[PATCH] remove usage of g_assert() in blockdev

Danny Kukawka danny.kukawka at web.de
Mon Nov 13 10:52:20 PST 2006


On Monday 13 November 2006 19:38, Andrey Borzenkov wrote:
> On Monday 13 November 2006 21:14, David Zeuthen wrote:
> > Danny Kukawka wrote:
[...]
> > Hmm, actually I think the use of the assert may be justified... but I
> > can see this causing problems maybe if uevents are happening during the
> > startup.. So I'd say OK to commit if you also use HAL_WARNING just
> > before the 'goto error'.
> >
> > Can you describe a bit more the circumstances when this bug happens?
> > E.g. what system, are uevents being triggered etc. etc.
>
> https://bugs.freedesktop.org/show_bug.cgi?id=8210

Hm ... didn't saw this bug as we get the problem at SUSE (but this is the 
same, as it looks to me): https://bugzilla.novell.com/show_bug.cgi?id=217563
Unfortunately the bug did not happen on every boot.

I think we don't need to exit/end the daemon only because we can't find a 
device, exceptionally not if we handle such errors in a complete othere way 
in the rest of the code. I can live with a error message and with missing one 
device instead of exit without a comment. ;-)

I add a HAL_WARNING (or should we better use a HAL_ERROR ?) and refer in the 
commit comment to #8210.

Danny


More information about the hal mailing list