[patch] asynchronous callouts

[David Zeuthen]

On Mon, 2004-04-26 at 22:12, Joe Shaw wrote:
> > Eh, it didn't work with hal-device-manager. 
> 
> What happened?  Traceback?  It worked ok here.
> 

I suppose different hardware will exercise different codepahts! Or
something; not sure. I mean; hald worked fine, so did lshal, but
hal-device-manager didn't work, most probably due to the way the device
is constructed.  At some point it would probably be good to stash a few
sysfs-trees in CVS we can do some regression testing against. Someday..

> > The patch also got a hal_device_store_print () utility. It would be nice
> > if you can add ChangeLog entries for this.
> 
> Yep, and the rest, of course.
> 

Thanks.

> I kept the signal connection there as-is, but I moved the add_to_gdl()
> from bus_device.c and class_device.c (which were identical) into
> common.c.  It didn't seem to make sense to me to abstract it out as long
> as it's pretty straightforward.
> 

Okay, makes sense.

> > And we need some async delay for when removing devices, but I suppose
> > this can wait a bit.
> 
> To protect removing and then immediately reinserting the same device?
> We'll need some way to detect that. :)
> 

Heh, that's funny :-) 

No, seriously some future set of callouts might do stuff to the system
on add that needs to be unrolled in the right order. 

Which reminds me that we should probably wait for children to disappear
before removing a device. It's not important to do right now, but I've
just added a note in doc/TODO about it.

Cheers,
David



_______________________________________________
hal mailing list
hal at freedesktop.org
http://freedesktop.org/mailman/listinfo/hal