Mirror of the gdb mailing list
 help / color / mirror / Atom feed
* Target Output Streaming in MI
@ 2002-05-02 14:11 chanskw
  2002-05-02 14:19 ` Daniel Jacobowitz
  0 siblings, 1 reply; 3+ messages in thread
From: chanskw @ 2002-05-02 14:11 UTC (permalink / raw)
  To: gdb

Hi,

As we are doing more investigation with MI, we having another question
regarding output streaming in MI.  As described in GDB manual:

`"@" STRING-OUTPUT'
     The target output stream contains any textual output from the
     running target.

Some of the examples in the manual show that program's output should be
prefixed with '@'.  However, we have never seen program output prefixed
with '@'.  After doing more experiments and looking into GDB source code,
we found that program output will only be prefixed with '@' during remote
debugging.  Is this really the case?

Thank you.

Regards,
Samantha Chan



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Target Output Streaming in MI
  2002-05-02 14:11 Target Output Streaming in MI chanskw
@ 2002-05-02 14:19 ` Daniel Jacobowitz
  2002-05-02 20:43   ` Brian Thomson
  0 siblings, 1 reply; 3+ messages in thread
From: Daniel Jacobowitz @ 2002-05-02 14:19 UTC (permalink / raw)
  To: chanskw; +Cc: gdb

On Thu, May 02, 2002 at 05:10:56PM -0400, chanskw@ca.ibm.com wrote:
> Hi,
> 
> As we are doing more investigation with MI, we having another question
> regarding output streaming in MI.  As described in GDB manual:
> 
> `"@" STRING-OUTPUT'
>      The target output stream contains any textual output from the
>      running target.
> 
> Some of the examples in the manual show that program's output should be
> prefixed with '@'.  However, we have never seen program output prefixed
> with '@'.  After doing more experiments and looking into GDB source code,
> we found that program output will only be prefixed with '@' during remote
> debugging.  Is this really the case?

Essentially, yes.  There's a known failure in mi-console.exp
corresponding to this; it requires that GDB manage pseudoterminals for
the inferior process, and we don't do that (nor have we really decided
that it is a good idea; I at least have doubts...).


IMVHO, the best thing to do for native would be to have MI traffic on a
separate file descriptor than the inferior's stdin/out/err.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Target Output Streaming in MI
  2002-05-02 14:19 ` Daniel Jacobowitz
@ 2002-05-02 20:43   ` Brian Thomson
  0 siblings, 0 replies; 3+ messages in thread
From: Brian Thomson @ 2002-05-02 20:43 UTC (permalink / raw)
  To: gdb


Daniel Jacobowitz wrote:

> IMVHO, the best thing to do for native would be to have MI traffic on a
> separate file descriptor than the inferior's stdin/out/err.
>
Could the external MI client pass the fd # on the command line?
Are there host platforms where an fd can't be externalized like that?

Brian Thomson
IBM




^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2002-05-03  3:43 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-05-02 14:11 Target Output Streaming in MI chanskw
2002-05-02 14:19 ` Daniel Jacobowitz
2002-05-02 20:43   ` Brian Thomson

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox