def execute(self, dumpfilename, directory):
"""
Executes the export.
@type dumpfilename: string
@param dumpfilename: Name of the svn dump file.
@type directory: string
@param directory: Directory to store the exported files.
"""
dump = SvnDumpFile()
dump.open(dumpfilename)
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if self.__exports.has_key(revnr):
for path, filename in self.__exports[revnr].iteritems():
print "r%-6d %s" % (revnr, path)
nodes = dump.get_nodes_by_path(path, "ACR")
saved = False
for node in nodes:
if node.has_text():
outfile = open(filename, "wb")
node.write_text_to_file(outfile)
outfile.close()
saved = True
print " saved as %s" % filename
if not saved:
if len(nodes) == 0:
print " not found"
else:
print " has no text"
dump.close()
return 0
def execute( self, dumpfilename, directory ):
"""
Executes the export.
@type dumpfilename: string
@param dumpfilename: Name of the svn dump file.
@type directory: string
@param directory: Directory to store the exported files.
"""
dump = SvnDumpFile()
dump.open( dumpfilename )
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if self.__exports.has_key( revnr ):
for path, filename in self.__exports[revnr].iteritems():
print "r%-6d %s" % ( revnr, path )
nodes = dump.get_nodes_by_path( path, "ACR" )
saved = False
for node in nodes:
if node.has_text():
outfile = open( filename, "wb" )
node.write_text_to_file( outfile )
outfile.close()
saved = True
print " saved as %s" % filename
if not saved:
if len(nodes) == 0:
print " not found"
else:
print " has no text"
dump.close()
return 0
def old_execute(self, dumpfilename):
"""
Print file list of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
dump = SvnDumpFile()
dump.open(dumpfilename)
actions = {"add": "A", "change": "M", "delete": "D", "replace": "R"}
lines = ""
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr == self.revNr or self.revNr == -1:
lines = ""
for node in dump.get_nodes_iter():
action = actions[node.get_action()]
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if node.has_copy_from():
fpath = node.get_copy_from_path()
frev = node.get_copy_from_rev()
if fpath == "" or fpath[0] != "/":
fpath = "/" + fpath
path += " (from %s:%d)" % (fpath, frev)
lines += " %s %s\n" % (action, path)
if revnr == self.revNr:
print lines,
lines = ""
break
if len(lines) > 0:
print lines,
dump.close()
return 0
def old_execute( self, dumpfilename ):
"""
Print file list of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
dump = SvnDumpFile()
dump.open( dumpfilename )
actions = { "add":"A", "change":"M", "delete":"D", "replace":"R" }
lines = ""
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr == self.revNr or self.revNr == -1:
lines = ""
for node in dump.get_nodes_iter():
action = actions[node.get_action()]
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if node.has_copy_from():
fpath = node.get_copy_from_path()
frev = node.get_copy_from_rev()
if fpath == "" or fpath[0] != "/":
fpath = "/" + fpath
path += " (from %s:%d)" % ( fpath, frev )
lines += " %s %s\n" % ( action, path )
if revnr == self.revNr:
print lines,
lines = ""
break
if len(lines) > 0:
print lines,
dump.close()
return 0
def execute(self):
"""
Executes the Edit.
"""
# +++ catch exception and return errorcode
srcdmp = SvnDumpFile()
srcdmp.open(self.__in_file)
dstdmp = None
hasrev = srcdmp.read_next_rev()
if hasrev:
if not self.dry_run:
dstdmp = SvnDumpFile()
if srcdmp.get_rev_nr() == 0:
# create new dump with revision 0
dstdmp.create_with_rev_0(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_date_str())
hasrev = srcdmp.read_next_rev()
else:
# create new dump starting with the same revNr
# as the input dump file
dstdmp.create_with_rev_n(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_nr())
# now copy all the revisions
while hasrev:
self.__process_rev(srcdmp, dstdmp)
hasrev = srcdmp.read_next_rev()
def split_dumpfiles( inputfilename, outlist ):
"""
Splits a dump file.
@type inputfilename: string
@param inputfilename: Name of the input file.
@type outlist: list
@param outlist: List of tuples containing start revnr, end revnr and filename.
@rtype: int
@return: 0 for success.
"""
if len(outlist) == 0:
return 0
outlist = outlist[:]
outlist.sort()
parallel = False
for i in range( 0, len(outlist)-1 ):
if outlist[i][1] > outlist[i+1][0]:
parallel = True
break
if not parallel:
indump = SvnDumpFile()
indump.open( inputfilename )
index = 0
startrev = outlist[index][0]
endrev = outlist[index][1]
outfile = outlist[index][2]
outdump = None
while indump.read_next_rev():
revnr = indump.get_rev_nr()
if outdump == None:
if revnr >= startrev:
outdump = SvnDumpFile()
if revnr == 0:
# create new dump with revision 0
outdump.create_with_rev_0( outfile, indump.get_uuid(),
indump.get_rev_date_str() )
else:
# create new dump starting with the
# same revNr as the original dump
outdump.create_with_rev_n( outfile, indump.get_uuid(),
indump.get_rev_nr() )
if outdump != None:
# have an output file, copy the revision if revnr > 0
if revnr > 0:
outdump.add_rev_from_dump( indump )
if revnr >= endrev:
# end revision reached
outdump.close()
outdump = None
index += 1
if index >= len(outlist):
# done.
break
# next range
startrev = outlist[index][0]
endrev = outlist[index][1]
outfile = outlist[index][2]
if outdump != None:
outdump.close()
indump.close()
else:
print "overlapping revision ranges not supported (yet)."
return 1
return 0
def join_dumpfiles( inputlist, outfilename ):
"""
Joins dump files.
@type inputlist: list
@param inputlist: A list containing the input filenames.
@type outfilename: string
@param outfilename: Name of the output dump file.
@rtype: int
@return: 0 for success.
"""
outdump = None
noutrev = 0
lastrev = -1
for filename in inputlist:
print "reading %s ..." % filename
ninrev = 0
indump = SvnDumpFile()
indump.open( filename )
hasrev = indump.read_next_rev()
if hasrev:
if outdump == None:
outdump = SvnDumpFile()
if indump.get_rev_nr() == 0:
# create new dump with revision 0
outdump.create_with_rev_0( outfilename, indump.get_uuid(),
indump.get_rev_date_str() )
hasrev = indump.read_next_rev()
else:
# create new dump starting with the
# same revNr as the original dump
outdump.create_with_rev_n( outfilename, indump.get_uuid(),
indump.get_rev_nr() )
else:
# check rev number
if indump.get_rev_nr() == 0:
hasrev = indump.read_next_rev()
if hasrev:
if (lastrev + 1) != indump.get_rev_nr():
print "renumbering of revisions not supported."
print "last rev was %d, next is %d." % ( lastrev,
indump.get_rev_nr() )
indump.close()
outdump.close()
return 1
while hasrev:
outdump.add_rev_from_dump( indump )
ninrev += 1
lastrev = indump.get_rev_nr()
hasrev = indump.read_next_rev()
indump.close()
print " copied %d revisions." % ninrev
noutrev += ninrev
outdump.close()
print "wrote %d revisions, last was r%d." % ( noutrev, lastrev )
def execute(self, callback):
"""
Execute the diff.
@type callback: SvnDumpDiffCallback
@param callback: Callback object for diffs found.
