[Python-modules-commits] [python-fastimport] annotated tag debian/0.9.6-3 created (now 1d0cb12)
Mattia Rizzolo
mattia at debian.org
Sun Apr 23 06:45:49 UTC 2017
This is an automated email from the git hooks/post-receive script.
mattia pushed a change to annotated tag debian/0.9.6-3
in repository python-fastimport.
at 1d0cb12 (tag)
tagging dc1a31a6948708463e288a6e83b5047c1af4b5d5 (commit)
tagged by Mattia Rizzolo
on Sun Apr 23 08:45:33 2017 +0200
- Log -----------------------------------------------------------------
python-fastimport Debian release 0.9.6-3
-----BEGIN PGP SIGNATURE-----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=mpZt
-----END PGP SIGNATURE-----
Adeodato Simó (1):
Support kind changes in bzr-fast-export.
Alexander Belchenko (2):
fast-import-filter should produce binary output.
fast-export should force binary sys.stdout on win32
Alexey Stukalov (1):
set stdout to binary mode on windows
Anders Kaseorg (2):
git-darcs: Import to a real ref (refs/remotes/darcs/$remote).
git-darcs: Fix exit/return confusion.
Brian Warner (4):
git-darcs: remember and pass options to darcs-fast-export
git-darcs: change #!/bin/sh to /bin/bash, '$? == 0' is a bash-ism
darcs-fast-import: fix timezone processing
darcs-fast-export: allow hyphens in tag names, I think they're ok
Brian de Alwis (2):
Add support for multiple branches by supporting the 'reset' command.
Reset takes a <commitsh> and not just a revid; added note to
Cécile Tonglet (1):
Get handlers from class object using getattr() for possible inheritance
D Herring (4):
added a few git-darcs commands
git-darcs: minor bugfix
fix timezones in darcs-fast-export
fix timezones in darcs-fast-export, take 2
David Reitter (1):
fix bug #348038 (call to warning() with two arguments) and handle malformed revisions gracefully by not generating any output
Elliot Murphy (1):
Fix typo in online help.
Erik Hesselink (1):
darcs-fast-export: no longer fail on empty patch name
Félix Mattrat (11):
python 3 compatibility, futurize stage 1
python 3 compatibility, futurize stage 2
make the tests work again on python2
sorted out string/unicode problems, updated the unicode test, renamed assertEquals method to assertEqual (deprecated in py3)
removed unused/useless statements
added a tox configuration file to simplify multiple versions testing
added the future package in the deps of travis conf
added python 3.5 and PyPy3
Explicitly declared all the byte strings. Renamed variables using reserved keywords
remove support of python3.2/python3.2/python3.4
Fixed the test_copy_to_outside
Gonéri Le Bouder (4):
avoid STDERR crash
add the missing ":" since revid_to_mark are "committish"
really put ':' in the commit mark
LocalGitBzrDirFormat has no repository_format
Harry Hirsch (2):
Merged fix for directory handling in plain format
Don't emit directory info for renames operations when using plain format
Ian Clatworthy (399):
1st cut: gfi parser + --info processing method
use Bazaar file kinds
tweak parser for better git-fast-export compatibility
abstract CommitHandler
first cut at generic processing method
generic processing method working for one revision in one branch
start of multiple commit handling
custom parameters for processors
dump parameter for info processor
1st cut are dequoting paths
tag support
lightweight tags, filter processor and param validation
commit of new files working
commit of modified files working
begin notes on known limitations, etc.
