| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
multi-threading implementation of all odb functions
|
| |
|
|
| |
everything. Next is to implement pack-file reading, then alternates which should allow to resolve everything
|
| | |
|
| |
|
|
|
|
| |
parsing which truncated newlines although it was ilegitimate. Its up to the reader to truncate therse, nowhere in the git code I could find anyone adding newlines to commits where it is written
Added performance tests for serialization, it does about 5k commits per second if writing to tmpfs
|
| | |
|
| |
|
|
| |
missing ) and added performance tests which are extremely promising
|
| |
|
|
| |
efficiently considering that it copies string buffers all the time
|
| | |
|
| |
|
|
| |
appears to be working
|
| |
|
|
|
|
| |
objects will be written using our utilities, and certain object retrieval functionality moves into the GitObjectDatabase which is used by the repo instance
Added performance test for object database access, which shows quite respectable tree parsing performance, and okay blob access. Nonetheless, it will be hard to beat the c performance using a pure python implementation, but it can be a nice practice to write it anyway to allow more direct pack manipulations. Some could benefit from the ability to write packs as these can serve as local cache if alternates are used
|
| |
|
|
| |
bug of course which just didn't kick in yet
|
| |
|
|
|
|
| |
from their object information directly. This is faster, and resolves issues with the rev-list format and empty commit messages
Adjusted many tests to go with the changes, as they were still mocked. The mock was removed if necessary and replaced by code that actually executes
|
| |
|
|
|
|
|
| |
performance is slightly better
git.cmd: added method to provide access to the content stream directly. This is more efficient if large objects are handled, if it is actually used
test.helpers: removed unnecessary code
|
| |
|
|
| |
objects if it could serialize itself
|
| |
|
|
| |
subproces.Popen invocation. It could be used to pass custom environments, without changing the own one (#26)
|
| |
|
|
| |
the child less cluttered, and make it easier to debug as it will only have the file descriptors we set. It appears to be more stable regarding the stdin-is-closed-but-child-doesn't-realize-this issue
|
| |
|
|
|
|
| |
has unified its way it reads from stdin, now it wants all items to be terminated by a newline usually. Previously, it could have been that it really didn't want to have a termination character when the last item was written to the file. Bumped the minimum requirements to 1.7.0 to be sure it is working as I think it will.
Still, I have to admit that sometime it just appears the closed pipe will not stop git from waiting for more input, at least with the previous implementation
|
| |
|
|
| |
convenience)
|
| |
|
|
| |
containing the lock was removed. This is unlikely to happen in a production envrironment, but may happen during testing, as folders are moved/deleted once the test is complete. Daemons might still be waiting for something, and they should be allowed to terminate instead of waiting for a possibly long time
|
| |
|
|
|
|
| |
newline at the end of the split line was not split away automatically. Added test for this, and the trivial fix
Wow, at least two people reviewd the code, but it slipped through anyway :)
|
| |
|
|
| |
Cmd: AutoInterrupt handles boundary cases more gracefully as it can be that the os module suddenly becomes None if the interpreter is going down
|
| |
|
|
|
|
|
|
| |
and absolute paths were given
Commit.create_from_tree: fixed critical bug that would cause it to create a branch named master by default, instead of the reference actually set ( which is master in many, but not all cases )
- in fact it could be detached as well, we would fail ungracefully although we could assume master then ... although we cant really make the decision
Repo.is_dirty: improved its abiility to deal with empty repositories and a missing head. Weird thing is that the test always worked fine with the previous code, but it didn't work for me in a similar situation without this change at least
|
| | |
|
| |
|
|
|
|
|
| |
different path in the index than the actual one on disk ( from which the object will be created )
Fixed bug the way newlines were handled, which hopefully fixes occasional hangs as well. It works fine with git 1.7.1
Most of the changes are due to the tab-space conversion - its weird once more as I thought it was all in spaces before ... .
|
| |
|
|
|
|
|
|
| |
repo.is_dirty: Will not fail on empty repo ( anymore )
index.entries: will just be empty if the repository is empty
refs: added to_full_path method which can be used to create fully synthetic instances of Reference types, added a test for it
Converted all touched files to spaces, which is why git reports so many changed files. Actually I was thinking every file would use spaces, but apparently not
|
| |
|
|
|
|
|
| |
on lighthouse.
README/intro.rst: added information about the new repository at github
tree: added marker to indicate that submodules would have to be returned there
|
| |
|
|
| |
if an unset slot is being deleted.
|
| |
|
|
| |
Fixed test which required to be on master to work - it now uses the HEAD symbolic ref instead.
|
| |
|
|
|
|
|
| |
This commit includes
- an update to git.objects.utils:parse_actor_and_date to parse the timezone offset
- updates to the git.objects.Commit and git.objects.Tag objects to support *_tz_offset attributes
- updates to tests in test.git.test_commit and test.git.test_refs to check for appropriate *_tz_offset attributes
|
| | |
|
| |
|
|
|
|
| |
to paths anymore as it will only append '--' if paths are actually given.
Added unittest to verify this
|
| |
|
|
| |
previous version was making assumptions that would only be true for old git repositories it sesms. The new version of the algorithm deals with this gracefully.
|
| |
|
|
| |
data, which becomes important if git-write-tree is supposed to be used
|
| |
|
|
| |
test.helpers: temporary rw repository creators now set the working dir of the program, easing working with relative paths a lot
|
| |
|
|
| |
exists. Previously it would just fail
|
| |
|
|
| |
regarding its parents
|
| |
|
|
| |
order to make it impossible for threads to overwrite each other if we are talking about miliseconds. This renders plenty of 'special' handling redundant
|
| | |
|
| |
|
|
| |
Repo.init: fixed incorrect use of the path which was to optionally specify where to initialize the repository. Update test as well
|
| |
|
|
|
|
| |
available at some point natively, within the git suite
Progress parsing now deals properly with Ascii_Escape characters that are meant for the tty - git might stop sending this at some point, but we can deal with it no matter what
|
| |
|
|
| |
throughput )
|
| |
|
|
| |
allowing the keys to contain slashes; adjusted test to check for this
|
| |
|
|
| |
locks will check whether the lock they wrote truly is theirs - in case threads are racing, this might not be the case. Now this issue will be detected and results in a proper failure
|
| |
|
|
| |
uses a 'failing' lock file, but now its possible to easily put a blocking lock file in its place
|
| | |
|
| |
|
|
|
|
| |
caused lockups while waiting for stdout
NOTE: This does not have the desired effect, the issue appears to be somewhere within git, possibly, as git simply does not terminate waiting for something, even if stdout is closed.
|
| | |
|
| | |
|
| |
|
|
| |
based git operations in concurrent environments
|
| |
|
|
|
|
| |
exist more gracefully
Added BlockingLockFile which may be used to wait for the lock according to some criteria
|