"""
# diff started
callback.comparing(self.__filename1, self.__filename2)
# open files
dump1 = SvnDumpFile()
dump2 = SvnDumpFile()
dump1.open(self.__filename1)
dump2.open(self.__filename2)
# compare uuid
if dump1.get_uuid() != dump2.get_uuid():
callback.rev_diff("UUID", dump1.get_uuid(), dump2.get_uuid())
hasrev1 = dump1.read_next_rev()
hasrev2 = dump2.read_next_rev()
while hasrev1 and hasrev2:
# compare rev numbers
if dump1.get_rev_nr() != dump2.get_rev_nr():
callback.rev_diff("RevNr", dump1.get_rev_nr(),
dump2.get_rev_nr())
# error has been reported so set both flags to false
hasrev1 = False
hasrev2 = False
break
# next revision...
callback.next_revision(dump1.get_rev_nr(), dump2.get_rev_nr())
# compare rev date
if dump1.get_rev_date() != dump2.get_rev_date():
callback.rev_diff("RevDate", str(dump1.get_rev_date()),
str(dump2.get_rev_date()))
if dump1.get_rev_date_str() != dump2.get_rev_date_str():
callback.rev_diff("RevDateStr", dump1.get_rev_date_str(),
dump2.get_rev_date_str())
# compare rev author
# compare rev log
# compare rev props
self.__compare_properties(True, dump1.get_rev_props(),
dump2.get_rev_props(), callback)
# compare nodes
self.__compare_nodes(dump1, dump2, callback)
# read next revision
hasrev1 = dump1.read_next_rev()
hasrev2 = dump2.read_next_rev()
if hasrev1 or hasrev2:
print "random error ;-) (different rev nr or EOF of one file)"
# done.
callback.compare_done()
def merge(self):
"""
Executes the merge.
"""
if len(self.__in_files) == 0:
print("merge: no input files specified")
return
if len(self.__out_file) == 0:
print("merge: no output file specified")
return
# open input dump files
for inFile in self.__in_files:
inDump = SvnDumpFile()
inDump.open(inFile)
inDump.read_next_rev()
self.__in_dumps = self.__in_dumps + [inDump]
if inDump.get_rev_date_str() < self.__out_r0_date:
self.__out_r0_date = inDump.get_rev_date_str()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
return
# open output file
self.outDump = SvnDumpFile()
if self.outStartRev == 0:
self.outDump.create_with_rev_0(self.__out_file,
self.__in_dumps[0].get_uuid(), self.__out_r0_date)
else:
self.outDump.create_with_rev_n(self.__out_file,
self.__in_dumps[0].get_uuid(), self.outStartRev)
# skip revision 0 of all dumps
for inDump in self.__in_dumps:
if inDump.get_rev_nr() == 0:
# +++ what about r0 revprops?
inDump.read_next_rev()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
self.outDump.close()
return
# get revision dates
oldest = None
oldestStr = ""
for index in range(len(self.__in_dumps)):
revDat = self.__in_dumps[index].get_rev_date()
self.__in_rev_dates.append(revDat)
if oldest is None or revDat < oldest:
oldest = revDat
oldestStr = self.__in_dumps[index].get_rev_date_str()
# add additional directories
if len(self.__out_dirs) > 0:
self.outDump.add_rev({"svn:log": self.__out_message,
"svn:author": self.__out_author,
"svn:date": oldestStr})
for dirName in self.__out_dirs:
node = SvnDumpNode(dirName, "add", "dir")
self.outDump.add_node(node)
# loop over all revisions
while dumpCount > 0:
# find index of the oldest revision
oldestIndex = 0
for index in range(1, dumpCount):
if self.__in_rev_dates[index] < self.__in_rev_dates[oldestIndex]:
oldestIndex = index
# copy revision
self.__copy_revision(oldestIndex)
print("Revision: %-8d from r%-8d %s" % (self.outDump.get_rev_nr(),
self.__in_dumps[oldestIndex].get_rev_nr(),
self.__in_files[oldestIndex]))
# read next revision
srcDump = self.__in_dumps[oldestIndex]
if srcDump.read_next_rev():
self.__in_rev_dates[oldestIndex] = srcDump.get_rev_date()
else:
dumpCount = self.__remove_empty_dumps()
# close output
print("created %d revisions" % self.outDump.get_rev_nr())
self.outDump.close()
def execute(self):
"""
Executes the EolFix.
"""
# +++ catch exception and return errorcode
srcdmp = SvnDumpFile()
srcdmp.open(self.__in_file)
dstdmp = None
hasrev = srcdmp.read_next_rev()
if hasrev:
if not self.__dry_run:
dstdmp = SvnDumpFile()
if srcdmp.get_rev_nr() == 0:
# create new dump with revision 0
dstdmp.create_with_rev_0(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_date_str())
hasrev = srcdmp.read_next_rev()
else:
# create new dump starting with the same revNr
# as the input dump file
dstdmp.create_with_rev_n(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_nr())
# now copy all the revisions
while hasrev:
print("\n\n*** r%d ***\n" % srcdmp.get_rev_nr())
self.__process_rev(srcdmp, dstdmp)
hasrev = srcdmp.read_next_rev()
if self.__warning_file is not None:
self.__warning_file.write(
"\n\n# %d warnings\n" % self.__warning_count)
self.__warning_file.close()
self.__warning_file = None
self.__warning_count = 0
def execute(self, dumpfilename):
"""
Print file list of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
# pass 1: search copy-from revisions
dump = SvnDumpFile()
dump.open(dumpfilename)
copyfromrevs = {}
filedict = {}
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr > self.revNr:
break
for node in dump.get_nodes_iter():
#action = actions[node.get_action()]
action = node.get_action()
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if action == "add" and node.has_copy_from():
copyfromrevs[node.get_copy_from_rev()] = True
dump.close()
# pass 2: do the work
dump = SvnDumpFile()
dump.open(dumpfilename)
filedict = {}
prevrevnr = 0
while dump.read_next_rev():
revnr = dump.get_rev_nr()
# loop over missing revisions
prevrevnr += 1
while prevrevnr < revnr:
if copyfromrevs.has_key(prevrevnr):
copyfromrevs[prevrevnr] = filedict.keys()[:]
prevrevnr += 1
if revnr > self.revNr:
break
for node in dump.get_nodes_iter():
action = node.get_action()
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if action == "add":
filedict[path] = path
if node.has_copy_from():
frompath = node.get_copy_from_path() + "/"
if frompath[0] != "/":
frompath = "/" + frompath
fromlen = len(frompath)
topath = path + "/"
for path in copyfromrevs[node.get_copy_from_rev()]:
if path.startswith(frompath):
newpath = topath + path[fromlen:]
filedict[newpath] = newpath
elif action == "delete":
del filedict[path]
if path[-1] != "/":
path = path + "/"
for subpath in filedict.keys()[:]:
if subpath.startswith(path):
del filedict[subpath]
if copyfromrevs.has_key(revnr):
copyfromrevs[revnr] = filedict.keys()[:]
dump.close()
filelist = []
for path in filedict:
filelist.append(path)
filelist.sort()
for path in filelist:
print path
return 0
def execute( self, inputfile, outputfile ):
"""
Fix a cvs2svn created dump file.
@type inputfile: string
@param inputfile: Name of the cvs2svn created dump file.
@type outputfile: string
@param outputfile: Name of the fixed dump file.