safe processing tweaks
escape commit messages, diff author to committer and cache fixes
timestamp loaded commit messages
filtering enhancements: selected fields, filecommands, non-verbose format
clean-up fixes after filtering enhancements
fix one inventory lookup bug
fix more inventory lookup bugs
blob tracking analysis and verbose mode for info processor
smart blob caching using analysis done by --info
slightly better progress reporting
more progress reporting tweaks
1st cut at performance tuning
checkpoint and count params to generic processor
improve explanation of faster check in revisionloader
add heads analysis to info processor
fix branch updating for the single branch case
move single_plural into helpers
make tree updating optional and minor UI improvements
report lost branches
identify unmarked blobs and commits by line numbers
fix head tracking when unmarked commits used
create branches as required
clean-up doc ready for initial release
fix blob tracking when marks are reused
always use heads/master as the trunk if it is present
update multiple working trees if requested
removed parent not found warnings as not a problem
verbose mode cleanup
smart caching of serialised inventories
fix compatibility with Python 2.4
abort write_group when an unexpected exception occurs so we see the latter
add option for enabling experimental stuff
one revision loader instance
skip check re fulltext storage better than delta for inventories when in experimental mode
cleanly restart after an interruption - basic mirroring
disable autopacking
switch on experimental mode by default
fix line number tracking
handle existing branches and only count the branches really updated
fix head tracking when from clause implied
Integrate lots of fixes from James Westby
integrate dato's bzr-fast-export
bundle svn-fast-export and hg-fast-export
support merges when from clause implicit
fix missing revisions bug
improve hg-fast-export integration
remove warning about delete iff file is in a merge parent
save tags known about in each branch
update notes on what's done and what's not
fix a duplicate file-id after rename bug
Merge Elliot's help fix
Support kind changes in bzr-fast-export (merge dato's patch)
remove call to inv_vf.enable_cache - API is going away
fix tests to use new CommitCommand constructor
reduce memory usage while reading large blobs
fix symlink importing
if checkpointed, pack repository and delete obsolete_packs
fix utf-8 decoding bugs
add inv-fulltext option and improve speed
fix from Pieter de Bie - hack around broken front-ends
support new Repository API
added NEWS and setup.py
'bzr help fastimport' now provides useful help
Merge Pieter de Bie's export-fixes branch
Merge Mirko Friedenhagen's regex improvement to svn-fast-export.py
release 0.6
prepare for 0.7 development
hg-fast-export should use binary mode on Windows (Alexey Stukalov)
bzr-fast-export exports rm+mv correctly now (Jonas)
Fix recursive rename handling (Pieter de Bie)
setup.py support fixed (Jelmer Vernooij)
setup.py no longer requires python2.4
Add deleteall support (Miklos Vajna)
fix last-modified revision after rename (John Arbash Meinel)
minor comment clean-ups
ignore lightweight tags without a from clause
only output warning about missing from clause for lightweight tags in verbose mode
fix support for rich-root (and subtree) repositories
update NEWS, making latest changes the start of 0.8dev
fix os import as needed on Windows
remove --inv-fulltext option
bundle latest hg-fast-export
reapply local enhancements to hg-fast-export
Handle a directory becoming a file and subsequent child deletes
handle empty emails & names/paths that aren't utf8 encoded
handle multiple merges in the one merge clause
handle > 16 parents in fast-import-info
let the id-map file have more revisions than the repository
warn on tags with a missing from clause
recursively delete children when a directory is deleted
initial cut at reset support
make deleteall less agressive in the files it tries to delete
rename fast-import-filter to fast-import-query
fix unicode-related exception in fast-import-query
initial cut at fast-import-filter
code & tests for formatting Commands as file-import stream strings
get fast-import-filter formatting commands in the output correctly
tweak from and merge formatting
add _paths to option & params names as other types of filtering may be added later
no filtering tests
include single file tests
include subdir & multiple files tests
common_directory tests & tweaks
add exclude_paths tests
fix parsing of filecopy command
fix parsing of filecopy command
rename and copy tests & tweaks
deleteall tests & tweaks
tag & reset tests & tweaks
ensure output is utf8 encoded when required/recommended
added NEWS item
clean-up tests for GenericProcessor
code & tests for file copying
symlink copying tests
added NEWS item
support copying of files and symlinks
update NEWS with info about filecopy bug
fast-import-filter for including/excluding paths from a stream
move note/message/debug into higher level classes
move escape_commit_message into helpers
move GenericCacheManager into its own module
move GenericBranchUpdater into its own module
move import/export of marks into a module
generic_processor.py refactoring
turn bzr-fast-export into a fast-export command
fix lru_cache import
extend & use marks_file API
remove unnecessary checkpoint at 0 commits
use note and warning APIs
incorporate git-fast-import tweaks from bzr-fast-export script
refactor bzr_exporter to use Command objects
trivial bzr_exporter clean-ups
fix mark numbering so parents always get lower numbers
fix branch of first commit to not be refs/heads/tmp
fix missing os import in generic_processor
