Re: [RFA]: File-I/O patch, testsuite

[Andrew Cagney <ac131313@redhat.com>]

Someone [me] finally back to looking at this testsuite stuff :-(

> the following patch adds two files, fileio.c and fileio.exp, which test
> the new File-I/O protocol.
> 
> However, this tests are currently running always, regardless if the
> target implements this functionality or not.  I'm not quite sure
> about the best way to skip these tests for those targets...

I was going to suggest:

> # test only on a remote target board
> if {! [is_remote target]} {
>     return
> }

but it doesn't hurt to have this run native (should work?).

What about having the expect script recognize the failure mode and then 
just unsupported out?  Something like trying a file open / read first?

What about (from interrupt.exp):

> if [target_info exists gdb,noinferiorio] {
>     verbose "Skipping interrupt.exp because of noinferiorio."
>     return
> }

Doesn't dejagnu's remote config need updating?  interrupt.exp, for 
instance, should now work so the gdb,noinferiorio option should not be set).

Rather than
	gdb_send
	gdb_expect
can:
	gdb_test
or similar be used?

I don't understand why:

+catch "system \"chmod -f +w dir2.fileio.test\""
+catch "system \"rm -rf *.fileio.test\""

appears twice.

Andrew