| Age | Commit message (Collapse) | Author |
|---|
|
+ HookException thrown on Popen, and were missed on Windows.
+ No SHELL on Popen??
+ Minor fixes:
+ Try harder to delete trees - no remorses.
+ Simplify exception reprs.
+ Unittest-ize test_index assertions.
|
|
+ CAUSE: In Windows, Diffs freeze while reading Popen streams,
probably buffers smaller; good-thin(TM) in this case because reading a
Popen-proc from the launching-thread freezes GIL. The alternative to
use `proc.communicate()` also relies on big buffers.
+ SOLUTION: Use `cmd.handle_process_output()` to consume Diff-proc
streams.
+ Retroffited `handle_process_output()` code to support also
byte-streams, both Threading(Windows) and Select/Poll (Posix) paths
updated.
- TODO: Unfortunately, `Diff._index_from_patch_format()` still slurps
input; need to re-phrase header-regexes linewise to resolve it.
|
|
+ DataDriven TCs for identifying which separate case failed.
+ appveyor: rework matrix, conda3.4 cannot install in develop mode
|
|
+ Stop using gitdb's respective helper.
+ Fix files chmod(555) which CANNOT DELETE on Windows (but do on Linux).
|
|
|
|
+ FIX all Diff TCs.
|
|
+ TCs: unittest-asserts for git-tests.
|
|
+ PY3 means "PY3 or later" (TODO: fix also for *gitdb* project).
|
|
+ Simplify call_process, no win-code case, no `make_call()` nested func.
+ Del needless WinError try..catch, in `_call_process()` already
converted as GitCommandNotFound by `execute()`.
+ pyism: kw-loop-->comprehension, facilitate debug-stepping
|
|
+ On MINGW-git, daemon exists but if invoked as 'git daemon', DAEMON
CANNOT DIE!
+ So, launch `git-daemon` on Apveyor, but
- remote TCs fail due to paths problems.
+ Updated README instructions on Windows.
+ Restore disabled remote TCs on Windows.
+ Disable failures on daemon-tests only the last moment
(raise SkipTest) so when ready, it will also pass.
|
|
|
|
+ see
http://help.appveyor.com/discussions/problems/5334-nosetests-finsih-bu-build-stuck-and-next-job-dealys-to-start
+ Use `io.DEFAULT_BUFFER_SIZE`.
+ test_commit: replace asserts with unittest-asserts.
- TRY Popen() NO universal_newlines: NO, reverted in next commits.
+
[travisci skip]
|
|
+ Retrofit try...finally blocks to ensure killing the daemon - now
vulnerable also on Windows due to Popen() + CREATE_NEW_PROCESS_GROUP
- BUT `test_base.test_with_rw_remote_and_rw_repo()` TC fails in MINGW
due to invalid remote-URL in fetching-repo's config. Another day.
- NEXT FREEZE to solve: test-diff_interface() under MINGW!
|
|
|
|
+ test_diff: replace asserts with unittest-asserts.
|
|
+ FIXED most hangs BUT no more `git-daemon` un-killable!
+ Use logger for utils to replace stray print().
|
|
|
|
+ More win-fixes:
+ Do not check unicode files in < py3.
+ util, #519: x4 timeout of lock-file blocking, failing in Appveyor.
|
|
|
|
+ The code in `_read_lines_from_fno()` was reading the stream only once
per invocation, so when input was larger than `mmap.PAGESIZE`, bytes
were forgotten in the stream.
+ Replaced buffer-building code with iterate-on-file-descriptors.
+ Also set deamon-threads.
|
|
|
|
+ Fixed the hangs at `test_git:TestGit.test_handle_process_output()`.
[travisci skip]
|
|
+ Del extra spaces, import os.path as osp
|
|
Thanks to @yarikoptic for catching this one !
|
|
It ain't pretty, but should do the job.
Related to #514
|
|
Relying on the destructor will not work, even though
the code used to rely on it.
Now we handle failures more explicitly. Far from perfect,
but a good start for a fix.
Fixes #514
|
|
by demanding a space after a word remote, and space is explicitly described as
a delimiter in man git-config:
To begin a subsection put its name in double quotes, separated by space from the section name,
otherwise e.g.
File "/usr/lib/python2.7/dist-packages/datalad/support/gitrepo.py", line 836, in get_remote_branches
for remote in self.repo.remotes:
File "/home/yoh/deb/gits/python-git/git/repo/base.py", line 271, in remotes
return Remote.list_items(self)
File "/home/yoh/deb/gits/python-git/git/util.py", line 745, in list_items
out_list.extend(cls.iter_items(repo, *args, **kwargs))
File "/home/yoh/deb/gits/python-git/git/remote.py", line 453, in iter_items
raise ValueError("Remote-Section has invalid format: %r" % section)
ValueError: Remote-Section has invalid format: u'remotes'
|
|
Fixes #508
|
|
It's not optimal, as we can now return one of two types
which are only compatible in the most basic ways.
However, it is better than before, I presume.
Fixes #510
|
|
It's entirely untested if this repo still does the right thing,
but I'd think it does.
Fixes #504
|
|
As TagObjects can point to other TagObjects, we need
to keep going in order to resolve the final commit.
Fixes #503
|
|
Update Remotes section
The Remotes section was missing some pretty important info.
|
|
Assure that gpgsig is not initialized with None
to allow the automatic deserialization to kick in.
Fixes #500
|
|
|
|
|
|
|
|
is_dirty supports path. Fixes #482.
|
|
|
|
|
|
|
|
- add a second line to commit messages with the "BAD MESSAGE" text
- read in the log and confirm that the seond line is not in the log file
|
|
|
|
|
|
This fixes a UI problem with using GitPython from a GUI python probgram.
Each repo that is opened creates a git cat-file processs and that provess will create
a console window with out this change.
|
|
|
|
Add missing newline when writing a symbolic ref.
|
|
That way, we do not have to figure the change type out by
examining the diff object.
It's implemented in a way that should yield more desireable results
as we keep the change-type that git is providing us with.
Fixes #493
|
|
|
|
|
|
That way, we will not only get the summary line
contained in the blame, but fetch the full message.
This is more costly than the previous implementation
allowed it to be, but being less surprising/correct
certainly is the preferred behaviour here.
Fixes #485
|