unhide fast-import and fast-import-info commands
fast-import-filter should produce binary output (Alexander Belchenko)
documentation clean-ups
fast-export command replacing bzr-fast-export script
move GenericCommitHandler into its own module in prep for a delta-based one
refactor InventoryCommitHandler general stuff into parent class
enhance RevisionLoader to try inventory deltas & decide on rich-roots
fix osutils call in delete_recursive
generalise RevisionLoader to RevisionStore as a repo abstraction
basic DeltaCommitHandler generating deltas
basic DeltaCommitHandler mostly going bar rename
basic units tests for filemodify
basic delete tests
file <-> symlink change tests
file executable off <-> on tests
file/symlink <-> directory change tests & fix
file/symlink <-> directory rename tests
add directory delete test
add lots of import tests
merge import tests and tweaks to make them pass
refactor rename_item
refactor delete_item
get DeltaCommitHandler passing all tests
extend fast-import-info with merges, rename old paths & copy source paths
use inventory deltas in experimental mode & enhanced fast-import-info
ensure version appears in bzr plugins output (Alexander Belchenko)
add total merges to fast-import-info & record when old/src paths no longer required
fix encoding issue in bzr_exporter (Teemu Likonen)
nicer and round-trippable mapping of git ref names to bzr branch names
nicer and round-trippable mapping of git ref names to bzr branch names
head tracking tests and fix
update fast-import-info to use same head-tracking code as fast-import
fix head tracking when stream includes resets
remove other-resets frm fast-import-info output
fix fast-export's encoding of user names
Fix str + int concat in bzr-fast-export (Stéphane Raimbault)
fix marks importing in fast-export
improve error message when .bzr directory not found
fix restart when using CHKInventories
improve error msg when bad timezone encountered
make sure fast-export uses a binary stream on Windows
fix inventory copying when using deltas
improved-but-not-yet-working CHKInventory support
_ensure_directory to return parent_id, not parent_ie
set maximum_size & key_width for initial parent_id_basename_to_file_id map
CHKInventory support for non rich-root repos working, for simple imports at least
ensure the chk stuff is only used on formats actually supporting it
improve error msg when bad timezone encountered
make sure fast-export uses a binary stream on Windows
get non-chk formats working again & combine delta entries when required
fix TREE_ROOT delta entry after 1st revision & tweak _delete_item usage
fix ie reference in _get_file_lines()
initial working version for CHKInventory-based formats
use iter_non_root_entries if it exists
lookup directories on demand in CHKInventories, not all upfront
smarter RevisionStore.chk_load()
performance tweaks for chk formats
use new propagate_caches parameter on create_by_apply_delta
use new propagate_caches parameter on create_by_apply_delta
speed up blob parsing
make groupcompress _FAST during import
handle delete of missing files for chk formats
fix first file is in a subdirectory bug for chk formats
error on zero length paths in the input stream
handle delete of unknown file in chk formats & reduce noise
larger default inventory cache for chk formats
show commit rate rather than meaningless ETA in verbose mode
parent_id_to_basename_index is no longer a serializer attribute - always required now
miscellaneous progress reporting fixes
clear caches before packing; show cache stats in verbose mode
make groupcompress space efficient before packing
store empty texts for non-files for chk formats, not just other formats
minor revision_store clean-ups
extend tests to test the chk code path
extend tests to test the chk code path
improve per-file graph generation
improve per-file graph generation
make the file-id cache optional and branch-ref aware
make per-file parents tuples and fix text loading in chk formats
fix per-graph parent handling for adds and renames
always repack the repository on completion
binary output on Windows (Alexander Belchenko)
back out cache propagation until more reliable
handle adding a file to a dir deleted in the same commit
graceful handling of faulty revisions (David Reitter)
incremental packing for chk formats
blob reference counting, not just sticky vs otherwise
fix blob tracking when -v not given
add autopack option to fast-import
use inv deltas by default for all formats now: --classic to get old algorithm for packs
handle commit commands with no from clause and no sub-commands
add -r option to fast-export
fix rename adjustment & kind change logic in fast-export
fix parsing when a rename old-path has spaces in it
faster export of revision range & improved diagnostics in fast-export
fix round-tripping of committer & author when name is an email
improve fast-export's handling of rename+delete combinations
upgrade tests to use --development6-rich-root
report triggering commit when exception occurs
Ignore quotes around second path in a pair (Dmitri Paduchikh)
Here-document style commit messages (Samuel Bronson)
release 0.8
Open 0.9 development
Integate latest (25-May-2009) hg-fast-export
fix inv-delta generation when deleting directories
update NEWS
merge darcs-fast-export (Miklos Jayna)
NEWS item wrt darcs-fast-export
Force parser to generate commit messages in unicode - even when not in utf8 in the input stream
delegate commit message escaping to the serializer if it's a modern one
Smarter selection of branch format based on shared repository format
Fix dump_str to ensure CommitCommand.file_iter is an interator, not a callable
prune directories that become empty after a delete or rename
get tests passing again
