gnu-octave
CI: still seeing "fatal: unsafe repository"
fatal: unsafe repository ('/home/jovyan/octsympy' is owned by someone else)
To add an exception for this directory, call:
git config --global --add safe.directory /home/jovyan/octsympy
Maybe we need to git config as well (or instead) of sudo git config... At least I don't think that "global" thing means "all users". More like "all repos"
The current situation is kind of awkward and complicated. Since https://github.com/actions/checkout/issues/760 is now closed (it was still opened when we introduced the workarounds if I recalled correctly), one might think the issue should have been resolved.
Unfortunately, testing seems to suggest otherwise. The situation can be summarized using the following tables:
| Previous results | bist | doctests | *pythonic* |
|---|---|---|---|
| w/o workarounds | failed to build with "fatal: unsafe repository" | failed to build with "fatal: unsafe repository" | (no such tests) |
| w/ workarounds | build successfully w/o "fatal: unsafe repository" | build successfully but show "fatal: unsafe repository" | (no such tests) |
Current results (before merging ci_pkg_uninst) |
bist, doctests | bist-pythonic, doctests-pythonic |
|---|---|---|
| w/o workarounds | build successfully but show "fatal: unsafe repository" | failed to build with "fatal: unsafe repository" |
| w/ workarounds | build successfully but show "fatal: unsafe repository" | build successfully w/o "fatal: unsafe repository" |
Current results (after merging ci_pkg_uninst) |
bist | doctests | pythonic-dist-doctest |
|---|---|---|---|
| w/o workarounds | failed to build with "fatal: unsafe repository" | build successfully but show "fatal: unsafe repository" | failed to build with "fatal: unsafe repository" |
| w/ workarounds | build successfully w/o "fatal: unsafe repository" | build successfully w/o "fatal: unsafe repository" | build successfully w/o "fatal: unsafe repository" |
Was that before or after c6edc2ad311813485a01ca33c5fd5ea5bafc7302? I thought that cleaned up some of the messages.
Yeah, there was before c6edc2ad311813485a01ca33c5fd5ea5bafc7302, I was looking at the main branch only. I'll update the tables.