Thanks for your PR, love it. We just had a quick discuison about this PR and we would like to see a solaris host_system check to decided which way to go. Could you please rewrite it?

Could you please check whether #326 helps a bit?

Could you please check whether #326 helps?

It doesn't change anything, since our ld doesn't understand --undefined-version either.

Thanks for your PR, love it. We just had a quick discuison about this PR and we would like to see a solaris host_system check to decided which way to go. Could you please rewrite it?

Sure, I can wrap it with a host_system check.
I see a suggestion for a slightly more elegant implementation of the empty mapfile test in #320 so I'll try to update this one as well.

Simplified the test and wrapped it in a build_machine.system() check. I also added a symbol auto-reduction directive local: * to the real mapfile which removes the need to pass -B local or -B eliminate to ld.

Fine for me. Any objections @DimStar77 ?

Ok, lets fly then :) Thanks!

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 69.37%. Comparing base (de9c069) to head (e85c52d).
Report is 3 commits behind head on main.

@@            Coverage Diff             @@
##             main     #324      +/-   ##
==========================================
+ Coverage   69.30%   69.37%   +0.06%     
==========================================
  Files          18       18              
  Lines         935      937       +2     
  Branches      265      267       +2     
==========================================
+ Hits          648      650       +2     
  Misses        174      174              
  Partials      113      113              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.