5 * Cope with different zlib buffer sizes in sha1 file parser.
8 * Fix get_transport_and_path for HTTP/HTTPS URLs.
11 * Avoid calling free_objects() on NULL in error cases. (Chris Eberle)
13 * Fix use --bare argument to 'dulwich init'. (Chris Eberle)
17 * Add support for retrieving tarballs from remote servers.
18 (Jelmer Vernooij, #379087)
24 * Repo.do_commit has a new argument 'ref'.
26 * Repo.do_commit has a new argument 'merge_heads'. (Jelmer Vernooij)
28 * New ``Repo.get_walker`` method. (Jelmer Vernooij)
30 * New ``Repo.clone`` method. (Jelmer Vernooij, #725369)
32 * ``GitClient.send_pack`` now supports the 'side-band-64k' capability.
35 * ``HttpGitClient`` which supports the smart server protocol over
36 HTTP. "dumb" access is not yet supported. (Jelmer Vernooij, #373688)
38 * Add basic support for alternates. (Jelmer Vernooij, #810429)
42 * unittest2 or python >= 2.7 is now required for the testsuite.
43 testtools is no longer supported. (Jelmer Vernooij, #830713)
47 * Fix compilation with older versions of MSVC. (Martin gz)
49 * Special case 'refs/stash' as a valid ref. (Jelmer Vernooij, #695577)
51 * Smart protocol clients can now change refs even if they are
52 not uploading new data. (Jelmer Vernooij, #855993)
54 * Don't compile C extensions when running in pypy.
55 (Ronny Pfannschmidt, #881546)
57 * Use different name for strnlen replacement function to avoid clashing
58 with system strnlen. (Jelmer Vernooij, #880362)
62 * ``Repo.revision_history`` is now deprecated in favor of ``Repo.get_walker``.
69 * New DeltaChainIterator abstract class for quickly iterating all objects in
70 a pack, with implementations for pack indexing and inflation.
73 * New walk module with a Walker class for customizable commit walking.
76 * New tree_changes_for_merge function in diff_tree. (Dave Borowitz)
78 * Easy rename detection in RenameDetector even without find_copies_harder.
83 * Avoid storing all objects in memory when writing pack.
84 (Jelmer Vernooij, #813268)
86 * Support IPv6 for git:// connections. (Jelmer Vernooij, #801543)
88 * Improve performance of Repo.revision_history(). (Timo Schmid, #535118)
90 * Fix use of SubprocessWrapper on Windows. (Paulo Madeira, #670035)
92 * Fix compilation on newer versions of Mac OS X (Lion and up). (Ryan McKern, #794543)
94 * Prevent raising ValueError for correct refs in RefContainer.__delitem__.
96 * Correctly return a tuple from MemoryObjectStore.get_raw. (Dave Borowitz)
98 * Fix a bug in reading the pack checksum when there are fewer than 20 bytes
99 left in the buffer. (Dave Borowitz)
101 * Support ~ in git:// URL paths. (Jelmer Vernooij, #813555)
103 * Make ShaFile.__eq__ work when other is not a ShaFile. (Dave Borowitz)
105 * ObjectStore.get_graph_walker() now no longer yields the same
106 revision more than once. This has a significant improvement for
107 performance when wide revision graphs are involved.
108 (Jelmer Vernooij, #818168)
110 * Teach ReceivePackHandler how to read empty packs. (Dave Borowitz)
112 * Don't send a pack with duplicates of the same object. (Dave Borowitz)
114 * Teach the server how to serve a clone of an empty repo. (Dave Borowitz)
116 * Correctly advertise capabilities during receive-pack. (Dave Borowitz)
118 * Fix add/add and add/rename conflicts in tree_changes_for_merge.
121 * Use correct MIME types in web server. (Dave Borowitz)
125 * write_pack no longer takes the num_objects argument and requires an object
126 to be passed in that is iterable (rather than an iterator) and that
127 provides __len__. (Jelmer Vernooij)
129 * write_pack_data has been renamed to write_pack_objects and no longer takes a
130 num_objects argument. (Jelmer Vernooij)
132 * take_msb_bytes, read_zlib_chunks, unpack_objects, and
133 PackStreamReader.read_objects now take an additional argument indicating a
134 crc32 to compute. (Dave Borowitz)
136 * PackObjectIterator was removed; its functionality is still exposed by
137 PackData.iterobjects. (Dave Borowitz)
139 * Add a sha arg to write_pack_object to incrementally compute a SHA.
142 * Include offset in PackStreamReader results. (Dave Borowitz)
144 * Move PackStreamReader from server to pack. (Dave Borowitz)
146 * Extract a check_length_and_checksum, compute_file_sha, and
147 pack_object_header pack helper functions. (Dave Borowitz)
149 * Extract a compute_file_sha function. (Dave Borowitz)
151 * Remove move_in_thin_pack as a separate method; add_thin_pack now completes
152 the thin pack and moves it in in one step. Remove ThinPackData as well.
155 * Custom buffer size in read_zlib_chunks. (Dave Borowitz)
157 * New UnpackedObject data class that replaces ad-hoc tuples in the return
158 value of unpack_object and various DeltaChainIterator methods.
161 * Add a lookup_path convenience method to Tree. (Dave Borowitz)
163 * Optionally create RenameDetectors without passing in tree SHAs.
166 * Optionally include unchanged entries in RenameDetectors. (Dave Borowitz)
168 * Optionally pass a RenameDetector to tree_changes. (Dave Borowitz)
170 * Optionally pass a request object through to server handlers. (Dave Borowitz)
174 * If setuptools is installed, "python setup.py test" will now run the testsuite.
177 * Add a new build_pack test utility for building packs from a simple spec.
180 * Add a new build_commit_graph test utility for building commits from a
181 simple spec. (Dave Borowitz)
187 * Fix double decref in _diff_tree.c. (Ted Horst, #715528)
189 * Fix the build on Windows. (Pascal Quantin)
191 * Fix get_transport_and_path compatibility with pre-2.6.5 versions of Python.
192 (Max Bowsher, #707438)
194 * BaseObjectStore.determine_wants_all no longer breaks on zero SHAs.
197 * write_tree_diff() now supports submodules.
200 * Fix compilation for XCode 4 and older versions of distutils.sysconfig.
205 * Sphinxified documentation. (Lukasz Balcerzak)
207 * Add Pack.keep.(Marc Brinkmann)
211 * The order of the parameters to Tree.add(name, mode, sha) has changed, and
212 is now consistent with the rest of Dulwich. Existing code will still
213 work but print a DeprecationWarning. (Jelmer Vernooij, #663550)
215 * Tree.entries() is now deprecated in favour of Tree.items() and
216 Tree.iteritems(). (Jelmer Vernooij)
222 * New `dulwich.diff_tree` module for simple content-based rename detection.
225 * Add Tree.items(). (Jelmer Vernooij)
227 * Add eof() and unread_pkt_line() methods to Protocol. (Dave Borowitz)
229 * Add write_tree_diff(). (Jelmer Vernooij)
231 * Add `serve_command` function for git server commands as executables.
234 * dulwich.client.get_transport_and_path now supports rsync-style repository URLs.
235 (Dave Borowitz, #568493)
239 * Correct short-circuiting operation for no-op fetches in the server.
242 * Support parsing git mbox patches without a version tail, as generated by
243 Mercurial. (Jelmer Vernooij)
245 * Fix dul-receive-pack and dul-upload-pack. (Jelmer Vernooij)
247 * Zero-padded file modes in Tree objects no longer trigger an exception but
248 the check code warns about them. (Augie Fackler, #581064)
250 * Repo.init() now honors the mkdir flag. (#671159)
252 * The ref format is now checked when setting a ref rather than when reading it back.
253 (Dave Borowitz, #653527)
255 * Make sure pack files are closed correctly. (Tay Ray Chuan)
259 * Run the tutorial inside the test suite. (Jelmer Vernooij)
261 * Reorganized and updated the tutorial. (Jelmer Vernooij, Dave Borowitz, #610550,
269 * HTTP server correctly handles empty CONTENT_LENGTH. (Dave Borowitz)
271 * Don't error when creating GitFiles with the default mode. (Dave Borowitz)
273 * ThinPackData.from_file now works with resolve_ext_ref callback.
276 * Provide strnlen() on mingw32 which doesn't have it. (Hans Kolek)
278 * Set bare=true in the configuratin for bare repositories. (Dirk Neumann)
282 * Use slots for core objects to save up on memory. (Jelmer Vernooij)
284 * Web server supports streaming progress/pack output. (Dave Borowitz)
286 * New public function dulwich.pack.write_pack_header. (Dave Borowitz)
288 * Distinguish between missing files and read errors in HTTP server.
291 * Initial work on support for fastimport using python-fastimport.
