Re: Help : bonobo-activation, ORBit2, linc problems
- From: Michael Meeks <michael trna ximian com>
- To: Louise Miller - Sun Microsystems Ireland - Computer Systems - Engineer <Louise Miller Sun COM>
- Cc: <gnome-components-list gnome org>
- Subject: Re: Help : bonobo-activation, ORBit2, linc problems
- Date: Wed, 21 Nov 2001 19:51:53 -0500 (EST)
Hi Louise,
Thanks for your _most_ helpful bug report; this is really
useful and gives me a handle on a nasty bug that's been worrying me (
as well as fixing a load of things, and beefing up the regression
tests :-).
> (process:19124): ** ERROR **: file linc-connection.c: line 123
> (linc_connection_connected): assertion failed: (cnx->priv->tag ==
> NULL)
> aborting...
> FAIL: bonobo-activation-test
> testing async interfaces
You must get an EINPROGRESS from the connect in
linc_connection_initiate, this doesn't tend to happen on Linux; I've
just cleaned that path.
> (process:26267): ** ERROR **: file linc-protocols.c: line 218
> (linc_protocol_get_sockaddr_ipv4): assertion failed: (hostname)
> aborting...
> FAIL: ./test.sh
This seems to suggest that your local comms are not using Unix
domain sockets - which is prolly a major performance issue. This
exposed a bug in the IOP_profile copy for the IOP profile; now fixed.
Anyway; to ensure that we are checking all such things in the
tests I've added some flags to the regression tests - so we do two runs
of test/everything with UDS & IPv4.
Currently though it seems that a bug I've been seeing in
nautilus with the two ends seemingly blocking indefinately is showing
up in test/everything - which is rather good in some ways. I'll commit
the above fixes and get to that.
Regards,
Michael.
--
mmeeks gnu org <><, Pseudo Engineer, itinerant idiot
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]