[AG-TECH] rat dying

[Don Morton]

Robert Olson wrote:
> 
> what do you have your sample rate set at? I think that's a crash in the
> sample rate conversion code. Might hvae been someone else with a non-std
> sample rate as well.

I'm sure I'm using the default 16KHz.  At any rate, yeh, looks like
that problem is mentioned.  Do you have a reference for source
code?  I'll just compile my own.  I don't like rpm's - too little
control, like Windoze :)


> 
> Looks like this is fixed in the latest rat, time to burn new rpms:
> 
> >4.2.13  - Fix spurious assert in G726 decoder
> >         - Fix G726-40 unpack (bits per sample mis-represented as 3!)
> >         - Fix sinc converter selection bug.  Should fix:
> >                 (a) Crash in convert_sinc.c:
> >                         convert_sinc.c:134: failed assertion
> >                         convert_sinc.c:122: failed assertion
> >                         util.c:87: _block_alloc: Assertion `size > 0'
> > failed.
> >                     reported by Greg Herlein.
> >                 (b) Another crash in convert_sinc.c:
> >                         convert_sinc.c:1107: sinc_upsample_stereo:
> >                         Assertion `out == dst + dst_len' failed.
> >                     reported by <thavaqos at yahoo.com>
> 
> --bob
> 
> At 07:46 PM 3/8/2001 -0700, Don Morton wrote:
> 
> >Hi all, starting today I've had some frequent problems with
> >my rat dying when logged on as normal user (ag).  The message
> >I get is
> >
> >rat-4.2.10-media: convert_sinc.c1107: sinc_upsample_sterio: Assertion
> >`out == dst + dst_len' failed
> >Command rat done
> >
> >It was happening initially in the Windmer room (I'm still using
> >a bridge), but then happened in the Lobby, too, and happened when
> >I changed to a different bridge.  Typically, rat would come up,
> >the other users would start showing up on it, then it would die.
> >Typically died within 30 seconds, often within 5.
> >
> >Just for the halibut, I tried it as root user after all that,
> >and it stayed up for over an hour (in fact, it didn't die
> >at all until I physically killed it).
> >
> >It's definitely an intermittent sort of thing - it happened
> >to me earlier tonight (as user "ag") a couple of times, but
> >my most recent execution of it has been stable for up to
> >5 minutes now.
> >
> >Any ideas?  Without understanding the rat code, I'm having
> >a wild guess that maybe it was a result of something that
> >somebody else was sending, that my rat didn't like?????
> >Again, completely wild guess....
> >
> >I'm off to Vegas on Saturday, so will likely shut our whole
> >AGN down till I'm back on Friday.
> >
> >Thanks, Don
> >
> >    Don Morton                  http://www-scinet.umt.edu/~morton/
> >    Department of Computer Science       The University of Montana
> >    Missoula, MT 59812 | Voice (406) 243-4975 | Fax (406) 243-5139

-- 
   Don Morton                     http://www.cs.umt.edu/u/morton/
   Department of Computer Science       The University of Montana
   Missoula, MT 59812 | Voice (406) 243-4975 | Fax (406) 243-5139