diff mbox

Regression for attach from stdin [Re: [pushed] Re: [PATCH v6 0/2] enable target-async by default]

Message ID 1404291574.3766.35.camel@bordewijk.wildebeest.org
State Not applicable
Headers

Commit Message

Mark Wielaard July 2, 2014, 8:59 a.m. UTC 
  On Tue, 2014-07-01 at 18:28 +0200, Jan Kratochvil wrote:
> On Thu, 29 May 2014 15:44:02 +0200, Pedro Alves wrote:
> > I went ahead and pushed this all in, with Eli's comments addressed.
> 
> 329ea57934a9d4b250a0b417af1ec47bc2d0ceb6 is the first bad commit
> commit 329ea57934a9d4b250a0b417af1ec47bc2d0ceb6
> Author: Pedro Alves <palves@redhat.com>
> Date:   Thu May 29 19:58:57 2014 +0100
>     enable target async by default; separate MI and target notions of async
> 
> This affects GCC testsuite where guality.exp does not start at all with:
> 	gdb: took too long to attach

Thanks for tracking this down Jan. Should this patch be reverted or
should gcc guality.h be patched to work around the new async default
somehow?

Currently I am using the following patch to guality.h in gcc which seems
to work around it, but I don't really understand why:


Thanks,

Mark

Comments

Pedro Alves July 2, 2014, 9:15 a.m. UTC | #1 
On 07/02/2014 09:59 AM, Mark Wielaard wrote:
> Thanks for tracking this down Jan. Should this patch be reverted 

Unless it turns into a deep rat hole, I'd like to fix the issue
instead.  I haven't investigated this particular issue yet, but I've
been busy fixing PR17072, and I ended up fixing several issues
related to execution commands entered directly in the command line,
before GDB's main event loop starts.  This one sounds related.

> or
> should gcc guality.h be patched to work around the new async default
> somehow?

No, the async default in an internal implementation detail.  It should
cause no visible change of behavior, other than enabling features.
Pedro Alves July 3, 2014, 3:38 p.m. UTC | #2 
On 07/02/2014 10:15 AM, Pedro Alves wrote:
> On 07/02/2014 09:59 AM, Mark Wielaard wrote:
>> Thanks for tracking this down Jan. Should this patch be reverted 
> 
> Unless it turns into a deep rat hole, I'd like to fix the issue
> instead.  I haven't investigated this particular issue yet, but I've
> been busy fixing PR17072, and I ended up fixing several issues
> related to execution commands entered directly in the command line,
> before GDB's main event loop starts.  This one sounds related.

Looking at this now.  The other series unfortunately doesn't fix this.
diff mbox

Patch

diff --git a/gcc/testsuite/gcc.dg/guality/guality.h
b/gcc/testsuite/gcc.dg/guality/guality.h
index 8b657f2..d83e270 100644
--- a/gcc/testsuite/gcc.dg/guality/guality.h
+++ b/gcc/testsuite/gcc.dg/guality/guality.h
@@ -243,6 +243,7 @@  main (int argc, char *argv[])
          || fprintf (guality_gdb_input, "\
 set height 0\n\
 attach %i\n\
+print guality_attached\n\
 set guality_attached = 1\n\
 b %i\n\
 continue\n\