git_remote_helpers/fastimport/commands.py

   1 # Copyright (C) 2008 Canonical Ltd
   2 #
   3 # This program is free software; you can redistribute it and/or modify
   4 # it under the terms of the GNU General Public License as published by
   5 # the Free Software Foundation; either version 2 of the License, or
   6 # (at your option) any later version.
   7 #
   8 # This program is distributed in the hope that it will be useful,
   9 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  10 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  11 # GNU General Public License for more details.
  12 #
  13 # You should have received a copy of the GNU General Public License
  14 # along with this program; if not, write to the Free Software
  15 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  16
  17 """Import command classes."""
  18
  19 import os
  20
  21 # There is a bug in git 1.5.4.3 and older by which unquoting a string consumes
  22 # one extra character. Set this variable to True to work-around it. It only
  23 # happens when renaming a file whose name contains spaces and/or quotes, and
  24 # the symptom is:
  25 #   % git-fast-import
  26 #   fatal: Missing space after source: R "file 1.txt" file 2.txt
  27 # http://git.kernel.org/?p=git/git.git;a=commit;h=c8744d6a8b27115503565041566d97c21e722584
  28 GIT_FAST_IMPORT_NEEDS_EXTRA_SPACE_AFTER_QUOTE = False
  29
  30
  31 # Lists of command names
  32 COMMAND_NAMES = ['blob', 'checkpoint', 'commit', 'feature', 'progress',
  33     'reset', 'tag']
  34 FILE_COMMAND_NAMES = ['filemodify', 'filedelete', 'filecopy', 'filerename',
  35     'filedeleteall']
  36
  37
  38 # Feature names
  39 MULTIPLE_AUTHORS_FEATURE = "multiple-authors"
  40 COMMIT_PROPERTIES_FEATURE = "commit-properties"
  41 EMPTY_DIRS_FEATURE = "empty-directories"
  42 FEATURE_NAMES = [
  43     MULTIPLE_AUTHORS_FEATURE,
  44     COMMIT_PROPERTIES_FEATURE,
  45     EMPTY_DIRS_FEATURE,
  46     ]
  47
  48
  49 # for classes with no meaningful __str__()
  50 def _simplerepr(self):
  51     return "<%s at 0x%x>" % (self.__class__.__name__, id(self))
  52
  53 # classes that define __str__() should use this instead
  54 def _detailrepr(self):
  55     return ("<%s at 0x%x: %s>"
  56             % (self.__class__.__name__, id(self), str(self)))
  57
  58
  59 class ImportCommand(object):
  60     """Base class for import commands."""
  61
  62     def __init__(self, name):
  63         self.name = name
  64         # List of field names not to display
  65         self._binary = []
  66
  67     __repr__ = _simplerepr
  68
  69     def format(self):
  70         """Format this command as a fastimport dump fragment.
  71
  72         Returns a (possibly multiline) string that, if seen in a
  73         fastimport stream, would parse to an equivalent command object.
  74         """
  75         raise NotImplementedError("abstract method")
  76
  77     def dump_str(self, names=None, child_lists=None, verbose=False):
  78         """Dump fields as a string.
  79
  80         :param names: the list of fields to include or
  81             None for all public fields
  82         :param child_lists: dictionary of child command names to
  83             fields for that child command to include
  84         :param verbose: if True, prefix each line with the command class and
  85             display fields as a dictionary; if False, dump just the field
  86             values with tabs between them
  87         """
  88         interesting = {}
  89         if names is None:
  90             fields = [k for k in self.__dict__.keys() if not k.startswith('_')]
  91         else:
  92             fields = names
  93         for field in fields:
  94             value = self.__dict__.get(field)
  95             if field in self._binary and value is not None:
  96                 value = '(...)'
  97             interesting[field] = value
  98         if verbose:
  99             return "%s: %s" % (self.__class__.__name__, interesting)
 100         else:
 101             return "\t".join([repr(interesting[k]) for k in fields])
 102
 103
 104 class _MarkMixin(object):
 105     """mixin for fastimport commands with a mark: blob, commit."""
