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 repo,hg,cmdutil,util,ui,revlog,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
   9 from tempfile import mkstemp
  10 from optparse import OptionParser
  11 import re
  12 import sys
  13 import os
  14
  15 # silly regex to catch Signed-off-by lines in log message
  16 sob_re=re.compile('^Signed-[Oo]ff-[Bb]y: (.+)$')
  17 # insert 'checkpoint' command after this many commits or none at all if 0
  18 cfg_checkpoint_count=0
  19 # write some progress message every this many file contents written
  20 cfg_export_boundary=1000
  21
  22 def gitmode(flags):
  23   return 'l' in flags and '120000' or 'x' in flags and '100755' or '100644'
  24
  25 def wr(msg=''):
  26   if msg == None:
  27     msg = ''
  28   print msg
  29   #map(lambda x: sys.stderr.write('\t[%s]\n' % x),msg.split('\n'))
  30
  31 def checkpoint(count):
  32   count=count+1
  33   if cfg_checkpoint_count>0 and count%cfg_checkpoint_count==0:
  34     sys.stderr.write("Checkpoint after %d commits\n" % count)
  35     wr('checkpoint')
  36     wr()
  37   return count
  38
  39 def get_parent_mark(parent,marks):
  40   """Get the mark for some parent.
  41   If we saw it in the current session, return :%d syntax and
  42   otherwise the SHA1 from the cache."""
  43   return marks.get(str(parent),':%d' % (parent+1))
  44
  45 def file_mismatch(f1,f2):
  46   """See if two revisions of a file are not equal."""
  47   return node.hex(f1)!=node.hex(f2)
  48
  49 def split_dict(dleft,dright,l=[],c=[],r=[],match=file_mismatch):
  50   """Loop over our repository and find all changed and missing files."""
  51   for left in dleft.keys():
  52     right=dright.get(left,None)
  53     if right==None:
  54       # we have the file but our parent hasn't: add to left set
  55       l.append(left)
  56     elif match(dleft[left],right):
  57       # we have it but checksums mismatch: add to center set
  58       c.append(left)
  59   for right in dright.keys():
  60     left=dleft.get(right,None)
  61     if left==None:
  62       # if parent has file but we don't: add to right set
  63       r.append(right)
  64     # change is already handled when comparing child against parent
  65   return l,c,r
  66
  67 def get_filechanges(repo,revision,parents,mleft):
  68   """Given some repository and revision, find all changed/deleted files."""
  69   l,c,r=[],[],[]
  70   for p in parents:
  71     if p<0: continue
  72     mright=repo.changectx(p).manifest()
  73     l,c,r=split_dict(mleft,mright,l,c,r)
  74   l.sort()
  75   c.sort()
  76   r.sort()
  77   return l,c,r
  78
  79 def get_author(logmessage,committer,authors):
  80   """As git distincts between author and committer of a patch, try to
  81   extract author by detecting Signed-off-by lines.
  82
  83   This walks from the end of the log message towards the top skipping
  84   empty lines. Upon the first non-empty line, it walks all Signed-off-by
  85   lines upwards to find the first one. For that (if found), it extracts
  86   authorship information the usual way (authors table, cleaning, etc.)
  87
  88   If no Signed-off-by line is found, this defaults to the committer.
  89
  90   This may sound stupid (and it somehow is), but in log messages we
  91   accidentially may have lines in the middle starting with
  92   "Signed-off-by: foo" and thus matching our detection regex. Prevent
  93   that."""
