various fixes for RubyParser

[zenspider]

See individual commits for details.

One thing this doesn't have is the ACTUAL fixes I want for parser changes, because that would cause failures in the tests against RubyParser.

Is there a plan for how to deal with drift if I'm hoping to EOL RubyParser? Add them to known_failures?

Also, it looks like an import hasn't been for some time, and maybe the last one was done against git and not against a release... I'm hesitant to add/change those files as I'm not sure I'm doing the process correctly:

$ git clean -qf -- test/prism/fixtures ; rake seattlerb:import

[zenspider]

Added an optional commit to update the tests to current. I can pull it to a different PR to keep this discussion scoped.

[Earlopain]

Generally adding failures to excludes in ruby_parser_test.rb somewhere is totally fine. It would be nice not to if possible but there are so many edge-cases, and often the result can also be different without actually affecting any consumer. I guess it depends on what drift you mean, that's not clear to me yet.

If I run the import, I also get pct_q_backslash_nl.txt, I think you might have forgotten to commit that one?

[zenspider]

@Earlopain prolly gets swallowed as I'm on APFS and it is case insensitive by default... but since that commit broke EVERYTHING I should prolly punt it for now

[kddnewton]

I'm seeing a couple of changes in here:

1. Explicitly requiring timeout — ✅ makes sense
2. Allowing passing scopes into the translator — ✅ makes sense
3. Fixing the rake task regexp — ✅ makes sense
4. Allowing subclassing the compiler — ?? this one I'm curious about. Not a big deal I suppose, but generally I try to avoid inheritance as much as possible so I'm curious about the use case

[zenspider]

4. is simple... it allows me to make changes to prism locally until they get upstreamed... with the compiler marked private and not duplicating the same thing when calling copy_compiler that makes it harder.

I can get around it with a prepended module, but that starts to feel clever to me.

Oh, and for some context... In flog/flay/etc I've switched over to NotRubyParser which is a subclass of Prism's compat layer. I pop things in there to override until they get upstreamed.

[kddnewton]

Okay sounds good

[kddnewton]

Actually only thing I'd like to change is to make scopes the same type as everywhere else in prism. That is, not automatically wrapping it in a [], since that only allows one scope. But I'll change that after merge.