[RFC] Add call to gdb_exit prior to gdb_start in gdb.cp/ref-params.exp

[RFC] Add call to gdb_exit prior to gdb_start in gdb.cp/ref-params.exp

[Kevin Buettner]

I've been testing a mep-elf toolchain using the SID simulator.  I found
it necessary to add the call to gdb_exit as indicated in the patch below.
Without this patch, I get 3 passes and 16 unresolved testcases.  With the
patch, I see 14 passes (and no failures or unresolved testcases).

I took a look at invcations of gdb_start in other tests and found
that it is customary to invoke gdb_exit prior to gdb_start.  (It is
not _always_ done, but nearly so.)  Thus it seems to me that the
missing call to gdb_exit in ref-params.exp is merely an oversight.

Comments?

	* gdb.cp/ref-params.exp (gdb_start_again): Invoke `gdb_exit'
	prior to `gdb_start'.

Index: testsuite/gdb.cp/ref-params.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.cp/ref-params.exp,v
retrieving revision 1.3
diff -u -p -r1.3 ref-params.exp
--- testsuite/gdb.cp/ref-params.exp	9 Jan 2007 17:59:12 -0000	1.3
+++ testsuite/gdb.cp/ref-params.exp	13 Jul 2007 23:34:03 -0000
@@ -46,6 +46,7 @@ proc gdb_start_again { text } {
     global binfile
     global srcfile
 
+    gdb_exit
     gdb_start
     gdb_reinitialize_dir $srcdir/$subdir
     gdb_load ${binfile}

Re: [RFC] Add call to gdb_exit prior to gdb_start in gdb.cp/ref-params.exp

[Daniel Jacobowitz]

On Fri, Jul 13, 2007 at 04:53:41PM -0700, Kevin Buettner wrote:
> I've been testing a mep-elf toolchain using the SID simulator.  I found
> it necessary to add the call to gdb_exit as indicated in the patch below.
> Without this patch, I get 3 passes and 16 unresolved testcases.  With the
> patch, I see 14 passes (and no failures or unresolved testcases).
> 
> I took a look at invcations of gdb_start in other tests and found
> that it is customary to invoke gdb_exit prior to gdb_start.  (It is
> not _always_ done, but nearly so.)  Thus it seems to me that the
> missing call to gdb_exit in ref-params.exp is merely an oversight.
> 
> Comments?

Hmm, my runs of ref-params.exp don't show GDB being restarted.  I
think it's just reusing the same GDB.  Still, your patch looks right.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [RFC] Add call to gdb_exit prior to gdb_start in gdb.cp/ref-params.exp

[Daniel Jacobowitz]

On Fri, Jul 13, 2007 at 04:53:41PM -0700, Kevin Buettner wrote:
> I've been testing a mep-elf toolchain using the SID simulator.  I found
> it necessary to add the call to gdb_exit as indicated in the patch below.
> Without this patch, I get 3 passes and 16 unresolved testcases.  With the
> patch, I see 14 passes (and no failures or unresolved testcases).
> 
> I took a look at invcations of gdb_start in other tests and found
> that it is customary to invoke gdb_exit prior to gdb_start.  (It is
> not _always_ done, but nearly so.)  Thus it seems to me that the
> missing call to gdb_exit in ref-params.exp is merely an oversight.
> 
> Comments?

Hmm, my runs of ref-params.exp don't show GDB being restarted.  I
think it's just reusing the same GDB.  Still, your patch looks right.

-- 
Daniel Jacobowitz
CodeSourcery

Re: [RFC] Add call to gdb_exit prior to gdb_start in gdb.cp/ref-params.exp

[Kevin Buettner]

On Fri, 13 Jul 2007 16:53:41 -0700
Kevin Buettner <kevinb@redhat.com> wrote:

> 	* gdb.cp/ref-params.exp (gdb_start_again): Invoke `gdb_exit'
> 	prior to `gdb_start'.

Committed.

Kevin