Gracefully handle email addresses with unicode characters
update NEWS
initial cut at a simplified ui to xxx-fast-export
tweak required svn product name
implement fast-export-from-git
first cut at fast-export-from-hg - not working yet
fast-export-from-darcs - working but chardet usage/packaging still to be sorted out
get fast-export-from-hg working
first cut at .gz implicit compression
merge darcs-fast-export fixes and better doc
add encoding option to fast-export-from-darcs
tweak fast-export-from-hg to use the same prefix for the mapping and status files
back out implicit gz compression - can come later
initial cut at fast-export-from-svn
NEWS item and doco tweaks
first batch of fast-export-from-xxx commands
tweak help formatting
more help tweaks
fix reporting of non-utf8 commits to go to stderr
added fast-export-from-mnt command
added fast-export-to-cvs
Merge fast-export-from-p4 (Matt McClure)
Create a repository implicitly if one doesn't already exist
update help
Update the working tree for trunk implicitly
fast-import destination argument, implicit repo creation & trunk working tree updating
minor doc tweaks
Fix NameError running fast-export-from-git
eliminate the reference to my temporary cvs2svn branch
implicitly uncompress if a fast-import source ends in .gz
configrable output stream for InfoProcessor
Smarter blob tracking by implicitly collecting statistics before starting the import
update existing tests to reflect implicit directory pruning bahaviour
test and fix for implicit directory delete recursing up
make implicit directory deletion recurse upwards
tweak GenericProcessor __init__ method
handle an empty input stream (Gonéri Le Bouder)
update NEWS
Fix git-bzr mark compatibility issue (Gonéri Le Bouder)
lookup file-ids in inventories instead of a cache
Eliminate the existing file-id cache altogether
More robust implicit delete logic when file-id not found
Only count implicit checkpoints when deciding when to auto-pack
Improve diagnostics collection when inconsistent delta found
Merge fixes from trunk, rev 220.
handle adding the same file twice in the one commit
Handle rename of a just added file
handle a delete of a newly added file
handle copy of a newly added file
better diagnostics on inconsistent delta
commented out draft of CopyModified tests
Change tests to use 2a instead of development6-rich-root
backout git-bzr fix as it was breaking fast-export
Support an explicit output destination for bzr fast-export
merge fast-export fixes from trunk (rev 222)
Handle delete then add of a file/symlink in the one commit
Add tests for renaming to a deleted destination
Add tests for copying to a path deleted in the same commit
add tests for symlink renaming
Handle rename of a file/symlink modified already in this commit
Handle copy of a file/symlink already modified in this commit
Support displaying a CommitCommand without inline blob data
add --commit-mark option to fast-import-query to show just one commit (without file contents)
Handle rename then modification of the new path
merge enhancement for easy lookup of a commit (trunk rev 224)
Handle deleting a directory then adding a file within it in the same commit
Merge fixes for inconsistent deltas
Stop fast-export from exceeding the maximum recursion depth
merge recursion depth exceeded fix
Fix launching of git on Windows
Use the git on Windows path for the check and not just the run itself
Handle git submodules in the stream by warning about + ignoring them
Definition and output of feature commands
Definition and formatting of multiple authors and revision properties
First cut at exporting additional metadata via 'features'
tweak proposed property formatting
Define feature names in one place
Fix ghost handling and improve progress tracking in fast-export
fix broken tests
merge ghost handling fix from trunk
feature parsing
parsing of multiple authors and commit properties
Store multiple authors and revision properties when defined
More intelligent empty directory removal (Tom Widmer)
Validate features are known before importing
add --plain option to fast-export
Fix feature checking
export and import empty directories
add revision count to 'Starting export ...' message
tweak author formatting to use same smart rule as used for committer
Teach fast-import-filter and fast-import-query about feature commands
Tweak some diagnostic messages
Merge feature support including generation/parsing of commit-properties, multiple-authors and empty-directories
Handle delete, rename then modify all in the one commit
Make sure renamed directories are found in file-id lookups
Sanitize None revision properties to empty string
Ensure empty string property values stay as empty strings
implicitly rename children on export when directory renamed
--forward, not --reverse, in fast-export help
Handle formatting of unicode emails
Fix unicode email address parsing
fix warning messages on non-utf8 names & emails
merge directory handling fix for darcs-fast-import
merge fixes for darcs-fast-import for better hg 1.3 interoperability
merge Harry's fix so directories aren't output in plain format
merge fix for empty patch name handling in darcs-fast-import
fix date parsing bug found while importing samba
base file-ids on the basename, not path, as jam suggested. This improves the samba import from 565M to 353M.
Merge fix for tags off the mainline
Merge Ted's improvements to svn-fast-export
ignore generated executables
Add a README explaining how to compile svn-fast-export.c
Add --user-map option to both fast-import and fast-import-filter
update NEWS with recent changes
Change the default inventory cache size to 1. For large projects, this reduces memory overhead and also speeds up conversion.
