3 import sys, math, random, os, re, signal, tempfile, stat, errno, traceback
4 from heapq import heappush, heappop
7 sys.path.append('''@@GIT_PYTHON_PATH@@''')
8 from gitMergeCommon import *
10 originalIndexFile = os.environ.get('GIT_INDEX_FILE',
11 os.environ.get('GIT_DIR', '.git') + '/index')
12 temporaryIndexFile = os.environ.get('GIT_DIR', '.git') + \
13 '/merge-recursive-tmp-index'
14 def setupIndex(temporary):
16 os.unlink(temporaryIndexFile)
20 newIndex = temporaryIndexFile
22 newIndex = originalIndexFile
23 os.environ['GIT_INDEX_FILE'] = newIndex
25 # This is a global variable which is used in a number of places but
26 # only written to in the 'merge' function.
28 # cacheOnly == True => Don't leave any non-stage 0 entries in the cache and
29 # don't update the working directory.
30 # False => Leave unmerged entries in the cache and update
31 # the working directory.
35 # The entry point to the merge code
36 # ---------------------------------
38 def merge(h1, h2, branch1Name, branch2Name, graph, callDepth=0):
39 '''Merge the commits h1 and h2, return the resulting virtual
40 commit object and a flag indicating the cleaness of the merge.'''
41 assert(isinstance(h1, Commit) and isinstance(h2, Commit))
42 assert(isinstance(graph, Graph))
45 sys.stdout.write(' '*callDepth)
53 ca = getCommonAncestors(graph, h1, h2)
54 infoMsg('found', len(ca), 'common ancestor(s):')
61 [mergedCA, dummy] = merge(mergedCA, h,
62 'Temporary shared merge branch 1',
63 'Temporary shared merge branch 2',
65 assert(isinstance(mergedCA, Commit))
73 runProgram(['git-read-tree', h1.tree()])
76 [shaRes, clean] = mergeTrees(h1.tree(), h2.tree(), mergedCA.tree(),
77 branch1Name, branch2Name)
79 if clean or cacheOnly:
80 res = Commit(None, [h1, h2], tree=shaRes)
87 getFilesRE = re.compile(r'^([0-7]+) (\S+) ([0-9a-f]{40})\t(.*)$', re.S)
88 def getFilesAndDirs(tree):
91 out = runProgram(['git-ls-tree', '-r', '-z', tree])
92 for l in out.split('\0'):
93 m = getFilesRE.match(l)
95 if m.group(2) == 'tree':
97 elif m.group(2) == 'blob':
102 # Those two global variables are used in a number of places but only
103 # written to in 'mergeTrees' and 'uniquePath'. They keep track of
104 # every file and directory in the two branches that are about to be
106 currentFileSet = None
107 currentDirectorySet = None
109 def mergeTrees(head, merge, common, branch1Name, branch2Name):
110 '''Merge the trees 'head' and 'merge' with the common ancestor
111 'common'. The name of the head branch is 'branch1Name' and the name of
112 the merge branch is 'branch2Name'. Return a tuple (tree, cleanMerge)
113 where tree is the resulting tree and cleanMerge is True iff the
116 assert(isSha(head) and isSha(merge) and isSha(common))
119 print 'Already uptodate!'
127 [out, code] = runProgram(['git-read-tree', updateArg, '-m',
128 common, head, merge], returnCode = True)
130 die('git-read-tree:', out)
132 [tree, code] = runProgram('git-write-tree', returnCode=True)
135 global currentFileSet, currentDirectorySet
136 [currentFileSet, currentDirectorySet] = getFilesAndDirs(head)
137 [filesM, dirsM] = getFilesAndDirs(merge)
138 currentFileSet.union_update(filesM)
139 currentDirectorySet.union_update(dirsM)
141 entries = unmergedCacheEntries()
142 renamesHead = getRenames(head, common, head, merge, entries)
143 renamesMerge = getRenames(merge, common, head, merge, entries)
145 cleanMerge = processRenames(renamesHead, renamesMerge,
146 branch1Name, branch2Name)
147 for entry in entries:
150 if not processEntry(entry, branch1Name, branch2Name):
153 if cleanMerge or cacheOnly:
154 tree = runProgram('git-write-tree').rstrip()
160 return [tree, cleanMerge]
162 # Low level file merging, update and removal
163 # ------------------------------------------
165 def mergeFile(oPath, oSha, oMode, aPath, aSha, aMode, bPath, bSha, bMode,
166 branch1Name, branch2Name):
171 if stat.S_IFMT(aMode) != stat.S_IFMT(bMode):
173 if stat.S_ISREG(aMode):
180 if aSha != oSha and bSha != oSha:
192 elif stat.S_ISREG(aMode):
193 assert(stat.S_ISREG(bMode))
195 orig = runProgram(['git-unpack-file', oSha]).rstrip()
196 src1 = runProgram(['git-unpack-file', aSha]).rstrip()
197 src2 = runProgram(['git-unpack-file', bSha]).rstrip()
198 [out, code] = runProgram(['merge',
199 '-L', branch1Name + '/' + aPath,
200 '-L', 'orig/' + oPath,
201 '-L', branch2Name + '/' + bPath,
202 src1, orig, src2], returnCode=True)
204 sha = runProgram(['git-hash-object', '-t', 'blob', '-w',
213 assert(stat.S_ISLNK(aMode) and stat.S_ISLNK(bMode))
219 return [sha, mode, clean, merge]
221 def updateFile(clean, sha, mode, path):
222 updateCache = cacheOnly or clean
223 updateWd = not cacheOnly
225 return updateFileExt(sha, mode, path, updateCache, updateWd)
227 def updateFileExt(sha, mode, path, updateCache, updateWd):
232 pathComponents = path.split('/')
233 for x in xrange(1, len(pathComponents)):
234 p = '/'.join(pathComponents[0:x])
237 createDir = not stat.S_ISDIR(os.lstat(p).st_mode)
245 die("Couldn't create directory", p, e.strerror)
247 prog = ['git-cat-file', 'blob', sha]
248 if stat.S_ISREG(mode):
257 fd = os.open(path, os.O_WRONLY | os.O_TRUNC | os.O_CREAT, mode)
258 proc = subprocess.Popen(prog, stdout=fd)
261 elif stat.S_ISLNK(mode):
262 linkTarget = runProgram(prog)
263 os.symlink(linkTarget, path)
267 if updateWd and updateCache:
268 runProgram(['git-update-index', '--add', '--', path])
270 runProgram(['git-update-index', '--add', '--cacheinfo',
271 '0%o' % mode, sha, path])
273 def removeFile(clean, path):
274 updateCache = cacheOnly or clean
275 updateWd = not cacheOnly
278 runProgram(['git-update-index', '--force-remove', '--', path])
284 if e.errno != errno.ENOENT and e.errno != errno.EISDIR:
287 def uniquePath(path, branch):
288 def fileExists(path):
293 if e.errno == errno.ENOENT:
298 newPath = path + '_' + branch
300 while newPath in currentFileSet or \
301 newPath in currentDirectorySet or \
304 newPath = path + '_' + branch + '_' + str(suffix)
305 currentFileSet.add(newPath)
308 # Cache entry management
309 # ----------------------
312 def __init__(self, path):
318 # Used for debugging only
320 if self.mode != None:
321 m = '0%o' % self.mode
329 return 'sha1: ' + sha1 + ' mode: ' + m
331 self.stages = [Stage(), Stage(), Stage(), Stage()]
333 self.processed = False
336 return 'path: ' + self.path + ' stages: ' + repr([str(x) for x in self.stages])
338 class CacheEntryContainer:
342 def add(self, entry):
343 self.entries[entry.path] = entry
346 return self.entries.get(path)
349 return self.entries.itervalues()
351 unmergedRE = re.compile(r'^([0-7]+) ([0-9a-f]{40}) ([1-3])\t(.*)$', re.S)
352 def unmergedCacheEntries():
353 '''Create a dictionary mapping file names to CacheEntry
354 objects. The dictionary contains one entry for every path with a
355 non-zero stage entry.'''
