Date: Mon, 12 Nov 2007 09:35:14 -0500 From: Alexander Kabaev <kabaev@gmail.com> To: =?iso-8859-1?Q?Bj=F6rn_K=F6nig?= <bkoenig@alpha-tierchen.de> Cc: freebsd-java@freebsd.org Subject: Re: jdk16 and -fno-tree-vrp Message-ID: <20071112093514.01e247e7@kan.dnsalias.net> In-Reply-To: <57823.192.168.1.2.1194874662.squirrel@webmail.alpha-tierchen.de> References: <57823.192.168.1.2.1194874662.squirrel@webmail.alpha-tierchen.de>
index | next in thread | previous in thread | raw e-mail
[-- Attachment #1 --] On Mon, 12 Nov 2007 13:37:42 -0000 (UTC) Björn König <bkoenig@alpha-tierchen.de> wrote: > Hello, > > I tried to investigate the recently discussed problem with jdk16 > where you suggested to add -fno-tree-vrp to CFLAGS, but I can't > reconstruct it. I used 7.0-CURRENT (GENERIC without debugging stuff) > and ports both from October 1st 2007. jdk-6u1 built and installed > without problems for me. I also tried jdk-6u2 without "success"; it > worked fine. > > So I strongly disapprove the change of GCC to not use tree-vrp by > default and I want to suggest that you change it back. GCC 4.2 seem > to have many bugs and I think it's the wrong way to change it's > default behaviour to work around these bugs, because this reveals > other bugs. > > If you are still able to reproduce the error (e.g. with explicit > -ftree-vrp) then I would be glad if you can provide detailed > information about the environment. I want to get the same error, > because I'm searching for a more suitable solution, e.g. a better GCC > patch. > > Regards > Björn 1. jdk15 is missing from the list of jdk's you tried to build, and that was the primary tree-vrp victim; 2. You did not mention architecture you did your experiments on. This particular optimization step was implicated in a number if miscompiles reported by several developers and just taking it out of the picture takes care of majority. This is right decision too given where we are in the release cycle and unclear position of the project towards GPL3. Reverting the patch is task easy enough to anybody. Only after someone takes an active interest in tracking down the miscompiles, creating test cases and submits them to GCC folks to get them fixed in FSF tree, the patch will be removed. Not before. -- Alexander Kabaev [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (FreeBSD) iD8DBQFHOGSiQ6z1jMm+XZYRAvoPAJ9X3HIP0MoevHuUV9avM+O+nalz0gCgk4bQ ncVOjbv5TYofa94wsQhHppY= =pQ9K -----END PGP SIGNATURE-----help
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20071112093514.01e247e7>