 106     def __init__(self, mark, location):
 107         self.mark= mark
 108         self.location = location
 109
 110         # Provide a unique id in case the mark is missing
 111         if mark is None:
 112             self.id = '%s@%d' % (os.path.basename(location[0]), location[1])
 113         else:
 114             self.id = ':%s' % mark
 115
 116     def __str__(self):
 117         return self.id
 118
 119     __repr__ = _detailrepr
 120
 121
 122 class BlobCommand(ImportCommand, _MarkMixin):
 123
 124     def __init__(self, mark, data, location):
 125         ImportCommand.__init__(self, 'blob')
 126         _MarkMixin.__init__(self, mark, location)
 127         self.data = data
 128         self._binary = ['data']
 129
 130     def format(self):
 131         if self.mark is None:
 132             mark_line = ""
 133         else:
 134             mark_line = "\nmark :%s" % self.mark
 135         return "blob%s\ndata %d\n%s" % (mark_line, len(self.data), self.data)
 136
 137
 138 class CheckpointCommand(ImportCommand):
 139
 140     def __init__(self):
 141         ImportCommand.__init__(self, 'checkpoint')
 142
 143     def format(self):
 144         return "checkpoint"
 145
 146
 147 class CommitCommand(ImportCommand, _MarkMixin):
 148
 149     def __init__(self, ref, mark, author, committer, message, from_,
 150         merges, file_cmds, location=None, more_authors=None, properties=None):
 151         ImportCommand.__init__(self, 'commit')
 152         _MarkMixin.__init__(self, mark, location)
 153         self.ref = ref
 154         self.author = author
 155         self.committer = committer
 156         self.message = message
 157         self.from_ = from_
 158         self.merges = merges
 159         self.file_cmds = file_cmds
 160         self.more_authors = more_authors
 161         self.properties = properties
 162         self._binary = ['file_cmds']
 163
 164     def format(self, use_features=True, include_file_contents=True):
 165         if self.mark is None:
 166             mark_line = ""
 167         else:
 168             mark_line = "\nmark :%s" % self.mark
 169         if self.author is None:
 170             author_section = ""
 171         else:
 172             author_section = "\nauthor %s" % format_who_when(self.author)
 173             if use_features and self.more_authors:
 174                 for author in self.more_authors:
 175                     author_section += "\nauthor %s" % format_who_when(author)
 176         committer = "committer %s" % format_who_when(self.committer)
 177         if self.message is None:
 178             msg_section = ""
 179         else:
 180             msg = self.message.encode('utf8')
 181             msg_section = "\ndata %d\n%s" % (len(msg), msg)
 182         if self.from_ is None:
 183             from_line = ""
 184         else:
 185             from_line = "\nfrom %s" % self.from_
 186         if self.merges is None:
 187             merge_lines = ""
 188         else:
 189             merge_lines = "".join(["\nmerge %s" % (m,)
 190                 for m in self.merges])
 191         if use_features and self.properties:
 192             property_lines = []
 193             for name in sorted(self.properties):
 194                 value = self.properties[name]
 195                 property_lines.append("\n" + format_property(name, value))
 196             properties_section = "".join(property_lines)
 197         else:
 198             properties_section = ""
 199         if self.file_cmds is None:
 200             filecommands = ""
 201         else:
 202             if include_file_contents:
 203                 format_str = "\n%r"
 204             else:
 205                 format_str = "\n%s"
 206             filecommands = "".join(
 207                 ["\n" + fc.format() for fc in self.file_cmds])
 208         return "commit %s%s%s\n%s%s%s%s%s%s" % (self.ref, mark_line,
 209             author_section, committer, msg_section, from_line, merge_lines,
 210             properties_section, filecommands)
 211
 212     def dump_str(self, names=None, child_lists=None, verbose=False):
 213         result = [ImportCommand.dump_str(self, names, verbose=verbose)]
 214         for f in self.file_cmds:
 215             if child_lists is None:
 216                 continue
 217             try:
 218                 child_names = child_lists[f.name]
 219             except KeyError:
 220                 continue
 221             result.append("\t%s" % f.dump_str(child_names, verbose=verbose))
 222         return '\n'.join(result)
 223
 224
 225 class FeatureCommand(ImportCommand):
 226
 227     def __init__(self, feature_name, value=None, location=None):
 228         ImportCommand.__init__(self, 'feature')
 229         self.feature_name = feature_name
 230         self.value = value
 231         self.location = location
 232
 233     def format(self):
 234         if self.value is None:
 235             value_text = ""
 236         else:
 237             value_text = "=%s" % self.value
 238         return "feature %s%s" % (self.feature_name, value_text)
 239
 240
 241 class ProgressCommand(ImportCommand):
 242
 243     def __init__(self, message):
 244         ImportCommand.__init__(self, 'progress')
 245         self.message = message
 246
 247     def format(self):
 248         return "progress %s" % (self.message,)
 249
 250
 251 class ResetCommand(ImportCommand):
 252
 253     def __init__(self, ref, from_):
 254         ImportCommand.__init__(self, 'reset')
 255         self.ref = ref
 256         self.from_ = from_
 257
 258     def format(self):
 259         if self.from_ is None:
 260             from_line = ""
 261         else:
 262             # According to git-fast-import(1), the extra LF is optional here;
 263             # however, versions of git up to 1.5.4.3 had a bug by which the LF
 264             # was needed. Always emit it, since it doesn't hurt and maintains
 265             # compatibility with older versions.
 266             # http://git.kernel.org/?p=git/git.git;a=commit;h=655e8515f279c01f525745d443f509f97cd805ab
 267             from_line = "\nfrom %s\n" % self.from_
 268         return "reset %s%s" % (self.ref, from_line)
 269
 270
 271 class TagCommand(ImportCommand):
 272
 273     def __init__(self, id, from_, tagger, message):
 274         ImportCommand.__init__(self, 'tag')
 275         self.id = id
 276         self.from_ = from_
 277         self.tagger = tagger
 278         self.message = message
 279
 280     def __str__(self):
 281         return self.id
 282
 283     __repr__ = _detailrepr
 284
 285     def format(self):
 286         if self.from_ is None:
 287             from_line = ""
 288         else:
 289             from_line = "\nfrom %s" % self.from_
 290         if self.tagger is None:
 291             tagger_line = ""
 292         else:
 293             tagger_line = "\ntagger %s" % format_who_when(self.tagger)
 294         if self.message is None:
 295             msg_section = ""
 296         else:
 297             msg = self.message.encode('utf8')
 298             msg_section = "\ndata %d\n%s" % (len(msg), msg)
 299         return "tag %s%s%s%s" % (self.id, from_line, tagger_line, msg_section)
 300
 301
 302 class FileCommand(ImportCommand):
 303     """Base class for file commands."""
 304     pass
 305
 306
 307 class FileModifyCommand(FileCommand):
 308
 309     def __init__(self, path, mode, dataref, data):
 310         # Either dataref or data should be null
 311         FileCommand.__init__(self, 'filemodify')
 312         self.path = check_path(path)
 313         self.mode = mode
 314         self.dataref = dataref
 315         self.data = data
 316         self._binary = ['data']
 317
 318     def __str__(self):
 319         return self.path
 320
 321     __repr__ = _detailrepr
 322
 323     def format(self, include_file_contents=True):
 324         datastr = ""
 325         if self.dataref is None:
 326             dataref = "inline"
 327             if include_file_contents:
 328                 datastr = "\ndata %d\n%s" % (len(self.data), self.data)
 329         else:
 330             dataref = "%s" % (self.dataref,)
 331         path = format_path(self.path)
 332         return "M %s %s %s%s" % (self.mode, dataref, path, datastr)
 333
 334     def is_regular(self):
 335         """Return true if this is a regular file (mode 644)."""
 336         return self.mode.endswith("644")
 337
 338     def is_executable(self):
 339         """Return true if this is an executable file (mode 755)."""
 340         return self.mode.endswith("755")
 341
 342     def is_symlink(self):
 343         """Return true if this is a symlink (mode 120000)."""
 344         return self.mode == "120000"
 345
 346     def is_gitlink(self):
 347         """Return true if this is a gitlink (mode 160000)."""
 348         return self.mode == "160000"
 349
 350
 351 class FileDeleteCommand(FileCommand):
 352
 353     def __init__(self, path):
 354         FileCommand.__init__(self, 'filedelete')
 355         self.path = check_path(path)
 356
 357     def __str__(self):
 358         return self.path
 359
 360     __repr__ = _detailrepr
 361
 362     def format(self):
 363         return "D %s" % (format_path(self.path),)
 364
 365
 366 class FileCopyCommand(FileCommand):
 367
 368     def __init__(self, src_path, dest_path):
 369         FileCommand.__init__(self, 'filecopy')
 370         self.src_path = check_path(src_path)
 371         self.dest_path = check_path(dest_path)
 372
 373     def __str__(self):
 374         return "%s -> %s" % (self.src_path, self.dest_path)
 375
 376     __repr__ = _detailrepr
 377
 378     def format(self):
 379         return "C %s %s" % (
 380             format_path(self.src_path, quote_spaces=True),
 381             format_path(self.dest_path))
 382
 383
 384 class FileRenameCommand(FileCommand):
 385
 386     def __init__(self, old_path, new_path):
 387         FileCommand.__init__(self, 'filerename')
 388         self.old_path = check_path(old_path)
 389         self.new_path = check_path(new_path)
 390
 391     def __str__(self):
 392         return "%s -> %s" % (self.old_path, self.new_path)
 393
 394     __repr__ = _detailrepr
 395
 396     def format(self):
 397         return "R %s %s" % (
 398             format_path(self.old_path, quote_spaces=True),
 399             format_path(self.new_path))
 400
 401
 402 class FileDeleteAllCommand(FileCommand):
 403
 404     def __init__(self):
 405         FileCommand.__init__(self, 'filedeleteall')
 406
 407     def format(self):
 408         return "deleteall"
 409
 410
 411 def check_path(path):
 412     """Check that a path is legal.
 413
 414     :return: the path if all is OK
 415     :raise ValueError: if the path is illegal
 416     """
 417     if path is None or path == '':
 418         raise ValueError("illegal path '%s'" % path)
 419     return path
 420
 421
 422 def format_path(p, quote_spaces=False):
 423     """Format a path in utf8, quoting it if necessary."""
 424     if '\n' in p:
 425         import re
 426         p = re.sub('\n', '\\n', p)
 427         quote = True
 428     else:
 429         quote = p[0] == '"' or (quote_spaces and ' ' in p)
 430     if quote:
 431         extra = GIT_FAST_IMPORT_NEEDS_EXTRA_SPACE_AFTER_QUOTE and ' ' or ''
 432         p = '"%s"%s' % (p, extra)
 433     return p.encode('utf8')
 434
 435
 436 def format_who_when(fields):
 437     """Format a tuple of name,email,secs-since-epoch,utc-offset-secs as a string."""
 438     offset = fields[3]
 439     if offset < 0:
 440         offset_sign = '-'
 441         offset = abs(offset)
 442     else:
 443         offset_sign = '+'
 444     offset_hours = offset / 3600
 445     offset_minutes = offset / 60 - offset_hours * 60
 446     offset_str = "%s%02d%02d" % (offset_sign, offset_hours, offset_minutes)
 447     name = fields[0]
 448     if name == '':
 449         sep = ''
 450     else:
 451         sep = ' '
 452     if isinstance(name, unicode):
 453         name = name.encode('utf8')
 454     email = fields[1]
 455     if isinstance(email, unicode):
 456         email = email.encode('utf8')
 457     result = "%s%s<%s> %d %s" % (name, sep, email, fields[2], offset_str)
 458     return result
 459
 460
 461 def format_property(name, value):
 462     """Format the name and value (both unicode) of a property as a string."""
 463     utf8_name = name.encode('utf8')
 464     if value is not None:
 465         utf8_value = value.encode('utf8')
 466         result = "property %s %d %s" % (utf8_name, len(utf8_value), utf8_value)
 467     else:
 468         result = "property %s" % (utf8_name,)
 469     return result