Fix parsing error when a property is found after a multi-line one
handle git ref names with slashes in them
Get fastimport working on non-chk repositories again for bzr versions after 2.0.0
Handle multi-level branches
http read support for darcs-fast-export
Merge Max's fixes for tags created via commit commands
Default branch nick to mapped git ref name (Max Bowsher)
Merge Jelmer's fix for exporters installation
Faster post-processing of the inventory delta
Merge John's smarter caching of blobs to improve memory footprint
Merge John's improvements to info_processor.py
Merge John's test refactoring - uses multiplication now
Merge Python 2.5 fix for darcs-fast-export
Fix Monotone prefix and command name: mnt -> mtn
Ignore progress messages unless in verbose mode
Add explorer integration
Fallback to system configobj if none in bzrib.util
Don't bail for http://path/to/repo darcs remotes
Release 0.9
Merge patches from Miklos for darcs-fast-export
James Westby (9):
The data sections have an optional LF at the end in the byte count format.
The space between the author and email is optional in committer.
ImportProcessor.params is a dict, not a list.
Make the rename handling more robust.
Make the parser handle multiple words in the committer name.
Open the input in binary mode, as suggested by Paul Moore.
Set stdout to binary mode on Windows to make it use the correct line endings.
Add support for -Dfast-import.
Correct typo: InventoryLnk -> InventoryLink
Jelmer Vernooij (134):
Fix setup.py
Fix installation of the exporters module.
Cope with Debian's bzr using the system configobj, not shipping with its own.
Merge fixes from Miklos for darcs fast import.
Move pure-fastimport code into its own directory, in preparation of splitting it into a separate package.
Split python-fastimport into its own separate package.
Split out fastimport, import it from the system.
Remove bzr-specific code, whitespace.
Add testrepository configuration, test_suite() function.
Initial release.
Fix some imports.
Ignore testrepository directories.
Only require passing BzrDir to bzr-specific processors.
Remove bzr-specific bits from processor.
remove bzrisms from Processor.
Fix typo, handle bzr-specific locking in GenericProcessor.
Split out reftracker.
Import processors from bzr-fastimport.
Use modes directly rather than bzr file kinds and executable flags.
Don't rely on bzrlib.trace.
remove dependency on bzrlib.
Initial release. Closes: #598490
merge new upstream snapshot.
Fix vcs url.
Install fastimport.processors.
Import upstream version 0.9.0~bzr290
Merge revision that installs fastimport.processors.
Cope with FileModifyCommand now having mode rather than is_executable and kind.
Import upstream version 0.9.0~bzr291
Import another bugfix.
Support out of band processing of commands.
Import upstream version 0.9.0~bzr292
Fix tests.
* New upstream snapshot.
Import upstream version 0.9.0~bzr293
Upload to unstable.
Avoid file_iter, use iter_files instead.
Import upstream version 0.9.0~bzr294
* New upstream snapshot.
If there is no mark then don't track it.
Import upstream version 0.9.0~bzr295
* New upstream snapshot.
Stop utf-8 decoding several strings as they are not guaranteed to be unicode.
Check for leading slashes in paths, invalid path types.
Avoid attempting to utf-8 decode/encode committer and author information.
Import upstream version 0.9.0~bzr298
Require testtools 0.9.2 to run the test suite.
Depend on testtools 0.9.2.
Merge entry for unuploaded version.
Import upstream version 0.9.0~bzr299
* New upstream snapshot.
release 0.9.0
Start on 0.9.1.
Import upstream version 0.9.0
New upstream release.
Merge packaging branch.
Fix changelog.
* Add watch file.
Switch to dh_python2.
Switch to debhelper 7, drop cdbs.
Remove unused Breaks line.
Remove old FSF address and point at FSF website instead.
Only fix references to commits that have been squashed rather than removing everything that we haven't touched yet
Merge support for squash_empty_commit argument to FilterProcessor.
Implement 'done' feature.
Rename tarball for the benefit of pip.
Release 0.9.1.
Start on 0.9.2.
remove .pc directory.
Update watch file.
add bzr-builddeb magic
Add machine-parseable copyright file.
Update watch file for new release location.
Add basic tests for autopkgtest.
Import upstream version 0.9.1
* New upstream release.
releasing version 0.9.1-1
Clarify AUTHORS.
Remove bzr-specific reftracker and idmapfile modules.
remove info_processor too (uses reftracker).
Add basic tests for dates.
Cope with invalid timezones like +51800 a little bit better.
Allow less strict parsing of fastimport streams.
Release 0.9.2.
Start on 0.9.3.
Fix watch file.
