This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [RFA/testsuite] Fix linux-dp failure for multiple targets
- From: Fernando Nasser <fnasser at redhat dot com>
- To: Daniel Jacobowitz <drow at mvista dot com>
- Cc: gdb-patches at sources dot redhat dot com, Michael Snyder <msnyder at redhat dot com>
- Date: Wed, 30 Jan 2002 10:04:01 -0500
- Subject: Re: [RFA/testsuite] Fix linux-dp failure for multiple targets
- Organization: Red Hat Canada
- References: <20020109183926.A17757@nevyn.them.org> <20020130005733.C28963@nevyn.them.org>
Daniel Jacobowitz wrote:
>
> On Wed, Jan 09, 2002 at 06:39:26PM -0500, Daniel Jacobowitz wrote:
> > Two things I noticed in my testsuite runs. We were never finding the
> > manager thread, and in gcc3 (which apparently unlike GCC2 produces reliable
> > and stable addresses for the data pointers; I don't know why...) the second
> > target board to run would find the old value of the seen array. 'array set'
> > doesn't do what the author of the testcase expected.
> >
> > Is this OK?
>
> I'd appreciate it if someone would review this.
>
Michael Snyder has the final word on this one.
I don't have any objections.
Fernando
> > 2002-01-09 Daniel Jacobowitz <drow@mvista.com>
> >
> > * gdb.threads/linux-dp.exp: Use 'array unset', not 'array set'.
> > (check_philosopher_stack): Check for manager thread before checking
> > for a just-starting thread.
> >
> > Index: linux-dp.exp
> > ===================================================================
> > RCS file: /cvs/src/src/gdb/testsuite/gdb.threads/linux-dp.exp,v
> > retrieving revision 1.4
> > diff -u -p -r1.4 linux-dp.exp
> > --- linux-dp.exp 2001/05/25 01:29:01 1.4
> > +++ linux-dp.exp 2002/01/09 23:35:01
> > @@ -172,6 +172,15 @@ proc check_philosopher_stack {thread see
> > }
> > set interesting 1
> > }
> > + -re ".* in __pthread_manager \\(.*$gdb_prompt $" {
> > + if {$manager_seen == 1} {
> > + fail "manager thread is distinct: $thread"
> > + } else {
> > + set manager_seen 1
> > + pass "manager thread is distinct: $thread"
> > + }
> > + set interesting 1
> > + }
> > -re "pthread_start_thread.*\r\n$gdb_prompt $" {
> > ## Maybe the thread hasn't started yet.
> > pass $name
> > @@ -185,15 +194,6 @@ proc check_philosopher_stack {thread see
> > }
> > set interesting 1
> > }
> > - -re ".* in __pthread_manager \\(.*$gdb_prompt $" {
> > - if {$manager_seen == 1} {
> > - fail "manager thread is distinct: $thread"
> > - } else {
> > - set manager_seen 1
> > - pass "manager thread is distinct: $thread"
> > - }
> > - set interesting 1
> > - }
> > -re " in \\?\\?.*\r\n$gdb_prompt $" {
> > ## Sometimes we can't get a backtrace. I'm going to call
> > ## this a pass, since we do verify that at least one
> > @@ -214,7 +214,7 @@ proc check_philosopher_stack {thread see
> > }
> >
> > set any_interesting 0
> > -array set seen {}
> > +array unset seen
> > for {set i 1} {$i <= 7} {incr i} {
> > if [check_philosopher_stack $i seen] {
> > set any_interesting 1
> >
>
> --
> Daniel Jacobowitz Carnegie Mellon University
> MontaVista Software Debian GNU/Linux Developer
--
Fernando Nasser
Red Hat Canada Ltd. E-Mail: fnasser@redhat.com
2323 Yonge Street, Suite #300
Toronto, Ontario M4P 2C9