This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: DWARF2 and COFF?
- To: Michael Meissner <meissner at cygnus dot com>
- Subject: Re: DWARF2 and COFF?
- From: Jeffrey A Law <law at redhat dot com>
- Date: Wed, 06 Dec 2000 17:51:15 -0700
- cc: "Mark E." <snowball3 at bigfoot dot com>, binutils at sources dot redhat dot com
- Reply-To: law at redhat dot com
In message <20001206165036.A3361@cse.cygnus.com>you write:
> > It is my understanding that while dwarf2 happens to be used by
environmen
> ts
> > that use ELF, it is not dependant on ELF. Assuming my understanding is
> > correct, can bfd/dwarf2.c be made generic or should I just give up?
>
> Depends on the COFF varient. Dwarf2 needs:
>
> 1) The ability to create arbitrary sections (some COFF's, notably
> Mips &
> Alpha ECOFF can't create any sections not specified in the ABI);
>
> 2) The ability to to do address relocations non-byte boundaries;
>
> 3) The use of .8byte, .4byte, .2byte and .section pseudo ops;
>
> 4) The ability to mark sections not to be loaded into memory.
I would be all for generalizing bfd/dwarf2.c -- there are other object
formats that do have these kind of abilities (which we do want to use at
some point). SOM for the PA comes to mind :-)
jeff