CI: don't rely on primary GNU file server

[fingolfin]

Use a mirror is more resilient and often faster, as one close to you is selected.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.68%. Comparing base (c218ec1) to head (756c723).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6388      +/-   ##
==========================================
- Coverage   78.69%   78.68%   -0.02%     
==========================================
  Files         683      683              
  Lines      292902   292902              
  Branches     8658     8658              
==========================================
- Hits       230511   230462      -49     
- Misses      60570    60621      +51     
+ Partials     1821     1819       -2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[reiniscirpons]

Looks good, on my system this is faster (about 6 mb/s for the mirror vs 2 mb/s without). However it seems that one of the CI jobs is failing now:

Run wget https://ftpmirror.gnu.org/gnu/autoconf/autoconf-2.72.tar.xz
--2026-05-11 10:20:19--  https://ftpmirror.gnu.org/gnu/autoconf/autoconf-2.72.tar.xz
Resolving ftpmirror.gnu.org (ftpmirror.gnu.org)... 209.51.188.200, 2001:470:142:5::200
Connecting to ftpmirror.gnu.org (ftpmirror.gnu.org)|209.51.188.200|:443... connected.
HTTP request sent, awaiting response... 502 Bad Gateway
2026-05-11 10:20:20 ERROR 502: Bad Gateway.

Could it be some sort of bot protection? Might also just be a random failure I suppose.

[fingolfin]

Dunno -- we had several such errors today, with the original URL, which motivated me to make this change.

It has certainly happened in the past that file servers blocked GitHub, at least temporary, to deal with extra load imposed on them. And I can related...

Maybe we can somehow cache this file, e.g via actions/cache