cvs2git-trunk.options

# (Be in -*- mode: python; coding: utf-8 -*- mode.)
#
# ====================================================================
# Copyright (c) 2006-2010 CollabNet.  All rights reserved.
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution.  The terms
# are also available at http://subversion.tigris.org/license-1.html.
# If newer versions of this license are posted there, you may use a
# newer version instead, at your option.
#
# This software consists of voluntary contributions made by many
# individuals.  For exact contribution history, see the revision
# history and logs, available at http://cvs2git.tigris.org/.
# ====================================================================

#                  #####################
#                  ## PLEASE READ ME! ##
#                  #####################
#
# This is a template for an options file that can be used to configure
# cvs2git to convert to git rather than to Subversion.  See
# www/cvs2git.html and www/cvs2git.html for general information, and
# see the comments in this file for information about what options are
# available and how they can be set.
#
# The program that is run to convert from CVS to git is called
# cvs2git.  Run it with the --options option, passing it this file
# like this:
#
#     cvs2git --options=cvs2git-example.options
#
# The output of cvs2git is a blob file and a dump file that can be
# loaded into git using the "git fast-import" command.  Please read
# www/cvs2git.html for more information.
#
# Many options do not have defaults, so it is easier to copy this file
# and modify what you need rather than creating a new options file
# from scratch.  This file is in Python syntax, but you don't need to
# know Python to modify it.  But if you *do* know Python, then you
# will be happy to know that you can use arbitary Python constructs to
# do fancy configuration tricks.
#
# But please be aware of the following:
#
# * In many places, leading whitespace is significant in Python (it is
#   used instead of curly braces to group statements together).
#   Therefore, if you don't know what you are doing, it is best to
#   leave the whitespace as it is.
#
# * In normal strings, Python treats a backslash ("\") as an escape
#   character.  Therefore, if you want to specify a string that
#   contains a backslash, you need either to escape the backslash with
#   another backslash ("\\"), or use a "raw string", as in one if the
#   following equivalent examples:
#
#       cvs_executable = 'c:\\windows\\system32\\cvs.exe'
#       cvs_executable = r'c:\windows\system32\cvs.exe'
#
#   See http://docs.python.org/tutorial/introduction.html#strings for
#   more information.
#
# Two identifiers will have been defined before this file is executed,
# and can be used freely within this file:
#
#     ctx -- a Ctx object (see cvs2svn_lib/context.py), which holds
#         many configuration options
#
#     run_options -- an instance of the GitRunOptions class (see
#         cvs2svn_lib/git_run_options.py), which holds some variables
#         governing how cvs2git is run


# Import some modules that are used in setting the options:
import os

from cvs2svn_lib import config
from cvs2svn_lib import changeset_database
from cvs2svn_lib.common import CVSTextDecoder
from cvs2svn_lib.log import logger
from cvs2svn_lib.project import Project
from cvs2svn_lib.git_revision_collector import GitRevisionCollector
from cvs2svn_lib.external_blob_generator import ExternalBlobGenerator
from cvs2svn_lib.git_output_option import GitRevisionMarkWriter
from cvs2svn_lib.git_output_option import GitOutputOption
from cvs2svn_lib.dvcs_common import KeywordHandlingPropertySetter
from cvs2svn_lib.revision_manager import NullRevisionCollector
from cvs2svn_lib.rcs_revision_manager import RCSRevisionReader
from cvs2svn_lib.cvs_revision_manager import CVSRevisionReader
from cvs2svn_lib.symbol_strategy import AllBranchRule
from cvs2svn_lib.symbol_strategy import AllTagRule
from cvs2svn_lib.symbol_strategy import BranchIfCommitsRule
from cvs2svn_lib.symbol_strategy import ExcludeRegexpStrategyRule
from cvs2svn_lib.symbol_strategy import ForceBranchRegexpStrategyRule
from cvs2svn_lib.symbol_strategy import ForceTagRegexpStrategyRule
from cvs2svn_lib.symbol_strategy import ExcludeTrivialImportBranchRule
from cvs2svn_lib.symbol_strategy import ExcludeVendorBranchRule
from cvs2svn_lib.symbol_strategy import HeuristicStrategyRule
from cvs2svn_lib.symbol_strategy import UnambiguousUsageRule
from cvs2svn_lib.symbol_strategy import HeuristicPreferredParentRule
from cvs2svn_lib.symbol_strategy import SymbolHintsFileRule
from cvs2svn_lib.symbol_transform import ReplaceSubstringsSymbolTransform
from cvs2svn_lib.symbol_transform import RegexpSymbolTransform
from cvs2svn_lib.symbol_transform import IgnoreSymbolTransform
from cvs2svn_lib.symbol_transform import NormalizePathsSymbolTransform
from cvs2svn_lib.property_setters import AutoPropsPropertySetter
from cvs2svn_lib.property_setters import CVSBinaryFileDefaultMimeTypeSetter
from cvs2svn_lib.property_setters import CVSBinaryFileEOLStyleSetter
from cvs2svn_lib.property_setters import DefaultEOLStyleSetter
from cvs2svn_lib.property_setters import EOLStyleFromMimeTypeSetter
from cvs2svn_lib.property_setters import ExecutablePropertySetter
from cvs2svn_lib.property_setters import KeywordsPropertySetter
from cvs2svn_lib.property_setters import MimeMapper
from cvs2svn_lib.property_setters import SVNBinaryFileKeywordsPropertySetter
import csv

