This is the mail archive of the
mailing list for the GCC project.
Re: type consistency of gimple
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: David Edelsohn <dje at watson dot ibm dot com>
- Cc: Richard Guenther <richard dot guenther at gmail dot com>, Kenneth Zadeck <zadeck at naturalbridge dot com>, GCC <gcc at gcc dot gnu dot org>, "Mitchell, Mark" <mark at codesourcery dot com>, "Novillo, Diego" <dnovillo at redhat dot com>, "Hubicha, Jan" <jh at suse dot cz>, Andrew Pinski <pinskia at physics dot uc dot edu>
- Date: Fri, 11 Aug 2006 17:31:11 -0400
- Subject: Re: type consistency of gimple
- References: <44DCEAAA.firstname.lastname@example.org> <email@example.com> <200608112103.k7BL3R833292@makai.watson.ibm.com>
David Edelsohn wrote:
> Some historical discussions as a refresher:
I honestly don't have the doc anymore, but i did send it to some people
before i stopped working on it.
I had guessed that nobody would really care enough to review the patches
that would be necessary to fix all the type conversion problems (IE
implicit converts between lhs and rhs that we hvae in some places) until
something like LTO forced it, and at the time, we weren't anywhere close.
Watching the amount of pinging Andrew Pinski has had to do with his type
correctness patches, I can't say this was the wrong decision.
Until someone just adds checks and forces them on, and we have a flag
day about it (IE where we declared we have caught all the obvious
places, and the checks go on, and anything broken still broken is just
fixed as it is found), i have a hard time believing it will get done.
There are just too many nooks and crannies that may get type correctness
wrong for it to be possible to get them all even with bootstraps and
tests on multiple platforms.