[developers] PET XML input and [incr tsdb()]

[Stephan Oepen]

hi again,

> I've attached the debug output. It's not saying much to me, so I hope 
> you can help. At the point I copied that, [incr tsdb()] was still saying 
> 2% processed and doing nothing.

thanks for the detailed logs.  they confirm that the input string has
been sent off to the [incr tsdb()] client, so next you would have to
look into PET to see why it apparently does not parse or return.  if
you were so inclined, i would recommend (re-)compiling a binary with
optimization turned off, launching it as an [incr tsdb()] client, and
then attaching to the process in gdb(1).  i suspect francis is right,
and the PET XML input layer is not quite happy with the input string
the way it arrives in [incr tsdb()] client mode.  but maybe the recent
code changes by bernd actually have resolved this already?

--- using the raw [incr tsdb()] dump mode may be a short-term option,
but in general i am not very fond of it: whenever there is a change to
the DB schema, PET source code needs adaptation.  also, in dump mode,
[incr tsdb()] facilities like parallelization, forced time-outs, auto-
restarts of failed clients, et al. are not available.

                                                         best  -  oe

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+++ Universitetet i Oslo (IFI); Boks 1080 Blindern; 0316 Oslo; (+47) 2284 0125
+++     CSLI Stanford; Ventura Hall; Stanford, CA 94305; (+1 650) 723 0515
+++       --- oe at csli.stanford.edu; oe at ifi.uio.no; stephan at oepen.net ---
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++