hg-fast-export.py

   1 #!/usr/bin/env python
   2
   3 # Copyright (c) 2007, 2008 Rocco Rutte <pdmef@gmx.net> and others.
   4 # License: MIT <http://www.opensource.org/licenses/mit-license.php>
   5
   6 from mercurial import node
   7 from hg2git import setup_repo,fixup_user,get_branch,get_changeset
   8 from hg2git import load_cache,save_cache,get_git_sha1,set_default_branch,set_origin_name
   9 from optparse import OptionParser
  10 import re
  11 import sys
  12 import os
  13
  14 if sys.platform == "win32":
  15   # On Windows, sys.stdout is initially opened in text mode, which means that
  16   # when a LF (\n) character is written to sys.stdout, it will be converted
  17   # into CRLF (\r\n).  That makes git blow up, so use this platform-specific
  18   # code to change the mode of sys.stdout to binary.
  19   import msvcrt
  20   msvcrt.setmode(sys.stdout.fileno(), os.O_BINARY)
  21
  22 # silly regex to catch Signed-off-by lines in log message
  23 sob_re=re.compile('^Signed-[Oo]ff-[Bb]y: (.+)$')
  24 # insert 'checkpoint' command after this many commits or none at all if 0
  25 cfg_checkpoint_count=0
  26 # write some progress message every this many file contents written
  27 cfg_export_boundary=1000
  28
  29 def gitmode(flags):
  30   return 'l' in flags and '120000' or 'x' in flags and '100755' or '100644'
  31
  32 def wr(msg=''):
  33   if msg:
  34     sys.stdout.write(msg)
  35   sys.stdout.write('\n')
  36   #map(lambda x: sys.stderr.write('\t[%s]\n' % x),msg.split('\n'))
  37
  38 def checkpoint(count):
  39   count=count+1
  40   if cfg_checkpoint_count>0 and count%cfg_checkpoint_count==0:
  41     sys.stderr.write("Checkpoint after %d commits\n" % count)
  42     wr('checkpoint')
  43     wr()
  44   return count
  45
  46 def revnum_to_revref(rev, old_marks):
  47   """Convert an hg revnum to a git-fast-import rev reference (an SHA1
  48   or a mark)"""
  49   return old_marks.get(rev) or ':%d' % (rev+1)
  50
  51 def file_mismatch(f1,f2):
  52   """See if two revisions of a file are not equal."""
  53   return node.hex(f1)!=node.hex(f2)
  54
  55 def split_dict(dleft,dright,l=[],c=[],r=[],match=file_mismatch):
  56   """Loop over our repository and find all changed and missing files."""
  57   for left in dleft.keys():
  58     right=dright.get(left,None)
  59     if right==None:
  60       # we have the file but our parent hasn't: add to left set
  61       l.append(left)
  62     elif match(dleft[left],right):
  63       # we have it but checksums mismatch: add to center set
  64       c.append(left)
  65   for right in dright.keys():
  66     left=dleft.get(right,None)
  67     if left==None:
  68       # if parent has file but we don't: add to right set
  69       r.append(right)
  70     # change is already handled when comparing child against parent
  71   return l,c,r
  72
  73 def get_filechanges(repo,revision,parents,mleft):
  74   """Given some repository and revision, find all changed/deleted files."""
  75   l,c,r=[],[],[]
  76   for p in parents:
  77     if p<0: continue
  78     mright=repo.changectx(p).manifest()
  79     l,c,r=split_dict(mleft,mright,l,c,r)
  80   l.sort()
  81   c.sort()
  82   r.sort()
  83   return l,c,r
  84
  85 def get_author(logmessage,committer,authors):
  86   """As git distincts between author and committer of a patch, try to
  87   extract author by detecting Signed-off-by lines.
  88
  89   This walks from the end of the log message towards the top skipping
  90   empty lines. Upon the first non-empty line, it walks all Signed-off-by
  91   lines upwards to find the first one. For that (if found), it extracts
  92   authorship information the usual way (authors table, cleaning, etc.)
  93
  94   If no Signed-off-by line is found, this defaults to the committer.
  95
  96   This may sound stupid (and it somehow is), but in log messages we
  97   accidentially may have lines in the middle starting with
  98   "Signed-off-by: foo" and thus matching our detection regex. Prevent
  99   that."""
