summaryrefslogtreecommitdiff
path: root/gitdb
Commit message (Collapse)AuthorAgeFilesLines
* chore(ver): bump v2.1.1.dev0 --> dev1Kostis Anagnostopoulos2016-10-281-1/+1
|
* feat(mman): drop PY25 support - always use sliding global-mmanKostis Anagnostopoulos2016-10-281-10/+3
|
* fix(leaks): close "leaky" `memoryview` on mmaps after `to_bytes()`Kostis Anagnostopoulos2016-10-281-0/+3
| | | | | All 3 gitpython projects stopped using memoryviews, but it's probably safe to release also user's memviews it after it has been converted to bytes.
* refact(elapsed): improve no div0 when time-elapsed too smallKostis Anagnostopoulos2016-10-282-16/+16
|
* fix(compat): PY3-check must hold even for PY4Kostis Anagnostopoulos2016-10-281-1/+1
|
* chore(ver): bump v2.1.0.dev1 --> 2.1.1.dev0v2.1.1.dev0Kostis Anagnostopoulos2016-10-281-1/+1
| | | | | | | + travis: depend on smmap2-v2.1.1.dev0 + fix leaks_memview: + smmap2: yes + gitdb2: no
* fix(io): BREAKING, wrap more out-stream usagesKostis Anagnostopoulos2016-10-2715-191/+189
| | | | | | + chrore(deps): depend on *contextlib2* for `ExitStack` in PY2. + refact(util): BREAKING API move consts out of utils. + style(pep8): fixe all sources.
* refact(win_errs): move HIDE_WINDOWS_KNOWN_ERRORS from main-code to testKostis Anagnostopoulos2016-10-254-10/+18
|
* test(travis): enable all tests (inc perf) on TravisCIKostis Anagnostopoulos2016-10-253-9/+0
|
* style(listuple): use literals for empty lists/tuplesKostis Anagnostopoulos2016-10-257-11/+11
|
* chore(ver): bump 2.0.0.dev0-->2.1.0.dev2 (yes, last ver was old)v2.1.0.dev2Kostis Anagnostopoulos2016-10-241-1/+1
|
* fix(pack): restore packers as LazyMixinsKostis Anagnostopoulos2016-10-241-88/+75
| | | | | | | Speed-up for non-error commits: + no-leak contextlibing(v.2.0.0, 941b6c7e): 5.20 + Prev commit, no LazyMixin(v.2.1.0.dev1): 7.70 + This commit, LazyMixin: 5.50
* fix(win): FIX and HIDE 2 win-errors remainingv2.1.0.dev1Kostis Anagnostopoulos2016-10-2413-88/+117
| | | | | + File-in-use errors were fixed with `gitdb.util.mman.collect()`! + This call is disabled `gitdb.util.HIDE_WINDOWS_KNOWN_ERRORS == False`. + Depend on latest smmp `v2.1.0.dev1` tag
* chore(ver): bump 2.0.0-->2.1.0.dev0v2.1.0.dev0Kostis Anagnostopoulos2016-10-241-1/+1
|
* chore(gitdb): actually delete submodule from sourcesKostis Anagnostopoulos2016-10-241-0/+0
|
* chore(ci): depend on "leaks" smmap branchKostis Anagnostopoulos2016-10-241-19/+0
| | | + stop importing git-submodules for gitdb & smmap
* feat(io): breaking API: retrofit Packers as context-managers!Kostis Anagnostopoulos2016-10-2410-295/+458
| | | | | | | | | | + Packers MUST be invoked inside `Withh...` blocks, or `_cursor` won't exist! + Had to drop NotLazy for their hierarchy :-( + Count entrances/exits. + feat(util: add `rmtree()` for READ_ONLY files on Windows. 3-->2 Windows TCs now fail.
* feat(io): Retrofit streams as context-managers.Kostis Anagnostopoulos2016-10-247-29/+147
| | | | | + feat(util): add logger. + feat(util): add suppress-ex context-handler (from PY3 sources).
* feat(src): subst `reduce` with `sum` for size calcsKostis Anagnostopoulos2016-10-245-9/+5
| | | | + fix(loose-db): fix bad-attr in ex-message
* tc: HALF FIX of `test_pack_entity ()`Kostis Anagnostopoulos2016-10-221-3/+6
| | | | | + On Windows, you cannot write onto a file held by another live file-pointer (test_pack.py:#L204). + The TC fails later, on clean up (the usual).
* TCs: fix div-by-zero on elapsed times (appveyor CPU is fast!)Kostis Anagnostopoulos2016-10-223-11/+11
|
* chore(rename): gitdb2 v2.02.0.0Sebastian Thiel2016-10-162-1/+1
| | | | v2 is chosen to better match the name.