357 lines = runProgram(['git-ls-files', '-z', '--unmerged']).split('\0')
360 res = CacheEntryContainer()
362 m = unmergedRE.match(l)
364 mode = int(m.group(1), 8)
366 stage = int(m.group(3))
374 e.stages[stage].mode = mode
375 e.stages[stage].sha1 = sha1
377 die('Error: Merge program failed: Unexpected output from',
381 lsTreeRE = re.compile(r'^([0-7]+) (\S+) ([0-9a-f]{40})\t(.*)\n$', re.S)
382 def getCacheEntry(path, origTree, aTree, bTree):
383 '''Returns a CacheEntry object which doesn't have to correspond to
384 a real cache entry in Git's index.'''
390 m = lsTreeRE.match(out)
392 die('Unexpected output from git-ls-tree:', out)
393 elif m.group(2) == 'blob':
394 return [m.group(3), int(m.group(1), 8)]
398 res = CacheEntry(path)
400 [oSha, oMode] = parse(runProgram(['git-ls-tree', origTree, '--', path]))
401 [aSha, aMode] = parse(runProgram(['git-ls-tree', aTree, '--', path]))
402 [bSha, bMode] = parse(runProgram(['git-ls-tree', bTree, '--', path]))
404 res.stages[1].sha1 = oSha
405 res.stages[1].mode = oMode
406 res.stages[2].sha1 = aSha
407 res.stages[2].mode = aMode
408 res.stages[3].sha1 = bSha
409 res.stages[3].mode = bMode
413 # Rename detection and handling
414 # -----------------------------
418 src, srcSha, srcMode, srcCacheEntry,
419 dst, dstSha, dstMode, dstCacheEntry,
423 self.srcMode = srcMode
424 self.srcCacheEntry = srcCacheEntry
427 self.dstMode = dstMode
428 self.dstCacheEntry = dstCacheEntry
431 self.processed = False
433 class RenameEntryContainer:
438 def add(self, entry):
439 self.entriesSrc[entry.srcName] = entry
440 self.entriesDst[entry.dstName] = entry
442 def getSrc(self, path):
443 return self.entriesSrc.get(path)
445 def getDst(self, path):
446 return self.entriesDst.get(path)
449 return self.entriesSrc.itervalues()
451 parseDiffRenamesRE = re.compile('^:([0-7]+) ([0-7]+) ([0-9a-f]{40}) ([0-9a-f]{40}) R([0-9]*)$')
452 def getRenames(tree, oTree, aTree, bTree, cacheEntries):
453 '''Get information of all renames which occured between 'oTree' and
454 'tree'. We need the three trees in the merge ('oTree', 'aTree' and
455 'bTree') to be able to associate the correct cache entries with
456 the rename information. 'tree' is always equal to either aTree or bTree.'''
