Closed fingolfin closed 3 years ago
OK, now that checks run, it looks OK (but let's wait for them to finish).
The reason for my suspicion was that there still is a CI job "CI / ${{ matrix.gap-branch }} ${{ matrix.ABI }} (pull_request) " listed on PR #17, which I still don't understand how it could happen... Weird.
Merging #20 (11be459) into master (a02a430) will not change coverage. The diff coverage is
n/a
.
@@ Coverage Diff @@
## master #20 +/- ##
=======================================
Coverage 73.89% 73.89%
=======================================
Files 343 343
Lines 590 590
=======================================
Hits 436 436
Misses 154 154
This PR has an ulterior motive: I suspect there still might something wrong with CI tests for PRs, but perhaps only for PRs submitted from a separate fork (my previous forks were from branches which I had put intogap-packages/unitlib
, this one is fromfingolfin/unitlib
.So please do not merge for now until this is resolved.
OK, I think the CI tests here are good for now. So perhaps it is the reverse from what I thought, and the
CI / ${{ matrix.gap-branch }} ${{ matrix.ABI }} (pull_request)
only appears for "internal" PRs which originate from a branch on the main repo... Theory: perhaps it happens because theif:...
expression skipping the tests is outside the matrix and thus it skips the test before ever expanding the matrix. That might explain it (it doesn't explain the weirdness we saw before, but, oh well...)