Re: C++ testsuite changes

[Daniel Jacobowitz <drow at mvista dot com>]

On Thu, Jan 01, 2004 at 05:15:45PM -0500, Michael Chastain wrote:
> drow> Can you identify what expect feature was causing the problem?  I'm
> drow> quite curious.
> 
> I'm working on it.  Something like this is happening:
> 
>   virtfunc.exp calls gdb_test_multiple
>   gdb_test_multiple gets the "eof" case
>   virtfunc.exp calls gdb_test_multiple with nearly the same pattern
>   gdb_test_multiple gets the "eof" case
>   virtfunc.exp calls gdb_test_multiple with nearly the same pattern
>   gdb_test_multiple gets the "eof" case
>   virtfunc.exp calls gdb_test_multiple with nearly the same pattern
>   gdb_test_multiple gets the "eof" case
>   virtfunc.exp calls gdb_test_multiple with a different pattern
>   gdb_test_multiple recovers and gets normal output!
> 
> I'll try futzing with the problem pattern and see if that makes the
> problem go away.
> 
> drow> I suspect we need to do an expect update...
> 
> Well, tcl 8.4.5 + expect 5.3.9 + dejagnu 1.4.3 works fine on
> hppa2.0w-hp-hpux11.11 (as well as i686-pc-linux-gnu of course).
> 
> What would it take to do an expect update?  In particular, what is the
> list of hosts that have to be tested with a new version of 'expect'
> in order to do an expect update?

I haven't the faintest idea.  But I think it may be better to simply do
the update and deal with the consequences.  The real question is, does
anything other than dejagnu use the in-tree expect?

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer