This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Your patch breaks i386-pe gas
- To: binutils at sourceware dot cygnus dot com
- Subject: Re: Your patch breaks i386-pe gas
- From: "Mark E." <snowball3 at bigfoot dot com>
- Date: Wed, 16 Aug 2000 17:49:11 -0400
- References: <20000816122309.A25775@valinux.com> (hjl@valinux.com)
> Date: Wed, 16 Aug 2000 12:23:09 -0700
> From: "H . J . Lu" <hjl@valinux.com>
>
> This patch breaks gas/testsuite/gas/all/cofftag.s under i386-pe:
[snip]
> I'm pretty sure that the cofftag test captures the way historical COFF
> assemblers behave. So changing this would make gas different from
> historical COFF assemblers. Whether that matters, I don't know.
>
I believe this change of mine is correct for all COFF targets. The change is
definitely what DJGPP COFF expects, or else debug symbols are merged when
they shouldn't.
FWIW, the online COFF manual that is mirrored everywhere it seems supports
the change (copied from
http://osr5doc.sco.com:1996/topics/COFF_SectNumFld.html):
A special section number (-2) marks symbolic debugging symbols, including
structure/union/enumeration tag names,
typedefs, and the name of the file. A section number of -1 indicates that the
symbol has a value but is not relocatable.
Examples of absolute-valued symbols include automatic and register variables,
function arguments, and .eos symbols.
If you guys accept this, I take it I need to change cofftag.[sd]?
Mark