This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Binutils checkin policy (Re: i386 gas is broken)
- From: Hans-Peter Nilsson <hp at bitrange dot com>
- To: "H . J . Lu" <hjl at lucon dot org>
- Cc: <binutils at sourceware dot cygnus dot com>
- Date: Thu, 15 Nov 2001 03:34:53 -0500 (EST)
- Subject: Re: Binutils checkin policy (Re: i386 gas is broken)
On Thu, 15 Nov 2001, H . J . Lu wrote:
> On Thu, Nov 15, 2001 at 06:30:50PM +1030, Alan Modra wrote:
> > What fun! Corrupted frag chain, so something is scribbling where it
> > shouldn't. More fallout from 2001-11-14 Daniel Jacobowitz, I think.
There were no gas patches in there, so it's less likely than the
last days gas changes.
> I'd like to suggest a policy that any non-minor changes to binutils
> should pass
>
> # .../configure --enable-targets=all --enable-64-bit-bfd
> # make
> # make check
>
> on Linux/x86.
(I guess you mean --target=i686-pc-linux-gnu if that's not the
host system.)
A couple of cross-builds with make check before and after the
patch won't hurt either, in addition to or in place of a native
make check. When the patch causing breakage is identified, I
suggest we revert it, and let the author debug it.
FWIW, I see major breakdown in simulator builds which can be
explained by a fragged frag chain (as crashes in those ways).
Only h8300-hms seems to have survived a complete build.
There's another coff target that didn't, i960-coff.
brgds, H-P