This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [C++ Patch / RFC] PR 46206
- From: Paolo Carlini <paolo dot carlini at oracle dot com>
- To: David Edelsohn <dje dot gcc at gmail dot com>
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>, Jason Merrill <jason at redhat dot com>
- Date: Thu, 08 Aug 2013 20:31:32 +0200
- Subject: Re: [C++ Patch / RFC] PR 46206
- References: <CAGWvny=6Nx6W7u47KjisEu8uM_gGUZ8=GjFRyvDM1SrF9HdwYQ at mail dot gmail dot com> <5203DBEE dot 4000103 at oracle dot com> <CAGWvnymGgJaCX_d7L+wL+VH4h0m2-foUZwW7hrMt9yXWCz2HMA at mail dot gmail dot com>
On 08/08/2013 08:18 PM, David Edelsohn wrote:
Why does the patch and fix have any architecture or OS-dependency?
It's tricky, but it depends on the way are internally stored and handled
*multiple* TYPE_DECL for essentially the same typedef. The class layout
must involve both such typedef and a few data members. I had a look to
testresults, and the patch appears to work for other -linux targets too,
not just x86_64-linux, and it's elementary, doesn't add complexity to
the code, thus unless it proves to cause regressions on the targets
where it appears to work, it would be a pity to revert it. But as I said
the issue is weird, after all remained unfixed for 3-4 years,
realistically, I don't know if together with all the other issues I have
in my TODO, I can promise to understand it in such detail to fix it on
AIX too.
Paolo.