On Sat, Nov 15, 2025 at 11:08:31PM +0530, Jason Merrill wrote:
> Tested x86_64-pc-linux-gnu. Any opinions? Bikeshedding of the flag name?
>
> -- 8< --
>
> For simple testcases that want to use the std module, it would be useful to
> have a reasonably short way to request building the binary module form
> before the testcase. So with this patch users can write
>
> g++ -std=c++20 -fmodules --compile-std-module mytest.C
>
> gcc/c-family/ChangeLog:
>
> * c.opt: Add --compile-std-module.
>
> gcc/cp/ChangeLog:
>
> * lang-specs.h: Add @c++-system-module.
> * g++spec.cc (lang_specific_driver): Handle --compile-std-module.
>
> gcc/ChangeLog:
>
> * opts-common.cc (generate_canonical_option): Handle
> OPT_SPECIAL_input_file.
If this would be solely for GCC testsuite and nothing else, my preference
would be to do this in DejaGNU with // { dg-compile-std-module
But if the intent is also users compiling simple C++ sources, then that
doesn't work.
I wonder if expansion of --compile-std-module into
-xc++-system-header bits/stdc++.h -xc++-system-module bits/std.cc
-xc++-system-module bits/std.compat.cc -xnone
won't cause surprising result for users which are using -x <language>.
E.g. if their C++ source is *.c and they compile with
g++ -xc++ --compile-std-module foo.c
won't that compile it using cc1 because -xnone cancelled the
earlier language setting?
I.e. don't we need -xpush and -xpop or -xsave and -xrestore
which would save/restore existing -x <language> setting?
Jakub
