Paolo.
.. I should honestly add that now I don't really believe - as stated by
Volker - that the issue, in any substantive way, is a recent, post
6.1.0, regression: for the GNU extension build_conditional_expr was
definitely copying arg2 = arg1 before that release and using save_expr
on arg1 and the comment in cp_save_expr was already true.
- [C++ Patch] PR 84333 ("[6/7/8 Regression] ICE with tern... Paolo Carlini
- Re: [C++ Patch] PR 84333 ("[6/7/8 Regression] ICE ... Jason Merrill
- Re: [C++ Patch] PR 84333 ("[6/7/8 Regression] ... Paolo Carlini
- Re: [C++ Patch] PR 84333 ("[6/7/8 Regressi... Paolo Carlini
- Re: [C++ Patch] PR 84333 ("[6/7/8 Regressi... Jason Merrill