Not sure - I don't really have much c knowledge, and we haven't currently identified a reproducible test case for it - it's happened in two meetings where the immediate concern was to get back up and running!
Richard > -----Original Message----- > From: Colin Perkins [mailto:c...@csperkins.org] > Sent: 09 October 2004 10:33 > To: Richard Lawley > Cc: 'Access Grid Technical mailing list' > Subject: Re: [AG-TECH] strange rat error > > On 8 Oct 2004, at 11:03, Richard Lawley wrote: > > I've just seen this same problem again in another meeting, > again with > > an Insors node causing the problems. They were showing up > initially > > in rat with two entries. The operator there stopped the > audio tool, > > and one of them went away, but the other remained. Rat > still crashed > > immediately upon connection for everyone else. I tried rat > 4.2.22 and > > 4.2.16 with the same behaviour. We changed venues, and then > > everything was fine (even with the insors node present). > > > > Rat was crashing with two different errors: > > > > rat-4.2.22-media: cc_vanilla.c:298: vanilla_decoder_peek: Assertion > > `done <= len' failed. > > Looking at the code surrounding it, that assert looks bogus. > If you try recompiling with it changed to "assert(done >= > len)" do the problems go away? > > Colin >
