[driftregion]

Hello cppcheck maintainer. Thank you for your work.

The #line preprocessor directive (standard since C90) was missing from misra.py, leading to false positives in my analysis. I have added it here.

See:
https://en.cppreference.com/c/preprocessor/line

[danmar]

This is an AI review. Take it with a grain of salt and feel free to reject it by resolving the comment.

The fix looks right to me: #line is a standard directive, and with this change the 20.13 false positive is gone.

Could you add a regression test? One option that doesn't shift any line numbers in addons/test/misra/misra-test.c is to replace the empty line after #else1 // 20.13 (line 1946) with:

#line 1947

I tried it with the same commands CI uses (cppcheck --dump -DDUMMY --suppress=uninitvar --inline-suppr misra/misra-test.c --std=c89 --platform=unix64 + misra.py -verify). Without your fix it reports Not expected: misra/misra-test.c:1946 20.13, and with your fix it passes.

[driftregion]

Hi Daniel, thanks for your work maintaining cppcheck. I had a look at the AI reviews on #8879, #8876, and cppcheck-opensource/simplecpp#704 just now. They seem reasonable. Before I go about addressing them, I'd like to request your human opinion on whether you endorse the AI-proposed changes. This statement makes it clear that the review was AI-generated (great idea btw), but leaves me with some confusion about whether the human in the loop endorses its content:

This is an AI review. Take it with a grain of salt and feel free to reject it by resolving the comment.

Thanks!

[danmar]

I looked quickly. Adding that test seems reasonable to me.

I will mostly trust peoples judgments about AI reviews.

[dmcppcheck]

AI review

I found no new problems. With this change, #line directives are no longer reported as misra-c2012-20.13, and I checked all the ways of spelling the directive that I tried. A malformed directive such as #lines 3 is still reported. The PR still doesn't add the regression test that was asked for in comments.md (#line 1947 in addons/test/misra/misra-test.c). That has already been raised, so I'm not repeating it as a separate comment.

Testing

• I built cppcheck in pr/ with make -j8 cppcheck.
• I wrote scratch/t.c with several spellings of the directive:

  #line 10
  # line 20 "x.c"
  #	line 30
  #line	40
  #  line 50
  #lines 3
  int x;

  I ran cppcheck --dump -q t.c and then python3 misra.py t.c.dump with both the PR's and the base's misra.py.
  • PR: 20.13 is reported only for #lines 3 (x.c:50), which is correct.
  • Base: 20.13 is reported for all 6 directives.
  • simplecpp turns tabs into single spaces in the directive string in the dump, so the space-only regex in misra_20_13 handles those cases correctly.
• Existing MISRA verify test: I ran cppcheck --dump -DDUMMY --suppress=uninitvar --inline-suppr misra/misra-test.c --std=c89 --platform=unix64 followed by misra.py -verify misra/misra-test.c.dump. It reported no Not expected or Not detected lines.
• pytest: python3 -m pytest addons/test/misra_test.py stopped with a collection error before running any tests, so that suite was not run. I did not look into the cause.

---

Automated review of d4aa765 by Claude. The pull request was built and tested locally.