 100
 101   loglines=logmessage.split('\n')
 102   i=len(loglines)
 103   # from tail walk to top skipping empty lines
 104   while i>=0:
 105     i-=1
 106     if len(loglines[i].strip())==0: continue
 107     break
 108   if i>=0:
 109     # walk further upwards to find first sob line, store in 'first'
 110     first=None
 111     while i>=0:
 112       m=sob_re.match(loglines[i])
 113       if m==None: break
 114       first=m
 115       i-=1
 116     # if the last non-empty line matches our Signed-Off-by regex: extract username
 117     if first!=None:
 118       r=fixup_user(first.group(1),authors)
 119       return r
 120   return committer
 121
 122 def export_file_contents(ctx,manifest,files):
 123   count=0
 124   max=len(files)
 125   for file in files:
 126     # Skip .hgtags files. They only get us in trouble.
 127     if file == ".hgtags":
 128       sys.stderr.write('Skip %s\n' % (file))
 129       continue
 130     d=ctx.filectx(file).data()
 131     wr('M %s inline %s' % (gitmode(manifest.flags(file)),file))
 132     wr('data %d' % len(d)) # had some trouble with size()
 133     wr(d)
 134     count+=1
 135     if count%cfg_export_boundary==0:
 136       sys.stderr.write('Exported %d/%d files\n' % (count,max))
 137   if max>cfg_export_boundary:
 138     sys.stderr.write('Exported %d/%d files\n' % (count,max))
 139
 140 def sanitize_name(name,what="branch"):
 141   """Sanitize input roughly according to git-check-ref-format(1)"""
 142
 143   def dot(name):
 144     if name[0] == '.': return '_'+name[1:]
 145     return name
 146
 147   n=name
 148   p=re.compile('([[ ~^:?*]|\.\.)')
 149   n=p.sub('_', n)
 150   if n[-1] in ('/', '.'): n=n[:-1]+'_'
 151   n='/'.join(map(dot,n.split('/')))
 152   p=re.compile('_+')
 153   n=p.sub('_', n)
 154
 155   if n!=name:
 156     sys.stderr.write('Warning: sanitized %s [%s] to [%s]\n' % (what,name,n))
 157   return n
 158
 159 def export_commit(ui,repo,revision,old_marks,max,count,authors,sob,brmap):
 160   def get_branchname(name):
 161     if brmap.has_key(name):
 162       return brmap[name]
 163     n=sanitize_name(name)
 164     brmap[name]=n
 165     return n
 166
 167   (revnode,_,user,(time,timezone),files,desc,branch,_)=get_changeset(ui,repo,revision,authors)
 168
 169   branch=get_branchname(branch)
 170
 171   parents = [p for p in repo.changelog.parentrevs(revision) if p >= 0]
 172
 173   if len(parents)==0 and revision != 0:
 174     wr('reset refs/heads/%s' % branch)
 175
 176   wr('commit refs/heads/%s' % branch)
 177   wr('mark :%d' % (revision+1))
 178   if sob:
 179     wr('author %s %d %s' % (get_author(desc,user,authors),time,timezone))
 180   wr('committer %s %d %s' % (user,time,timezone))
 181   wr('data %d' % (len(desc)+1)) # wtf?
 182   wr(desc)
 183   wr()
 184
 185
 186   # Sort the parents based on revision ids so that we always get the
 187   # same resulting git repo, no matter how the revisions were
 188   # numbered.
