Re: Make remote-sim target always have a thread

[Daniel Jacobowitz <drow at false dot org>]

On Wed, Jul 09, 2008 at 12:11:15PM +0100, Pedro Alves wrote:
> A Monday 07 July 2008 20:15:18, Daniel Jacobowitz wrote:
> > If you'd like, I believe arm*-*-eabi and arm*-*-elf should XFAIL the
> > args.exp and fileio.exp failures.  I believe they're inherent; ARM
> > defined the semihosting interface (long ago), and it does not include
> > the necessary bits for some of this.  For instance, the command line
> > is passed as a string.  It's possible some of the fileio.exp failures
> > are bugs in the simulator though - I haven't checked them exhaustively.
> 
> I'd prefer to investigate better the failures than XFAILing them
> blindly.  Not all the args.exp tests fail for instance, and I just
> noticed that there are more failures in other tests than I
> originally reported.  I'll have to pass for now, and defer to someone
> who's more caring for the sim target.

It shows up in hardware ARM testing too.  The args.exp failures are
because the command-line-as-string bits do not handle single and
double quotes in the way GDB expects (i.e. the way a Bourne shell
does); so tests for empty arguments in particular do not work.

The fileio tests, yes, those should be checked more precisely.

Thanks for fixing up sim, anyway :-)

-- 
Daniel Jacobowitz
CodeSourcery