[NO-TICKET] Minor: Fix checking for dladdr in profiling

[ivoanjo]

What does this PR do?

This PR fixes the checks for the dladdr/dladdr1 function added in #4745 to actually check the values of the defines.

Specifically, the have_... helpers in extconf.rb behave differently from the other defines we set up.

That is, have_func("dladdr") will always define HAVE_DLADDR, even though it might define it to 0 (not available). (Our other checks only define things when they should be used, e.g.
$defs << "-DNO_POSTPONED_TRIGGER" if RUBY_VERSION < "3.3").

This means that it's not correct to use #ifdef HAVE_DLADDR to check if HAVE_DLADDR exists and is == 1, we actually need to test it separately.

In this very specific case, in practice, I suspect we won't trigger issues because our extconf.rb is

if have_header("dlfcn.h")
  (have_struct_member("struct link_map", "l_name", "link.h") && have_func("dladdr1")) ||
    have_func("dladdr")
end

...and thus if the header doesn't exist, we won't run the other checks, so I think only in a situation where the header exists and doesn't have the function would there be a problem (and I'm not sure that can happen on Linux).

Anyway, I believe it's still worth fixing as this is a footgun waiting to happen, and compilation errors are always mega annoying.

Motivation:

Fix potential compilation error when installing profiling.

Change log entry

Yes. Fix potential profiler compilation issue.

Additional Notes:

I've actually ran into this in the past in
ruby/ruby@7472fff

How to test the change?

If CI is green, this should be good!

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.52%. Comparing base (d903ac5) to head (05a6cea).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #4783      +/-   ##
==========================================
- Coverage   97.54%   97.52%   -0.02%     
==========================================
  Files        1483     1483              
  Lines       88562    88561       -1     
  Branches     4591     4591              
==========================================
- Hits        86384    86370      -14     
- Misses       2178     2191      +13     

☔ 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.

[pr-commenter]

Benchmarks

Benchmark execution time: 2025-07-07 10:59:28

Comparing candidate commit 05a6cea in PR branch ivoanjo/minor-fix-checking-for-dladdr with baseline commit d903ac5 in branch master.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 43 metrics, 6 unstable metrics.

[p-datadog]

Would it be less error-prone to undef these defines if they are set to zero?

[ivoanjo]

Would it be less error-prone to undef these defines if they are set to zero?

I think undef would be a bit confusing TBH... If we went down that route, I think it would be preferable to perhaps ignore Ruby's defines and use our own, e.g. do something like

if have_func("dladdr1")
  $defs << "-DUSE_DLADDR1" 
elsif have_func("dladdr")
  $defs << "-DUSE_DLADDR" 
end

But it's not amazingly clean either. Any specific preference?

[ivoanjo]

I'm going to go ahead and merge this one -- happy to change the strategy on a follow-up once we decide what option to use.