Re: COFF questions

"Randall Hyde" <>
1 Mar 2001 02:34:59 -0500

          From comp.compilers

Related articles
COFF questions (Xavier Leclercq) (2001-02-23)
Re: COFF questions (Hans-Bernhard Broeker) (2001-02-25)
Re: COFF questions (John Whaley) (2001-02-25)
Re: COFF questions (Randall Hyde) (2001-03-01)
Re: COFF questions (Norman Black) (2001-03-01)
COFF questions (Jeffrey B. Messinger) (1991-02-11)
| List of all articles for this month |

From: "Randall Hyde" <>
Newsgroups: comp.compilers
Date: 1 Mar 2001 02:34:59 -0500
Organization: Posted via Supernews,
References: 01-02-111 01-02-131
Keywords: linker
Posted-Date: 01 Mar 2001 02:34:59 EST

"Hans-Bernhard Broeker" <> wrote in message
> Well, for the start, you should note that what Microsoft calls 'COFF'
> is not the same as what that word meant before they came in to spoil
> it. Another case of "embrace and extend", that is. That's one reason
> why the GNU BFD library needs a separate backend for 'pe_coff'.

I was under the impression that MS changed the fields that are allowed
to be changed in the COFF definition (which seems perfectly reasonable
to me). Can you provide examples where MS actually violated the COFF
standard? The fact that GNU's tools need to handle MS output
specifically probably has to do with debug information. The word
COFF, prior to MS, left the definition of debug info up to the vendor.
You can hardly blame MS for choosing their own variants here. Randy

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.