  94
  95   loglines=logmessage.split('\n')
  96   i=len(loglines)
  97   # from tail walk to top skipping empty lines
  98   while i>=0:
  99     i-=1
 100     if len(loglines[i].strip())==0: continue
 101     break
 102   if i>=0:
 103     # walk further upwards to find first sob line, store in 'first'
 104     first=None
 105     while i>=0:
 106       m=sob_re.match(loglines[i])
 107       if m==None: break
 108       first=m
 109       i-=1
 110     # if the last non-empty line matches our Signed-Off-by regex: extract username
 111     if first!=None:
 112       r=fixup_user(first.group(1),authors)
 113       return r
 114   return committer
 115
 116 def export_file_contents(ctx,manifest,files):
 117   count=0
 118   max=len(files)
 119   for file in files:
 120     d=ctx.filectx(file).data()
 121     wr('M %s inline %s' % (gitmode(manifest.flags(file)),file))
 122     wr('data %d' % len(d)) # had some trouble with size()
 123     wr(d)
 124     count+=1
 125     if count%cfg_export_boundary==0:
 126       sys.stderr.write('Exported %d/%d files\n' % (count,max))
 127   if max>cfg_export_boundary:
 128     sys.stderr.write('Exported %d/%d files\n' % (count,max))
 129
 130 def is_merge(parents):
 131   c=0
 132   for parent in parents:
 133     if parent>=0:
 134       c+=1
 135   return c>1
 136
 137 def sanitize_name(name,what="branch"):
 138   """Sanitize input roughly according to git-check-ref-format(1)"""
 139
 140   def dot(name):
 141     if name[0] == '.': return '_'+name[1:]
 142     return name
 143
 144   n=name
 145   p=re.compile('([[ ~^:?*]|\.\.)')
 146   n=p.sub('_', n)
 147   if n[-1] == '/': n=n[:-1]+'_'
 148   n='/'.join(map(dot,n.split('/')))
 149   p=re.compile('_+')
 150   n=p.sub('_', n)
 151
 152   if n!=name:
 153     sys.stderr.write('Warning: sanitized %s [%s] to [%s]\n' % (what,name,n))
 154   return n
 155
 156 def export_commit(ui,repo,revision,marks,heads,last,max,count,authors,sob,brmap):
 157   def get_branchname(name):
 158     if brmap.has_key(name):
 159       return brmap[name]
 160     n=sanitize_name(name)
 161     brmap[name]=n
 162     return n
 163
 164   (revnode,_,user,(time,timezone),files,desc,branch,_)=get_changeset(ui,repo,revision,authors)
 165   parents=repo.changelog.parentrevs(revision)
 166
 167   branch=get_branchname(branch)
 168
 169   wr('commit refs/heads/%s' % branch)
 170   wr('mark :%d' % (revision+1))
 171   if sob:
 172     wr('author %s %d %s' % (get_author(desc,user,authors),time,timezone))
 173   wr('committer %s %d %s' % (user,time,timezone))
 174   wr('data %d' % (len(desc)+1)) # wtf?
 175   wr(desc)
 176   wr()
 177
 178   pidx1, pidx2 = 0, 1
 179   if parents[0] < parents[1]:
 180     pidx1, pidx2 = 1, 0
 181
 182   src=heads.get(branch,'')
 183   link=''
 184   if src!='':
 185     # if we have a cached head, this is an incremental import: initialize it
 186     # and kill reference so we won't init it again
 187     wr('from %s' % src)
 188     heads[branch]=''
 189     sys.stderr.write('%s: Initializing to parent [%s]\n' %
 190         (branch,src))
 191     link=src # avoid making a merge commit for incremental import
 192   elif link=='' and not heads.has_key(branch) and revision>0:
 193     # newly created branch and not the first one: connect to parent
 194     tmp=get_parent_mark(parents[0],marks)
 195     wr('from %s' % tmp)
 196     sys.stderr.write('%s: Link new branch to parent [%s]\n' %
 197         (branch,tmp))
 198     link=tmp # avoid making a merge commit for branch fork
 199   elif last.get(branch,revision) != parents[pidx1] and parents[pidx1] > 0 and revision > 0:
 200     pm=get_parent_mark(parents[pidx1],marks)