458 assert(tree == aTree or tree == bTree)
459 inp = runProgram(['git-diff-tree', '-M', '--diff-filter=R', '-r',
462 ret = RenameEntryContainer()
464 recs = inp.split("\0")
465 recs.pop() # remove last entry (which is '')
469 m = parseDiffRenamesRE.match(rec)
472 die('Unexpected output from git-diff-tree:', rec)
474 srcMode = int(m.group(1), 8)
475 dstMode = int(m.group(2), 8)
482 srcCacheEntry = cacheEntries.get(src)
483 if not srcCacheEntry:
484 srcCacheEntry = getCacheEntry(src, oTree, aTree, bTree)
485 cacheEntries.add(srcCacheEntry)
487 dstCacheEntry = cacheEntries.get(dst)
488 if not dstCacheEntry:
489 dstCacheEntry = getCacheEntry(dst, oTree, aTree, bTree)
490 cacheEntries.add(dstCacheEntry)
492 ret.add(RenameEntry(src, srcSha, srcMode, srcCacheEntry,
493 dst, dstSha, dstMode, dstCacheEntry,
495 except StopIteration:
499 def fmtRename(src, dst):
500 srcPath = src.split('/')
501 dstPath = dst.split('/')
503 endIndex = min(len(srcPath), len(dstPath)) - 1
504 for x in range(0, endIndex):
505 if srcPath[x] == dstPath[x]:
506 path.append(srcPath[x])
512 return '/'.join(path) + \
513 '/{' + '/'.join(srcPath[endIndex:]) + ' => ' + \
514 '/'.join(dstPath[endIndex:]) + '}'
516 return src + ' => ' + dst
518 def processRenames(renamesA, renamesB, branchNameA, branchNameB):
521 srcNames.add(x.srcName)
523 srcNames.add(x.srcName)
526 for path in srcNames:
527 if renamesA.getSrc(path):
530 branchName1 = branchNameA
531 branchName2 = branchNameB
535 branchName1 = branchNameB
536 branchName2 = branchNameA
538 ren1 = renames1.getSrc(path)
539 ren2 = renames2.getSrc(path)
541 ren1.dstCacheEntry.processed = True
542 ren1.srcCacheEntry.processed = True
547 ren1.processed = True
548 removeFile(True, ren1.srcName)
550 # Renamed in 1 and renamed in 2
551 assert(ren1.srcName == ren2.srcName)
552 ren2.dstCacheEntry.processed = True
553 ren2.processed = True
555 if ren1.dstName != ren2.dstName:
556 print 'CONFLICT (rename/rename): Rename', \
557 fmtRename(path, ren1.dstName), 'in branch', branchName1, \
558 'rename', fmtRename(path, ren2.dstName), 'in', branchName2
561 if ren1.dstName in currentDirectorySet:
562 dstName1 = uniquePath(ren1.dstName, branchName1)
563 print ren1.dstName, 'is a directory in', branchName2, \
564 'adding as', dstName1, 'instead.'
565 removeFile(False, ren1.dstName)
567 dstName1 = ren1.dstName
569 if ren2.dstName in currentDirectorySet:
570 dstName2 = uniquePath(ren2.dstName, branchName2)
571 print ren2.dstName, 'is a directory in', branchName1, \
572 'adding as', dstName2, 'instead.'
573 removeFile(False, ren2.dstName)
575 dstName2 = ren1.dstName
577 updateFile(False, ren1.dstSha, ren1.dstMode, dstName1)
578 updateFile(False, ren2.dstSha, ren2.dstMode, dstName2)
580 print 'Renaming', fmtRename(path, ren1.dstName)
581 [resSha, resMode, clean, merge] = \
582 mergeFile(ren1.srcName, ren1.srcSha, ren1.srcMode,
583 ren1.dstName, ren1.dstSha, ren1.dstMode,
584 ren2.dstName, ren2.dstSha, ren2.dstMode,
585 branchName1, branchName2)
588 print 'Auto-merging', ren1.dstName
591 print 'CONFLICT (content): merge conflict in', ren1.dstName
595 updateFileExt(ren1.dstSha, ren1.dstMode, ren1.dstName,
596 updateCache=True, updateWd=False)
597 updateFile(clean, resSha, resMode, ren1.dstName)
599 # Renamed in 1, maybe changed in 2
600 if renamesA == renames1:
605 srcShaOtherBranch = ren1.srcCacheEntry.stages[stage].sha1
606 srcModeOtherBranch = ren1.srcCacheEntry.stages[stage].mode
608 dstShaOtherBranch = ren1.dstCacheEntry.stages[stage].sha1
