enable building proper with OTP-29

[mikpe]

• eliminate old-style catches
• resolve is_record/1 ambiguity
• fix variable exported from subexpression

[codecov]

Codecov Report

❌ Patch coverage is 66.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 85.41%. Comparing base (f511d29) to head (757db6f).
⚠️ Report is 13 commits behind head on master.

Files with missing lines	Patch %	Lines
src/proper_erlang_abstract_code.erl	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #324      +/-   ##
==========================================
+ Coverage   85.27%   85.41%   +0.14%     
==========================================
  Files          14       14              
  Lines        4590     4594       +4     
==========================================
+ Hits         3914     3924      +10     
+ Misses        676      670       -6     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[kostis]

I want to understand the rationale for this change and what exactly is wrong with this code in general and in the upcoming OTP 29. To me, this code besides being "valid Erlang", is perfectly understandable in what it does and how it achieves it. Moreover, it's the last statement in the body of this function, so there is not even a chance that variables will be used after it. So why is it better to have to write a new function of six lines for it?

Also, I've noticed that the release notes for the pre-releases of 29 advertise a "new language feature" that encourages such code. I quote from that page:

By enabling the compr_assign feature, it is now possible to bind variables in a comprehensions. For example: [H || E <- List, H = erlang:phash2(E), H rem 10 =:= 0]

How does the PropEr code differ from this example and what am I missing?

[mikpe]

It's due to an unfortunate combination of two language changes:

1. Exporting bindings from subexpressions is deprecated. In this case the lhs of (Inst = ...) =/= ok triggers that. Plain matches don't count as subexpressions.
2. Previously you could write [Y || X <- L, Y = f(X), predicate(Y)] and it would compile but give a bad filter exception at runtime. The other change, enabled by compr_assign, makes those bindings work.

I've experimented a bit and found a shorter workaround: instead of having a match (Y = f(X)) to bind the intermediate value, use a single-element generator (Y <- [f(X)]). I'm pushing an update with that change.

[kostis]

Thanks for the PR. I've left a comment concerning one of the changes.