Import upstream version 0.9.2
New upstream release.
releasing version 0.9.2-1
releasing version 0.9.2-1
Add XS-Testsuite header. Closes: #692669
Change fileid of processors/
Ignore some files.
Remove download URL.
Add Makefile.
Add travis config.
Remove dependency on testtools.
Install unittest2 for python2.6.
Build with pypy in travis.
Actually run tests with unittest2.
Fix compatibility with python2.6, which does not have assertIsInstance.
Remove unused helpers.
Remove binary_stream helper.
Add MANIFEST file.
Add basic readme file.
Release 0.9.3.
Start on 0.9.4.
Add upstream signing key.
Update watch file.
Add gbp config.
Merge tag 'upstream/0.9.3' into debian
Imported Upstream version 0.9.3
New upstream release.
Fix autopkgtest.
Bump standards version to 3.9.5 (no changes).
Migrate packaging to Git.
releasing package python-fastimport version 0.9.3-1
Merge branch 'commit-copy' of git://github.com/masklinn/python-fastimport
Update NEWS.
Merge branch 'authorship-namedtuple' of git://github.com/masklinn/python-fastimport
Update NEWS.
Add pypy package.
Bump standards version to 3.9.6 (no changes).
releasing package python-fastimport version 0.9.4-2
Update Vcs-Git, Vcs-Browser and Maintainer fields to reflect the dulwich package is now maintained by the Debian Python Modules packaging team.
Fix watch URL.
releasing package python-fastimport version 0.9.4-3
'include ' in MANIFEST.
Fix license names in debian/copyright.
Drop '-b unstable' from Vcs-Git header.
debian/watch: Fix pypi URL.
Remove unused bzr-builddeb.conf.
releasing package python-fastimport version 0.9.4-4
Simplify handling of errors.
Jelmer Vernooij (33):
Add Breaks: for bzr-fastimport without version, as all versions that have been packaged in Debian break with the current version of python-fastimport. Closes: #747900
releasing package python-fastimport version 0.9.3-2
Fix in Vcs-Git header.
Release 0.9.4.
Start on fastimport 0.9.5.
Imported Upstream version 0.9.4
Merge tag 'upstream/0.9.4' into debian
New upstream release.
releasing package python-fastimport version 0.9.4-1
Drop future dependency.
Add NEWS entry.
Collapse import statements, remove unnecessary helpers by inlining them.
Add Félix to AUTHORS.
Release 0.9.5.
Ignore some files.
Imported Upstream version 0.9.5
Merge tag 'upstream/0.9.5' into unstable
New upstream release.
Bump standards version (no changes).
Ignore build stamp.
Add python3-fastimport package, now that upstream has added Python3 support.
releasing package python-fastimport version 0.9.5-1
Improve compatibility with python3.4.
Add python3.4 support.
Set maintainer field, update homepage.
Release 0.9.6.
Imported Upstream version 0.9.6
Merge tag 'upstream/0.9.6' into unstable
New upstream release.
Adds support for python 3.4.
releasing package python-fastimport version 0.9.6-1
Only mark as breaking older versions of bzr-fastimport << 0.13.0+bzr361-1.
releasing package python-fastimport version 0.9.6-2
John Arbash Meinel (18):
Change the rename code to create a new text entry.
Change (.keys(), .values()) to using (.iteritems())
Change to multiplying tests rather than manually.
add the failing test that we preserve the last-modified revision
(broken) Start working towards using CommitBuilder rather than using a custom implementation.
Found a bug in CommitBuilder.finish_inventory().
We need to handle when the object has been deleted.
Fall back to the repository for cases where the content is not present in the stream yet.
Dump sticky blobs to disk when memory pressure gets high.
Switch to closing the large-content blobs that we store to disk.
bzr *does* run atexit functions when exiting, but doesn't run deconstructors.
The fix that landed in bzr did not set builder.new_inventory. :(
Use the new KnownGraph.add_node() functionality.
When post-processing the delta stream, don't ask to generate a full inventory to check for deletions.
Some debugging code. It looks like the main bugs involve files that are deleted and restored.
Add a bunch of direct tests for the _TreeShim interface.
Switch to using CommitBuilder
Include the known_graph code.
John Whitley (1):
Don't emit directory info when plain format is specified.
Jonathan Nieder (2):
Fix typo in setup.py
python-fastimport: overview documentation
Justin Bogner (1):
Don't bail for http://path/to/repo darcs remotes
Lars Hoss (1):
Add missing import
Matt McClure (2):
Adds a fast-export-from-p4 command.
Verify the presence of the p4 command-line tool and the Perforce Python
Matthias Andree (14):
Don't guess darcs-2 repo format, but look at _darcs/format.
