Martin von Zweigbergk <martinvonz@google.com> [Sat, 27 Apr 2019 23:14:49 -0700] rev 42245
test: demonstrate failure to follow rename with shadowed linkrev
This shows a difference in handling of copies between `hg st`
(pathcopies()) and `hg co -m`. The issue here is that mergecopies()
uses the unadjusted linkrev() for determining when to stop walking
ancestors.
Differential Revision: https://phab.mercurial-scm.org/D6322
Martin von Zweigbergk <martinvonz@google.com> [Sat, 27 Apr 2019 22:57:15 -0700] rev 42244
tests: slightly modify a linkrev test to prepare for expanding it
The test case checks that the copy tracing code doesn't get confused
by linkrevs when walking a file's ancestors. This patch chnages the
test slightly so a second commit is grafted, thus producing a second
"bad" linkrev. I'll use this in the next patch to demonstrate a bug.
Differential Revision: https://phab.mercurial-scm.org/D6321
Martin von Zweigbergk <martinvonz@google.com> [Sat, 27 Apr 2019 22:55:54 -0700] rev 42243
copies: process files in deterministic order for stable tests
I also fixed a typo while at it.
Differential Revision: https://phab.mercurial-scm.org/D6320
Ludovic Chabant <ludovic@chabant.com> [Fri, 19 Apr 2019 14:26:32 +0000] rev 42242
py3: properly reject non-encoded strings given to hgweb