# To choose the level of logging output, uncomment one of the
# following lines:
logger.log_level = logger.WARN
#logger.log_level = logger.QUIET
#logger.log_level = logger.NORMAL
#logger.log_level = logger.VERBOSE
#logger.log_level = logger.DEBUG


# The directory to use for temporary files:
ctx.tmpdir = r'tmp-cvs2git'

# During FilterSymbolsPass, cvs2git records the contents of file
# revisions into a "blob" file in git-fast-import format.  The
# ctx.revision_collector option configures that process.  Choose one of the two ersions and customize its options.

# This first alternative is much slower but is better tested and has a
# chance of working with CVSNT repositories.  It invokes CVS or RCS to
# reconstuct the contents of CVS file revisions:
ctx.revision_collector = GitRevisionCollector(
    # The file in which to write the git-fast-import stream that
    # contains the file revision contents:
    './tmp-cvs2git/git-blob.dat',

    # The following option specifies how the revision contents of the
    # RCS files should be read.
    #
    # RCSRevisionReader uses RCS's "co" program to extract the
    # revision contents of the RCS files during CollectRevsPass.  The
    # constructor argument specifies how to invoke the "co"
    # executable.
    #
    # CVSRevisionReader uses the "cvs" program to extract the revision
    # contents out of the RCS files during OutputPass.  This option is
    # considerably slower than RCSRevisionReader because "cvs" is
    # considerably slower than "co".  However, it works in some
    # situations where RCSRevisionReader fails; see the HTML
    # documentation of the "--use-cvs" option for details.  The
    # constructor argument specifies how to invoke the "co"
    # executable.
    #
    # Uncomment one of the two following lines:
    #RCSRevisionReader(co_executable=r'co'),
    CVSRevisionReader(cvs_executable=r'cvs'),
    )
# This second alternative is vastly faster than the version above.  It
# uses an external Python program to reconstruct the contents of CVS
# file revisions:
#ctx.revision_collector = ExternalBlobGenerator('./tmp-cvs2git/git-blob.dat')

# cvs2git doesn't need a revision reader because OutputPass only
# refers to blobs that were output during CollectRevsPass, so leave
# this option set to None.
ctx.revision_reader = None

# Change the following line to True if the conversion should only
# include the trunk of the repository (i.e., all branches and tags
# should be omitted from the conversion):
ctx.trunk_only = False

# How to convert CVS author names, log messages, and filenames to
# Unicode.  The first argument to CVSTextDecoder is a list of encoders
# that are tried in order in 'strict' mode until one of them succeeds.
# If none of those succeeds, then fallback_encoder (if it is
# specified) is used in lossy 'replace' mode.  Setting a fallback
# encoder ensures that the encoder always succeeds, but it can cause
# information loss.
ctx.cvs_author_decoder = CVSTextDecoder(
    [
        'utf8',
        'latin1',
        'ascii',
        ],
    #fallback_encoding='ascii'
    )
ctx.cvs_log_decoder = CVSTextDecoder(
    [
        'utf8',
        'latin1',
        'ascii',
        ],
    #fallback_encoding='ascii'
    )
# You might want to be especially strict when converting filenames to
# Unicode (e.g., maybe not specify a fallback_encoding).
ctx.cvs_filename_decoder = CVSTextDecoder(
    [
        'utf8',
        'latin1',
        'ascii',
        ],
    #fallback_encoding='ascii'
    )

