Pierre-Yves David <pierre-yves.david@ens-lyon.org> [Thu, 13 Oct 2016 13:47:47 +0200] rev 30164
eol: make sure we always release the wlock when writing cache
If any exception were to happen after we acquired the wlock, we could leave it
unreleased. We move the wlock release in a 'finally:' close as it should be.
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 13 Oct 2016 21:42:11 +0200] rev 30163
pathencode: use assert() for PyBytes_Check()
This should have been added in a8c948ee3668. I sent the patch to the
list prematurely.
Mads Kiilerich <madski@unity3d.com> [Wed, 12 Oct 2016 12:22:18 +0200] rev 30162
merge: clarify warning for (not) merging flags without ancestor
Give hints why it can't merge and what it will do instead.
Mads Kiilerich <madski@unity3d.com> [Wed, 12 Oct 2016 12:22:18 +0200] rev 30161
merge: only show "cannot merge flags for %s" warning if flags are different
Mads Kiilerich <madski@unity3d.com> [Wed, 12 Oct 2016 12:22:18 +0200] rev 30160
tests: add test coverage of merging x flag without ancestor
It is more noisy than necessary - we will fix that later.
Gregory Szorc <gregory.szorc@gmail.com> [Sat, 08 Oct 2016 17:07:43 +0200] rev 30159
dirs: document Py_SIZE weirdness
Assigning to what looks like a function is clown shoes. Document that
it is a macro referring to a struct member.