildopts))
2324
2325 def diff(repo, node1=None, node2=None, match=None, changes=None,
2326 opts=None, losedatafn=None, prefix='', relroot='', copy=None,
2327 hunksfilterfn=None):
2328 '''yields diff of changes to files between two nodes, or node and
2329 working directory.
2330
2331 if node1 is None, use first dirstate parent instead.
2332 if node2 is None, compare node1 with working directory.
2333
2334 losedatafn(**kwarg) is a callable run when opts.upgrade=True and
2335 every time some change cannot be represented with the current
2336 patch format. Return False to upgrade to git patch format, True to
2337 accept the loss or raise an exception to abort the diff. It is
2338 called with the name of current file being diffed as 'fn'. If set
2339 to None, patches will always be upgraded to git format when
2340 necessary.
2341
2342 prefix is a filename prefix that is prepended to all filenames on
2343 display (used for subrepos).
2344
2345 relroot, if not empty, must be normalized with a trailing /. Any match
2346 patterns that fall outside it will be ignored.
2347
2348 copy, if not empty, should contain mappings {dst@y: src@x} of copy
2349 information.
2350
2351 hunksfilterfn, if not None, should be a function taking a filectx and
2352 hunks generator that may yield filtered hunks.
2353 '''
2354 for fctx1, fctx2, hdr, hunks in diffhunks(
2355 repo, node1=node1, node2=node2,
2356 match=match, changes=changes, opts=opts,
2357 losedatafn=losedatafn, prefix=prefix, relroot=relroot, copy=copy,
2358 ):
2359 if hunksfilterfn is not None:
2360 # If the file has been removed, fctx2 is None; but this should
2361 # not occur here since we catch removed files early in
2362 # logcmdutil.getlinerangerevs() for 'hg log -L'.
2363 assert fctx2 is not None, \
2364 'fctx2 unexpectly None in diff hunks filtering'
2365 hunks = hunksfilterfn(fctx2, hunks)
2366 text = ''.join(sum((list(hlines) for hrange, hlines in hunks), []))
2367 if hdr and (text or len(hdr) > 1):
2368 yield '\n'.join(hdr) + '\n'
2369 if text:
2370 yield text
2371
2372 def diffhunks(repo, node1=None, node2=None, match=None, changes=None,
2373 opts=None, losedatafn=None, prefix='', relroot='', copy=None):
2374 """Yield diff of changes to files in the form of (`header`, `hunks`) tuples
2375 where `header` is a list of diff headers and `hunks` is an iterable of
2376 (`hunkrange`, `hunklines`) tuples.
2377
2378 See diff() for the meaning of parameters.
2379 """
2380
2381 if opts is None:
2382 opts = mdiff.defaultopts
2383
2384 if not node1 and not node2:
2385 node1 = repo.dirstate.p1()
2386
2387 def lrugetfilectx():
2388 cache = {}
2389 order = collections.deque()
2390 def getfilectx(f, ctx):
2391 fctx = ctx.filectx(f, filelog=cache.get(f))
2392 if f not in cache:
2393 if len(cache) > 20:
2394 del cache[order.popleft()]
2395 cache[f] = fctx.filelog()
2396 else:
2397 order.remove(f)
2398 order.append(f)
2399 return fctx
2400 return getfilectx
2401 getfilectx = lrugetfilectx()
2402
2403 ctx1 = repo[node1]
2404 ctx2 = repo[node2]
2405
2406 relfiltered = False
2407 if relroot != '' and match.always():
2408 # as a special case, create a new matcher with just the relroot
2409 pats = [relroot]
2410 match = scmutil.match(ctx2, pats, default='path')
2411 relfiltered = True
2412
2413 if not changes:
2414 changes = repo.status(ctx1, ctx2, match=match)
2415 modified, added, removed = changes[:3]
2416
2417 if not modified and not added and not removed:
2418 return []
2419
2420 if repo.ui.debugflag:
2421 hexfunc = hex
2422 else:
2423 hexfunc = short
2424 revs = [hexfunc(node) for node in [ctx1.node(), ctx2.node()] if node]
2425
2426 if copy is None:
2427 copy = {}
2428 if opts.git or opts.upgrade:
2429 copy = copies.pathcopies(ctx1, ctx2, match=match)
2430
2431 if relroot is not None:
2432 if not relfiltered:
2433 # XXX this would ideally be done in the matcher, but that is
2434 # generally meant to 'or' patterns, not 'and' them. In this case we
2435 # need to 'and' all the patterns from the matcher with relroot.
