Fix two issues with tuple functions

[WardBrian]

1. Accidentally returning a reference
2. Integer arguments failed under certain circumstances, are now templated

These were found because the Stan tests were raising a -Wreturn-stack-address warning. We should try to catch those types of things in stanc going forward.

Submission Checklist

• Run unit tests
• Documentation
  • If a user-facing facing change was made, the documentation PR is here:
  • OR, no user-facing changes were made

Release notes

Fixed a few issues with the C++ generated for functions which accept tuples as arguments.

Copyright and Licensing

By submitting this pull request, the copyright holder is agreeing to
license the submitted work under the BSD 3-clause license (https://opensource.org/licenses/BSD-3-Clause)

[codecov]

Codecov Report

Merging #1356 (45e4b82) into master (743d0dd) will increase coverage by 0.04%.
Report is 2 commits behind head on master.
The diff coverage is 97.89%.

❗ Current head 45e4b82 differs from pull request most recent head ca286a2. Consider uploading reports for the commit ca286a2 to get more accurate results

@@            Coverage Diff             @@
##           master    #1356      +/-   ##
==========================================
+ Coverage   89.39%   89.44%   +0.04%     
==========================================
  Files          65       65              
  Lines       10607    10644      +37     
==========================================
+ Hits         9482     9520      +38     
+ Misses       1125     1124       -1     

Files Changed	Coverage Δ
src/stan_math_backend/Cpp.ml	86.18% <90.90%> (+0.21%)	⬆️
src/stan_math_backend/Lower_functions.ml	98.80% <98.21%> (+0.13%)	⬆️
src/analysis_and_optimization/Memory_patterns.ml	90.57% <100.00%> (ø)
src/frontend/Ast_to_Mir.ml	94.25% <100.00%> (+0.07%)	⬆️
src/stan_math_backend/Lower_expr.ml	92.85% <100.00%> (ø)
src/stan_math_backend/Lower_program.ml	99.18% <100.00%> (ø)

... and 1 file with indirect coverage changes

[WardBrian]

IMO this will justify a 2.33.1, but the nice thing is the only thing that would be different between the two versions is the Stanc executables.

@rok-cesnovar @serban-nicusor-toptal

[rok-cesnovar]

Great catch. Do we wait for a few days or just do the patch release?

[WardBrian]

@SteveBronder and I still want to take a slightly deeper look at this, so it might be nice to also see if anything else comes in in the next couple days

[WardBrian]

Ok I think this is finally ready for review @SteveBronder

It's a bit tricky but honestly I think I like the code better now than before.

[SteveBronder]

Two minor comments but imo looks good!

[SteveBronder]

Lgtm!

[WardBrian]

@rok-cesnovar @serban-nicusor-toptal I think we should try to do 2.33.1 in the middle of next week, giving a few more days for new issues to come in.

At the moment it would just be the 2.33.0 branches from all the repos except for stanc3