[Bug mi/32571] GDB Mi crashes when setting a disabled breakpoint

https://sourceware.org/bugzilla/show_bug.cgi?id=3D32571

--- Comment #49 from Jarmo Tiitto <jarmo.tiitto at gmail dot com> ---
(In reply to Sam James from comment #47)
> Or, well, I'm even more lost now: you even confirmed it worked in
> https://sourceware.org/bugzilla/show_bug.cgi?id=3D32571#c36, so why would=
 this
> be the same bug? The only possible reason would be that both GCC + gdb
> changes got reverted later, and why would that happen?

All of the context was already under here and this looked *very* similar to
what I have been seeing previously so far. So I wanted to ask here first be=
fore
creating a new bug.

(In reply to Sam James from comment #48)
> One possible explanation is that the bug wasn't fully fixed in GCC, maybe
> because the testcase was over-reduced.
>=20
> 0e8dbf5c5d6ae1d2b78322e5fed70200bc999cba does:
>=20
> +#if defined (__GNUC__) && !defined (__clang__) \
> +  && ((__GNUC__ >=3D 12 && __GNUC__ <=3D 15)       \
> +      || (__GNUC__ =3D=3D 16 && __GNUC_MINOR__ < 1))
> +/* Work around PR gcc/120987 starting gcc 12, and assume it will be fixe=
d in
> +   the gcc 16.1 release.  */
> +#pragma GCC optimize("-fno-ipa-modref")
> +#endif
> +
>=20=20
> If you change it '1' to '2', does it then work again?

I'll try build GDB from the source, since I need to do this anyway now that
it's broken for me.

--=20
You are receiving this mail because:
You are on the CC list for the bug.=