Re: [RFA] gdb.mi/*.exp, send_gdb vs. gdb_test

[Michael Snyder <msnyder at vmware dot com>]

Vladimir Prus wrote:
> On Wednesday 26 May 2010 04:36:02 Michael Snyder wrote:
>
> > -    send_gdb "-gdb-version\n"
> > -    gdb_expect {
> > -           -re "GNU gdb .*\r\n$mi_gdb_prompt$" \
> > -               { pass "acceptance of MI operations" 
> > -                 return 1}
> > -           -re ".*\r\n$mi_gdb_prompt$" \
> > -               { fail "acceptance of MI operations"
> > -                 note "Skipping all other MI tests." }
> > -           -re "Undefined command.*$gdb_prompt $" \
> > -               { fail "acceptance of MI operations"
> > -                 note "Skipping all other MI tests." }
> > -           -re ".*$gdb_prompt $" \
> > -               { fail "acceptance of MI operations"
> > -                 note "Skipping all other MI tests." }
> > -           timeout { fail "acceptance of MI operations (timeout)"
> > -                     note "Skipping all other MI tests." }
> > +    gdb_test_multiple "-gdb-version" "acceptance of MI operations" {
>
> You replace send_gdb with gdb_test_multiple, while changelog say you
> replace with gdb_test.
>
> > +    note "Skipping all other MI tests."
>
> This seem to be emitted unconditionally now? It was previously emitted
> on timeout only. There are two such issues in -- mi-basics and mi2-basics.
> Seems OK otherwise.

It's not unconditional.  The "pass" clause includes a return, so this
note will only be emitted on failure.