Am 22.07.24 um 22:20 schrieb Sebastian Kuzminsky:
A recent commit in 2.9 triggered a bug in po4a and repeatedly crashed the
buildbot.
You can trigger the problem using this process:
1. Start a debian:buster container
2. Check out commit fadb1c28 (current head of 2.9)
3. Install build dependencies like usual
4. Install po4a 0.67 fromhttp://linuxcnc.org/dists/buster/base
5. Configure with --enable-build-documentation
6. Build
All versions of po4a that I've tested are affected by this bug, including 0.67
(which we backported to buster and bullseye), 0.69 (in bookworm), and 0.73 (in
trixie).
I bisected the recent history of 2.9 and determined the commit that introduced
the problem is d1414e36. I've reverted that commit in 2.9, and I'm merging to
master now.
Before reintroducing that commit, please test on the affected platforms as
outlined above.
Currently we have two problems with the automated builds:
1. The GitHub CI build fails for all PRs for 2.9 with:
Runtest: 260 tests run, 258 successful, 2 failed + 0 expected, 1 skipped
Failed:
./tests/mdi-queue/oword-queue-buster
./tests/mdi-queue/simple-queue-buster
2. The documents for master (2.9 also?, I don't know) are still not
updated on the website (https://github.com/LinuxCNC/linuxcnc/issues/3031)
Could this be related to the earlier problem? Or does anyone has an idea
how to resolve this?
_______________________________________________
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers
