Yongqiang Tian <[email protected]> writes: > Thanks very much for reviewing and testing the patch. > > I do not have GCC commit access. Could you please help me to merge it? > > Thank you very much and have a great day. > > Thanks, > Yongqiang > > On Wed, 21 Oct 2026 at 21:38, Gaius Mulley <[email protected]> wrote: >> >> Yongqiang Tian <[email protected]> writes: >> >> > m2decl_DeclareKnownConstant creates an IDENTIFIER_NODE without setting >> > its string, even though the constant is intended to be anonymous. Tree >> > dumps treat the nonnull DECL_NAME as printable and dereference its null >> > string, causing a crash for anonymous set constants. >> > >> > Build the CONST_DECL with a NULL_TREE name instead. The dumper then >> > uses the declaration's generated name. Add a dump regression test. >> > >> > The test case was found by a non-AI automated testing tool. >> > The investigation was assisted by OpenAI Codex 5.6 Sol and Claude. >> > >> > PR modula2/127991 >> > >> > gcc/m2/ChangeLog: >> > >> > * gm2-gcc/m2decl.cc (m2decl_DeclareKnownConstant): Build the >> > anonymous constant declaration without a name. >> > >> > gcc/testsuite/ChangeLog: >> > >> > * gm2.dg/dump/pim/pass/anonymousconst.mod: New test. >> > * gm2.dg/dump/pim/pass/dg-dump-pim-pass.exp: New test. >> > >> > Signed-off-by: Yongqiang Tian <[email protected]>
Certainly, no problem, I'll commmit the patch using the changelog above. Thanks for the patch and testcode regards, Gaius
