Please, make CI caching more tolerant

[Anton-Latukha]

In #2503 the caching workflow for the most important part (GHC 9.2) - would fail. Because cabal v2-build all would not work there.

Think about, would it be reasonable to cache all OR main+test OR main. That would allow acquiring relevant partial cache, which would allow a faster CI loop to solve problems. But that also would happen when freeze is not possible (since build all has the exact same requirement) & when freeze is not possible - under certain conditions we do not have an effective way to detect when to save new cache.

At the same time we really can not expand cache more pretty much due to pool limit & for the same reason, we reaching a limit on expanding the key matrix more.

[Anton-Latukha]

Somehow I evaluate it as still being worth it. But need to think a bit more.

[Anton-Latukha]

Since this is a report for ongoing work & to not open a number of reports for different nuances - this report can be about tolerance of caching workflow at all - rescoping this report to tolerance as a whole.

[Anton-Latukha]

New information is: https://github.com/haskell/haskell-language-server/runs/4627750464?check_suite_focus=true

[Anton-Latukha]

I propose to have:

cabal $cabalBuild || cabal $cabalBuild || cabal $cabalBuild || true

true is needed because new information shows that if any of continue-on-error: true step fails - caching action does not runs post hook, so despite having continue-on-error: true we need to make that step always return true.