2436 def filterrel(l):
2437 return [f for f in l if f.startswith(relroot)]
2438 modified = filterrel(modified)
2439 added = filterrel(added)
2440 removed = filterrel(removed)
2441 relfiltered = True
2442 # filter out copies where either side isn't inside the relative root
2443 copy = dict(((dst, src) for (dst, src) in copy.iteritems()
2444 if dst.startswith(relroot)
2445 and src.startswith(relroot)))
2446
2447 modifiedset = set(modified)
2448 addedset = set(added)
2449 removedset = set(removed)
2450 for f in modified:
2451 if f not in ctx1:
2452 # Fix up added, since merged-in additions appear as
2453 # modifications during merges
2454 modifiedset.remove(f)
2455 addedset.add(f)
2456 for f in removed:
2457 if f not in ctx1:
2458 # Merged-in additions that are then removed are reported as removed.
2459 # They are not in ctx1, so We don't want to show them in the diff.
2460 removedset.remove(f)
2461 modified = sorted(modifiedset)
2462 added = sorted(addedset)
2463 removed = sorted(removedset)
2464 for dst, src in list(copy.items()):
2465 if src not in ctx1:
2466 # Files merged in during a merge and then copied/renamed are
2467 # reported as copies. We want to show them in the diff as additions.
2468 del copy[dst]
2469
2470 prefetchmatch = scmutil.matchfiles(
2471 repo, list(modifiedset | addedset | removedset))
2472 scmutil.prefetchfiles(repo, [ctx1.rev(), ctx2.rev()], prefetchmatch)
2473
2474 def difffn(opts, losedata):
2475 return trydiff(repo, revs, ctx1, ctx2, modified, added, removed,
2476 copy, getfilectx, opts, losedata, prefix, relroot)
2477 if opts.upgrade and not opts.git:
2478 try:
2479 def losedata(fn):
2480 if not losedatafn or not losedatafn(fn=fn):
2481 raise GitDiffRequired
2482 # Buffer the whole output until we are sure it can be generated
2483 return list(difffn(opts.copy(git=False), losedata))
2484 except GitDiffRequired:
2485 return difffn(opts.copy(git=True), None)
2486 else:
2487 return difffn(opts, None)
2488
2489 def diffsinglehunk(hunklines):
2490 """yield tokens for a list of lines in a single hunk"""
2491 for line in hunklines:
2492 # chomp
2493 chompline = line.rstrip('\n')
2494 # highlight tabs and trailing whitespace
2495 stripline = chompline.rstrip()
2496 if line.startswith('-'):
2497 label = 'diff.deleted'
2498 elif line.startswith('+'):
2499 label = 'diff.inserted'
2500 else:
2501 raise error.ProgrammingError('unexpected hunk line: %s' % line)
2502 for token in tabsplitter.findall(stripline):
2503 if token.startswith('\t'):
2504 yield (token, 'diff.tab')
2505 else:
2506 yield (token, label)
2507
2508 if chompline != stripline:
2509 yield (chompline[len(stripline):], 'diff.trailingwhitespace')
2510 if chompline != line:
2511 yield (line[len(chompline):], '')
2512
2513 def diffsinglehunkinline(hunklines):
2514 """yield tokens for a list of lines in a single hunk, with inline colors"""
2515 # prepare deleted, and inserted content
2516 a = ''
2517 b = ''
2518 for line in hunklines:
2519 if line[0] == '-':
2520 a += line[1:]
2521 elif line[0] == '+':
2522 b += line[1:]
2523 else:
2524 raise error.ProgrammingError('unexpected hunk line: %s' % line)
2525 # fast path: if either side is empty, use diffsinglehunk
2526 if not a or not b:
2527 for t in diffsinglehunk(hunklines):
2528 yield t
2529 return
2530 # re-split the content into words
2531 al = wordsplitter.findall(a)
2532 bl = wordsplitter.findall(b)
2533 # re-arrange the words to lines since the diff algorithm is line-based
2534 aln = [s if s == '\n' else s + '\n' for s in al]
2535 bln = [s if s == '\n' else s + '\n' for s in bl]
2536 an = ''.join(aln)
2537 bn = ''.join(bln)
2538 # run the diff algorithm, prepare atokens and btokens
2539 atokens = []
2540 btokens = []
2541 blocks = mdiff.allblocks(an, bn, lines1=aln, lines2=bln)
2542 for (a1, a2, b1, b2), btype in blocks:
2543 changed = btype == '!'