294 * New dulwich.pack.MemoryPackIndex class. (Jelmer Vernooij)
296 * Delegate SHA peeling to the object store. (Dave Borowitz)
300 * Use GitFile when modifying packed-refs in tests. (Dave Borowitz)
302 * New tests in test_web with better coverage and fewer ad-hoc mocks.
305 * Standardize quote delimiters in test_protocol. (Dave Borowitz)
307 * Fix use when testtools is installed. (Jelmer Vernooij)
309 * Add trivial test for write_pack_header. (Jelmer Vernooij)
311 * Refactor some of dulwich.tests.compat.server_utils. (Dave Borowitz)
313 * Allow overwriting id property of objects in test utils. (Dave Borowitz)
315 * Use real in-memory objects rather than stubs for server tests.
318 * Clean up MissingObjectFinder. (Dave Borowitz)
322 * ObjectStore.iter_tree_contents now walks contents in depth-first, sorted
323 order. (Dave Borowitz)
325 * ObjectStore.iter_tree_contents can optionally yield tree objects as well.
328 * Add side-band-64k support to ReceivePackHandler. (Dave Borowitz)
330 * Change server capabilities methods to classmethods. (Dave Borowitz)
332 * Tweak server handler injection. (Dave Borowitz)
334 * PackIndex1 and PackIndex2 now subclass FilePackIndex, which is
335 itself a subclass of PackIndex. (Jelmer Vernooij)
339 * Add docstrings for various functions in dulwich.objects. (Jelmer Vernooij)
341 * Clean up docstrings in dulwich.protocol. (Dave Borowitz)
343 * Explicitly specify allowed protocol commands to
344 ProtocolGraphWalker.read_proto_line. (Dave Borowitz)
346 * Add utility functions to DictRefsContainer. (Dave Borowitz)
353 * Fix memory leak in C implementation of sorted_tree_items. (Dave Borowitz)
355 * Use correct path separators for named repo files. (Dave Borowitz)
357 * python > 2.7 and testtools-based test runners will now also pick up skipped
358 tests correctly. (Jelmer Vernooij)
362 * Move named file initilization to BaseRepo. (Dave Borowitz)
364 * Add logging utilities and git/HTTP server logging. (Dave Borowitz)
366 * The GitClient interface has been cleaned up and instances are now reusable.
369 * Allow overriding paths to executables in GitSSHClient.
370 (Ross Light, Jelmer Vernooij, #585204)
372 * Add PackBasedObjectStore.pack_loose_objects(). (Jelmer Vernooij)
376 * Add tests for sorted_tree_items and C implementation. (Dave Borowitz)
378 * Add a MemoryRepo that stores everything in memory. (Dave Borowitz)
380 * Quiet logging output from web tests. (Dave Borowitz)
382 * More flexible version checking for compat tests. (Dave Borowitz)
384 * Compat tests for servers with and without side-band-64k. (Dave Borowitz)
388 * Clean up file headers. (Dave Borowitz)
392 * Use GitFile when modifying packed-refs in tests. (Dave Borowitz)
396 * dulwich.pack.write_pack_index_v{1,2} now take a file-like object
397 rather than a filename. (Jelmer Vernooij)
399 * Make dul-daemon/dul-web trivial wrappers around server functionality.
402 * Move reference WSGI handler to web.py. (Dave Borowitz)
404 * Factor out _report_status in ReceivePackHandler. (Dave Borowitz)
406 * Factor out a function to convert a line to a pkt-line. (Dave Borowitz)
411 note: This list is most likely incomplete for 0.6.0.
415 * Fix ReceivePackHandler to disallow removing refs without delete-refs.
418 * Deal with capabilities required by the client, even if they
419 can not be disabled in the server. (Dave Borowitz)
421 * Fix trailing newlines in generated patch files.
424 * Implement RefsContainer.__contains__. (Jelmer Vernooij)
426 * Cope with \r in ref files on Windows. (
427 http://github.com/jelmer/dulwich/issues/#issue/13, Jelmer Vernooij)
429 * Fix GitFile breakage on Windows. (Anatoly Techtonik, #557585)
431 * Support packed ref deletion with no peeled refs. (Augie Fackler)
433 * Fix send pack when there is nothing to fetch. (Augie Fackler)
435 * Fix fetch if no progress function is specified. (Augie Fackler)
437 * Allow double-staging of files that are deleted in the index.
440 * Fix RefsContainer.add_if_new to support dangling symrefs.
443 * Non-existant index files in non-bare repositories are now treated as
444 empty. (Dave Borowitz)
446 * Always update ShaFile.id when the contents of the object get changed.
