This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Fix gnat.dg/aliasing2.adb regression
- From: Richard Guenther <richard dot guenther at gmail dot com>
- To: Eric Botcazou <ebotcazou at adacore dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 2 Dec 2010 14:36:53 +0100
- Subject: Re: [patch] Fix gnat.dg/aliasing2.adb regression
- References: <201011290824.07093.ebotcazou@adacore.com> <201011292124.13304.ebotcazou@adacore.com> <AANLkTimwbvHeaihD9LH3AF9wKQ+1sdd7N+-rhQPjF774@mail.gmail.com> <201012021225.35279.ebotcazou@adacore.com>
On Thu, Dec 2, 2010 at 12:25 PM, Eric Botcazou <ebotcazou@adacore.com> wrote:
>> So your original patch is ok.
>
> I missed that this original patch also breaks C++, because TYPE_MIN/MAX_VALUE
> for C++ types can contain C++-specific tree nodes that CONTAINS_PLACEHOLDER_P
> isn't prepared to handle. ?So we would need to teach the latter about them
> and this would slow down both the C++ and the Ada compilers.
>
> I'm proposing the following trick instead: do not fiddle with TYPE_CANONICAL
> if a non-shared type is requested. ?The Ada compiler doesn't want these types
> to be merged, so having different TYPE_CANONICALs is just fine I think.
>
> Only LTO-bootstrapped (for C/Ada) on x86_64-suse-linux for now, I'll do a full
> regular bootstrap if accepted. ?OK for mainline?
Yes.
Thanks,
Richard.
>
> 2010-12-02 ?Eric Botcazou ?<ebotcazou@adacore.com>
>
> ? ? ? ?* tree.c (build_range_type_1): Do not SET_TYPE_STRUCTURAL_EQUALITY
> ? ? ? ?for a non-shared type.
>
>
> --
> Eric Botcazou
>