Internal Happy errors are not helpful

[Kleidukos]

The Internal Happy error exception gives a CallStack that is most unhelpful, as it only gives the location of the error function used in the generated code. This could be slightly improved by having HasCallStack pragmas earlier in the program.

Moreover, more information about has failed would be appreciated. Currently I get

λ❯ runAlex "let example = 42" parseMiniML 
*** Exception: Internal Happy error

CallStack (from HasCallStack):
  error, called at /home/hecate/Code/Compiler Experiments/qamar/dist-newstyle/build/x86_64-linux/ghc-9.2.4/qamar-0.1.0.0/build/Parser.hs:1441:17 in qamar-0.1.0.0-inplace:Parser

and it's quite unbecoming.

[sgraf812]

Frankly, if you see this error then there's a bug in happy that you should report. You can't do anything about it and you are not using happy in a wrong way (at least I don't think that you are).

Did you see this error in conjunction with -XStrictData? Then this issue the same as #273, for which I provide a fix in #289.

[Kleidukos]

Yes it must have been in conjunction with -XStrictData. See https://gist.github.com/heitor-lassarote/9cc4203990352f5df5f6b84da9291df9?permalink_comment_id=4306546#gistcomment-4306546