d-bus head

[Havoc Pennington]

On Mon, 2003-10-27 at 15:08, David Zeuthen wrote:
> 
> I think there is an issue with DBUS head at the moment, or maybe there
> is something I don't grok. dbus-daemon-1 --system starts fine in the
> background but when I connect with dbus-monitor --system I get the
> 'Disconnected prior to receiving reply' error message. 
> 

"make check" doesn't pass on HEAD at the moment, because a corrupt
message gets in the message stream somehow. If you DBUS_VERBOSE=1 make
check in the "bus" subdir you can get a dump of the broken message.
I fixed 1 other make check problem but didn't get to this one yet.

> Using a CVS snapshot from October 13 works fine (later versions may work
> as well, but I haven't done 'cvs update' after that since I have been on
> vacation for the past two weeks)

If you could narrow down the exact ChangeLog where I broke it that'd be
helpful.

dbus-message.c is fairly seriously messed up at the moment, the 1-byte
field name change and my signatures change were both a bit incomplete.

> Oh, and I get a segfault when launching dbus-daemon-1 on a big-endian
> system (Powerbook G4) - I'll try to debug it as far as I can - maybe my
> own bad.

Let me know if you figure this one out.

I want to make a release as soon as make check is working again, I just
need a bit of free time to sort out that corrupt message issue.

Havoc