# Template for the commit message to be used for initial project
# commits.
ctx.initial_project_commit_message = (
    'Standard project directories initialized by cvs2git.'
    )

# Template for the commit message to be used for post commits, in
# which modifications to a vendor branch are copied back to trunk.
# This message can use '%(revnum)d' to include the SVN revision number
# of the revision that included the change to the vendor branch
# (admittedly rather pointless in a cvs2git conversion).
ctx.post_commit_message = (
    'This commit was generated as part of Drupal\'s Great Git Migration to track changes on a CVS '
    'vendor branch.'
    )

# Template for the commit message to be used for commits in which
# symbols are created.  This message can use '%(symbol_type)s' to
# include the type of the symbol ('branch' or 'tag') or
# '%(symbol_name)s' to include the name of the symbol.
ctx.symbol_commit_message = (
    "This commit was manufactured as part of Drupal's Great Git Migration to create %(symbol_type)s "
    "'%(symbol_name)s'."
    )

# Template for the commit message to be used for commits in which
# tags are pseudo-merged back to their source branch.  This message can
# use '%(symbol_name)s' to include the name of the symbol.
# (Not used by default unless you enable tie_tag_fixup_branches on
# GitOutputOption.)
ctx.tie_tag_ancestry_message = (
    "This commit was manufactured as part of Drupal's Great Git Migration to tie ancestry for "
    "tag '%(symbol_name)s' back to the source branch."
    )

# Some CVS clients for MacOS store resource fork data into CVS along
# with the file contents itself by wrapping it all up in a container
# format called "AppleSingle".  Subversion currently does not support
# MacOS resource forks.  Nevertheless, sometimes the resource fork
# information is not necessary and can be discarded.  Set the
# following option to True if you would like cvs2git to identify files
# whose contents are encoded in AppleSingle format, and discard all
# but the data fork for such files before committing them to
# Subversion.  (Please note that AppleSingle contents are identified
# by the AppleSingle magic number as the first four bytes of the file.
# This check is not failproof, so only set this option if you think
# you need it.)
ctx.decode_apple_single = False

# This option can be set to the name of a filename to which are stored
# statistics and conversion decisions about the CVS symbols.
#ctx.symbol_info_filename = None
ctx.symbol_info_filename = 'symbol-info.txt'

# cvs2git uses "symbol strategy rules" to help decide how to handle
# CVS symbols.  The rules in a project's symbol_strategy_rules are
# applied in order, and each rule is allowed to modify the symbol.
# The result (after each of the rules has been applied) is used for
# the conversion.
#
# 1. A CVS symbol might be used as a tag in one file and as a branch
#    in another file.  cvs2git has to decide whether to convert such a
#    symbol as a tag or as a branch.  cvs2git uses a series of
#    heuristic rules to decide how to convert a symbol.  The user can
#    override the default rules for specific symbols or symbols
#    matching regular expressions.
#
# 2. cvs2git is also capable of excluding symbols from the conversion
#    (provided no other symbols depend on them.
#
# 3. CVS does not record unambiguously the line of development from
#    which a symbol sprouted.  cvs2git uses a heuristic to choose a
#    symbol's "preferred parents".
#
# The standard branch/tag/exclude StrategyRules do not change a symbol
# that has already been processed by an earlier rule, so in effect the
# first matching rule is the one that is used.

global_symbol_strategy_rules = [
    # It is possible to specify manually exactly how symbols should be
    # converted and what line of development should be used as the
    # preferred parent.  To do so, create a file containing the symbol
    # hints and enable the following option.
    #
    # The format of the hints file is described in the documentation
    # for the --symbol-hints command-line option.  The file output by
    # the --write-symbol-info (i.e., ctx.symbol_info_filename) option
    # is in the same format.  The simplest way to use this option is
    # to run the conversion through CollateSymbolsPass with
    # --write-symbol-info option, copy the symbol info and edit it to
    # create a hints file, then re-start the conversion at
    # CollateSymbolsPass with this option enabled.
    #SymbolHintsFileRule('symbol-hints.txt'),

