✅ Deploy Preview for opentelemetry-cpp-api-docs canceled.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 89.69%. Comparing base (7801cd9) to head (c510a4a).
Report is 1 commits behind head on main.

@@            Coverage Diff             @@
##             main    #3355      +/-   ##
==========================================
+ Coverage   89.67%   89.69%   +0.02%     
==========================================
  Files         211      211              
  Lines        6832     6832              
==========================================
+ Hits         6126     6127       +1     
+ Misses        706      705       -1     

The PR looks good overall!
Just one small suggestion — instead of deleting config.h, should we move these changes there?
config.h is a public header, and some users might still have a dependency on it (directly or indirectly).

looks good overall!
Just one small suggestion — instead of deleting config.h, should we move these changes there?
config.h is a public header, and some users might still have a dependency on it (directly or indirectly).

Is the original file name also kind of misleading, because the config.h is usually used to store the status of configuration process? If so, probably it is a good time to remove it then it will not conflict the name if we add it later.

Is the original file name also kind of misleading, because the config.h is usually used to store the status of configuration process? If so, probably it is a good time to remove it then it will not conflict the name if we add it later.

I see your point about the naming — agree that config.h could have been better named originally.

However, since config.h is already part of the public API in stable release, removing it would technically be a breaking change for downstream users. Even if it was meant to be used internally, we never documented that. I would recommend keeping config.h for now, even if it's mostly empty, and cleaning it up properly in separate PR as part of ABIv2.

Is the original file name also kind of misleading, because the config.h is usually used to store the status of configuration process? If so, probably it is a good time to remove it then it will not conflict the name if we add it later.

I see your point about the naming — agree that config.h could have been better named originally.

However, since config.h is already part of the public API in stable release, removing it would technically be a breaking change for downstream users. Even if it was meant to be used internally, we never documented that. I would recommend keeping config.h for now, even if it's mostly empty, and cleaning it up properly in separate PR as part of ABIv2.

Good point, I restore config.h, but only add error message for ABI v2 and warning messages for v1. How do you think about it?