* Test failure in mi-cli.exp
@ 2007-08-03 13:28 Robert Norton
2007-08-03 13:50 ` Daniel Jacobowitz
0 siblings, 1 reply; 2+ messages in thread
From: Robert Norton @ 2007-08-03 13:28 UTC (permalink / raw)
To: gdb
I observed the following when running the gdb testsuite on our
(proprietary) target:
888-interpreter-exec console "set $pc=0x0"
~"Current language: auto; currently asm\n"
888^done
(gdb)
FAIL: gdb.mi/mi-cli.exp: -interpreter-exec console "set $pc=0x0"
The problem is that setting $pc to 0 puts us in an asm file causing the
~"Current language:" message which doesn't match the regex. The same
failure occurs in mi2-cli.exp. Since setting the pc is quite likely to
cause messages of this type on various targets may I propose changing
the test to something like:
mi_gdb_test "888-interpreter-exec console \"set \$pc=0x0\"" \
{.*888\^done} \
"-interpreter-exec console \"set \$pc=0x0\""
(Note the .* before 888\^done)
An alternative would be to set something other than the pc.
This has caused problems in the past:
http://www.ecos.sourceware.org/ml/gdb/2004-04/msg00101.html
Cheers,
Robert
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: Test failure in mi-cli.exp
2007-08-03 13:28 Test failure in mi-cli.exp Robert Norton
@ 2007-08-03 13:50 ` Daniel Jacobowitz
0 siblings, 0 replies; 2+ messages in thread
From: Daniel Jacobowitz @ 2007-08-03 13:50 UTC (permalink / raw)
To: Robert Norton; +Cc: gdb
On Fri, Aug 03, 2007 at 06:28:37AM -0700, Robert Norton wrote:
> The problem is that setting $pc to 0 puts us in an asm file causing the
> ~"Current language:" message which doesn't match the regex. The same
> failure occurs in mi2-cli.exp. Since setting the pc is quite likely to
> cause messages of this type on various targets may I propose changing
> the test to something like:
>
> mi_gdb_test "888-interpreter-exec console \"set \$pc=0x0\"" \
> {.*888\^done} \
> "-interpreter-exec console \"set \$pc=0x0\""
>
> (Note the .* before 888\^done)
Or just recognize the "Current language" message specifically. I
would rather avoid .* in MI tests where we don't need it.
--
Daniel Jacobowitz
CodeSourcery
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2007-08-03 13:50 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-08-03 13:28 Test failure in mi-cli.exp Robert Norton
2007-08-03 13:50 ` Daniel Jacobowitz
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox