SLTozer wrote:

> A bigger question though is whether this can be tested.

That's a good question - in theory yes, but it sounds brittle. Assuming we're 
testing the output of backtrace/symbolize, we'd have a test that could fail if 
any of a variety of function names changed. Another argument against testing is 
that this code exists to generate formatted, human readable output - most of 
the complicated parts are in `llvm-symbolizer` and `libbacktrace`, while these 
functions are essentially just invoking them and formatting the output.

All the same, testing is generally good; I'll consider how would be best to 
test this, but I will also happily take suggestions if you/others have any.

https://github.com/llvm/llvm-project/pull/143591
_______________________________________________
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits

Reply via email to