* support optional gitdb_speedupsThomas Grainger2016-10-046-1270/+3
|
* Merge pull request #28 from scop/withSebastian Thiel2016-07-304-15/+10
|\ | | | | Handle more file open/close with "with"
| * Handle more file open/close with "with"Ville Skyttä2016-07-274-15/+10
| |
* | Spelling fixesVille Skyttä2016-07-279-10/+10
|/
* fix(loose): avoid unnecessary file rename on windowsSebastian Thiel2015-10-042-4/+9
| | | | | | This should workaround possible permission issues. Related to https://github.com/gitpython-developers/GitPython/issues/353
* docs(gitdb): discourage usage of GitDB typeSebastian Thiel2015-08-221-1/+5
|
* Duplicate `const` fixedGermán M. Bravo2015-04-071-5/+5
| | | | Remove duplicate `const` to stop the warning: "duplicate 'const' declaration specifier"
* Bumped version to 0.6.40.6.4Sebastian Thiel2015-01-121-1/+1
|
* Adjusted stream logic to make it work on all tested platforms ... .Sebastian Thiel2015-01-123-4/+11
| | | | | | | | | | | | | | | | | As taken from https://github.com/gitpython-developers/gitdb/blob/master/gitdb/stream.py#L292 -> NOTE: Behavior changed in PY2.7 onward, which requires special handling to make the tests work properly. They are thorough, and I assume it is truly working. Why is this logic as convoluted as it is ? Please look at the table in https://github.com/gitpython-developers/gitdb/issues/19 to learn about the test-results. Bascially, on py2.6, you want to use branch 1, whereas on all other python version, the second branch will be the one that works. However, the zlib VERSIONs as well as the platform check is used to further match the entries in the table in the github issue. This is it ... it was the only way I could make this work everywhere. IT's CERTAINLY GOING TO BITE US IN THE FUTURE ... . <- Fixes #19
* minor spell fixes + empty line unification + comparison for python 2.6Yaroslav Halchenko2015-01-084-9/+7
|
* Make tests independent of actual repository dataSebastian Thiel2015-01-082-4/+5
| | | | | | | Therefore, hardcoded sha's are not allowed anymore, as the contents of the repository is unknown. Fixes #16, for real this time ;)
* Removed unused imports - should have been in the last commit obviouslySebastian Thiel2015-01-082-2/+0
|
* Fixed up tests to use the GITDB_TEST_GIT_REPO_BASE at all timesSebastian Thiel2015-01-086-41/+43
| | | | | | | I have verified that all tests are working, even without a parent git repository, as long as the said environment variable is set. Fixes #16
* Protected stream closure against possibilty of being a bytesSebastian Thiel2015-01-072-2/+4
| | | | | | For some reason, it gets bytes where it did expect a stream ... . Probably I should have figured out where this was input, instead of fixing it the brutal way
* Fixed possible file-handle leakSebastian Thiel2015-01-071-3/+11
| | | | | Configured travis to artificially restrict handle count to protect from regression in that regard
* Bumped version to 0.6.30.6.3Sebastian Thiel2015-01-061-1/+1
|
* Fixed yet another issue with smmap's latest changesSebastian Thiel2015-01-062-2/+15
| | | | Now we deal with memory views as well ...
* Fixed python 3 compatibility issue that only showed on windows0.6.2Sebastian Thiel2015-01-062-7/+7
| | | | And bumped version to 0.6.2
* A minor fix after porting git-python over to PY3Sebastian Thiel2015-01-061-1/+1
| | | | It doesn't do anything (in terms of fixing an issue), but it should be more correct than what was there previously
* Applied autopep8Sebastian Thiel2015-01-0434-283/+366
| | | | autopep8 -v -j 8 --max-line-length 120 --in-place --recursive
* Improved decompression test to scan the entire git repository, instead of ↵Sebastian Thiel2015-01-011-2/+4
| | | | | | | | just packs This should make it easier to assert the issue is truly fixed now [skip ci]
* Bumped new versionSebastian Thiel2015-01-012-3/+3
| | | | Fixed tiny issue in python 3
* Added integrity test for loose objects to search large datasets forSebastian Thiel2015-01-011-0/+31
| | | | | | the issue described in https://github.com/gitpython-developers/GitPython/issues/220 See test notes for proper usage, it all depends on a useful dataset with high entropy
* Slightly improved loose object decompression testSebastian Thiel2015-01-011-1/+9
|
* Increased initial size of decompressed data to obtain loose object header ↵Sebastian Thiel2015-01-014-8/+15
| | | | | | | | | information This appears to fix https://github.com/gitpython-developers/GitPython/issues/220 , in this particular case. Nonetheless, we might just have gotten lucky here, and the actual issue is not yet solved and can thus re-occour. It would certainly be best to churn through plenty of loose objects to assure this truly works now. Maybe the pack could be recompressed as loose objects to get a sufficiently large data set
* Added readthedocs badgeSebastian Thiel2014-11-191-0/+0
|
* It seems zlib works differently in py26, and thus requires special handling.py2n3Sebastian Thiel2014-11-142-17/+11
| | | | This also explains why the tests suddenly stopped working - after all, the interpreter changed ... .
* Now I am skipping a problematic test on travis CI.Sebastian Thiel2014-11-141-2/+12
| | | | Maybe I can find a py 2.6 interpreter somewhere to reproduce it.
View Lorry Controller Status