hald-addon-acpi dying

[Danny Kukawka]

On Donnerstag, 19. Juli 2007, Hannu Lyytinen wrote:
> On Wed, 18 Jul 2007, Danny Kukawka wrote:
> > 1) I can only point to my patch from 19.03. this year, to cleanup the
> > dbus_error handling within this addon. See attached patch again.
>
> Yes, your patch does the trick as well. Now the weird part is that
> I am unable to locate your patch either in git or in the hal or
> hal-commit lists. I must confess I am a total newbie when it
> comes to git, but at least I have found out this branch I pulled
> out is "master".

The patch is not part of git, the patch was rejected (search the mailing list 
for more).

> > 2) Are you sure it crash with dbus_error_free() and not by call
> > dbus_error_is_set() ?
>
> When I was investigating the problem, the GDB backtrace showed
> that it did crash in dbus_error_free() right after examining the
> error status with dbus_error_is_set(). However, I think this is
> natural since if "error" was not initialized, it might have
> contained anything and dbus_error_is_set() succeeded.

It is may logical, but IMO the whole DBus stuff should be more robust and 
should be more tolerate against such coding errors. It should not kill the 
programm ... but this is an other discussion.  

> > 3) how did you compile DBus?
>
> It's 1.0.2 shipped in Fedora 7 updates.

okay

Danny