summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* git cmd implementation of repository appears to work, at least this is what ↵Sebastian Thiel2011-05-3021-83/+118
| | | | the test suggests. Pure python implementation still has some trouble, but this should be very fixable
* Fixed tests far enough to allow basic repository tests to be applied to any ↵Sebastian Thiel2011-05-3021-65/+152
| | | | of the new database types. This reduces code duplication to the mere minimum, but allows custom tests to be added on top easily and flexibly
* Finished moving all repository methods to the respective interfaces and ↵Sebastian Thiel2011-05-3012-863/+936
| | | | | | implementations. It seems theoretically work together now, although it clearly is much more complex than ever before. The repo package was slimmed down to being a module once again, which is only there for compatability actually
* Intermediate commit with quite some progress in order to put all extra ↵Sebastian Thiel2011-05-2911-689/+650
| | | | methods on the default Repo implementation into interfaces or something that can be abstracted. It shows that it would indeed be good to keep the differentiation between Repositories which contain an object database as it is clearly easier to setup any combination of repositories that use git and those that do not, with just the addition of one more level of indirection. Lets see how it will end up
* Made repository paths methods a property to be compatible with the existing ↵Sebastian Thiel2011-05-1014-80/+234
| | | | repo interface. Added submodule interface ... goal is to provide all of the extra repo functionality in custom interfaces
* Started to fix imports - tests still have no chance to work as database ↵Sebastian Thiel2011-05-0642-91/+1461
| | | | changed drastically. Now the actual work begins
* replaced all gitdb strings with gitSebastian Thiel2011-05-0652-144/+148
|
* removed gitdb submoduleSebastian Thiel2011-05-062-4/+0
|
* Added all code from gitdb to gitpython. Next is to make it generally work. ↵Sebastian Thiel2011-05-0564-197/+9917
| | | | Then the tests will need some work
* Added async submodule, formerly required by gitdb (which is still a valid ↵Sebastian Thiel2011-05-053-3/+6
| | | | submodule just yet)
* Made most primal imports work, but stopped here as there are many more ↵Sebastian Thiel2011-05-059-15/+20
| | | | changes when doing the merge
* Moved rev_parse code into gitdb, this probably broke pretty much here, which ↵Sebastian Thiel2011-04-2910-332/+20
| | | | is still to be fixed of course
* configparser is now part of gitdb. Test still remains until it was replaced ↵Sebastian Thiel2011-04-253-1785/+3
| | | | by its gitdb counterpart through inheritance. All items which are part of gitdb but used by gitpython need their own verification within gitpython
* Updated objects to use the ones defined in gitdb as basis. Only the ↵Sebastian Thiel2011-04-0714-1068/+37
| | | | submodule implementation is left in git-python as it requires some advanced features. No tests where run yet
* Removed all parts of the reference implementation which doesn't require the ↵Sebastian Thiel2011-04-0713-1473/+86
| | | | git command. everything else was moved to GitDB. None of the tests is yet expected to run, although git-python should have less trouble getting the tests back up running than GitDB. plenty of code needs to be de-duplicated though in case of the tests, which will be some work
* Moved push and fetch methods partly from remote to the gitcmdobjdb ↵Sebastian Thiel2011-04-055-337/+393
| | | | implementation, including related types. It doesn't yet work, and the omnipresence of the repo imposes a problem right now, as the required ref functionality is not yet part of the gitdb specification. theoretically, the whole ref implementation has to move into gitdb
* Added downloadable versions of the respective documentation, which should ↵Sebastian Thiel2011-04-0234-4/+11057
| | | | aid packagers a bit
* Fixed incorrect handling of rev_parse if it was fed with branches that ↵Sebastian Thiel2011-02-232-2/+11
| | | | looked like hexshas. Thanks to Ning (frostyplanet) who pointed this issue out
* setup.py: fixed incorrect handling of toplevel package path - I must have ↵0.3.1-beta2Sebastian Thiel2011-01-171-5/+4
| | | | been very confused when setting it up that way
* git.repo.base: substituted all usages of the active_branch property with ↵Sebastian Thiel2011-01-171-4/+4
| | | | self.head.* in order to more safely get the current commit. active_branch would only work on non-detached heads
* OSX Signal handling: removed previous hack to fix it somehow using sleep - ↵Sebastian Thiel2011-01-123-12/+1
| | | | the error was caused by PySide which, when instantiating a QApplication, changed the signal handlers to interrupt, instead of retry operating system calls, which caused git-python to fail subsequently. signal.siginterrupt can be used to fix this behaviour after an QApplication was instatiated.
