This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] PR 15075 dprintf interferes with "next"


Ping http://sourceware.org/ml/gdb-patches/2013-06/msg00139.html

Thanks,
Hui

On Fri, Jun 7, 2013 at 11:15 AM, Hui Zhu <teawater@gmail.com> wrote:
> Hi Pedro,
>
> Thanks for your review.
>
> On Tue, Jun 4, 2013 at 1:48 AM, Pedro Alves <palves@redhat.com> wrote:
>> On 06/03/2013 05:06 AM, Hui Zhu wrote:
>>> Ping http://sourceware.org/ml/gdb-patches/2013-05/msg00958.html
>>>
>>
>> As this exposes the non-stop racy failure, we should fix it that one
>> first.  Failing that, we should kfail or skip the test with
>> remote targets.  Let's consider the latter option later if we don't
>> manage to address the race timely.  As I said on:
>>
>>  http://sourceware.org/ml/gdb-patches/2013-05/msg01111.html
>>
>> I'm investigating this.  I have a prototype patch, but I need
>> a bit more to handle some details, like what to do with
>> signal catchpoints when we find threads had been stopped with
>> a signal (I'm currently thinking of skipping the catchpoints).
>> I'm composing a test to exercise/expose this kind of stuff,
>> for a better RFC.
>
> I add a check in the begin of dprintf-non-stop.exp.  Then if the
> target in remote, this test will not support.
>
> Please help me review it.
>
> Thanks,
> Hui
>
> 2013-06-07  Yao Qi  <yao@codesourcery.com>
>             Hui Zhu  <hui@codesourcery.com>
>             Pedro Alves  <palves@redhat.com>
>
>         PR breakpoints/15075
>         PR breakpoints/15434
>         * breakpoint.c (bpstat_stop_status): Call
>         b->ops->after_condition_true.
>         (update_dprintf_command_list): Don't append "continue" command
>         to the command list of dprintf breakpoint.
>         (base_breakpoint_after_condition_true): New function.
>         (base_breakpoint_ops): Add base_breakpoint_after_condition_true.
>         (dprintf_create_breakpoints_sal,
>         dprintf_after_condition_true): New functions.
>         (initialize_breakpoint_ops): Set dprintf_create_breakpoints_sal
>         and dprintf_after_condition_true.
>         * breakpoint.h (breakpoint_ops): Add after_condition_true.
>
> 2013-06-07  Yao Qi  <yao@codesourcery.com>
>             Hui Zhu  <hui@codesourcery.com>
>
>         PR breakpoints/15075
>         PR breakpoints/15434
>         * gdb.base/dprintf-next.c: New file.
>         * gdb.base/dprintf-next.exp: New file.
>         * gdb.base/dprintf-non-stop.c: New file.
>         * gdb.base/dprintf-non-stop.exp: New file.
>         * gdb.base/dprintf.exp: Don't check "continue" in the output
>         of "info breakpoints".
>         * gdb.mi/mi-breakpoint-changed.exp (test_insert_delete_modify):
>         Don't check "continue" in script field.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]