RE: [RFA] Submit process record and replay third time, 3/9

["Marc Khouzam" <marc.khouzam@ericsson.com>]

Hi,

again, thanks for the great work on Process Record and Replay.
I almost have the Eclipse DSF-GDB frontent for it.
However I ran into a small issue where a frontend is unable to turn
off Record.  I'm not sure if there should be an MI command corresponding
to the 'record' command, but if I have to use 'record' directly to
turn *off* reverse debugging, I get a prompt

"Process record target already running, do you want to delete the old
record log?(y or [n])"

The prompt is not the problem.  The problem is that the default is 'no'.
I think that when running in a frontend (at least in eclipse), the
default
choice automatically gets selected. Therefore, the frontend cannot turn
off
recording.

Could we make the default to be 'yes'?
I get the feeling there is a better way to handle this... I just don't
know how.

Thanks

Marc


> -----Original Message-----
> From: gdb-patches-owner@sourceware.org 
> [mailto:gdb-patches-owner@sourceware.org] On Behalf Of teawater
> Sent: Monday, January 12, 2009 10:06 PM
> To: gdb-patches@sourceware.org
> Subject: Re: [RFA] Submit process record and replay third time, 3/9
> 
> Update follow the cvs-head.
> 
> On Thu, Jan 8, 2009 at 13:46, teawater <teawater@gmail.com> wrote:
> >
> > This patch add the process record and replay target.  This 
> is the core
> > part of process record and replay.
> >
> > 2009-01-08  Hui Zhu  <teawater@gmail.com>
> >
> >        Process record and replay target.
> >
> >        * Makefile.in (record.c): New file.
> >        * record.c, record.h: New file.
> >
> >        2008-12-28  Michael Snyder  <msnyder@vmware.com>
> >        * Comments, spelling, white space clean-ups.
> >
> >        2008-12-26  Michael Snyder  <msnyder@vmware.com>
> >        * record.h: Don't export record_not_record.
> >        * record.c (record_not_record): Rename to in_record_wait.
> >        (record_not_record_set): Rename to in_record_wait_set.
> >        (record_not_record_cleanup): Rename to 
> in_record_wait_cleanup.
> >        (record_store_registers): Check in_record_wait flag.
> >        (record_xfer_partial): Ditto.
> >
> >        2008-10-07  Michael Snyder  <msnyder@vmware.com>
> >        * record.h (record_exec_direction): Delete.
> >        (RECORD_IS_REPLAY): Consult infrun global direction variable.
> >        * record.c: (record_wait_cleanups): Use infrun state 
> variable.
> >        (record_wait): Ditto.
> >        (record_get_exec_direction, 
> record_set_exec_direction): Remove.
> >        (record_can_execute_reverse): New target method.
> >
> >        2008-10-06  Michael Snyder  <msnyder@vmware.com>
> >        * record.c (displaced_step_fixup): Remove.
> >        (record_message_cleanups): Remove displaced step handling.
> >        (record_message): Remove displaced step handling.
> >
> >        2008-10-06  Michael Snyder  <msnyder@vmware.com>
> >        * record.c (record_regcache_raw_write_regnum): Remove.
> >
> >        2008-10-05  Michael Snyder  <msnyder@vmware.com>
> >        * record.c, record.h: Rename execdir to exec_direction.
> >
> >        2008-10-02  Michael Snyder  <msnyder@vmware.com>
> >        * record.c (record_open): Call target_can_async_p() instead
> >        of relying on a global variable.
> >        * record.h (record_linux_async_permitted): Delete.
> >
> >        2008-09-19  Michael Snyder  <msnyder@vmware.com>
> >        * record.c (trivial): Fix two commas in a comment.
> >        * record.c (record_wait): On end of record log, return
> >        TARGET_WAITKIND_NO_HISTORY and let infrun decide what to do.
> >
> >        2008-09-06  Michael Snyder  <msnyder@vmware.com>
> >        * record.c: Comment and message string cleanup.
> >        Add some function header comments.
> >
> >        2008-08-01  Michael Snyder  <msnyder@specifix.com>
> >        * record.c (_initialize_record): Clarify language in help
> >        strings.
> >        Fix up comment format (period must be followed by 
> two spaces).
>