    # To force all symbols matching a regular expression to be
    # converted as branches, add rules like the following:
    #ForceBranchRegexpStrategyRule(r'branch.*'),

    # To force all symbols matching a regular expression to be
    # converted as tags, add rules like the following:
    #ForceTagRegexpStrategyRule(r'tag.*'),

    # To force all symbols matching a regular expression to be
    # excluded from the conversion, add rules like the following:
    #ExcludeRegexpStrategyRule(r'unknown-.*'),

    # Sometimes people use "cvs import" to get their own source code
    # into CVS.  This practice creates a vendor branch 1.1.1 and
    # imports the code onto the vendor branch as 1.1.1.1, then copies
    # the same content to the trunk as version 1.1.  Normally, such
    # vendor branches are useless and they complicate the SVN history
    # unnecessarily.  The following rule excludes any branches that
    # only existed as a vendor branch with a single import (leaving
    # only the 1.1 revision).  If you want to retain such branches,
    # comment out the following line.  (Please note that this rule
    # does not exclude vendor *tags*, as they are not so easy to
    # identify.)
    ExcludeTrivialImportBranchRule(),

    # To exclude all vendor branches (branches that had "cvs import"s
    # on them but no other kinds of commits), uncomment the following
    # line:
    #ExcludeVendorBranchRule(),

    # Usually you want this rule, to convert unambiguous symbols
    # (symbols that were only ever used as tags or only ever used as
    # branches in CVS) the same way they were used in CVS:
    UnambiguousUsageRule(),

    # If there was ever a commit on a symbol, then it cannot be
    # converted as a tag.  This rule causes all such symbols to be
    # converted as branches.  If you would like to resolve such
    # ambiguities manually, comment out the following line:
    BranchIfCommitsRule(),

    # Last in the list can be a catch-all rule that is used for
    # symbols that were not matched by any of the more specific rules
    # above.  (Assuming that BranchIfCommitsRule() was included above,
    # then the symbols that are still indeterminate at this point can
    # sensibly be converted as branches or tags.)  Include at most one
    # of these lines.  If none of these catch-all rules are included,
    # then the presence of any ambiguous symbols (that haven't been
    # disambiguated above) is an error:

    # Convert ambiguous symbols based on whether they were used more
    # often as branches or as tags:
    HeuristicStrategyRule(),
    # Convert all ambiguous symbols as branches:
    #AllBranchRule(),
    # Convert all ambiguous symbols as tags:
    #AllTagRule(),

    # The last rule is here to choose the preferred parent of branches
    # and tags, that is, the line of development from which the symbol
    # sprouts.
    HeuristicPreferredParentRule(),
    ]

# Specify a username to be used for commits for which CVS doesn't
# record the original author (for example, the creation of a branch).
# This should be a simple (unix-style) username, but it can be
# translated into a git-style name by the author_transforms map.
ctx.username = 'cvs2git'

# ctx.file_property_setters and ctx.revision_property_setters contain
# rules used to set the svn properties on files in the converted
# archive.  For each file, the rules are tried one by one.  Any rule
# can add or suppress one or more svn properties.  Typically the rules
# will not overwrite properties set by a previous rule (though they
# are free to do so).  ctx.file_property_setters should be used for
# properties that remain the same for the life of the file; these
# should implement FilePropertySetter.  ctx.revision_property_setters
# should be used for properties that are allowed to vary from revision
# to revision; these should implement RevisionPropertySetter.
#
# Obviously, SVN properties per se are not interesting for a cvs2git
# conversion, but some of these properties have side-effects that do
# affect the git output.  FIXME: Document this in more detail.
ctx.file_property_setters.extend([
    # To read auto-props rules from a file, uncomment the following line
    # and specify a filename.  The boolean argument specifies whether
    # case should be ignored when matching filenames to the filename
    # patterns found in the auto-props file:
    #AutoPropsPropertySetter(
    #    r'/home/username/.subversion/config',
    #    ignore_case=True,
    #    ),

    # To read mime types from a file and use them to set svn:mime-type
    # based on the filename extensions, uncomment the following line
    # and specify a filename (see
    # http://en.wikipedia.org/wiki/Mime.types for information about
    # mime.types files):
    #MimeMapper(r'/etc/mime.types', ignore_case=False),

    # Omit the svn:eol-style property from any files that are listed
    # as binary (i.e., mode '-kb') in CVS:
    CVSBinaryFileEOLStyleSetter(),

    # If the file is binary and its svn:mime-type property is not yet
    # set, set svn:mime-type to 'application/octet-stream'.
    CVSBinaryFileDefaultMimeTypeSetter(),

    # To try to determine the eol-style from the mime type, uncomment
    # the following line:
    #EOLStyleFromMimeTypeSetter(),

    # Choose one of the following lines to set the default
    # svn:eol-style if none of the above rules applied.  The argument
    # is the svn:eol-style that should be applied, or None if no
    # svn:eol-style should be set (i.e., the file should be treated as
    # binary).
    #
    # The default is to treat all files as binary unless one of the
    # previous rules has determined otherwise, because this is the
    # safest approach.  However, if you have been diligent about
    # marking binary files with -kb in CVS and/or you have used the
    # above rules to definitely mark binary files as binary, then you
    # might prefer to use 'native' as the default, as it is usually
    # the most convenient setting for text files.  Other possible
    # options: 'CRLF', 'CR', 'LF'.
    DefaultEOLStyleSetter(None),
    #DefaultEOLStyleSetter('native'),

    # Prevent svn:keywords from being set on files that have
    # svn:eol-style unset.
    SVNBinaryFileKeywordsPropertySetter(),

    # If svn:keywords has not been set yet, set it based on the file's
    # CVS mode:
    KeywordsPropertySetter(config.SVN_KEYWORDS_VALUE),

    # Set the svn:executable flag on any files that are marked in CVS as
    # being executable:
    ExecutablePropertySetter(),

    # The following causes keywords to be collapsed in all text to be
    # committed:
    KeywordHandlingPropertySetter('collapsed'),

    ])
ctx.revision_property_setters.extend([
    ])

# To skip the cleanup of temporary files, uncomment the following
# option:
#ctx.skip_cleanup = True


# In CVS, it is perfectly possible to make a single commit that
# affects more than one project or more than one branch of a single
# project.  Subversion also allows such commits.  Therefore, by
# default, when cvs2git sees what looks like a cross-project or
# cross-branch CVS commit, it converts it into a
# cross-project/cross-branch Subversion commit.
#
# However, other tools and SCMs have trouble representing
# cross-project or cross-branch commits.  (For example, Trac's Revtree
# plugin, http://www.trac-hacks.org/wiki/RevtreePlugin is confused by
# such commits.)  Therefore, we provide the following two options to
# allow cross-project/cross-branch commits to be suppressed.

# cvs2git only supports single-project conversions (multiple-project
# conversions wouldn't really make sense for git anyway).  So this
# option must be set to False:
ctx.cross_project_commits = False

# git itself doesn't allow commits that affect more than one branch,
# so this option must be set to False:
ctx.cross_branch_commits = False

# cvs2git does not yet handle translating .cvsignore files into
# .gitignore files, so by default, the .cvsignore files are included
# in the conversion output.  If you would like to omit the .cvsignore
# files from the output, set this option to False:
ctx.keep_cvsignore = False

# By default, it is a fatal error for a CVS ",v" file to appear both
# inside and outside of an "Attic" subdirectory (this should never
# happen, but frequently occurs due to botched repository
# administration).  If you would like to retain both versions of such
# files, change the following option to True, and the attic version of
# the file will be written to a subdirectory called "Attic" in the
# output repository:
ctx.retain_conflicting_attic_files = True

# CVS uses unix login names as author names whereas git requires
# author names to be of the form "foo <bar>".  The default is to set
# the git author to "cvsauthor <cvsauthor>".  author_transforms can be
# used to map cvsauthor names (e.g., "jrandom") to a true name and
# email address (e.g., "J. Random <jrandom@example.com>" for the
# example shown).  All strings should be either Unicode strings (i.e.,
# with "u" as a prefix) or 8-bit strings in the utf-8 encoding.  The
# values can either be strings in the form "name <email>" or tuples
# (name, email).  Please substitute your own project's usernames here
# to use with the author_transforms option of GitOutputOption below.
author_transforms = {}
cvsReader = csv.reader(open('#CSV#', 'rb'), delimiter=',', quotechar='"', escapechar='\\')
for row in cvsReader:
    # uid  ,Full Name         ,CVS name ,email
    # "102","Jonathan W. Webb","webbj74","developer@example.com"
    #
    author_transforms[row[2]] = (row[1], row[3])

# This is the main option that causes cvs2git to output to a
# "fastimport"-format dumpfile rather than to Subversion:
ctx.output_option = GitOutputOption(
    # The file in which to write the git-fast-import stream that
    # contains the changesets and branch/tag information:
    os.path.join(ctx.tmpdir, 'git-dump.dat'),

    # The blobs will be written via the revision recorder, so in
    # OutputPass we only have to emit references to the blob marks:
    GitRevisionMarkWriter(),

    # Optional map from CVS author names to git author names:
    author_transforms=author_transforms,
    )

# Change this option to True to turn on profiling of cvs2git (for
# debugging purposes):
run_options.profiling = False


# Should CVSItem -> Changeset database files be memory mapped?  In
# some tests, using memory mapping speeded up the overall conversion
# by about 5%.  But this option can cause the conversion to fail with
# an out of memory error if the conversion computer runs out of
# virtual address space (e.g., when running a very large conversion on
# a 32-bit operating system).  Therefore it is disabled by default.
# Uncomment the following line to allow these database files to be
# memory mapped.
#changeset_database.use_mmap_for_cvs_item_to_changeset_table = True

# Now set the project to be converted to git.  cvs2git only supports
# single-project conversions, so this method must only be called
# once:
run_options.set_project(
    # The filesystem path to the part of the CVS repository (*not* a
    # CVS working copy) that should be converted.  This may be a
    # subdirectory (i.e., a module) within a larger CVS repository.
    r'#DIR#',

    # A list of symbol transformations that can be used to rename
    # symbols in this project.
    symbol_transforms=[
        # Use IgnoreSymbolTransforms like the following to completely
        # ignore symbols matching a regular expression when parsing
        # the CVS repository, for example to avoid warnings about
        # branches with two names and to choose the preferred name.
        # It is *not* recommended to use this instead of
        # ExcludeRegexpStrategyRule; though more efficient,
        # IgnoreSymbolTransforms are less flexible and don't exclude
        # branches correctly.  The argument is a Python-style regular
        # expression that has to match the *whole* CVS symbol name:
        #IgnoreSymbolTransform(r'nightly-build-tag-.*')

        # RegexpSymbolTransforms transform symbols textually using a
        # regular expression.  The first argument is a Python regular
        # expression pattern and the second is a replacement pattern.
        # The pattern is matched against each symbol name.  If it
        # matches the whole symbol name, then the symbol name is
        # replaced with the corresponding replacement text.  The
        # replacement can include substitution patterns (e.g., r'\1'
        # or r'\g<name>').  Typically you will want to use raw strings
        # (strings with a preceding 'r', like shown in the examples)
        # for the regexp and its replacement to avoid backslash
        # substitution within those strings.
        #RegexpSymbolTransform(r'release-(\d+)_(\d+)',
        #                      r'release-\1.\2'),
        #RegexpSymbolTransform(r'release-(\d+)_(\d+)_(\d+)',
        #                      r'release-\1.\2.\3'),

        # Simple 1:1 character replacements can also be done.  The
        # following transform, which converts backslashes into forward
        # slashes, should usually be included:
        ReplaceSubstringsSymbolTransform('\\','/'),

        # This last rule eliminates leading, trailing, and repeated
        # slashes within the output symbol names:
        NormalizePathsSymbolTransform(),
        ],

    # See the definition of global_symbol_strategy_rules above for a
    # description of this option:
    symbol_strategy_rules=global_symbol_strategy_rules,
    )