 201     sys.stderr.write('%s: Placing commit [r%d] in branch [%s] on top of [r%d]\n' %
 202         (branch,revision,branch,parents[pidx1]));
 203     wr('from %s' % pm)
 204
 205   if parents[pidx2] > 0:
 206     pm=get_parent_mark(parents[pidx2],marks)
 207     sys.stderr.write('%s: Merging with parent [%s] from [r%d]\n' %
 208         (branch,pm,parents[pidx2]))
 209     wr('merge %s' % pm)
 210
 211   last[branch]=revision
 212   heads[branch]=''
 213   # we need this later to write out tags
 214   marks[str(revision)]=':%d'%(revision+1)
 215
 216   ctx=repo.changectx(str(revision))
 217   man=ctx.manifest()
 218   added,changed,removed,type=[],[],[],''
 219
 220   if revision==0:
 221     # first revision: feed in full manifest
 222     added=man.keys()
 223     added.sort()
 224     type='full'
 225   elif is_merge(parents):
 226     # later merge revision: feed in changed manifest
 227     # for many files comparing checksums is expensive so only do it for
 228     # merges where we really need it due to hg's revlog logic
 229     added,changed,removed=get_filechanges(repo,revision,parents,man)
 230     type='thorough delta'
 231   else:
 232     # later non-merge revision: feed in changed manifest
 233     # if we have exactly one parent, just take the changes from the
 234     # manifest without expensively comparing checksums
 235     f=repo.status(repo.lookup(parents[0]),revnode)[:3]
 236     added,changed,removed=f[1],f[0],f[2]
 237     type='simple delta'
 238
 239   sys.stderr.write('%s: Exporting %s revision %d/%d with %d/%d/%d added/changed/removed files\n' %
 240       (branch,type,revision+1,max,len(added),len(changed),len(removed)))
 241
 242   map(lambda r: wr('D %s' % r),removed)
 243   export_file_contents(ctx,man,added)
 244   export_file_contents(ctx,man,changed)
 245   wr()
 246
 247   return checkpoint(count)
 248
 249 def export_tags(ui,repo,marks_cache,start,end,count,authors):
 250   l=repo.tagslist()
 251   for tag,node in l:
 252     tag=sanitize_name(tag,"tag")
 253     # ignore latest revision
 254     if tag=='tip': continue
 255     rev=repo.changelog.rev(node)
 256     # ignore those tags not in our import range
 257     if rev<start or rev>=end: continue
 258
 259     ref=get_parent_mark(rev,marks_cache)
 260     if ref==None:
 261       sys.stderr.write('Failed to find reference for creating tag'
 262           ' %s at r%d\n' % (tag,rev))
 263       continue
 264     sys.stderr.write('Exporting tag [%s] at [hg r%d] [git %s]\n' % (tag,rev,ref))
 265     wr('reset refs/tags/%s' % tag)
 266     wr('from %s' % ref)
 267     wr()
 268     count=checkpoint(count)
 269   return count
 270
 271 def load_authors(filename):
 272   cache={}
 273   if not os.path.exists(filename):
 274     return cache
 275   f=open(filename,'r')
 276   l=0
 277   lre=re.compile('^([^=]+)[ ]*=[ ]*(.+)$')
 278   for line in f.readlines():
 279     l+=1
 280     m=lre.match(line)
 281     if m==None:
 282       sys.stderr.write('Invalid file format in [%s], line %d\n' % (filename,l))
 283       continue
 284     # put key:value in cache, key without ^:
 285     cache[m.group(1).strip()]=m.group(2).strip()
 286   f.close()
 287   sys.stderr.write('Loaded %d authors\n' % l)
 288   return cache
 289
 290 def verify_heads(ui,repo,cache,force):
 291   branches=repo.branchtags()
 292   l=[(-repo.changelog.rev(n), n, t) for t, n in branches.items()]
 293   l.sort()
 294
 295   # get list of hg's branches to verify, don't take all git has
 296   for _,_,b in l:
 297     b=get_branch(b)
 298     sha1=get_git_sha1(b)
 299     c=cache.get(b)
 300     if sha1!=None and c!=None:
 301       sys.stderr.write('Verifying branch [%s]\n' % b)
 302     if sha1!=c:
 303       sys.stderr.write('Error: Branch [%s] modified outside hg-fast-export:'
 304         '\n%s (repo) != %s (cache)\n' % (b,sha1,c))
 305       if not force: return False
 306
 307   # verify that branch has exactly one head
 308   t={}
 309   for h in repo.heads():
 310     (_,_,_,_,_,_,branch,_)=get_changeset(ui,repo,h)
 311     if t.get(branch,False):
 312       sys.stderr.write('Error: repository has at least one unnamed head: hg r%s\n' %
 313           repo.changelog.rev(h))
 314       if not force: return False
 315     t[branch]=True
 316
 317   return True
 318
 319 def mangle_mark(mark):
 320   return str(int(mark)-1)
 321
 322 def hg2git(repourl,m,marksfile,headsfile,tipfile,authors={},sob=False,force=False):
 323   _max=int(m)
 324
 325   marks_cache=load_cache(marksfile,mangle_mark)
 326   heads_cache=load_cache(headsfile)
 327   state_cache=load_cache(tipfile)
 328
 329   ui,repo=setup_repo(repourl)
 330
 331   if not verify_heads(ui,repo,heads_cache,force):
 332     return 1
 333
 334   try:
 335     tip=repo.changelog.count()
 336   except AttributeError:
 337     tip=len(repo)
 338
 339   min=int(state_cache.get('tip',0))
 340   max=_max
 341   if _max<0 or max>tip:
 342     max=tip
 343
 344   c=0
 345   last={}
 346   brmap={}
 347   for rev in range(min,max):
 348     c=export_commit(ui,repo,rev,marks_cache,heads_cache,last,max,c,authors,sob,brmap)
 349
 350   c=export_tags(ui,repo,marks_cache,min,max,c,authors)
 351
 352   sys.stderr.write('Issued %d commands\n' % c)
 353
 354   state_cache['tip']=max
 355   state_cache['repo']=repourl
 356   save_cache(tipfile,state_cache)
 357
 358   return 0
 359
 360 if __name__=='__main__':
 361   def bail(parser,opt):
 362     sys.stderr.write('Error: No %s option given\n' % opt)
 363     parser.print_help()
 364     sys.exit(2)
 365
 366   parser=OptionParser()
 367
 368   parser.add_option("-m","--max",type="int",dest="max",
 369       help="Maximum hg revision to import")
 370   parser.add_option("--marks",dest="marksfile",
 371       help="File to read git-fast-import's marks from")
 372   parser.add_option("--heads",dest="headsfile",
 373       help="File to read last run's git heads from")
 374   parser.add_option("--status",dest="statusfile",
 375       help="File to read status from")
 376   parser.add_option("-r","--repo",dest="repourl",
 377       help="URL of repo to import")
 378   parser.add_option("-s",action="store_true",dest="sob",
 379       default=False,help="Enable parsing Signed-off-by lines")
 380   parser.add_option("-A","--authors",dest="authorfile",
 381       help="Read authormap from AUTHORFILE")
 382   parser.add_option("-f","--force",action="store_true",dest="force",
 383       default=False,help="Ignore validation errors by force")
 384   parser.add_option("-M","--default-branch",dest="default_branch",
 385       help="Set the default branch")
 386
 387   (options,args)=parser.parse_args()
 388
 389   m=-1
 390   if options.max!=None: m=options.max
 391
 392   if options.marksfile==None: bail(parser,'--marks')
 393   if options.headsfile==None: bail(parser,'--heads')
 394   if options.statusfile==None: bail(parser,'--status')
 395   if options.repourl==None: bail(parser,'--repo')
 396
 397   a={}
 398   if options.authorfile!=None:
 399     a=load_authors(options.authorfile)
 400
 401   if options.default_branch!=None:
 402     set_default_branch(options.default_branch)
 403
 404   sys.exit(hg2git(options.repourl,m,options.marksfile,options.headsfile,
 405     options.statusfile,authors=a,sob=options.sob,force=options.force))