MIPS gas testsuite now XPASSes 'MIPS16 reloc 2'?

MIPS gas testsuite now XPASSes 'MIPS16 reloc 2'?

[cgd]

To get a feel for the current state of the world, I just built
cross-compilation tools from sparc-solaris to the targets:

	mips-elf
	mips-linux
	mips64-elf
	mips-ecoff

I note that the 'MIPS16 reloc 2' test now XPASSes for mips-elf and
mips64-elf (the only two of those targets for which it's run).

Anybody know if this bug is really fixed?  I'd guess that the testcase
started passing after
http://sources.redhat.com/ml/binutils/2001-06/msg00164.html was
applied, but i've not determined that for sure.


(in case anybody wants to know what's up with the rest: as one might
expect, mips-linux currently passes.  mips-ecoff fails jal-empic,
la-empic, and lifloat-empic.  not looked into the causes of those
failures.)



cgd

Re: MIPS gas testsuite now XPASSes 'MIPS16 reloc 2'?

[Eric Christopher]

> Anybody know if this bug is really fixed?  I'd guess that the testcase
> started passing after
> http://sources.redhat.com/ml/binutils/2001-06/msg00164.html was
> applied, but i've not determined that for sure.
> 
> 

I'm pretty confident that it is fixed and that it happened with that
patch.  Someone would need to test to make sure.  Why is it XPASSED
anyhow?

-eric

-- 
Look out behind you!

Re: MIPS gas testsuite now XPASSes 'MIPS16 reloc 2'?

[Alan Modra]

On Sat, Jul 21, 2001 at 09:20:11AM +0000, Eric Christopher wrote:
> 
> > http://sources.redhat.com/ml/binutils/2001-06/msg00164.html was
> I'm pretty confident that it is fixed and that it happened with that
> patch.

I wasn't particularly trying to fix failing test cases when I made the
patch.  It was more a case of removing code that was no longer
necessary due to the symbol resolution changes, and fixing one or two
things I noticed in the process.

> Why is it XPASSED anyhow?

Someone decided it was too much trouble to fix the assembler and
instead marked the test as "expected to fail" (mips.exp:172).  XPASS
means an unexpected pass.

Alan

Re: MIPS gas testsuite now XPASSes 'MIPS16 reloc 2'?

[Eric Christopher]

> I wasn't particularly trying to fix failing test cases when I made the
> patch.  It was more a case of removing code that was no longer
> necessary due to the symbol resolution changes, and fixing one or two
> things I noticed in the process.

Great.  Thanks.

> 
> > Why is it XPASSED anyhow?
> 
> Someone decided it was too much trouble to fix the assembler and
> instead marked the test as "expected to fail" (mips.exp:172).  XPASS
> means an unexpected pass.
> 

Sometimes ;)  Anyhow, we can probably remove the xfail now that we know
it works.  Then we'll know if it starts failing again.

-eric


--
Look out behind you!