 189   parents.sort(key=repo.changelog.node, reverse=True)
 190
 191   ctx=repo.changectx(str(revision))
 192   man=ctx.manifest()
 193   added,changed,removed,type=[],[],[],''
 194
 195   if len(parents) == 0:
 196     # first revision: feed in full manifest
 197     added=man.keys()
 198     added.sort()
 199     type='full'
 200   else:
 201     wr('from %s' % revnum_to_revref(parents[0], old_marks))
 202     if len(parents) == 1:
 203       # later non-merge revision: feed in changed manifest
 204       # if we have exactly one parent, just take the changes from the
 205       # manifest without expensively comparing checksums
 206       f=repo.status(repo.lookup(parents[0]),revnode)[:3]
 207       added,changed,removed=f[1],f[0],f[2]
 208       type='simple delta'
 209     else: # a merge with two parents
 210       wr('merge %s' % revnum_to_revref(parents[1], old_marks))
 211       # later merge revision: feed in changed manifest
 212       # for many files comparing checksums is expensive so only do it for
 213       # merges where we really need it due to hg's revlog logic
 214       added,changed,removed=get_filechanges(repo,revision,parents,man)
 215       type='thorough delta'
 216
 217   sys.stderr.write('%s: Exporting %s revision %d/%d with %d/%d/%d added/changed/removed files\n' %
 218       (branch,type,revision+1,max,len(added),len(changed),len(removed)))
 219
 220   map(lambda r: wr('D %s' % r),removed)
 221   export_file_contents(ctx,man,added)
 222   export_file_contents(ctx,man,changed)
 223   wr()
 224
 225   return checkpoint(count)
 226
 227 def export_tags(ui,repo,old_marks,mapping_cache,count,authors):
 228   l=repo.tagslist()
 229   for tag,node in l:
 230     tag=sanitize_name(tag,"tag")
 231     # ignore latest revision
 232     if tag=='tip': continue
 233     # ignore tags to nodes that are missing (ie, 'in the future')
 234     if node.encode('hex_codec') not in mapping_cache:
 235       sys.stderr.write('Tag %s refers to unseen node %s\n' % (tag, node.encode('hex_codec')))
 236       continue
 237
 238     rev=int(mapping_cache[node.encode('hex_codec')])
 239
 240     ref=revnum_to_revref(rev, old_marks)
 241     if ref==None:
 242       sys.stderr.write('Failed to find reference for creating tag'
 243           ' %s at r%d\n' % (tag,rev))
 244       continue
 245     sys.stderr.write('Exporting tag [%s] at [hg r%d] [git %s]\n' % (tag,rev,ref))
 246     wr('reset refs/tags/%s' % tag)
 247     wr('from %s' % ref)
 248     wr()
 249     count=checkpoint(count)
 250   return count
 251
 252 def load_authors(filename):
 253   cache={}
 254   if not os.path.exists(filename):
 255     return cache
 256   f=open(filename,'r')
 257   l=0
 258   lre=re.compile('^([^=]+)[ ]*=[ ]*(.+)$')
 259   for line in f.readlines():
 260     l+=1
 261     m=lre.match(line)
 262     if m==None:
 263       sys.stderr.write('Invalid file format in [%s], line %d\n' % (filename,l))
 264       continue
 265     # put key:value in cache, key without ^:
 266     cache[m.group(1).strip()]=m.group(2).strip()
 267   f.close()
 268   sys.stderr.write('Loaded %d authors\n' % l)
 269   return cache
 270
 271 def branchtip(repo, heads):
 272   '''return the tipmost branch head in heads'''
 273   tip = heads[-1]
 274   for h in reversed(heads):
 275     if 'close' not in repo.changelog.read(h)[5]:
 276       tip = h
 277       break
 278   return tip
 279
 280 def verify_heads(ui,repo,cache,force):
 281   branches={}
 282   for bn, heads in repo.branchmap().iteritems():
 283     branches[bn] = branchtip(repo, heads)
 284   l=[(-repo.changelog.rev(n), n, t) for t, n in branches.items()]
 285   l.sort()
 286
 287   # get list of hg's branches to verify, don't take all git has
 288   for _,_,b in l:
 289     b=get_branch(b)
 290     sha1=get_git_sha1(b)
 291     c=cache.get(b)
 292     if sha1!=c:
 293       sys.stderr.write('Error: Branch [%s] modified outside hg-fast-export:'
 294         '\n%s (repo) != %s (cache)\n' % (b,sha1,c))
 295       if not force: return False
