On 11:28, Sat 24 Jan 09, Stuart Henderson wrote:
> On 2009/01/24 10:45, Michiel van Baak wrote:
> >
> > - The configure statement used disables a lot of things.
> > This is not needed because the configure script figures this out
> > itself. The statement here does the same thing:
>
> This is intentional, we want repeatable builds no matter what's
> installed. If somebody adds a new port for something that
> configure picks up, and it gets built and installed in a bulk
> build, the package may change and we don't want that.
Understood. Thanks for explaining why this is done.
>
> It's unlikely to be a big problem with Asterisk since it usually
> just results in a new module being built that won't end up in the
> package, but if it ends up with something linked to the main program
> that isn't recorded as a dependency it breaks the packages for some
> users, so we do it anyway as a precaution.
It causes totally no problem. I just noticed it because I wanted to
build a clean asterisk-trunk checkout with the same configure switches
as the port so I grabbed the configure line out of the config.log.
>
> It's also a useful reminder of things I want to enable when I have
> time :-)
:)
>
> > - chan_h323 wont compile. Something to do with channel.h included in a
> > cxx file. I'm not able to fix it. I tried but failed. The code in
> > question looks valid, so it probably has to do with gcc versions.
> > The digium developers had a look at well, but they are all using gcc4
> > and there it works great.
>
> with gcc4 on OpenBSD the build doesn't fail, but I don't end up
> with chan_h323.so, haven't had chance to check further yet.
> (we don't want to use gcc4 anyway though, the version in ports
> misses some important things from base gcc e.g. stack protector).
I dont even have the gcc4 port installed as I try to use the base gcc
for everything.
>
> I'll see if I can work out what's going on...
Thanks. the problem (I think) is in include/asterisk/channel.h at the
end of the file.
struct ast_group_info {
>
> > - codec_speex borks asterisk (reported before and confirmed by stuart)
>
> This brings a question - can anyone give me a pointer to somewhere
> that explains what can cause SIGBUS on non-strict-alignment arch?
> (amd64 in my case).
>
> > - asterisk 1.6.0.5 has been released to fix an AST- advisory. The code
> > is almost the same as 1.6.0.3, only a hole in chan_iax2.c has been
> > plugged. So use that version please.
>
> /me gives up on asterisk-announce and adds the distribution
> directory to his url checking script
>
> > Other then that the resulting build works great. Running in production
> > now without a lot of trouble.
> >
> > I would have come up with a patch, but too little time today. Maybe
> > later today I can create an updated patch.
>
> updated and against -current (I just committed 1.4.22.2)
cvs up running, will test later today. Thanks.
--
Michiel van Baak
[email protected]
http://michiel.vanbaak.eu
GnuPG key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x71C946BD
"Why is it drug addicts and computer aficionados are both called users?"
