[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: further kioslave queries
From:       Hamish Rodda <meddie () yoyo ! cc ! monash ! edu ! au>
Date:       2002-02-20 5:48:56
[Download RAW message or body]

>> Out the top of my head it can very well be that the code first tries a
>> stat() and that if that gives any errors those are ignored and a get() is
>> tried instead.
>
>Yes, due to KonqRun's "show error messages in the browser" code.
>This has been fixed though, or, well, will be fixed when Hamish Rodda's code
> in konq_run is enabled (and the corresponding handling of error:// in
> khtml_part).
>
>Seems more and more people writing kioslaves are hitting this problem,
>it looks like we should fix this before 3.0.
>
>Hamish: I remember you said you'd enable the more complete error messages
>after 3.1, for translations issues IIRC... but does that prevent us from
> enabling the handling of error://, with the current error messages ? This
> sounds like the best compromise to me. What are the known problems with the
> (currently commented out) error:// code ? I'd be glad to help.

There's nothing stopping the error:// code. Most of the work is easy (I'll do 
the cleanups suggested previously now), there's just that location setting 
problem, plus the code in konq_run which triggers on any error, not just when 
khtml isn't already loaded... problems which I guess are easy to fix when you 
know konq & khtmlpart. Plus I don't know where the get() after error on 
stat() code is to remove it.

The more detailed error messages are waiting moreso for accuracy and because 
of the freeze, there were some problems with them being too generic etc., so 
I'll do them up for review by others after 3.0.

Cheers,

Hamish

BTW, anyone else notice KAddressBook is unbearably slow with many (~1000) 
addresses?

>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]