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

List:       kde-core-devel
Subject:    Re: kconfiggroup and unhandled types
From:       "Aaron J. Seigo" <aseigo () kde ! org>
Date:       2007-10-29 16:23:53
Message-ID: 200710291023.54232.aseigo () kde ! org
[Download RAW message or body]


On Monday 29 October 2007, Oswald Buddenhagen wrote:
> On Sun, Oct 28, 2007 at 10:20:34PM -0600, Aaron J. Seigo wrote:
> > were there specific reasons for using kFatal that warrant keeping it
> > that way?
>
> well, it is a coding error, not a runtime error. we can't have

it's a coding error on the part of KConfig afaic. and dumping the entire 
config just because *one entry* is not supported is completely daft.

> compile-time checking, so at least kFatal should be used to avoid the
> error slipping through the cracks - seriously, who reads stdout/stderr
> unless an error is blatantly obvious and "suggests" that something might
> have been logged about it?

the problem is that this is the kind of thing that is easy to slip out into a 
production release. all it takes is one particular setting not being tested 
that happens to use an unsupported type (thank goodness i use QRectF in 
plasma, huh?) and there we have it: an app that will crash for 1000s of 
users.

moreover, i didn't even *notice* that the app was terminating prematurely 
because it was happening on app exit and there wasn't any further debug 
output expected at that point. it looked like the app terminated normally, to 
be honest.

i noticed the problem by. ...... reading the errors on stdout. so seriously 
who reads it? me. and because the error was obvious: it had decided not to 
save the configuration values.

at least now the debug output says what's actually happening versus the 
previously rather unhelpful "unhandle type" message ;)

-- 
Aaron J. Seigo
humru othro a kohnu se
GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA  EE75 D6B7 2EB1 A7F1 DB43

KDE core developer sponsored by Trolltech

[Attachment #3 (application/pgp-signature)]

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