Re: [evolution-patches] [Fwd: eds/calendar, crash import fix]

[Not Zed <notzed ximian com>]

On Tue, 2004-08-03 at 01:25 -0400, JP Rosevear wrote:
> On Mon, 2004-08-02 at 14:45 +0800, Not Zed wrote:
> > Email message/mailbox attachment, "Forwarded message - eds/calendar,
> > crash import fix"
> > On Mon, 2004-08-02 at 14:45 +0800, Not Zed wrote:
> > > 
> > > while looking at another bug i hit this crash on the first calendar
> > > list i tried to import i found on a web page.
> > > 
> > > it still doesn't import (or at least, if it does i can't find where
> > > it went), but at least e-d-s doesn't go into an infinite crash-loop
> > > trying to import a calendar file.
> > > 
> > > so, no idea if this is even closely related ot the right fix.
>
> If comp_uid is NULL after trying to generate it, we should still bail
> out with an error, invalid object or something.  We need the uid (and
> the spec demands it).
Ok, I wasn't sure on that.

But these external calendars seem to have this quite often, i just wonder if it will be needed for interoperability.  It seems odd to me that the importer isn't doing some pre-processing to cope with this though, rather than opening the data directly in the backend code.
> Also, you don't really need to lookup_component in practice if you
> create a uid since it should be unique and hence not found, this is nice
> and safe though.
True.

I'll do another patch which just fails, I guess.

-- 







Michael Zucchi <notzed ximian com>

"born to die, live to work, it's all downhill from here"

Novell's Evolution and Free Software Developer