"""
indump = SvnDumpFile()
indump.open( inputfile )
has_rev = indump.read_next_rev()
outdump = SvnDumpFile()
outdump.create_like( outputfile, indump )
rc = 0
self.__history = {}
while has_rev and rc == 0:
outdump.add_rev( indump.get_rev_props() )
for node in indump.get_nodes_iter():
msglist = self.__fix_node( indump.get_rev_nr(), node )
if msglist != None:
rc = 1
print "Error in r%d node %s:" % \
( indump.get_rev_nr(), node.get_path() )
for msg in msglist:
print " " + msg
break
outdump.add_node( node )
has_rev = indump.read_next_rev()
indump.close()
return rc
def execute(self, inputfile, outputfile):
"""
Fix a cvs2svn created dump file.
@type inputfile: string
@param inputfile: Name of the cvs2svn created dump file.
@type outputfile: string
@param outputfile: Name of the fixed dump file.
"""
indump = SvnDumpFile()
indump.open(inputfile)
has_rev = indump.read_next_rev()
outdump = SvnDumpFile()
outdump.create_like(outputfile, indump)
rc = 0
self.__history = {}
while has_rev and rc == 0:
outdump.add_rev(indump.get_rev_props())
for node in indump.get_nodes_iter():
msglist = self.__fix_node(indump.get_rev_nr(), node)
if msglist != None:
rc = 1
print "Error in r%d node %s:" % \
( indump.get_rev_nr(), node.get_path() )
for msg in msglist:
print " " + msg
break
outdump.add_node(node)
has_rev = indump.read_next_rev()
indump.close()
return rc
def execute(self, callback):
"""
Execute the diff.
@type callback: SvnDumpDiffCallback
@param callback: Callback object for diffs found.
"""
# diff started
callback.comparing(self.__filename1, self.__filename2)
# open files
dump1 = SvnDumpFile()
dump2 = SvnDumpFile()
dump1.open(self.__filename1)
dump2.open(self.__filename2)
# compare uuid
if dump1.get_uuid() != dump2.get_uuid():
callback.rev_diff("UUID", dump1.get_uuid(), dump2.get_uuid())
hasrev1 = dump1.read_next_rev()
hasrev2 = dump2.read_next_rev()
while hasrev1 and hasrev2:
# compare rev numbers
if dump1.get_rev_nr() != dump2.get_rev_nr():
callback.rev_diff("RevNr", dump1.get_rev_nr(), dump2.get_rev_nr())
# error has been reported so set both flags to false
hasrev1 = False
hasrev2 = False
break
# next revision...
callback.next_revision(dump1.get_rev_nr(), dump2.get_rev_nr())
# compare rev date
if dump1.get_rev_date() != dump2.get_rev_date():
callback.rev_diff("RevDate",
str(dump1.get_rev_date()),
str(dump2.get_rev_date()))
if dump1.get_rev_date_str() != dump2.get_rev_date_str():
callback.rev_diff("RevDateStr", dump1.get_rev_date_str(), dump2.get_rev_date_str())
# compare rev author
# compare rev log
# compare rev props
self.__compare_properties(True, dump1.get_rev_props(),
dump2.get_rev_props(), callback)
# compare nodes
self.__compare_nodes(dump1, dump2, callback)
# read next revision
hasrev1 = dump1.read_next_rev()
hasrev2 = dump2.read_next_rev()
if hasrev1 or hasrev2:
print("random error ;-) (different rev nr or EOF of one file)")
# done.
callback.compare_done()
def copy_dump_file( srcfile, dstfile, transformer=None ):
"""
Copy a dump file.
@type srcfile: string
@param srcfile: Source filename.
@type dstfile: string
@param dstfile: Destination filename.
@type transformer: class with method transform(dump)
@param transformer: A class to perform a transformation on each revision, or None.
"""
# SvnDumpFile classes for reading/writing dumps
srcdmp = SvnDumpFile()
dstdmp = SvnDumpFile()
# open source file
srcdmp.open( srcfile )
hasrev = srcdmp.read_next_rev()
if hasrev:
# create the dump file
dstdmp.create_like( dstfile, srcdmp )
# now copy all the revisions
while hasrev:
if transformer != None:
transformer.transform( srcdmp )
dstdmp.add_rev_from_dump( srcdmp )
hasrev = srcdmp.read_next_rev()
else:
print "no revisions in the source dump '%s' ???" % srcfile
# cleanup
srcdmp.close()
dstdmp.close()
class SvnDumpMerge:
"""
A class for merging svn dump files.
"""
# handle copyfrom-rev !!!
def __init__(self):
"""
Initialize.
"""
# output file name
self.__out_file = ""
# date for revision 0
self.__out_r0_date = "9999-99-99T99:99:99.999999Z"
# list additional directories
self.__out_dirs = []
# log message for directory creating revision
self.__out_message = ""
# author for the additional revision
self.__out_author = "svndumpmerge"
# variables used for input dump files
# file names
self.__in_files = []
# path renames [ [ ( from, to ), ... ], ... ]
self.__in_renames = []
# path regex substitutions [ [ ( search, replace ), ... ], ... ]
self.__in_regex_subs = []
# mkdir excludes [ {}, ... ]
self.__in_excludes = []
# revision number mappings [ {}, ... ]
self.__in_rev_nr_maps = []
# dump files (class SvnDumpFile)
self.__in_dumps = []
# revision dates of the dumps
self.__in_rev_dates = []
def set_output_file(self, filename, startRev=0):
"""
Sets the output file name and optional start revision.
@type filename: string
@param filename: Name of the output dump file.
@type startRev: integer
@param startRev: Start revision number, default is 0.
"""
self.__out_file = filename
self.outStartRev = startRev
def add_input_file(self, filename):
"""
Adds an input file and returns it's index.
@type filename: string
@param filename: Name of a input dump file.
@rtype: integer
@return: Index of the input file.
"""
index = len(self.__in_files)
self.__in_files = self.__in_files + [filename]
self.__in_renames = self.__in_renames + [[]]
self.__in_regex_subs = self.__in_regex_subs + [[]]
self.__in_excludes = self.__in_excludes + [{}]
self.__in_rev_nr_maps = self.__in_rev_nr_maps + [{}]
return index
def add_rename(self, index, prefixFrom, prefixTo):
"""
Adds a path prefix reanme.
@type index: integer
@param index: Index of the dump file.
@type prefixFrom: string
@param prefixFrom: From-path prefix (directory).
@type prefixTo: string
@param prefixTo: To-path prefix (directory).
"""
# make sure that prefixFrom starts and ends with a /
if prefixFrom[0:1] == "/":
prefixFrom = prefixFrom[1:]
if prefixFrom[len(prefixFrom) - 1:] != "/":
prefixFrom = prefixFrom + "/"
# make sure that prefixTo starts and ends with a /
if prefixTo[0:1] == "/":
prefixTo = prefixTo[1:]
if prefixTo[len(prefixTo) - 1:] != "/":
prefixTo = prefixTo + "/"
# add the rename
self.__in_renames[index] = self.__in_renames[index] + \
[ (prefixFrom, prefixTo ) ]
def add_regex_sub(self, index, reSearch, reReplace):
"""
Adds a path prefix rename.
@type index: integer
@param index: Index of the dump file.
@type reSearch: string
@param reSearch: Search regular expression.
@type reReplace: string
@param reReplace: Replace regular expression.