2544 for token in mdiff.splitnewlines(''.join(al[a1:a2])):
2545 atokens.append((changed, token))
2546 for token in mdiff.splitnewlines(''.join(bl[b1:b2])):
2547 btokens.append((changed, token))
2548
2549 # yield deleted tokens, then inserted ones
2550 for prefix, label, tokens in [('-', 'diff.deleted', atokens),
2551 ('+', 'diff.inserted', btokens)]:
2552 nextisnewline = True
2553 for changed, token in tokens:
2554 if nextisnewline:
2555 yield (prefix, label)
2556 nextisnewline = False
2557 # special handling line end
2558 isendofline = token.endswith('\n')
2559 if isendofline:
2560 chomp = token[:-1] # chomp
2561 token = chomp.rstrip() # detect spaces at the end
2562 endspaces = chomp[len(token):]
2563 # scan tabs
2564 for maybetab in tabsplitter.findall(token):
2565 if '\t' == maybetab[0]:
2566 currentlabel = 'diff.tab'
2567 else:
2568 if changed:
2569 currentlabel = label + '.changed'
2570 else:
2571 currentlabel = label + '.unchanged'
2572 yield (maybetab, currentlabel)
2573 if isendofline:
2574 if endspaces:
2575 yield (endspaces, 'diff.trailingwhitespace')
2576 yield ('\n', '')
2577 nextisnewline = True
2578
2579 def difflabel(func, *args, **kw):
2580 '''yields 2-tuples of (output, label) based on the output of func()'''
2581 if kw.get(r'opts') and kw[r'opts'].worddiff:
2582 dodiffhunk = diffsinglehunkinline
2583 else:
2584 dodiffhunk = diffsinglehunk
2585 headprefixes = [('diff', 'diff.diffline'),
2586 ('copy', 'diff.extended'),
2587 ('rename', 'diff.extended'),
2588 ('old', 'diff.extended'),
2589 ('new', 'diff.extended'),
2590 ('deleted', 'diff.extended'),
2591 ('index', 'diff.extended'),
2592 ('similarity', 'diff.extended'),
2593 ('---', 'diff.file_a'),
2594 ('+++', 'diff.file_b')]
2595 textprefixes = [('@', 'diff.hunk'),
2596 # - and + are handled by diffsinglehunk
2597 ]
2598 head = False
2599
2600 # buffers a hunk, i.e. adjacent "-", "+" lines without other changes.