449 * Various Python2.4-compatibility fixes. (Dave Borowitz)
451 * Fix thin pack handling. (Dave Borowitz)
455 * Add include-tag capability to server. (Dave Borowitz)
457 * New dulwich.fastexport module that can generate fastexport
458 streams. (Jelmer Vernooij)
460 * Implemented BaseRepo.__contains__. (Jelmer Vernooij)
462 * Add __setitem__ to DictRefsContainer. (Dave Borowitz)
464 * Overall improvements checking Git objects. (Dave Borowitz)
466 * Packs are now verified while they are received. (Dave Borowitz)
470 * Add framework for testing compatibility with C Git. (Dave Borowitz)
472 * Add various tests for the use of non-bare repositories. (Dave Borowitz)
474 * Cope with diffstat not being available on all platforms.
475 (Tay Ray Chuan, Jelmer Vernooij)
477 * Add make_object and make_commit convenience functions to test utils.
482 * The 'committer' and 'message' arguments to Repo.do_commit() have
483 been swapped. 'committer' is now optional. (Jelmer Vernooij)
485 * Repo.get_blob, Repo.commit, Repo.tag and Repo.tree are now deprecated.
488 * RefsContainer.set_ref() was renamed to RefsContainer.set_symbolic_ref(),
489 for clarity. (Jelmer Vernooij)
493 * The primary serialization APIs in dulwich.objects now work
494 with chunks of strings rather than with full-text strings.
501 * Support custom fields in commits (readonly). (Jelmer Vernooij)
503 * Improved ref handling. (Dave Borowitz)
505 * Rework server protocol to be smarter and interoperate with cgit client.
508 * Add a GitFile class that uses the same locking protocol for writes as
509 cgit. (Dave Borowitz)
511 * Cope with forward slashes correctly in the index on Windows.
512 (Jelmer Vernooij, #526793)
516 * --pure option to setup.py to allow building/installing without the C
517 extensions. (Hal Wine, Anatoly Techtonik, Jelmer Vernooij, #434326)
519 * Implement Repo.get_config(). (Jelmer Vernooij, Augie Fackler)
521 * HTTP dumb and smart server. (Dave Borowitz)
523 * Add abstract baseclass for Repo that does not require file system
524 operations. (Dave Borowitz)
530 * Add ObjectStore.iter_tree_contents(). (Jelmer Vernooij)
532 * Add Index.changes_from_tree(). (Jelmer Vernooij)
534 * Add ObjectStore.tree_changes(). (Jelmer Vernooij)
536 * Add functionality for writing patches in dulwich.patch.
547 * dulwich.object_store.tree_lookup_path will now return the mode and
548 sha of the object found rather than the object itself.
552 * Use binascii.hexlify / binascii.unhexlify for better performance.
554 * Cope with extra unknown data in index files by ignoring it (for now).
556 * Add proper error message when server unexpectedly hangs up. (#415843)
558 * Correctly write opcode for equal in create_delta.
564 * Implement ShaFile.__hash__().
566 * Implement Tree.__len__()
570 * Check for 'objects' and 'refs' directories
571 when looking for a Git repository. (#380818)
577 * Support the encoding field in Commits.
579 * Some Windows compatibility fixes.
581 * Fixed several issues in commit support.
585 * Basic support for handling submodules.
591 * Implemented Repo.__getitem__, Repo.__setitem__ and Repo.__delitem__ to
596 * Removed Repo.set_ref, Repo.remove_ref, Repo.tags, Repo.get_refs and
597 Repo.heads in favor of Repo.refs, a dictionary-like object for accessing
602 * Removed import of 'sha' module in objects.py, which was causing
603 deprecation warnings on Python 2.6.
609 * A new function 'commit_tree' has been added that can commit a tree
614 * The memory usage when generating indexes has been significantly reduced.
616 * A memory leak in the C implementation of parse_tree has been fixed.
618 * The send-pack smart server command now works. (Thanks Scott Chacon)
620 * The handling of short timestamps (less than 10 digits) has been fixed.
622 * The handling of timezones has been fixed.
628 * Fix compatibility with Python2.4.
634 * Support for activity reporting in smart protocol client.
636 * Optional C extensions for better performance in a couple of
637 places that are performance-critical.
643 * Fixed regression in Repo.find_missing_objects()
645 * Don't fetch ^{} objects from remote hosts, as requesting them
648 * Always write pack to disk completely before calculating checksum.
652 * Allow disabling thin packs when talking to remote hosts.