On Tuesday 18 November 2008 21:45:35, Daniel Jacobowitz wrote:
> On Fri, Nov 14, 2008 at 06:50:28PM +0000, Pedro Alves wrote:
> > I'm making use of this:
> > 
> >  /* Debugging level.  0 is off, and non-zero values mean to print some debug
> >     information (higher values, more information).  */
> >  extern int remote_debug;
> > 
> > ... to enable timestamps, you do "set debug remote 2".  level 1
> > still outputs the same as usual.
> 
> It's such a good idea, that it's already there :-)  "set debug
> timestamp".  However, I'd never tried it with remote; it clearly
> does not work right for multiple printf calls in a single debugging
> message.  I'd suggest fixing that, instead.

Ah, sneaky!  I missed that going in.

> 1227044669:904148 Sending packet: 1227044669:904155 $1227044669:904161
> q1227044669:904166 S1227044669:904172 y1227044669:904177
> m1227044669:904182 b1227044669:904188 o1227044669:904193
> l1227044669:904198 :1227044669:904205 :1227044669:904211
> #1227044669:904217 51227044669:904222 b1227044669:904228
> ...1227044669:904279 Ack

That attached patch make two changes to address this.

1) Makes sure that timestamped output always ends with a newline, so we
   get this,

  un-timestamped, like before, unchanged:

    Sending packet: $m80484d1,1#67... Packet received: 51

  and timestamped:

    1227058218:708583 Sending packet: $m80484d1,1#67...
    1227058218:708675 Packet received: 51

  [ If other output other than gdb_stdlog is mixed, then you still
  don't have the nice alignment, but, that doesn't seem as bad. ]

2) Makes sure a single fprintf_unfiltered is used to dump a packet,
   so we get a single timestamp for a packet, unlike your output
   above, where there's one timestamp per octet (that's from
   fputstrn_unfiltered ending up on fprintf_unfiltered once
   per char).

Something like this?

(I'm running it through the testsuite against gdbserver with
"set debug remote 1"/"set debug timestamp" just in case).

-- 
Pedro Alves