Can you please write a quick (but legible) self-contained explanation of why we want this?

@ulysses4ever philosophically I'm trying to convince myself to write a few dozen link tests 😅

the real problem is that when adding -pgm* flags we need to write at least 5(10) tests instead of one because the output contains the path to the C compiler

example from the pull request I'm referring to:

1. no -pgmc
single-2.out
single-2.test.hs
2. gcc is /usr/bin/gcc
single.out
single.test.hs
3. clang is /tools/ghc-<GHCVER>/lib/../mingw/bin/clang
win-single-4.out
win-single-4.test.hs
4. clang is /ghcup/ghc/<GHCVER>/lib/../mingw/bin/clang
win-single-8-6.out
win-single-8-6.test.hs
5. clang is /ghcup/ghc/<GHCVER>/lib/../mingw//bin/clang
win-single.out
win-single.test.hs

That sounds more like a bug / deficiency in the normalizer to me.

That sounds more like a bug / deficiency in the normalizer to me.

What is a normalizer?

It's a postprocessor for test output that produces those <UPPERCASE> things in place of mutable paths.
https://github.com/haskell/cabal/blob/master/cabal-testsuite/src/Test/Cabal/OutputNormalizer.hs

@zlonast: any progress? could we help?

@zlonast: a kind ping?

@Mikolaj Hi, sorry for not answering, I don't have time to continue here yet.

No worries, take care!