"""
# compile regex object
reSearch = re.compile(reSearch)
# add the rename
self.__in_regex_subs[index] = self.__in_regex_subs[index] + \
[ (reSearch, reReplace ) ]
def add_mkdir_exclude(self, index, dirName):
"""
Adds a mkdir exclude.
@type index: integer
@param index: Index of the dump file.
@type dirName: string
@param dirName: Name of the directory.
"""
# add the mkdir exclude
self.__in_excludes[index][dirName] = None
def add_directory(self, dirName):
"""
Adds an additional directory ('mkdir').
@type dirName: string
@param dirName: Name of the directory.
"""
if dirName[0:1] == "/":
dirName = dirName[1:]
if dirName[-1:] == "/":
dirName = dirName[:-1]
self.__out_dirs = self.__out_dirs + [dirName]
def set_log_message(self, msg):
"""
Set log message for additional dirs revision.
@type msg: string
@param msg: Log message.
"""
self.__out_message = msg
def merge(self):
"""
Executes the merge.
"""
if len(self.__in_files) == 0:
print "merge: no input files specified"
return
if len(self.__out_file) == 0:
print "merge: no output file specified"
return
# open input dump files
for inFile in self.__in_files:
inDump = SvnDumpFile()
inDump.open(inFile)
inDump.read_next_rev()
self.__in_dumps = self.__in_dumps + [inDump]
if inDump.get_rev_date_str() < self.__out_r0_date:
self.__out_r0_date = inDump.get_rev_date_str()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
return
# open output file
self.outDump = SvnDumpFile()
if self.outStartRev == 0:
self.outDump.create_with_rev_0(self.__out_file,
self.__in_dumps[0].get_uuid(),
self.__out_r0_date)
else:
self.outDump.create_with_rev_n(self.__out_file,
self.__in_dumps[0].get_uuid(),
self.outStartRev)
# skip revision 0 of all dumps
for inDump in self.__in_dumps:
if inDump.get_rev_nr() == 0:
# +++ what about r0 revprops?
inDump.read_next_rev()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
self.outDump.close()
return
# get revision dates
oldest = None
oldestStr = ""
for index in range(len(self.__in_dumps)):
revDat = self.__in_dumps[index].get_rev_date()
self.__in_rev_dates.append(revDat)
if oldest == None or revDat < oldest:
oldest = revDat
oldestStr = self.__in_dumps[index].get_rev_date_str()
# add additional directories
if len(self.__out_dirs) > 0:
self.outDump.add_rev({
"svn:log": self.__out_message,
"svn:author": self.__out_author,
"svn:date": oldestStr
})
for dirName in self.__out_dirs:
node = SvnDumpNode(dirName, "add", "dir")
self.outDump.add_node(node)
# loop over all revisions
while dumpCount > 0:
# find index of the oldest revision
oldestIndex = 0
for index in range(1, dumpCount):
if self.__in_rev_dates[index] < self.__in_rev_dates[
oldestIndex]:
oldestIndex = index
# copy revision
self.__copy_revision(oldestIndex)
print "Revision: %-8d from r%-8d %s" % (
self.outDump.get_rev_nr(),
self.__in_dumps[oldestIndex].get_rev_nr(),
self.__in_files[oldestIndex])
# read next revision
srcDump = self.__in_dumps[oldestIndex]
if srcDump.read_next_rev():
self.__in_rev_dates[oldestIndex] = srcDump.get_rev_date()
else:
dumpCount = self.__remove_empty_dumps()
# close output
print "created %d revisions" % self.outDump.get_rev_nr()
self.outDump.close()
def __copy_revision(self, dumpIndex):
"""
Copies a revision from inDump[dumpIndex] to outDump.
@type dumpIndex: integer
@param dumpIndex: Index of the input dump file.
"""
srcDump = self.__in_dumps[dumpIndex]
# add revision and revprops
self.outDump.add_rev(srcDump.get_rev_props())
# add nodes
index = 0
nodeCount = srcDump.get_node_count()
while index < nodeCount:
node = srcDump.get_node(index)
newNode = self.__change_node(dumpIndex, node)
if newNode != None:
self.outDump.add_node(newNode)
index = index + 1
# add revision info
self.__in_rev_nr_maps[dumpIndex][srcDump.get_rev_nr()] = \
self.outDump.get_rev_nr()
def __change_node(self, dumpIndex, node):
"""
Creates a new node if the path changed, else returns the old node.
@type dumpIndex: integer
@param dumpIndex: Index of the input dump file.
@type node: SvnDumpNode
@param node: A node.
"""
path = node.get_path()
# mkdir exclude check
if node.get_kind() == "dir" and node.get_action() == "add":
if path in self.__in_excludes[dumpIndex]:
return None
fromPath = ""
fromRev = 0
if node.has_copy_from():
fromPath = node.get_copy_from_path()
fromRev = node.get_copy_from_rev()
change = 0
newPath = self.__rename_path(path, dumpIndex)
newFromPath = fromPath
newFromRev = fromRev
if path != newPath:
change = 1
if fromRev > 0:
newFromPath = self.__rename_path(fromPath, dumpIndex)
if fromPath != newFromPath:
change = 1
newFromRev = self.__in_rev_nr_maps[dumpIndex][fromRev]
if fromRev != newFromRev:
change = 1
newMergeInfo = ""
if node.has_properties():
properties = node.get_properties()
if properties.has_key('svn:mergeinfo'):
mergeInfo = properties['svn:mergeinfo']
for line in mergeInfo.split('\n'):
m = re.match('^(.*):(\d+)-(\d+)', line)
if m != None:
mergePath = m.group(1)
mergeFrom = int(m.group(2))
mergeTo = int(m.group(3))
newMergePath = self.__rename_path(mergePath, dumpIndex)
newMergeFrom = self.__in_rev_nr_maps[dumpIndex][
mergeFrom]
newMergeTo = self.__in_rev_nr_maps[dumpIndex][mergeTo]
if len(newMergeInfo) != 0:
newMergeInfo = newMergeInfo + "\n"
newMergeInfo = newMergeInfo + newMergePath + ":" + str(
newMergeFrom) + "-" + str(newMergeTo)
if mergeInfo != newMergeInfo:
change = 1
if not change:
# no change needed
return node
# do the rename
newNode = SvnDumpNode(newPath, node.get_action(), node.get_kind())
if node.has_copy_from():
newNode.set_copy_from(newFromPath, newFromRev)
if node.has_properties():
newNode.set_properties(node.get_properties())
if len(newMergeInfo) > 0:
newNode.set_property('svn:mergeinfo', newMergeInfo)
if node.has_text():
newNode.set_text_node(node)
return newNode
def __rename_path(self, path, dumpIndex):
"""
Applies the renames to the path and returns the new path.
@type path: string
@param path: A path.
@type renames: list( ( string, string ) )
@param renames: List of rename tuples.
@rtype: string
@return Renamed path.
"""
# ensure that path does not have a leading slash
if len(path) > 1 and path[0:1] == "/":
path = path[1:]
sPath = path + "/"
for sPfx, dPfx in self.__in_renames[dumpIndex]:
sLen = len(sPfx)
if sPfx == "/":
return dPfx + path
elif sPath[:sLen] == sPfx:
if len(path) <= len(sPfx):
# it's the full path
return dPfx[0:len(dPfx) - 1]
else:
# there's a suffix
return dPfx + path[sLen:]
for reSearch, sReplace in self.__in_regex_subs[dumpIndex]:
path = reSearch.sub(sReplace, path, count=1)
return path
def __remove_empty_dumps(self):
"""
Removes dump files which reached EOF and returns the count of dumps.
@rtype: integer
@return: Count of remaining input dump files.
"""
index = 0
while index < len(self.__in_dumps):
inDump = self.__in_dumps[index]
if inDump.has_revision():
index = index + 1
else:
inDump.close()
eidx = index + 1
self.__in_files[index:eidx] = []
self.__in_renames[index:eidx] = []
self.__in_regex_subs[index:eidx] = []
self.__in_excludes[index:eidx] = []
self.__in_rev_nr_maps[index:eidx] = []
self.__in_dumps[index:eidx] = []
self.__in_rev_dates[index:eidx] = []
return index
def merge(self):
"""
Executes the merge.
"""
if len(self.__in_files) == 0:
print "merge: no input files specified"
return
if len(self.__out_file) == 0:
print "merge: no output file specified"
return
# open input dump files
for inFile in self.__in_files:
inDump = SvnDumpFile()
inDump.open(inFile)
inDump.read_next_rev()
self.__in_dumps = self.__in_dumps + [inDump]
if inDump.get_rev_date_str() < self.__out_r0_date:
self.__out_r0_date = inDump.get_rev_date_str()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
return
# open output file
self.outDump = SvnDumpFile()
if self.outStartRev == 0:
self.outDump.create_with_rev_0(self.__out_file,
self.__in_dumps[0].get_uuid(),
self.__out_r0_date)
else:
self.outDump.create_with_rev_n(self.__out_file,
self.__in_dumps[0].get_uuid(),
self.outStartRev)
# skip revision 0 of all dumps
for inDump in self.__in_dumps:
if inDump.get_rev_nr() == 0:
# +++ what about r0 revprops?
inDump.read_next_rev()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
self.outDump.close()
return
# get revision dates
oldest = None
oldestStr = ""
for index in range(len(self.__in_dumps)):
revDat = self.__in_dumps[index].get_rev_date()
self.__in_rev_dates.append(revDat)
if oldest == None or revDat < oldest:
oldest = revDat
oldestStr = self.__in_dumps[index].get_rev_date_str()
# add additional directories
if len(self.__out_dirs) > 0:
self.outDump.add_rev({
"svn:log": self.__out_message,
"svn:author": self.__out_author,
"svn:date": oldestStr
})
for dirName in self.__out_dirs:
node = SvnDumpNode(dirName, "add", "dir")
self.outDump.add_node(node)
# loop over all revisions
while dumpCount > 0:
# find index of the oldest revision
oldestIndex = 0
for index in range(1, dumpCount):
if self.__in_rev_dates[index] < self.__in_rev_dates[
oldestIndex]:
oldestIndex = index
# copy revision
self.__copy_revision(oldestIndex)
print "Revision: %-8d from r%-8d %s" % (
self.outDump.get_rev_nr(),
self.__in_dumps[oldestIndex].get_rev_nr(),
self.__in_files[oldestIndex])
# read next revision
srcDump = self.__in_dumps[oldestIndex]
if srcDump.read_next_rev():
self.__in_rev_dates[oldestIndex] = srcDump.get_rev_date()
else:
dumpCount = self.__remove_empty_dumps()
# close output
print "created %d revisions" % self.outDump.get_rev_nr()
self.outDump.close()
def copy_dump_file( srcfile, dstfile, transformer=None ):
"""
Copy a dump file.
@type srcfile: string
@param srcfile: Source filename.
@type dstfile: string
@param dstfile: Destination filename.
@type transformer: class with method transform(dump)
@param transformer: A class to perform a transformation on each revision, or None.
"""
# SvnDumpFile classes for reading/writing dumps
srcdmp = SvnDumpFile()
dstdmp = SvnDumpFile()
# Copy_from_rev casacading
oldRevToNewRev = dict()
# open source file
srcdmp.open( srcfile )
hasrev = srcdmp.read_next_rev()
if hasrev:
# create the dump file
dstdmp.create_like( dstfile, srcdmp )
# now copy all the revisions
while hasrev:
if transformer != None:
transformer.transform( srcdmp )
for node in srcdmp.get_nodes_iter():
if node.has_copy_from():
if oldRevToNewRev.has_key(node.get_copy_from_rev()):
node.set_copy_from_rev(oldRevToNewRev[node.get_copy_from_rev()])
else:
#We have a problem, the copy from revision is missing.
#We look for a previous revision containing the file
found = False
candidate = node.get_copy_from_rev()
while candidate > 0 and not found:
candidate = candidate - 1
found = oldRevToNewRev.has_key(candidate)
if found:
oldRevToNewRev[node.get_copy_from_rev()] = candidate
node.set_copy_from_rev(candidate)
dstdmp.add_rev_from_dump( srcdmp )
oldRevToNewRev[srcdmp.get_rev_nr()] = dstdmp.get_rev_nr()
hasrev = srcdmp.read_next_rev()
else:
print "no revisions in the source dump '%s' ???" % srcfile
# cleanup
srcdmp.close()
dstdmp.close()
def filter_dump_file(self, srcfile, dstfile ):
# SvnDumpFile classes for reading/writing dumps
srcdmp = SvnDumpFile()
dstdmp = SvnDumpFile()
# read a first time source file to find all referenced revisions (they should not be later removed)
srcdmp.open( srcfile )
hasrev = srcdmp.read_next_rev()
referenced_revs = set()
while hasrev:
for node in srcdmp.get_nodes_iter():
copy_from_rev = node.get_copy_from_rev()
if copy_from_rev != 0:
referenced_revs.add(copy_from_rev)
hasrev = srcdmp.read_next_rev()
srcdmp.close()
# start reading again source file
srcdmp.open( srcfile )
hasrev = srcdmp.read_next_rev()
if hasrev:
# create the dump file
dstdmp.create_like( dstfile, srcdmp )
# now copy all the revisions
while hasrev:
if self.verbosity > 0:
print "Revision-number: orig=%s, new=%s" % (srcdmp.get_rev_nr(), srcdmp.get_rev_nr()-self.skipped_revisions)
if srcdmp.get_rev_log() == "This is an empty revision for padding." and srcdmp.get_rev_nr() not in referenced_revs:
if self.verbosity > 0:
print " * skipped padding revison (number of skipped revisions = %s)" % self.skipped_revisions
self.skipped_revisions += 1
else:
self.transform( srcdmp )
self.rev_map[srcdmp.get_rev_nr()] = (srcdmp.get_rev_nr() - self.skipped_revisions)
dstdmp.add_rev_from_dump( srcdmp )
hasrev = srcdmp.read_next_rev()
else:
print "no revisions in the source dump '%s' ???" % srcfile
# cleanup
srcdmp.close()
dstdmp.close()
def execute(self, dumpfilename):
"""
Print log of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
print "\n\n" + "=" * 72
line = "-" * 72
print "Dumpfile: " + dumpfilename
dump = SvnDumpFile()
dump.open(dumpfilename)
actions = {"add": "A", "change": "M", "delete": "D", "replace": "R"}
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr >= self.__from_rev and revnr <= self.__to_rev:
author = dump.get_rev_author()
date = dump.get_rev_date_str()
log = dump.get_rev_log()
linecnt = len(log.split("\n"))
lines = "%d line" % linecnt
if linecnt > 1:
lines += "s"
print line
print "r%d | %s | %s | %s" % (revnr, author, date, lines)
if self.__verbose:
print "Changed paths:"
for node in dump.get_nodes_iter():
action = actions[node.get_action()]
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if node.has_copy_from():
fpath = node.get_copy_from_path()
frev = node.get_copy_from_rev()
if fpath == "" or fpath[0] != "/":
fpath = "/" + fpath
path += " (from %s:%d)" % (fpath, frev)
print " %s %s" % (action, path)
print "\n" + log.rstrip() + "\n"
print line
dump.close()
return 0
def split_dumpfiles(inputfilename, outlist):
"""
Splits a dump file.
@type inputfilename: string
@param inputfilename: Name of the input file.
@type outlist: list
@param outlist: List of tuples containing start revnr, end revnr and filename.
@rtype: int
@return: 0 for success.
"""
if len(outlist) == 0:
return 0
outlist = outlist[:]
outlist.sort()
parallel = False
for i in range(0, len(outlist) - 1):
if outlist[i][1] > outlist[i + 1][0]:
parallel = True
break
if not parallel:
indump = SvnDumpFile()
indump.open(inputfilename)
index = 0
startrev = outlist[index][0]
endrev = outlist[index][1]
outfile = outlist[index][2]
outdump = None
while indump.read_next_rev():
revnr = indump.get_rev_nr()
if outdump == None:
if revnr >= startrev:
outdump = SvnDumpFile()
if revnr == 0:
# create new dump with revision 0
outdump.create_with_rev_0(outfile, indump.get_uuid(),
indump.get_rev_date_str())
else:
# create new dump starting with the
# same revNr as the original dump
outdump.create_with_rev_n(outfile, indump.get_uuid(),
indump.get_rev_nr())
if outdump != None:
# have an output file, copy the revision if revnr > 0
if revnr > 0:
outdump.add_rev_from_dump(indump)
if revnr >= endrev:
# end revision reached
outdump.close()
outdump = None
index += 1
if index >= len(outlist):
# done.
break
# next range
startrev = outlist[index][0]
endrev = outlist[index][1]
outfile = outlist[index][2]
if outdump != None:
outdump.close()
indump.close()
else:
print "overlapping revision ranges not supported (yet)."
return 1
return 0
def copy_dump_file(srcfile, dstfile, transformer=None):
"""
Copy a dump file.
@type srcfile: string
@param srcfile: Source filename.
@type dstfile: string
@param dstfile: Destination filename.
@type transformer: class with method transform(dump)
@param transformer: A class to perform a transformation on each revision, or None.
"""
# SvnDumpFile classes for reading/writing dumps
srcdmp = SvnDumpFile()
dstdmp = SvnDumpFile()
# Copy_from_rev casacading
oldRevToNewRev = dict()
# open source file
srcdmp.open(srcfile)
hasrev = srcdmp.read_next_rev()
if hasrev:
# create the dump file
dstdmp.create_like(dstfile, srcdmp)
# now copy all the revisions
while hasrev:
if transformer != None:
transformer.transform(srcdmp)
for node in srcdmp.get_nodes_iter():
if node.has_copy_from():
if oldRevToNewRev.has_key(node.get_copy_from_rev()):
node.set_copy_from_rev(
oldRevToNewRev[node.get_copy_from_rev()])
else:
#We have a problem, the copy from revision is missing.
#We look for a previous revision containing the file
found = False
candidate = node.get_copy_from_rev()
while candidate > 0 and not found:
candidate = candidate - 1
found = oldRevToNewRev.has_key(candidate)
if found:
oldRevToNewRev[
node.get_copy_from_rev()] = candidate
node.set_copy_from_rev(candidate)
dstdmp.add_rev_from_dump(srcdmp)
oldRevToNewRev[srcdmp.get_rev_nr()] = dstdmp.get_rev_nr()
hasrev = srcdmp.read_next_rev()
else:
print "no revisions in the source dump '%s' ???" % srcfile
# cleanup
srcdmp.close()
dstdmp.close()
def join_dumpfiles(inputlist, outfilename):
"""
Joins dump files.
@type inputlist: list
@param inputlist: A list containing the input filenames.
@type outfilename: string
@param outfilename: Name of the output dump file.
@rtype: int
@return: 0 for success.
"""
outdump = None
noutrev = 0
lastrev = -1
for filename in inputlist:
print "reading %s ..." % filename
ninrev = 0
indump = SvnDumpFile()
indump.open(filename)
hasrev = indump.read_next_rev()
if hasrev:
if outdump == None:
outdump = SvnDumpFile()
if indump.get_rev_nr() == 0:
# create new dump with revision 0
outdump.create_with_rev_0(outfilename, indump.get_uuid(),
indump.get_rev_date_str())
hasrev = indump.read_next_rev()
else:
# create new dump starting with the
# same revNr as the original dump
outdump.create_with_rev_n(outfilename, indump.get_uuid(),
indump.get_rev_nr())
else:
# check rev number
if indump.get_rev_nr() == 0:
hasrev = indump.read_next_rev()
if hasrev:
if (lastrev + 1) != indump.get_rev_nr():
print "renumbering of revisions not supported."
print "last rev was %d, next is %d." % (
lastrev, indump.get_rev_nr())
indump.close()
outdump.close()
return 1
while hasrev:
outdump.add_rev_from_dump(indump)
ninrev += 1
lastrev = indump.get_rev_nr()
hasrev = indump.read_next_rev()
indump.close()
print " copied %d revisions." % ninrev
noutrev += ninrev
outdump.close()
print "wrote %d revisions, last was r%d." % (noutrev, lastrev)
def execute( self, dumpfilename ):
"""
Print file list of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
# pass 1: search copy-from revisions
dump = SvnDumpFile()
dump.open( dumpfilename )
copyfromrevs = {}
filedict = {}
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr > self.revNr:
break
for node in dump.get_nodes_iter():
#action = actions[node.get_action()]
action = node.get_action()
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if action == "add" and node.has_copy_from():
copyfromrevs[node.get_copy_from_rev()] = True
dump.close()
# pass 2: do the work
dump = SvnDumpFile()
dump.open( dumpfilename )
filedict = {}
prevrevnr = 0
while dump.read_next_rev():
revnr = dump.get_rev_nr()
# loop over missing revisions
prevrevnr += 1
while prevrevnr < revnr:
if copyfromrevs.has_key(prevrevnr):
copyfromrevs[prevrevnr] = filedict.keys()[:]
prevrevnr += 1
if revnr > self.revNr:
break
for node in dump.get_nodes_iter():
action = node.get_action()
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if action == "add":
filedict[path] = path
if node.has_copy_from():
frompath = node.get_copy_from_path() + "/"
if frompath[0] != "/":
frompath = "/" + frompath
fromlen = len(frompath)
topath = path + "/"
for path in copyfromrevs[node.get_copy_from_rev()]:
if path.startswith( frompath ):
newpath = topath + path[fromlen:]
filedict[newpath] = newpath
elif action == "delete":
del filedict[path]
if path[-1] != "/":
path = path + "/"
for subpath in filedict.keys()[:]:
if subpath.startswith( path ):
del filedict[subpath]
if copyfromrevs.has_key(revnr):
copyfromrevs[revnr] = filedict.keys()[:]
dump.close()
filelist = []
for path in filedict:
filelist.append( path )
filelist.sort()
for path in filelist:
print path
return 0
def execute(self):
"""
Executes the Edit.
