Patch: Only use strict COFF for djgpp

[Nick Clifton]

Hi H.J, Mark, Ian,

  Well I am inclined to agree with Mark on this one.  (ie that his
  original patch was the right way to go).  Especially since it makes
  DJGPP port work.  However I prefer H.Js's approach since it allows
  backwards compatibility.

: Since djgpp wants the strict COFF and other COFF targets are ok
: with the historical behavior, may I proprose this patch?
:
: 2000-08-19  H.J. Lu  <hjl@gnu.org>
: 
: 	* config.in (STRICTCOFF): New for strict COFF.
: 
: 	* configure.in: Define STRICTCOFF for i386-*-msdosdjgpp*,
: 	i386-*-go32* and i386-go32-rtems*.
: 	* configure: Rebuilt.
: 
: 	* config/obj-coff.c (obj_coff_endef): Follow the historical
: 	behavior if STRICTCOFF is not defined.

  I like the patch but with a few reservations:

    * Please could add documentation (to internals.texi) describing the
      configuration option.  In fact please add an entry to NEWS as
      well.

    * We need some way to ensure that the cofftag gas tests will pass
      when STRICTCOFF is defined.  We could just eliminate the test in
      the gas.exp script, (since lots of other ports do), but I feel
      that this is wrong.  I would suggest editing the cofftag.d file
      to allow -1 or -2 for the section numbers for entries 6, 7 and
      8.

    * Adding a comment in obj_coff_endef() describing why STRICTCOFF
      is used would be helpful.

  If you could resubmit a patch with these changes I would be happy to
  approve it.

  One day we may want to make the STRICTCOFF behavior the default,
  but for now lets leave it with DJGPP port.  If other COFF based
  ports start reporting problems then we can switch them over as well.

Cheers
	Nick