 296
 297   # verify that branch has exactly one head
 298   t={}
 299   for h in repo.heads():
 300     (_,_,_,_,_,_,branch,_)=get_changeset(ui,repo,h)
 301     if t.get(branch,False):
 302       sys.stderr.write('Error: repository has at least one unnamed head: hg r%s\n' %
 303           repo.changelog.rev(h))
 304       if not force: return False
 305     t[branch]=True
 306
 307   return True
 308
 309 def hg2git(repourl,m,marksfile,mappingfile,headsfile,tipfile,authors={},sob=False,force=False):
 310   _max=int(m)
 311
 312   old_marks=load_cache(marksfile,lambda s: int(s)-1)
 313   mapping_cache=load_cache(mappingfile)
 314   heads_cache=load_cache(headsfile)
 315   state_cache=load_cache(tipfile)
 316
 317   ui,repo=setup_repo(repourl)
 318
 319   if not verify_heads(ui,repo,heads_cache,force):
 320     return 1
 321
 322   try:
 323     tip=repo.changelog.count()
 324   except AttributeError:
 325     tip=len(repo)
 326
 327   min=int(state_cache.get('tip',0))
 328   max=_max
 329   if _max<0 or max>tip:
 330     max=tip
 331
 332   for rev in range(0,max):
 333         (revnode,_,_,_,_,_,_,_)=get_changeset(ui,repo,rev,authors)
 334         mapping_cache[revnode.encode('hex_codec')] = str(rev)
 335
 336
 337   c=0
 338   brmap={}
 339   for rev in range(min,max):
 340     c=export_commit(ui,repo,rev,old_marks,max,c,authors,sob,brmap)
 341
 342   state_cache['tip']=max
 343   state_cache['repo']=repourl
 344   save_cache(tipfile,state_cache)
 345   save_cache(mappingfile,mapping_cache)
 346
 347   c=export_tags(ui,repo,old_marks,mapping_cache,c,authors)
 348
 349   sys.stderr.write('Issued %d commands\n' % c)
 350
 351   return 0
 352
 353 if __name__=='__main__':
 354   def bail(parser,opt):
 355     sys.stderr.write('Error: No %s option given\n' % opt)
 356     parser.print_help()
 357     sys.exit(2)
 358
 359   parser=OptionParser()
 360
 361   parser.add_option("-m","--max",type="int",dest="max",
 362       help="Maximum hg revision to import")
 363   parser.add_option("--mapping",dest="mappingfile",
 364       help="File to read last run's hg-to-git SHA1 mapping")
 365   parser.add_option("--marks",dest="marksfile",
 366       help="File to read git-fast-import's marks from")
 367   parser.add_option("--heads",dest="headsfile",
 368       help="File to read last run's git heads from")
 369   parser.add_option("--status",dest="statusfile",
 370       help="File to read status from")
 371   parser.add_option("-r","--repo",dest="repourl",
 372       help="URL of repo to import")
 373   parser.add_option("-s",action="store_true",dest="sob",
 374       default=False,help="Enable parsing Signed-off-by lines")
 375   parser.add_option("-A","--authors",dest="authorfile",
 376       help="Read authormap from AUTHORFILE")
 377   parser.add_option("-f","--force",action="store_true",dest="force",
 378       default=False,help="Ignore validation errors by force")
 379   parser.add_option("-M","--default-branch",dest="default_branch",
 380       help="Set the default branch")
 381   parser.add_option("-o","--origin",dest="origin_name",
 382       help="use <name> as namespace to track upstream")
 383
 384   (options,args)=parser.parse_args()
 385
 386   m=-1
 387   if options.max!=None: m=options.max
 388
 389   if options.marksfile==None: bail(parser,'--marks')
 390   if options.mappingfile==None: bail(parser,'--mapping')
 391   if options.headsfile==None: bail(parser,'--heads')
 392   if options.statusfile==None: bail(parser,'--status')
 393   if options.repourl==None: bail(parser,'--repo')
 394
 395   a={}
 396   if options.authorfile!=None:
 397     a=load_authors(options.authorfile)
 398
 399   if options.default_branch!=None:
 400     set_default_branch(options.default_branch)
 401
 402   if options.origin_name!=None:
 403     set_origin_name(options.origin_name)
 404
 405   sys.exit(hg2git(options.repourl,m,options.marksfile,options.mappingfile,options.headsfile,
 406     options.statusfile,authors=a,sob=options.sob,force=options.force))