Fix subscript hoisting; correct coercion/matrix lowering

[t-kalinowski]

This PR fixes several lowering/translation issues hit by simplified lowering patterns:

• Fix [ ... ] so index expressions are compiled with the caller hoist context, avoiding invalid inline block ... end block expressions inside Fortran array designators (e.g. rev(seq_len(n)) in a multi-d subscript).
• Make as.double(<array>) and as.integer(<array>) drop dimensions (return a vector) by reshaping rank>1 results, matching base R semantics and fixing slice assignments like out[1:6] <- as.double(a).
• Fix matrix(<non-scalar expr>, nrow=, ncol=) lowering by emitting an explicit reshape(...) so the generated Fortran ranks match.
• Improve the bad rank error for rank-0 scalar declarations to point users at integer(1) / double(1).

Regression tests added:

• tests/testthat/test-subscript-hoisted-temporaries.R
• tests/testthat/test-as-double.R
• tests/testthat/test-matrix.R
• tests/testthat/test-c-wrapper-public-api-coverage.R

Tested: devtools::test()

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 02082b16b1

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Avoid double-evaluating matrix() data expression

The new matrix() lowering uses reshape(..., pad = {source}), which inserts the data expression twice. In Fortran, all actual arguments are evaluated before the intrinsic call, so a non-scalar data expression (e.g., runif(4) or any user-defined function with side effects) will be evaluated twice. That changes semantics versus R, which evaluates data once, and can produce different values or duplicated side effects. Consider hoisting data into a temporary and using that name in both reshape and pad, or only supplying pad when it’s needed and the source is already a plain symbol.

Useful? React with 👍 / 👎.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.78%. Comparing base (63d755e) to head (9d1ced5).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #94   +/-   ##
=======================================
  Coverage   92.78%   92.78%           
=======================================
  Files          27       27           
  Lines        5749     5754    +5     
=======================================
+ Hits         5334     5339    +5     
  Misses        415      415           

☔ 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.