609 dstModeOtherBranch = ren1.dstCacheEntry.stages[stage].mode
613 if ren1.dstName in currentDirectorySet:
614 newPath = uniquePath(ren1.dstName, branchName1)
615 print 'CONFLICT (rename/directory): Rename', \
616 fmtRename(ren1.srcName, ren1.dstName), 'in', branchName1,\
617 'directory', ren1.dstName, 'added in', branchName2
618 print 'Renaming', ren1.srcName, 'to', newPath, 'instead'
620 removeFile(False, ren1.dstName)
621 updateFile(False, ren1.dstSha, ren1.dstMode, newPath)
622 elif srcShaOtherBranch == None:
623 print 'CONFLICT (rename/delete): Rename', \
624 fmtRename(ren1.srcName, ren1.dstName), 'in', \
625 branchName1, 'and deleted in', branchName2
627 updateFile(False, ren1.dstSha, ren1.dstMode, ren1.dstName)
628 elif dstShaOtherBranch:
629 newPath = uniquePath(ren1.dstName, branchName2)
630 print 'CONFLICT (rename/add): Rename', \
631 fmtRename(ren1.srcName, ren1.dstName), 'in', \
632 branchName1 + '.', ren1.dstName, 'added in', branchName2
633 print 'Adding as', newPath, 'instead'
634 updateFile(False, dstShaOtherBranch, dstModeOtherBranch, newPath)
637 elif renames2.getDst(ren1.dstName):
638 dst2 = renames2.getDst(ren1.dstName)
639 newPath1 = uniquePath(ren1.dstName, branchName1)
640 newPath2 = uniquePath(dst2.dstName, branchName2)
641 print 'CONFLICT (rename/rename): Rename', \
642 fmtRename(ren1.srcName, ren1.dstName), 'in', \
643 branchName1+'. Rename', \
644 fmtRename(dst2.srcName, dst2.dstName), 'in', branchName2
645 print 'Renaming', ren1.srcName, 'to', newPath1, 'and', \
646 dst2.srcName, 'to', newPath2, 'instead'
647 removeFile(False, ren1.dstName)
648 updateFile(False, ren1.dstSha, ren1.dstMode, newPath1)
649 updateFile(False, dst2.dstSha, dst2.dstMode, newPath2)
650 dst2.processed = True
656 print 'Renaming', fmtRename(ren1.srcName, ren1.dstName)
657 [resSha, resMode, clean, merge] = \
658 mergeFile(ren1.srcName, ren1.srcSha, ren1.srcMode,
659 ren1.dstName, ren1.dstSha, ren1.dstMode,
660 ren1.srcName, srcShaOtherBranch, srcModeOtherBranch,
661 branchName1, branchName2)
664 print 'Auto-merging', ren1.dstName
667 print 'CONFLICT (rename/modify): Merge conflict in', ren1.dstName
671 updateFileExt(ren1.dstSha, ren1.dstMode, ren1.dstName,
672 updateCache=True, updateWd=False)
673 updateFile(clean, resSha, resMode, ren1.dstName)
677 # Per entry merge function
678 # ------------------------
680 def processEntry(entry, branch1Name, branch2Name):
681 '''Merge one cache entry.'''
683 debug('processing', entry.path, 'clean cache:', cacheOnly)
688 oSha = entry.stages[1].sha1
689 oMode = entry.stages[1].mode
690 aSha = entry.stages[2].sha1
691 aMode = entry.stages[2].mode
692 bSha = entry.stages[3].sha1
693 bMode = entry.stages[3].mode
695 assert(oSha == None or isSha(oSha))
696 assert(aSha == None or isSha(aSha))
697 assert(bSha == None or isSha(bSha))
699 assert(oMode == None or type(oMode) is int)
700 assert(aMode == None or type(aMode) is int)
701 assert(bMode == None or type(bMode) is int)
703 if (oSha and (not aSha or not bSha)):
705 # Case A: Deleted in one
707 if (not aSha and not bSha) or \
708 (aSha == oSha and not bSha) or \
709 (not aSha and bSha == oSha):
710 # Deleted in both or deleted in one and unchanged in the other
712 print 'Removing', path
713 removeFile(True, path)
715 # Deleted in one and changed in the other
718 print 'CONFLICT (delete/modify):', path, 'deleted in', \
719 branch1Name, 'and modified in', branch2Name + '.', \
720 'Version', branch2Name, 'of', path, 'left in tree.'
724 print 'CONFLICT (modify/delete):', path, 'deleted in', \
725 branch2Name, 'and modified in', branch1Name + '.', \
726 'Version', branch1Name, 'of', path, 'left in tree.'
730 updateFile(False, sha, mode, path)
732 elif (not oSha and aSha and not bSha) or \
733 (not oSha and not aSha and bSha):
735 # Case B: Added in one.
738 addBranch = branch1Name
739 otherBranch = branch2Name
742 conf = 'file/directory'
744 addBranch = branch2Name
745 otherBranch = branch1Name
748 conf = 'directory/file'
750 if path in currentDirectorySet:
752 newPath = uniquePath(path, addBranch)
753 print 'CONFLICT (' + conf + '):', \
754 'There is a directory with name', path, 'in', \
755 otherBranch + '. Adding', path, 'as', newPath
757 removeFile(False, path)
758 updateFile(False, sha, mode, newPath)
761 updateFile(True, sha, mode, path)
763 elif not oSha and aSha and bSha:
765 # Case C: Added in both (check for same permissions).
770 print 'CONFLICT: File', path, \
771 'added identically in both branches, but permissions', \
772 'conflict', '0%o' % aMode, '->', '0%o' % bMode
773 print 'CONFLICT: adding with permission:', '0%o' % aMode
775 updateFile(False, aSha, aMode, path)
777 # This case is handled by git-read-tree
781 newPath1 = uniquePath(path, branch1Name)
782 newPath2 = uniquePath(path, branch2Name)
783 print 'CONFLICT (add/add): File', path, \
784 'added non-identically in both branches. Adding as', \
785 newPath1, 'and', newPath2, 'instead.'
786 removeFile(False, path)
787 updateFile(False, aSha, aMode, newPath1)
788 updateFile(False, bSha, bMode, newPath2)
790 elif oSha and aSha and bSha:
792 # case D: Modified in both, but differently.
794 print 'Auto-merging', path
795 [sha, mode, clean, dummy] = \
796 mergeFile(path, oSha, oMode,
799 branch1Name, branch2Name)
801 updateFile(True, sha, mode, path)
804 print 'CONFLICT (content): Merge conflict in', path
807 updateFile(False, sha, mode, path)
809 updateFileExt(aSha, aMode, path,
810 updateCache=True, updateWd=False)
811 updateFileExt(sha, mode, path, updateCache=False, updateWd=True)
813 die("ERROR: Fatal merge failure, shouldn't happen.")
818 die('Usage:', sys.argv[0], ' <base>... -- <head> <remote>..')
820 # main entry point as merge strategy module
821 # The first parameters up to -- are merge bases, and the rest are heads.
822 # This strategy module figures out merge bases itself, so we only
825 if len(sys.argv) < 4:
828 for nextArg in xrange(1, len(sys.argv)):
829 if sys.argv[nextArg] == '--':
830 if len(sys.argv) != nextArg + 3:
831 die('Not handling anything other than two heads merge.')
833 h1 = firstBranch = sys.argv[nextArg + 1]
834 h2 = secondBranch = sys.argv[nextArg + 2]
839 print 'Merging', h1, 'with', h2
842 h1 = runProgram(['git-rev-parse', '--verify', h1 + '^0']).rstrip()
843 h2 = runProgram(['git-rev-parse', '--verify', h2 + '^0']).rstrip()
845 graph = buildGraph([h1, h2])
847 [dummy, clean] = merge(graph.shaMap[h1], graph.shaMap[h2],
848 firstBranch, secondBranch, graph)
852 if isinstance(sys.exc_info()[1], SystemExit):
855 traceback.print_exc(None, sys.stderr)