Refactor colnamesInt to provide context-specific error messages (#5039)

[ayushsingh9720]

This PR addresses issue #5039 by refactoring the internal colnamesInt function at both the C and R levels. By introducing a mandatory source argument, the internal error messages now provide specific context (e.g., pinpointing whether a failure occurred during an on= join, setkey, or frank), making debugging significantly more intuitive for the user.

Technical Changes

1. C Internal Logic (src/)
   src/utils.c: Modified the colnamesInt signature to accept a 5th argument, SEXP source. Updated all internal error() calls to include a %s placeholder, prepending the error with the provided context string.

src/data.table.h: Updated the function prototype to remain in sync with the new signature.

src/init.c: Updated the R-to-C registration from -1 (variable) to a strict 5 to ensure formal argument counting at the C interface level.

2. R Wrapper and Call Sites (R/)
   R/wrappers.R: Updated the R-side wrapper for colnamesInt to accept source=NULL and pass it through the .Call interface.

Global Refactor: Performed a comprehensive audit and update of 19 call sites across the following files to ensure the new 5th argument is passed correctly with descriptive hints:

R/data.table.R (Joins, setcolorder, setnames)

R/mergelist.R (Merge operations and internal column copying)

R/setkey.R (Key assignment)

R/frank.R (Ranking)

R/duplicated.R (Unique/Duplicated checks)

R/setops.R (Set operations)

Impact
Users will no longer see generic "inscrutable" errors. For example: Old Error: Error in colnamesInt(...) : argument specifying columns received non-existing column(s): cols[2]='z' New Error: Error in colnamesInt(...) : In x table's columns in on= join, argument specifying columns received non-existing column(s): cols[2]='z'

Checklist
[x] Reproduced the original "inscrutable" error.

[x] Successfully compiled using R CMD INSTALL ..

[x] Verified that all 19 call sites are updated to avoid "unused argument" or "argument mismatch" errors.

[x] Passed manual tests for joins, setkeys, and set operations.

Closes #5039

[jangorecki]

source is not a good name for an argument as it is widely used function name, I would use context instead.

Unit tests are missing.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.96%. Comparing base (3c06138) to head (2649ed9).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7554   +/-   ##
=======================================
  Coverage   98.96%   98.96%           
=======================================
  Files          87       87           
  Lines       16737    16743    +6     
=======================================
+ Hits        16563    16569    +6     
  Misses        174      174           

☔ 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.

[ayushsingh9720]

Thank you for the feedback, @jangorecki ! I will rename the argument to context across the C and R levels. I will also add a dedicated unit test file to verify that the context hints are being correctly prepended to the error messages. I'll push these updates shortly.

[jangorecki]

@ayushsingh9720 did you use AI when preparing this PR?

[ayushsingh9720]

Hi @jangorecki, yes, I've been using an AI as a , kind of a thought/learning partner to help me with the data.table codebase. as i am new to open source, it has been helping me understand the internl logic and debug build failures, but I am manually implementing the changes, running the local R CMD INSTALL builds, and verifying the error messages in my terminal to ensure everything works as intended.

[jangorecki]

Ok. Check how we define tests and try to follow that logic, you can read manual of internal function test.

[ayushsingh9720]

checks are passing now, i have updated the unit test(Test 2357) , to follow the internal test() , as suggested. really appreciate the advice

[ayushsingh9720]

@tdhock @Anirban166 @DorisAmoakohene @joshhwuu I have completed the contributor tests for issues #5039 and #6512. I've updated the Wiki and sent an introductory email to Toby regarding my interest in a mix of low-level C work and bug fixing for GSoC 2026. I've also joined the R-GSoC Slack for future coordination.