"""
# +++ catch exception and return errorcode
srcdmp = SvnDumpFile()
srcdmp.open(self.__in_file)
dstdmp = None
hasrev = srcdmp.read_next_rev()
if hasrev:
if not self.dry_run:
dstdmp = SvnDumpFile()
if srcdmp.get_rev_nr() == 0:
# create new dump with revision 0
dstdmp.create_with_rev_0(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_date_str())
hasrev = srcdmp.read_next_rev()
else:
# create new dump starting with the same revNr
# as the input dump file
dstdmp.create_with_rev_n(self.__out_file,
srcdmp.get_uuid(),
srcdmp.get_rev_nr())
# now copy all the revisions
while hasrev:
self.__process_rev(srcdmp, dstdmp)
hasrev = srcdmp.read_next_rev()
def execute( self, dumpfilename ):
"""
Print log of a dump file.
@type dumpfilename: string
@param dumpfilename: Name of the file to log.
"""
print "\n\n" + "=" * 72
line = "-" * 72
print "Dumpfile: " + dumpfilename
dump = SvnDumpFile()
dump.open( dumpfilename )
actions = { "add":"A", "change":"M", "delete":"D", "replace":"R" }
while dump.read_next_rev():
revnr = dump.get_rev_nr()
if revnr >= self.__from_rev and revnr <= self.__to_rev:
author = dump.get_rev_author()
date = dump.get_rev_date_str()
log = dump.get_rev_log()
linecnt = len( log.split( "\n" ) )
lines = "%d line" % linecnt
if linecnt > 1:
lines += "s"
print line
print "r%d | %s | %s | %s" % ( revnr, author, date, lines )
if self.__verbose:
print "Changed paths:"
for node in dump.get_nodes_iter():
action = actions[node.get_action()]
path = node.get_path()
if path == "" or path[0] != "/":
path = "/" + path
if node.has_copy_from():
fpath = node.get_copy_from_path()
frev = node.get_copy_from_rev()
if fpath == "" or fpath[0] != "/":
fpath = "/" + fpath
path += " (from %s:%d)" % ( fpath, frev )
print " %s %s" % ( action, path )
print "\n" + log.rstrip() + "\n"
print line
dump.close()
return 0
class SvnDumpMerge:
"""
A class for merging svn dump files.
"""
# handle copyfrom-rev !!!
def __init__(self):
"""
Initialize.
"""
# output file name
self.__out_file = ""
# date for revision 0
self.__out_r0_date = "9999-99-99T99:99:99.999999Z"
# list additional directories
self.__out_dirs = []
# log message for directory creating revision
self.__out_message = ""
# author for the additional revision
self.__out_author = "svndumpmerge"
# variables used for input dump files
# file names
self.__in_files = []
# path renames [ [ ( from, to ), ... ], ... ]
self.__in_renames = []
# path regex substitutions [ [ ( search, replace ), ... ], ... ]
self.__in_regex_subs = []
# mkdir excludes [ {}, ... ]
self.__in_excludes = []
# revision number mappings [ {}, ... ]
self.__in_rev_nr_maps = []
# dump files (class SvnDumpFile)
self.__in_dumps = []
# revision dates of the dumps
self.__in_rev_dates = []
def set_output_file(self, filename, startRev=0):
"""
Sets the output file name and optional start revision.
@type filename: string
@param filename: Name of the output dump file.
@type startRev: integer
@param startRev: Start revision number, default is 0.
"""
self.__out_file = filename
self.outStartRev = startRev
def add_input_file(self, filename):
"""
Adds an input file and returns it's index.
@type filename: string
@param filename: Name of a input dump file.
@rtype: integer
@return: Index of the input file.
"""
index = len(self.__in_files)
self.__in_files = self.__in_files + [filename]
self.__in_renames = self.__in_renames + [[]]
self.__in_regex_subs = self.__in_regex_subs + [[]]
self.__in_excludes = self.__in_excludes + [{}]
self.__in_rev_nr_maps = self.__in_rev_nr_maps + [{}]
return index
def add_rename(self, index, prefixFrom, prefixTo):
"""
Adds a path prefix reanme.
@type index: integer
@param index: Index of the dump file.
@type prefixFrom: string
@param prefixFrom: From-path prefix (directory).
@type prefixTo: string
@param prefixTo: To-path prefix (directory).
"""
# make sure that prefixFrom starts and ends with a /
if prefixFrom[0:1] == "/":
prefixFrom = prefixFrom[1:]
if prefixFrom[len(prefixFrom) - 1:] != "/":
prefixFrom = prefixFrom + "/"
# make sure that prefixTo starts and ends with a /
if prefixTo[0:1] == "/":
prefixTo = prefixTo[1:]
if prefixTo[len(prefixTo) - 1:] != "/":
prefixTo = prefixTo + "/"
# add the rename
self.__in_renames[index] = self.__in_renames[index] + \
[(prefixFrom, prefixTo)]
def add_regex_sub(self, index, reSearch, reReplace):
"""
Adds a path prefix rename.
@type index: integer
@param index: Index of the dump file.
@type reSearch: string
@param reSearch: Search regular expression.
@type reReplace: string
@param reReplace: Replace regular expression.
"""
# compile regex object
reSearch = re.compile(reSearch)
# add the rename
self.__in_regex_subs[index] = self.__in_regex_subs[index] + \
[(reSearch, reReplace)]
def add_mkdir_exclude(self, index, dirName):
"""
Adds a mkdir exclude.
@type index: integer
@param index: Index of the dump file.
@type dirName: string
@param dirName: Name of the directory.
"""
# add the mkdir exclude
self.__in_excludes[index][dirName] = None
def add_directory(self, dirName):
"""
Adds an additional directory ('mkdir').
@type dirName: string
@param dirName: Name of the directory.
"""
if dirName[0:1] == "/":
dirName = dirName[1:]
if dirName[-1:] == "/":
dirName = dirName[:-1]
self.__out_dirs = self.__out_dirs + [dirName]
def set_log_message(self, msg):
"""
Set log message for additional dirs revision.
@type msg: string
@param msg: Log message.
"""
self.__out_message = msg
def merge(self):
"""
Executes the merge.
"""
if len(self.__in_files) == 0:
print("merge: no input files specified")
return
if len(self.__out_file) == 0:
print("merge: no output file specified")
return
# open input dump files
for inFile in self.__in_files:
inDump = SvnDumpFile()
inDump.open(inFile)
inDump.read_next_rev()
self.__in_dumps = self.__in_dumps + [inDump]
if inDump.get_rev_date_str() < self.__out_r0_date:
self.__out_r0_date = inDump.get_rev_date_str()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
return
# open output file
self.outDump = SvnDumpFile()
if self.outStartRev == 0:
self.outDump.create_with_rev_0(self.__out_file,
self.__in_dumps[0].get_uuid(), self.__out_r0_date)
else:
self.outDump.create_with_rev_n(self.__out_file,
self.__in_dumps[0].get_uuid(), self.outStartRev)
# skip revision 0 of all dumps
for inDump in self.__in_dumps:
if inDump.get_rev_nr() == 0:
# +++ what about r0 revprops?
inDump.read_next_rev()
# remove empty dumps
dumpCount = self.__remove_empty_dumps()
if dumpCount == 0:
self.outDump.close()
return
# get revision dates
oldest = None
oldestStr = ""
for index in range(len(self.__in_dumps)):
revDat = self.__in_dumps[index].get_rev_date()
self.__in_rev_dates.append(revDat)
if oldest is None or revDat < oldest:
oldest = revDat
oldestStr = self.__in_dumps[index].get_rev_date_str()
# add additional directories
if len(self.__out_dirs) > 0:
self.outDump.add_rev({"svn:log": self.__out_message,
"svn:author": self.__out_author,
"svn:date": oldestStr})
for dirName in self.__out_dirs:
node = SvnDumpNode(dirName, "add", "dir")
self.outDump.add_node(node)
# loop over all revisions
while dumpCount > 0:
# find index of the oldest revision
oldestIndex = 0
for index in range(1, dumpCount):
if self.__in_rev_dates[index] < self.__in_rev_dates[oldestIndex]:
oldestIndex = index
# copy revision
self.__copy_revision(oldestIndex)
print("Revision: %-8d from r%-8d %s" % (self.outDump.get_rev_nr(),
self.__in_dumps[oldestIndex].get_rev_nr(),
self.__in_files[oldestIndex]))
# read next revision
srcDump = self.__in_dumps[oldestIndex]
if srcDump.read_next_rev():
self.__in_rev_dates[oldestIndex] = srcDump.get_rev_date()
else:
dumpCount = self.__remove_empty_dumps()
# close output
print("created %d revisions" % self.outDump.get_rev_nr())
self.outDump.close()
def __copy_revision(self, dumpIndex):
"""
Copies a revision from inDump[dumpIndex] to outDump.
@type dumpIndex: integer
@param dumpIndex: Index of the input dump file.
"""
srcDump = self.__in_dumps[dumpIndex]
# add revision and revprops
self.outDump.add_rev(srcDump.get_rev_props())
# add nodes
index = 0
nodeCount = srcDump.get_node_count()
while index < nodeCount:
node = srcDump.get_node(index)
newNode = self.__change_node(dumpIndex, node)
if newNode is not None:
self.outDump.add_node(newNode)
index = index + 1
# add revision info
self.__in_rev_nr_maps[dumpIndex][srcDump.get_rev_nr()] = \
self.outDump.get_rev_nr()
def __change_node(self, dumpIndex, node):
"""
Creates a new node if the path changed, else returns the old node.
@type dumpIndex: integer
@param dumpIndex: Index of the input dump file.
@type node: SvnDumpNode
@param node: A node.
"""
path = node.get_path()
# mkdir exclude check
if node.get_kind() == "dir" and node.get_action() == "add":
if path in self.__in_excludes[dumpIndex]:
return None
fromPath = ""
fromRev = 0
if node.has_copy_from():
fromPath = node.get_copy_from_path()
fromRev = node.get_copy_from_rev()
change = 0
newPath = self.__rename_path(path, dumpIndex)
newFromPath = fromPath
newFromRev = fromRev
if path != newPath:
change = 1
if fromRev > 0:
newFromPath = self.__rename_path(fromPath, dumpIndex)
if fromPath != newFromPath:
change = 1
newFromRev = self.__in_rev_nr_maps[dumpIndex][fromRev]
if fromRev != newFromRev:
change = 1
newMergeInfo = ""
if node.has_properties():
properties = node.get_properties()
if properties.has_key('svn:mergeinfo'):
mergeInfo = properties['svn:mergeinfo']
for line in mergeInfo.split('\n'):
m = re.match('^(.*):(.*)', line)
if m is not None:
mergePath = m.group(1)
revPart = m.group(2)
newMergePath = self.__rename_path(mergePath, dumpIndex)
if not newMergePath.startswith("/"):
newMergePath = "/" + newMergePath
if len(newMergeInfo) != 0:
newMergeInfo = newMergeInfo + "\n"
newMergeInfo = newMergeInfo + newMergePath + ":"
revSep = ""
for rm in re.finditer('(\d+)(?:-(\d+))?,?', revPart):
mergeFrom = int(rm.group(1))
newMergeFrom = self.__in_rev_nr_maps[dumpIndex][mergeFrom]
newMergeInfo = newMergeInfo + revSep + str(newMergeFrom)
revSep = ","
if rm.group(2) is not None:
mergeTo = int(rm.group(2))
newMergeTo = self.__in_rev_nr_maps[dumpIndex][mergeTo]
newMergeInfo = newMergeInfo + "-" + str(newMergeTo)
if mergeInfo != newMergeInfo:
change = 1
if not change:
# no change needed
return node
# do the rename
newNode = SvnDumpNode(newPath, node.get_action(), node.get_kind())
if node.has_copy_from():
newNode.set_copy_from(newFromPath, newFromRev)
if node.has_properties():
newNode.set_properties(node.get_properties())
if len(newMergeInfo) > 0:
newNode.set_property('svn:mergeinfo', newMergeInfo)
if node.has_text():
newNode.set_text_node(node)
return newNode
def __rename_path(self, path, dumpIndex):
"""
Applies the renames to the path and returns the new path.
@type path: string
@param path: A path.
@type renames: list( ( string, string ) )
@param renames: List of rename tuples.
@rtype: string
@return Renamed path.
"""
# ensure that path does not have a leading slash
if len(path) > 1 and path[0:1] == "/":
path = path[1:]
sPath = path + "/"
for sPfx, dPfx in self.__in_renames[dumpIndex]:
sLen = len(sPfx)
if sPfx == "/":
return dPfx + path
elif sPath[:sLen] == sPfx:
if len(path) <= len(sPfx):
# it's the full path
return dPfx[0:len(dPfx) - 1]
else:
# there's a suffix
return dPfx + path[sLen:]
for reSearch, sReplace in self.__in_regex_subs[dumpIndex]:
path = reSearch.sub(sReplace, path, count=1)
return path
def __remove_empty_dumps(self):
"""
Removes dump files which reached EOF and returns the count of dumps.
@rtype: integer
@return: Count of remaining input dump files.
"""
index = 0
while index < len(self.__in_dumps):
inDump = self.__in_dumps[index]
if inDump.has_revision():
index = index + 1
else:
inDump.close()
eidx = index + 1
self.__in_files[index:eidx] = []
self.__in_renames[index:eidx] = []
self.__in_regex_subs[index:eidx] = []
self.__in_excludes[index:eidx] = []
self.__in_rev_nr_maps[index:eidx] = []
self.__in_dumps[index:eidx] = []
self.__in_rev_dates[index:eidx] = []
return index
def copy_dump_file(srcfile, dstfile, transformer=None):
"""
Copy a dump file.
@type srcfile: string
@param srcfile: Source filename.
@type dstfile: string
@param dstfile: Destination filename.
@type transformer: class with method transform(dump)
@param transformer: A class to perform a transformation on each revision, or None.
"""
# SvnDumpFile classes for reading/writing dumps
srcdmp = SvnDumpFile()
dstdmp = SvnDumpFile()
# open source file
srcdmp.open(srcfile)
hasrev = srcdmp.read_next_rev()
if hasrev:
# create the dump file
dstdmp.create_like(dstfile, srcdmp)
# now copy all the revisions
while hasrev:
if transformer != None:
transformer.transform(srcdmp)
dstdmp.add_rev_from_dump(srcdmp)
hasrev = srcdmp.read_next_rev()
else:
print "no revisions in the source dump '%s' ???" % srcfile
# cleanup
srcdmp.close()
dstdmp.close()
