This adds a test case for ci tree iteration when there is a name conflict. This points out a behavior quirk in the current version that I'd like to fix - namely, all tree entries get mapped to one version of the case pattern in the ci code - i.e. even if you have A/1.txt and a/2.txt, both will be reported as a/1.txt and a/2.txt because we only copy the name of a file at a given frame once. It would be nice to fix this, but I'm worried about how complex that is if you get a/B/c/1.txt and A/b/C/2.txt. It may require a walk up the frames whenever you advance to the next item in a blended equivalence class.
Name |
Last commit
|
Last update |
---|---|---|
deps | Loading commit data... | |
docs | Loading commit data... | |
examples | Loading commit data... | |
include | Loading commit data... | |
packaging/rpm | Loading commit data... | |
src | Loading commit data... | |
tests-clar | Loading commit data... | |
.HEADER | Loading commit data... | |
.gitattributes | Loading commit data... | |
.gitignore | Loading commit data... | |
.mailmap | Loading commit data... | |
.travis.yml | Loading commit data... | |
AUTHORS | Loading commit data... | |
CMakeLists.txt | Loading commit data... | |
CONTRIBUTING.md | Loading commit data... | |
CONVENTIONS.md | Loading commit data... | |
COPYING | Loading commit data... | |
Makefile.embed | Loading commit data... | |
README.md | Loading commit data... | |
api.docurium | Loading commit data... | |
git.git-authors | Loading commit data... | |
libgit2.pc.in | Loading commit data... | |
libgit2_clar.supp | Loading commit data... |