Use proper ^[ sequence for escaping \x1b.
Revamp option parser, using optparse.
clean up close() calls
Only read _darcs/format if present.
Change to original wd before writing export marks.
Let user specify --encoding on command line.
Tweak progress output a bit.
Flush log regularly. (This isn't sync, so it's cheap enough.)
Mangle whitespace in tags to please git fast-import.
Revamp README.
Add support for hashed repository formats.
Add author mapping feature similar to git svn fetch --authors-file.
Detail Matthias Andree's contributions.
Mattia Rizzolo (1):
Add patch from upstream to fix a crash with UTF-8 encoded author names
Max Bowsher (8):
Store tags on entire branch ancestry, not just left-hand ancestry.
Set a tag when touching a refs/tags/ ref with a commit command.
Avoid spurious 'git-' being prefixed on branches whose names happen to end with 'trunk',
Remove unused and unimplemented BranchMapper.bzr_to_git, and update docstring.
Make BranchMapper just map one name per call.
Store the BranchMapper in the CacheManager so it can be got from other places.
Default branch-nick to mapped git ref name.
Also catch tagging via commit when resuming a crashed import.
Miklos Vajna (279):
initial version
add chardet support
add progress command
more progress messages
escape a bad char. more to come?
prefix progress with timestamp
remove fixme about the +x bit
add license header
add a readme
add a Makefile
typo in README
add goals section
add download section
use os.path.join()
log darcs apply output
link undarcs
handle empty author, bzr fast-export does not like it
Implement the 'deleteall' command.
mention how to darcs2bzr
add test scripts and a very simple repo
add some note about hg fastimport
mention hg fastimport
avoid deleteall
add bzr versions
t: add 2 new test patches
update hg infos
test-bzr: avoid unnecessary bzr branch
avoid hardcoding internal darcs backup paths
reorder darcs/git version infos
README: add a simple todo
avoid an unnecessary chdir()
test-git: avoid pristine, fix exit code
test-hg: avoid pristine fix exit code
test-bzr: avoid pristine fix exit code
import of t/Makefile
Revert "avoid hardcoding internal darcs backup paths"
add test2 - a darcs2 repo (doesn't work yet)
initial darcs2 support
darcs2: log the output
remove files which should be untracked
add darcs2->git test
update readme
add a NEWS file
start marks from 1
generate the darcs1 repo automatically
port test-bzr to the test lib
port test-hg.sh
remove static test2
port test2-git
eye candy
implement --export-marks
implement import_marks as well
t: typo fixes
t: cleanup
t: extract diff_git()
t: diff_git(): cleanup before clone
t: add die() to lib.sh
t: typos
t: lib: add upd_file_darcs()
incremental conversion is now working
add a new test to demonstrate incremental conversion
update news for 0.4
TAG 0.4
readme: try to be a bit less subjective
be nice to people who helped and who wrote backends
add incremental darcs2->bzr test
t: use diff_bzr in test-bzr.sh
t: add diff_hg()
update bzr version
update NEWS
README: avoid 'I', 'me' and 'my', I'm no longer the only contributor
handle old darcs patches where the format in the xml differs
fix for the case when tags have a second "Ignore-this:" line
add support for darcs1 incremental conversion
add darcs1 -> git incremental testcase
fix darcs2 incremental support
update NEWS
make workdir location configurable
add test for --working
make --working work with relative paths
update NEWS
darcs-fast-export.py -> darcs-fast-export
add "additional resources" to the readme
fix the case when we update an incremental import again
t: PATH tweaks
add d2x
d2x: add one-time hg support
document d2x
dfe: append a newline to the marks file
d2x: beautify output when no changes to pull
dfe: speed up incremental conversions a lot
d2x: cleanup
make the logfile path configurable
update NEWS
fix an error triggered by converting the tailor darcs repo
fix an error triggered by converting the haddock darcs repo
bump darcs requirement to a snapshot
update for 0.5
TAG 0.5
t: add create_git() to lib.sh
darcs-fast-import: initial import
add a TODO
handle progress command
t: remove unneded tmp branch after merging in create_git()
handle committer in case author is not given
implement skip_optional_lf()
make d-f-i a bit more oo-ish
make the last read line readable everywhere
new handle_mark()
add handle_data()
redesign handle_commit()
skip_optional_lf(): use self.read() as well
do not darcs add files which would be removed in this commit
d-f-i: support inline data
d-f-e: don't use global vars without a good reason
d-f-e: don't export ignore-this lines
initial tag support
d-f-e: fix exporting multiline commit messages
add new test: d-f-e | d-f-i
fix 'Format of STDIN stream' doc url
silently ignore checkout
