github: Test with musl

[cole-miller]

MicroK8s and Juju consume dqlite as a static library build against musl libc, so let's test this in our CI to avoid unexpected breaks like #679.

Signed-off-by: Cole Miller cole.miller@canonical.com

[codecov]

Codecov Report

Attention: Patch coverage is 39.39394% with 20 lines in your changes missing coverage. Please review.

Project coverage is 78.66%. Comparing base (2d0d71b) to head (4a512aa).
Report is 1 commits behind head on master.

Files	Patch %	Lines
test/raft/integration/test_uv_tcp_listen.c	0.00%	6 Missing and 6 partials ⚠️
test/raft/integration/test_uv_tcp_connect.c	16.66%	3 Missing and 2 partials ⚠️
src/raft/uv.c	60.00%	2 Missing ⚠️
test/lib/heap.c	83.33%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #682      +/-   ##
==========================================
+ Coverage   77.87%   78.66%   +0.79%     
==========================================
  Files         197      197              
  Lines       27953    27964      +11     
  Branches     5534     5541       +7     
==========================================
+ Hits        21768    21998     +230     
+ Misses       4346     4059     -287     
- Partials     1839     1907      +68     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[cole-miller]

Running into this issue it seems: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=764335.

[louiseschmidtgen]

Could you please include a test run against the musl tag used in our k8s distributions: https://github.com/canonical/k8s-dqlite/blob/master/hack/env.sh#L13? We might not always be on the same musl version as the juju team.

[SimonRichardson]

Juju will be happy to move to the later one, now we know the root cause. We intended to move that originally but ended up locating the segfault.

[cole-miller]

Sounds good.

[cole-miller]

I like how using a Dockerfile for this cleans things up. Looking into the segfaults.