Mail Archives: djgpp/2011/12/31/03:07:42
> > DJ did not suggest to use PE in DJGPP.
>
> "There is a way around this that Cygwin uses (PE is a coff-based format)
> [...]"
Please don't cut out the important parts when quoting me. The full
quote was:
"There is a way around this that Cygwin uses (PE is a coff-based
format) if someone wants to port it to djgpp..."
It would be unreasonable to think I meant "wants to port Cygwin to
DJGPP" so the only reasonable interpretation of what I actually said
was "if someone wants to port the way around it to DJGPP's COFF code".
The part in parens is a side-comment pointing out that Cygwin uses PE
file formats, and PE is a COFF variant, hence showing the relevence to
DJGPP's COFF limitation. It's in parens because it is *not* part of
the main sentence.
> > He suggested to make COFF used by DJGPP work around the
> > 64K limitation using the same technique as in the PE format.
>
> Well, I can only respond to what he actually wrote, and not what he,
> nor you, thought he said.
Please use common sense when listening to what people say. Everyone
else interpreted my comment right. If you know what PE format is,
you'd not think I'd want to use it as DJGPP's file format, and if you
don't know what PE is, you should find that out before making
assumptions.
(PE is the Windows native executable format, which is a wrapper around
COFF like DJGPP is. DJGPP can't produce PE executables because then
they'd be Windows programs, not DOS programs)
> It's not logical to conclude that a fix was present in a coff-based,
> PE format, used in Cygwin, given what he wrote and the way he wrote
> it.
How do you NOT get "Cygwin has a fix" from "There is a way around this
that Cygwin uses" ?
Also, as for your "Why not ELF?" question, the answer is: because I
said so. Using ELF will not magically fix COFF, and DJGPP will always
support COFF, even if it also supports ELF.
- Raw text -