2601 hunkbuffer = []
2602 def consumehunkbuffer():
2603 if hunkbuffer:
2604 for token in dodiffhunk(hunkbuffer):
2605 yield token
2606 hunkbuffer[:] = []
2607
2608 for chunk in func(*args, **kw):
2609 lines = chunk.split('\n')
2610 linecount = len(lines)
2611 for i, line in enumerate(lines):
2612 if head:
2613 if line.startswith('@'):
2614 head = False
2615 else:
2616 if line and not line.startswith((' ', '+', '-', '@', '\\')):
2617 head = True
2618 diffline = False
2619 if not head and line and line.startswith(('+', '-')):
2620 diffline = True
2621
2622 prefixes = textprefixes
2623 if head:
2624 prefixes = headprefixes
2625 if diffline:
2626 # buffered
2627 bufferedline = line
2628 if i + 1 < linecount:
2629 bufferedline += "\n"
2630 hunkbuffer.append(bufferedline)
2631 else:
2632 # unbuffered
2633 for token in consumehunkbuffer():
2634 yield token
2635 stripline = line.rstrip()
2636 for prefix, label in prefixes:
2637 if stripline.startswith(prefix):
2638 yield (stripline, label)
2639 if line != stripline:
2640 yield (line[len(stripline):],
2641 'diff.trailingwhitespace')
2642 break
2643 else:
2644 yield (line, '')
2645 if i + 1 < linecount:
2646 yield ('\n', '')
2647 for token in consumehunkbuffer():
2648 yield token
2649
2650 def diffui(*args, **kw):
2651 '''like diff(), but yields 2-tuples of (output, label) for ui.write()'''
2652 return difflabel(diff, *args, **kw)
2653
2654 def _filepairs(modified, added, removed, copy, opts):
2655 '''generates tuples (f1, f2, copyop), where f1 is the name of the file
2656 before and f2 is the the name after. For added files, f1 will be None,
2657 and for removed files, f2 will be None. copyop may be set to None, 'copy'
2658 or 'rename' (the latter two only if opts.git is set).'''
2659 gone = set()
2660
2661 copyto = dict([(v, k) for k, v in copy.items()])
2662
2663 addedset, removedset = set(added), set(removed)
2664
2665 for f in sorted(modified + added + removed):
2666 copyop = None
2667 f1, f2 = f, f
2668 if f in addedset:
2669 f1 = None
2670 if f in copy:
2671 if opts.git:
2672 f1 = copy[f]
2673 if f1 in removedset and f1 not in gone:
2674 copyop = 'rename'
2675 gone.add(f1)
2676 else:
2677 copyop = 'copy'
2678 elif f in removedset:
2679 f2 = None
2680 if opts.git:
2681 # have we already reported a copy above?
2682 if (f in copyto and copyto[f] in addedset
2683 and copy[copyto[f]] == f):
2684 continue
2685 yield f1, f2, copyop
2686
2687 def trydiff(repo, revs, ctx1, ctx2, modified, added, removed,
2688 copy, getfilectx, opts, losedatafn, prefix, relroot):
2689 '''given input data, generate a diff and yield it in blocks
2690
2691 If generating a diff would lose data like flags or binary data and
2692 losedatafn is not None, it will be called.
2693
2694 relroot is removed and prefix is added to every path in the diff output.
2695
2696 If relroot is not empty, this function expects every path in modified,
2697 added, removed and copy to start with it.'''
2698
2699 def gitindex(text):
2700 if not text:
2701 text = ""
2702 l = len(text)
2703 s = hashlib.sha1('blob %d\0' % l)
2704 s.update(text)
2705 return hex(s.digest())
2706
2707 if opts.noprefix:
2708 aprefix = bprefix = ''
2709 else:
2710 aprefix = 'a/'
2711 bprefix = 'b/'
2712
2713 def diffline(f, revs):
2714 revinfo = ' '.join(["-r %s" % rev for rev in revs])
2715 return 'diff %s %s' % (revinfo, f)
