Re: QUOTES_IN_INSN

[Hans-Peter Nilsson <hp@bitrange.com>]

On Wed, 11 May 2005, Jan Beulich wrote:

> >>> Hans-Peter Nilsson <hp@bitrange.com> 11.05.05 18:37:00 >>>
> >On Wed, 11 May 2005, Jan Beulich wrote:
> >> gas/testsuite/
> >> 2005-05-11  Jan Beulich  <jbeulich@novell.com>
> >>
> >> 	* gas/mmix/err-byte1.s: Don't expect specific error text on line 10.
> >
> >Can you explain this, please?  Your message said nothing about it.
> >Briefly, it looks like you introduce a regression in the error
> >message.  What is the exact change in the error output?
>
> There are now two messages on the same subject. I simply
> didn't want to check for both indivually; if you'd rather have
> me do it that way, I'll do.

Yes, please.  At least one of the checks should be explicitly
for something like "unterminated string".  Your patch changed
the test to match *any* error message on that line, which is
what I oppose.

Can you please quote the whole error output from the changed
test?

> This appears to be the only
> testcase
> that ever cared to check that unterminated strings are caught,
> which I found kind of surprising...

Sorry. O:-)

(And the specific reason I do it in the mmix-specific parts of
the testsuite is the munging I do in mmix_handle_mmixal called
for some sanity checking, IIRC.)

brgds, H-P