Jens Petersen
Okay I have one more question about my packaging of ghc for Fedora.
For long I always build the testsuite for every release perf build on all archs (for an example see the build.log links on https://koji.fedoraproject.org/koji/buildinfo?buildID=1086491).
Is this a useful meaningful thing to do?
I thought it good to have it as a reference for ghc builds on Fedora and EPEL, but it does add a considerable amount of time to builds (specially for the slower ARM arch's) so it is not without cost.
At the moment I wouold say that our testsuite is unreliable enough in non-validate configurations that this has relatively little value. This is something we are working on fixing and hopefully things will be more reliable in the future. I generally validate the tree prior to cutting a release. This of course won't catch environment- and distribution-specific issues, but I think it's a pretty good proxy for correctness. In other words, unlesss you find yourself looking at the testsuite output yourself, I think it would be fine to disable it.
So I am wondering how useful it is to continue running the testsuite for each "production" build I do. What do others and other distros, etc do for final releases?
As far as I know neither Debian nor NixOS run GHC's testsuite. Cheers, - Ben