2716
2717 def isempty(fctx):
2718 return fctx is None or fctx.size() == 0
2719
2720 date1 = dateutil.datestr(ctx1.date())
2721 date2 = dateutil.datestr(ctx2.date())
2722
2723 gitmode = {'l': '120000', 'x': '100755', '': '100644'}
2724
2725 if relroot != '' and (repo.ui.configbool('devel', 'all-warnings')
2726 or repo.ui.configbool('devel', 'check-relroot')):
2727 for f in modified + added + removed + list(copy) + list(copy.values()):
2728 if f is not None and not f.startswith(relroot):
2729 raise AssertionError(
2730 "file %s doesn't start with relroot %s" % (f, relroot))
2731
2732 for f1, f2, copyop in _filepairs(modified, added, removed, copy, opts):
2733 content1 = None
2734 content2 = None
2735 fctx1 = None
2736 fctx2 = None
2737 flag1 = None
2738 flag2 = None
2739 if f1:
2740 fctx1 = getfilectx(f1, ctx1)
2741 if opts.git or losedatafn:
2742 flag1 = ctx1.flags(f1)
2743 if f2:
2744 fctx2 = getfilectx(f2, ctx2)
2745 if opts.git or losedatafn:
2746 flag2 = ctx2.flags(f2)
2747 # if binary is True, output "summary" or "base85", but not "text diff"
2748 if opts.text:
2749 binary = False
2750 else:
2751 binary = any(f.isbinary() for f in [fctx1, fctx2] if f is not None)
2752
2753 if losedatafn and not opts.git:
2754 if (binary or
2755 # copy/rename
2756 f2 in copy or
2757 # empty file creation
2758 (not f1 and isempty(fctx2)) or
2759 # empty file deletion
2760 (isempty(fctx1) and not f2) or
2761 # create with flags
2762 (not f1 and flag2) or
2763 # change flags
2764 (f1 and f2 and flag1 != flag2)):
2765 losedatafn(f2 or f1)
2766
2767 path1 = f1 or f2
2768 path2 = f2 or f1
2769 path1 = posixpath.join(prefix, path1[len(relroot):])
2770 path2 = posixpath.join(prefix, path2[len(relroot):])
2771 header = []
2772 if opts.git:
2773 header.append('diff --git %s%s %s%s' %
2774 (aprefix, path1, bprefix, path2))
2775 if not f1: # added
2776 header.append('new file mode %s' % gitmode[flag2])
2777 elif not f2: # removed
2778 header.append('deleted file mode %s' % gitmode[flag1])
2779 else: # modified/copied/renamed
2780 mode1, mode2 = gitmode[flag1], gitmode[flag2]
2781 if mode1 != mode2:
2782 header.append('old mode %s' % mode1)
2783 header.append('new mode %s' % mode2)
2784 if copyop is not None:
2785 if opts.showsimilarity:
2786 sim = similar.score(ctx1[path1], ctx2[path2]) * 100
2787 header.append('similarity index %d%%' % sim)
2788 header.append('%s from %s' % (copyop, path1))
2789 header.append('%s to %s' % (copyop, path2))
2790 elif revs and not repo.ui.quiet:
2791 header.append(diffline(path1, revs))
2792
2793 # fctx.is | diffopts | what to | is fctx.data()
2794 # binary() | text nobinary git index | output? | outputted?
2795 # ------------------------------------|----------------------------
2796 # yes | no no no * | summary | no
2797 # yes | no no yes * | base85 | yes
2798 # yes | no yes no * | summary | no
2799 # yes | no yes yes 0 | summary | no
2800 # yes | no yes yes >0 | summary | semi [1]
2801 # yes | yes * * * | text diff | yes
2802 # no | * * * * | text diff | yes
2803 # [1]: hash(fctx.data()) is outputted. so fctx.data() cannot be faked
2804 if binary and (not opts.git or (opts.git and opts.nobinary and not
2805 opts.index)):
2806 # fast path: no binary content will be displayed, content1 and
2807 # content2 are only used for equivalent test. cmp() could have a
2808 # fast path.
2809 if fctx1 is not None:
2810 content1 = b'\0'
2811 if fctx2 is not None:
2812 if fctx1 is not None and not fctx1.cmp(fctx2):
2813 content2 = b'\0' # not different
2814 else:
2815 content2 = b'\0\0'
2816 else:
2817 # normal path: load contents
2818 if fctx1 is not None:
2819 content1 = fctx1.data()
2820 if fctx2 is not None:
2821 content2 = fctx2.data()
2822
2823 if binary and opts.git and not opts.nobinary:
2824 text = mdiff.b85diff(content1, content2)
2825 if text:
2826 header.append('index %s..%s' %
2827 (gitindex(content1), gitindex(content2)))
2828 hunks = (None, [text]),
2829 else:
2830 if opts.git and opts.index > 0:
2831 flag = flag1
2832 if flag is None:
2833 flag = flag2
2834 header.append('index %s..%s %s' %
2835 (gitindex(content1)[0:opts.index],
2836 gitindex(content2)[0:opts.index],
2837 gitmode[flag]))
2838
2839 uheaders, hunks = mdiff.unidiff(content1, date1,
2840 content2, date2,
2841 path1, path2,
2842 binary=binary, opts=opts)
2843 header.extend(uheaders)
2844 yield fctx1, fctx2, header, hunks
2845
2846 def diffstatsum(stats):
2847 maxfile, maxtotal, addtotal, removetotal, binary = 0, 0, 0, 0, False
2848 for f, a, r, b in stats:
2849 maxfile = max(maxfile, encoding.colwidth(f))
2850 maxtotal = max(maxtotal, a + r)
2851 addtotal += a
2852 removetotal += r
2853 binary = binary or b
2854
2855 return maxfile, maxtotal, addtotal, removetotal, binary
2856
2857 def diffstatdata(lines):
2858 diffre = re.compile('^diff .*-r [a-z0-9]+\s(.*)$')
2859
2860 results = []
2861 filename, adds, removes, isbinary = None, 0, 0, False
2862
2863 def addresult():
2864 if filename:
2865 results.append((filename, adds, removes, isbinary))
2866
2867 # inheader is used to track if a line is in the
2868 # header portion of the diff. This helps properly account
2869 # for lines that start with '--' or '++'
2870 inheader = False
2871
2872 for line in lines:
2873 if line.startswith('diff'):
2874 addresult()
2875 # starting a new file diff
2876 # set numbers to 0 and reset inheader
2877 inheader = True
2878 adds, removes, isbinary = 0, 0, False
2879 if line.startswith('diff --git a/'):
2880 filename = gitre.search(line).group(2)
2881 elif line.startswith('diff -r'):
2882 # format: "diff -r ... -r ... filename"
2883 filename = diffre.search(line).group(1)
2884 elif line.startswith('@@'):
2885 inheader = False
2886 elif line.startswith('+') and not inheader:
2887 adds += 1
2888 elif line.startswith('-') and not inheader:
2889 removes += 1
2890 elif (line.startswith('GIT binary patch') or
2891 line.startswith('Binary file')):
2892 isbinary = True
2893 addresult()
2894 return results
2895
2896 def diffstat(lines, width=80):
2897 output = []
2898 stats = diffstatdata(lines)
2899 maxname, maxtotal, totaladds, totalremoves, hasbinary = diffstatsum(stats)
2900
2901 countwidth = len(str(maxtotal))
2902 if hasbinary and countwidth < 3:
2903 countwidth = 3
2904 graphwidth = width - countwidth - maxname - 6
2905 if graphwidth < 10:
2906 graphwidth = 10
2907
2908 def scale(i):
2909 if maxtotal <= graphwidth:
2910 return i
2911 # If diffstat runs out of room it doesn't print anything,
2912 # which isn't very useful, so always print at least one + or -
2913 # if there were at least some changes.
2914 return max(i * graphwidth // maxtotal, int(bool(i)))
2915
2916 for filename, adds, removes, isbinary in stats:
2917 if isbinary:
2918 count = 'Bin'
2919 else:
2920 count = '%d' % (adds + removes)
2921 pluses = '+' * scale(adds)
2922 minuses = '-' * scale(removes)
2923 output.append(' %s%s | %*s %s%s\n' %
2924 (filename, ' ' * (maxname - encoding.colwidth(filename)),
2925 countwidth, count, pluses, minuses))
2926
2927 if stats:
2928 output.append(_(' %d files changed, %d insertions(+), '
2929 '%d deletions(-)\n')
2930 % (len(stats), totaladds, totalremoves))
2931
2932 return ''.join(output)
2933
2934 def diffstatui(*args, **kw):
2935 '''like diffstat(), but yields 2-tuples of (output, label) for
2936 ui.write()
2937 '''
2938
2939 for line in diffstat(*args, **kw).splitlines():
2940 if line and line[-1] in '+-':
2941 name, graph = line.rsplit(' ', 1)
2942 yield (name + ' ', '')
2943 m = re.search(br'\++', graph)
2944 if m:
2945 yield (m.group(0), 'diffstat.inserted')
2946 m = re.search(br'-+', graph)
2947 if m:
2948 yield (m.group(0), 'diffstat.deleted')
2949 else:
2950 yield (line, '')
2951 yield ('\n', '')
General Comments 0
You need to be logged in to leave comments. Login now