Ethereal development <ethereal-dev@xxxxxxxxxxxx> schrieb am 29.06.04 08:50:56:
>
> On Mon, Jun 28, 2004 at 10:51:09AM -0700, Guy Harris wrote:
> > ...and the code should do something more useful than crash if that
> > configuration appears in the configuration file as a result of being put
> > in there by hand, i.e. it shouldn't trust the value from the configuration
> > file, it should check it and, if it's not one of the valid values, ignore
> > the value from the configuration file and use the default value.
>
> I've checked in a change to do that.
Interestingly, you checked in a bugfix for a bug I just didn't saw :-)
There's another one, not that obvious. If you have:
gui.layout_content_1: PLIST
gui.layout_content_2: PLIST
putting the packet list into two panes, this will assert GTK, as a widget can't have two parents. The GUI will prevent you from doing this, as the prefs will not.
>
> Perhaps it should report the invalid value as an error, but the prefs
> code currently doesn't have a mechanism to report errors other than
> "syntax error" (which an invalid value wouldn't be, it's just an
> out-of-range number), "no such preference" (but there *is* such a
> preference), or "preference was in older releases but not this one
> (again, not so). That should probably be fixed at some point.
>
I'm unsure what's the best way to handle that.
> BTW, should we provide symbolic names for the layout types? The hard
> part is coming up with someting reasonable descriptive but shorter than
> "the layout with the packet list on the left and the packet detail and
> hex dump on the right, with the packet detail above the hex dump", etc..
>
Yes, I've thought about that while implementing, but didn't found short and descriptive names.
Regards, ULFL
________________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt neu bei WEB.DE FreeMail: http://freemail.web.de/?mc=021193