Commit message Age Author Refs
r53358:a0af298d
ci: automatically starts platform and py-version test for scheduled build The point of these nightly build is to catch error on case we do not systematically test. So we better test them.
0
r53357:23d51b52
ci: let branch job run for scheduled build too Otherwise the scheduled build are useless.
0
r53356:db85e2aa
test-sparse-revlog: fix macos sed invocation that old BSD sed does not like "\+"
0
r53355:a814534a
run-tests: add a "--tail-report" argument to analyse run parallelism This flag display the final running test for each thread and the timing of their conclusion. It help spot when some very slow test is hurting parallelism.
0
r53354:8dede0df
format: add pattern filtering to debugformat The number of format to consider increase, this will make the command simpler to use in tests and debug. See test changes for some direct benefits.
0
r53353:57a4adbe
rhg: test rhg status --change --copies without fallback This is a follow up to !1126 and !1137.
Mitchell Kember
0
r53352:4a28631f
rhg: add tests for rhg status without fallback This is a follow up to !1126 to add another test.
Mitchell Kember
0
r53351:7e5ed1e8
generate-churning-bundle: change dir before deleting temp dir
paugier
0
r53350:1a72e270
test-dirstate: add #if unix-permissions for a file mode check
paugier
0
r53349:9d79ffee
typing: use the `Status` protocol wherever `scmutil.status` was being used This likely isn't everything, but these were all of the places the latter was referenced in the generated *.pyi files, plus a few cases that were inferred as `Any`, but found in a module that was being changed anyway. We should figure out some sort of consistency as far as naming these Protocol classes (stdlib ones tend to be CamelCase and imported directly). The current convention of `from xxx.interfaces import foo as ifoo` is a little clever, but a little annoying to type out. Also, this package is likely to grow beyond just Protocol classes, where treating the types as interfaces is wrong (e.g. a theoretical `NodeT` type to represent the binary form of a node, instead of treating that and the incompatible hex form as both bytes). But that's a project for another day.
Matt Harbison
0
< 1 .. 7 8 9 10 11 .. 5341 >