Sanitize delete-cache-on-linker-errors make target

[supersven]

dist-newstyle: The directory name depends on the cabal version. cabal clean does the job (using the tool itself to manage it's cache).

~/.cabal: We shouldn't delete the whole folder as it may contain a manually tweaked config file. Deleting ~/.cabal/store should be sufficient.

E.g. my ~/.cabal/config contains this to get docs for all packages in Haskell Language Server:

program-default-options
  ghc-options: -haddock

Checklist

• The PR Title explains the impact of the change.
• The PR description provides context as to why the change should occur and what the code contributes to that effect. This could also be a link to a JIRA ticket or a Github issue, if there is one.
• If this PR changes development workflow or dependencies, they have been A) automated and B) documented under docs/developer/. All efforts have been taken to minimize development setup breakage or slowdown for co-workers.
• changelog.d contains the following bits of information (details):
  • A file with the changelog entry in one or more suitable sub-sections. The sub-sections are marked by directories inside changelog.d.

[sysvinit]

I have a soft nack against this make target, as I have my environment set up with the CABAL_DIR environment variable so that cabal stores its data somewhere other than ~/.cabal. However, I strongly suspect I'm a small minority here, so I won't block the PR, as it's easy enough for me to do a cache bust by hand when necessary.

[fisx]

I have a soft nack against this make target, as I have my environment set up with the CABAL_DIR environment variable so that cabal stores its data somewhere other than ~/.cabal. However, I strongly suspect I'm a small minority here, so I won't block the PR, as it's easy enough for me to do a cache bust by hand when necessary.

gotta protect minorities!! :)

also, isn't there a simple change that would honor $CABAL_DIR and do it right for all of us and for you?

[sysvinit]

also, isn't there a simple change that would honor $CABAL_DIR and do it right for all of us and for you?

For sure -- I was mostly meaning to point out that the make target as written is incompatible with setting CABAL_DIR, but it doesn't need to be perfect here and now.

[supersven]

Thanks for bringing this up, @sysvinit !

Would this work for you: 8101c5b ?

[smatting]

Sorry for the merge commit @supersven . I used GH's "resolve conflicts" feature after merging #2621 - feel free to throw away and rebase

[supersven]

Sorry for the merge commit @supersven . I used GH's "resolve conflicts" feature after merging #2621 - feel free to throw away and rebase

No worries, @smatting . I hope it will just vanish when I squash for merge.

[sysvinit]

Thanks for bringing this up, @sysvinit !

Would this work for you: 8101c5b ?

Yeah, that looks good.