Re: "make gas-test" broken?

[Doug Evans <dje at sebabeach dot org>]

Dave Korn wrote:
> Doug Evans wrote:
>
>   
> > Ya, that's the problem I was referring to in
> > http://sourceware.org/ml/cgen/2009-q3/msg00056.html.
> > I've changed method test-data of class <keyword> to flag a warning and
> > compensate.  It's not ideal, technically wrong, but for right now I
> > don't mind in this particular case.
> >     
>
>   Ah, gotcha; that's this bit, right?
>
>   
> > +	    ;; If an ifield has, e.g., 2 bits (values 0,1,2,3) and the keyword
> > +	    ;; only has two values, e.g. (foo 0) (bar 1), then we can get
> > +	    ;; invalid requests, i.e. for ifield values of 2 and 3.
> > +	    ;; It's not clear what to do here, but it seems like this is an
> > +	    ;; error in the description file.
> >     

Ya, that's it.

>   I also have some ifields like that, they're part of the opcode selection and
> not all combinations are valid.  (I'll test filling them out with dummy
> entries before I update my sandbox just to see if that would have made a
> difference.)
>   

Ifields can have assertions to flag invalid values as prohibited.  
gas-test should honor them but I think it currently doesn't.

>   
> > Ya, this is a change in Guile's behaviour (circa 1.4?).  string-append
> > use to accept symbols, it no longer does.
> >     
>
>   I was just about to ask about guile versions.  I've been noticing testsuite
> FAILs which could be related - I'll take a closer look at any that remain
> after I've updated.
>   

Today we support Guile 1.6.8 and 1.8.x.  I dropped 1.4.x awhile ago.  
When to drop 1.6.x is undecided.  If you're using 1.8.x I think you're fine.