This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA] gdb.base -- use gdb_test_no_output
- From: "Ulrich Weigand" <uweigand at de dot ibm dot com>
- To: jan dot kratochvil at redhat dot com (Jan Kratochvil)
- Cc: msnyder at vmware dot com (Michael Snyder), gdb-patches at sourceware dot org (gdb-patches at sourceware dot org)
- Date: Fri, 11 Jun 2010 21:35:00 +0200 (CEST)
- Subject: Re: [RFA] gdb.base -- use gdb_test_no_output
Jan Kratochvil wrote:
> FYI in the testcase the inferior is no loner running when `set language asm' is
> entered. There is preceding:
>
> (gdb) run^M
> Starting program: .../gdb.base/setshow foo bar blup baz bubble^M
> foo bar blup baz bubble ^M
> Program exited normally.^M
> (gdb) PASS: gdb.base/setshow.exp: passing args
Ah, that would explain it! This part of setshow.exp:
# Don't test if we can't pass args or if we're using a stub.
if { ![target_info exists use_gdb_stub] && ![target_info exists noargs] } {
#test passing args
gdb_test "cont" "Continuing.*" "continuing"
delete_breakpoints
gdb_test "run" "Starting program:.*foo bar blup baz bubble.*" "passing args"
}
is skipped on spu-elf because we cannot pass command line arguments
(well, at least it's not set up that way).
However, it seems to me that this is still a bug: I don't really think the
intention was that the remaining checks are to be executed on a running
inferior on some targets, and on a exited inferior on others ...
Bye,
Ulrich
--
Dr. Ulrich Weigand
GNU Toolchain for Linux on System z and Cell BE
Ulrich.Weigand@de.ibm.com