R6: Fix null pointer crash in tokenDelete function

[gaborbernat]

The tokenDelete() function in main/tokeninfo.c was being called with NULL pointers from the R6 parser error handling code, causing segmentation faults.

The R6 parser in parsers/r-r6class.c had a comment stating tokenDelete accepts NULL"but the function did not actually handle NULL pointers safely.
This occurred when parsing malformed R6 syntax such as:

• R6::SomethingElse() (wrong function after R6 namespace)
• R6:: (incomplete namespace reference)
• R6::R6Clas() (typo in R6Class)

The fix ensures ctags handles malformed R6 syntax gracefully by falling back to parsing assignments as regular global variables instead of crashing. Fixes segmentation fault when processing certain R6Class syntax patterns.

Signed-off-by: Bernát Gábor [email protected]

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.88%. Comparing base (ead1d93) to head (2bfc47b).
⚠️ Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4324      +/-   ##
==========================================
- Coverage   86.01%   85.88%   -0.13%     
==========================================
  Files         250      250              
  Lines       64159    62329    -1830     
==========================================
- Hits        55187    53533    -1654     
+ Misses       8972     8796     -176     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[masatake]

@gaborbernat Thank you. Valgrind says something. I will look into this.

[gaborbernat]

🤔 considering I did not touch the allocation surprised how could be any new leaks here 🤔

[masatake]

Yes! Internal token0 and token1 hide the outer ones.

[masatake]

CURRENT.tags may be used in debugging. Please don't include the commits.
Could you change the commit header "Fix null pointer crash in tokenDelete function" to "tokeninfo: fix null pointer crash in tokenDelete function"?

Could you change the commit header "Fix valgrind error" to "R6: fix valgrind error"?

[gaborbernat]

Done 👍 let me know if you'd like anything else.

[gaborbernat]

Once this lands, would it be possible to cut a new release? 🤔 This is actively crashing our OpenGrok indexing 😆 and otherwise we'd have to downgrade back to 5.9, but would rather upgrade to 6.3 if possible.

[masatake]

Once this lands, would it be possible to cut a new release? 🤔 This is actively crashing our OpenGrok indexing 😆 and otherwise we'd have to downgrade back to 5.9, but would rather upgrade to 6.3 if possible.

I will release 6.2.1, which may include the commits in this pull request, 556fc03, 0be19d0, and 33d1d1b.

I will not release 6.3 until #4266, #4267, and #4268 are merged.

[gaborbernat]

6.2.1 is perfect for me, thanks!

[masatake]

Thank you. I will work on releasing v6.2.1. It will take a week.

[gaborbernat]

Thanks!

[masatake]

@gaborbernat See https://github.com/universal-ctags/ctags/releases/tag/v6.2.1 .