Re: gas-980313 mips64orion problems

[Joel Sherrill <joel@OARcorp.com>]

> because apparently a chunk of the opcode table was also removed
> incorrectly.  (This and the tc-mips.c problem should be fixed in the
> next snapshot.)  I'm using an alpha-linux host.

You seem to have fixed the problem.  I am on a i686-pc-linux-gnu according
to config.guess (Redhat 5.0).  I am using the following versions:

BINUTILS        gas-980314
GCC             egcs-1.0.2-980309-prerelease
NEWLIB          newlib-1.8.0 + RTEMS patch

and building "one-tree" style.  In this configuration, I managed to
successfully build a full egcs (including objc and fortran).

Yeah!!!

FYI I managed to build this toolset for every other RTEMS target
using the 980310 snapshot so things look pretty good.  The RTEMS
targets are:

  hppa1.1-rtems, i386-rtems, i386-go32-rtems, i960-rtems,
  m68k-rtems, mips64orion-rtems, powerpc-rtems, sh-rtems,
  sparc-rtems

So things look pretty good.  How close does the 2.9 release look?

> If your problem persists with tonight's snapshot, please let me know.
> Also let me know what host platform and compiler you're using; the
> egcs (non-release) snapshots have recently had some problems with
> binutils, at least on my alpha.  (Oh, and send the .s file if it's not
> too much trouble.  I can try to re-create it, but I haven't been
> working much with the prerelease snapshots recently, so I may not have
> the sources handy.)

OK.  I should have remembered the host system part.  I did not sent
the .s along since it was obvious that code had been removed
accidentally.  Normally I would have sent more along more information.

Thanks.

--joel