- May 17, 2018
-
-
Paul Morelle authored
As stated in https://phab.mercurial-scm.org/D3037, repo.__getitem__ should be used instead of repo.changectx, which was removed.
-
Paul Morelle authored
-
- Mar 21, 2018
-
-
Paul Morelle authored
https://phab.mercurial-scm.org/D2003 Mercurial version 4.6 will introduce a refactoring of the utils.py file into smaller files, starting with dateutil.py. This change prepares this move so that the warnings won't break the tests.
-
- May 01, 2018
-
-
Kevin Bullock authored
-
- Feb 24, 2018
-
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
- Nov 29, 2017
-
-
Sean Farley authored
This is a large patch since we're copying over Mercurial's helpers-testrepo.sh and a slightly modified version of pyflakes.t
-
- Nov 28, 2017
-
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
- Nov 27, 2017
-
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
Sean Farley authored
-
- Feb 07, 2018
-
-
Tony Tung authored
-
- Feb 21, 2018
-
-
Sean Farley authored
This will be used in the following patch.
-
- Feb 07, 2018
-
-
Tony Tung authored
-
- Feb 06, 2018
-
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
- Feb 03, 2018
-
-
Kevin Bullock authored
The exchange.pull wrapper function was checking whether any paths were configured yet as a proxy to determine whether we're doing the initial clone or a subsequent pull. Core hg @ bdae51a83dfb (released in 4.5) now sets the 'default' path on the ui object before exchange.pull is called, but we can check a bit more directly by checking if we have any remote refs yet. This is the same as a check we do in GitHandler.fetch_pack().
-
- Feb 06, 2018
-
-
Tony Tung authored
-
- Feb 02, 2018
-
-
Kevin Bullock authored
Since we switched to using our own vendored copy of run-tests.py in ed168ce47715, the 'make test-x.y' rules to test against specific Mercurial versions weren't. Add --with-hg to the invocation to restore the correct behavior.
-
- Dec 24, 2017
-
-
Kevin Bullock authored
-
- Dec 22, 2017
-
-
Kevin Bullock authored
-
- Nov 30, 2017
-
-
Sean Farley authored
-
- Nov 28, 2017
-
-
Sean Farley authored
-
- Nov 30, 2017
-
-
Sean Farley authored
It seems fffe8883960b incorrectly did a match on the second parent (which is null) and that core Mercurial mistakenly matched the whole line with (re). This was recently fixed in 6d5718e39657 in core, so let's update the test with the correct match.
-
- Nov 27, 2017
-
-
Sean Farley authored
This allows us to code against the current API of dulwich so that one day we can drop the code in compat.py.
-
Sean Farley authored
New dulwich refactored the object it returns from fetch_pack (it use to be a dict but is now an object of two dicts, one for ref and the other for symrefs, and a string containing the agent.
-
Sean Farley authored
Beginning with dulwich 0.18, it now supports reporting the symrefs so we no longer need to monkey patch (will require future patches to use the new code, though).
-
Sean Farley authored
This makes a following refactor easier to review.
-
- Dec 02, 2017
-
-
Sean Farley authored
-
- Sep 05, 2016
-
-
Dan Villiom Podlaski Christiansen authored
Sharing repository data, but not Git metadata can lead to confusing scenarios, particularly in the face of history editing and/or forced updates. This adds a test for the share integration added in 1003994dd497.
-
- Dec 24, 2017
-
-
Kevin Bullock authored
-