* sleep ui hack fix: removed platform.system() to use sys.platform instead, as ↵Sebastian Thiel2011-01-122-4/+3
| | | | platform.system makes a system call itself
* blob: Added class level members to represent valid file modes that can be ↵Sebastian Thiel2011-01-062-0/+10
| | | | used if you create a new blob
* Merge remote branch 'ahgutsche/master'Sebastian Thiel2011-01-063-1/+13
|\
| * Added the bugfix code as a utility function to be reused wherever this ↵Andreas Gutsche2011-01-063-7/+14
| | | | | | | | specific subprocess bug occurs.
| * Hacked the wait function so that it works with pyside in OS X by using ↵Andreas Gutsche2011-01-061-1/+6
| | | | | | | | "sleep()".
* | Merge branch 'fix'Sebastian Thiel2010-12-031-6/+8
|\ \ | |/ |/|
| * submodule.add: handles addition of repositories with a specific branch ↵Sebastian Thiel2010-12-031-6/+8
| | | | | | | | correctly now - previously it would pass the full path of the branch, which in fact didn't work, unless the default branch is used to which it will default. The bug wasn't noticed because of this
* | setup .gitmodules to use a tracking branch automaticallySebastian Thiel2010-12-012-0/+1
|/
* Fixed bug, which was so obvious that I am wondering how none of the test ran ↵Sebastian Thiel2010-12-012-2/+2
| | | | into that, or at least triggered it
* Submodule tests are functional once again.Sebastian Thiel2010-12-012-3/+11
| | | | remote: Fixed bug that was caused by the unchecked deletion of an uncached attribute which did not necessarily exist
* Fixed import of gitdb, keeping changed gitdb commit so that the submodule ↵Sebastian Thiel2010-12-014-4/+10
| | | | | | tests may work. Switched gitdb submodule to new github url
* Renamed readme file to something github understands much better, adjusted ↵Sebastian Thiel2010-11-304-12/+17
| | | | documentation links to point to github instead of lighthouse/gitorious
* remote: fixed incorrect call signature when instantiating GitCommandErrorSebastian Thiel2010-11-301-1/+1
|
* Commit.create_from_tree used a now renamed keyword parameter - no test-case ↵Sebastian Thiel2010-11-262-2/+3
| | | | ever runs this code, which is still an issue
* -#######->WARNING<-####### Directory structure changed, see commit messageSebastian Thiel2010-11-25112-244/+33
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If you use git-python as a submodule of your own project, which alters the sys.path to import it, you will have to adjust your code to take the changed directory structure into consideration. Previously, you would put the path ./git-python/lib into your syspath. All modules moved one level up into the 'git' subdirectory, which means that the 'git-python' directory now contains the 'git' root package. To allow git to be found, add ./git-python into your path. To finalize your update, run the following commands git submodule update --init --recursive As there will be left-over directories, consider running git-clean
| * update setup.py to allow binary builds, which is required for installation ↵Sebastian Thiel2010-11-251-1/+7
| | | | | | | | and distribution
| * Updated information filesSebastian Thiel2010-11-253-13/+14
| |
| * Fixed setup toolsSebastian Thiel2010-11-253-223/+3
| |
| * Fixed docsSebastian Thiel2010-11-251-1/+2
| |
| * Fixed all testcasesSebastian Thiel2010-11-254-5/+6
| |
| * Moved everything into the git subdirectory - some tests still need to be ↵Sebastian Thiel2010-11-25107-1/+1
|/ | | | adjusted
* submodule.update: the actual update will now mention the revision to which ↵Sebastian Thiel2010-11-251-2/+2
| | | | | | it updates to_latest_revision would now be honored if possible in dry_run mode
* Merge branch 'submodupdate'Sebastian Thiel2010-11-256-295/+478
|\
| * Implemented dry_run mode including tests for RootModule.update and SubmoduleSebastian Thiel2010-11-253-170/+261
| |
| * Integrated progress updates into all submodule.update methods.Sebastian Thiel2010-11-252-17/+86
| | | | | | | | Submodule.update now supports dry_run as well, which is still to be implemented in all cases
| * Submodule: Added dry_run and progress parameter to the update method. It is ↵Sebastian Thiel2010-11-245-135/+158
|/ | | | copatible to the RemoteProgress and should satisfy all progress needs. Dryrun will be useful in conjunction with the progress to verify the changes to be done
* Merge branch 'reflogintegration'Sebastian Thiel2010-11-249-32/+61
|\
| * submodule.update: previous_commit is now set according to the stored reflog ↵Sebastian Thiel2010-11-242-12/+11
| | | | | | | | value, and is not using ORIG_HEAD anymore
| * Adjusted remaining usages of set_reference and set_commit to set a logmessageSebastian Thiel2010-11-244-6/+18
| |
View Lorry Controller Status