don't try to unlink non-existing files
implement deleteall command
reorder to match spec order
add rename command
implement filecopy
todo update
add --export-marks option
add --import-marks option
add incremental import test
d-f-i: avoid adding re-adding already added files in incremental mode
update TODO
t: implement create_bzr() and diff_importbzr()
d-f-i: ignore the from line after the reset one as well
add bzr->darcs test
reset ignore: but only in case it's a "from" line
add hg->darcs test
update todo
d-f-i: add --logfile
d-f-e: handle the case when we just continue an incremental conversion
initial two-way converter test
t: testimport-git-twoway.sh: do a bit more stress-test
update doc
update TODO
one more todo
avoid hardcoded python path
d2x: typo
initial import of x2d
add bzr x2d testcase
add hg x2d test
update todo
syntax fix
update doc
bzr fastimport seem not to work here fine for two-way sync
document how to install the hg plugin
add d-f-e --git-branch
git-darcs: initial import
add git-darcs test
g-d: make push behave like git-push
g-d: be more explicit to the user when nothing fetched
document git-darcs in the README
update NEWS
g-d: add pull subcommand
update doc
add d2x manpage
add d-f-e manpage
add d-f-i manpage
add g-d manpage
add x2d manpage
add makefile rules for html/man docs
fix asciidoc man warnings
add install target
i want dashes in manpage names
update todo
the right fix for the dash problem
doc update
fix git-darcs manpage example
add dist and release targets
updates for 0.6
TAG 0.6
todo update
add a benchmark script
bench: be a bit more verbose
add HACKING
don't do the re-add dance in every commit for unchanged files
add verify support to the bench script
bench: leave everything there if the conversion is corrupted so it's possible to debug it
be a bit more verbose in the log file so we can see what takes so much time
fix typo when doing an continuing an incremental export and the workdir is removed
bench: default to all repos
document status (outside versions)
rename directpatch to oldfashionedpatch
initial darcs2 inventory parser
cleanup
fix testimport-git-twoway.sh
do not try to rename non-existing files
extract inline_path()
add a new simplify_nodes() function
d2x: allow specifying extra d-f-e options
do renames first, we need it now that we simplify nodes
update NEWS
d-f-e: simplify add A, remove A in the same commit
bench: specify git version in log filename
bench: allow benchmarking a single repo
add bench script for tailor
d-f-e: add --progress option
d-f-e: avoid firstChild and toxml()
updates for 0.7
TAG 0.7
bench-tailor: clean up config
bench-tailor: ignore charset problems
bench-tailor: fix char err ignore opt
docu update
update numbers and add Lele
fix renaming of empty dirs
don't try to list non-dirs
fix renaming multiple empty dirs
d-f-e: simplify 'adddir A, rename A B' to -> noop
d-f-e: do not duplicate lists with no good reason
d-f-e: fix inverted patch prefix
hack for the inverted patch + file modify combo
add new tests to demonstrate a breakage
Revert "don't do the re-add dance in every commit for unchanged files"
update HACKING
whoops, that would have to be part of the revert (Lele is still in the xml branch)
asciidoc: fix refentryinfo position
t/bench: log the number of patches as well
add bench graph code
update NEWS
add link to the benchmark details
add big-zoo reference
update VERSION to 0.8
TAG 0.8
new todo items
evaluate darcs-to-git
update undarcs' entry
add new darcs bug id
fix python2.6 warning by replace the popen2 module with the subprocess one
fast-export: escape anything which is [^a-zA-Z0-9.] with an undescore
add test for having accents in tag names
update NEWS
README: at least we can now require a stable darcs
t: exclude more hg files
bzr-fast-export -> bzr-fast-export.py
Merge branch 'add-dfe-opts' of git://github.com/warner/darcs-fast-export
port d667ba5 to darcs-fast-import
updates for 0.9
TAG 0.9
update bzr and hg versions
improve typical workflow example
gd: make find-git output similar to the find-darcs one
get rid of HACKING, as we won't merge the side branches to bzr-fastimport
Merge darcs-fast-export
darcs-fast-export: bash4 fixes
darcs-fast-export: use --include-merges option to bzr-log in x2d
darcs-fast-export: document the --authors-file option
darcs-fast-export: document operands to options which have one
darcs-fast-export: give a more reasonable error message when chardet is not found
darcs: fix importing files in subdirs
darcs: make sure the working directory has no changes when verifying imports
darcs: handle timezones like +1300 in the importer
... 80 lines suppressed ...
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/python-modules/packages/python-fastimport.git
More information about the Python-modules-commits
mailing list