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

List:       kde-devel
Subject:    Re: Focus Policy
From:       Don Sanders <dsanders () cch ! com ! au>
Date:       1999-12-30 1:05:47
[Download RAW message or body]

On Thu, 30 Dec 1999, Stephan Litsch wrote:
> > > I am reading the mailinglists "kde-devel" and "kde-core-devel" but
> > > where can I find a primary site or README about the actual situation
> > > of porting, changes and bugs? It would already be helpfull to get
> > > informations about the actual CVS-status: broken, may be or try it!
> 
> > CVS should compile at all times. If not, someone needs to fix it :-)
> 
> And NT should behave as a real operation system! What do you want to say by
> "SHOULD"! And who is "SOMEONE"?

It should compile unless someone has stated on kde-devel or kde-core that it
will be broken. It may be broken at any time due to changes.

If an unplanned compile error persists for say several days you are free to
complain to the list. I normally spend half an hour trying to fix stuff myself
just in my version, or just leave out optional parts that don't compile. (I
tend to update once every three weeks, this is a real burden and takes up a lot
of me free time, I choose my time to update very carefully).

Normally core developers fix things that are wrong as they find them so I don't
bother to post error reports.

> The link "kdecvs.stud.th.heilbronn.de" in "www.kde.org" is down to determine
> who is "SOMEONE".
Probably "someone", is someone reachable via this list. 

> I have no CVS-access -- or is there a possiblity to get
> read-only-access now?
This is what cvsup is for http://www.kde.org/cvsup.html

> The developers completly ignore the "Changes"- or
> "PORTING"-files!
That's a shame.

> The site "developer.kde.org" also seems to be down for some
> days (weeks)! Or have you tried to reach the site before answering? No? Ok!
http://developer.kde.org/ is working from here.

> Dear Waldo, your answer is VERY helpfull! Thanks!!! 
> 
> > > It's the big loss of time, I complain about. Wouldn't it be nice if
> > > there would be a location, I can read about bigger changes in
> > > progress? And about: WHAT HAS CHANGED!
> > 
> > Maybe we can start a mailinglist where developers can announce
> > changes / new features.
> 
> That would be nice. Please start with the actual changes in KFileDialog and why
> it is necessary to replace a functioning class with the actuall bullshit in
> CVS. 
Take it easy. The KRASH statement that Waldo made specifically stated that the
KFileDialog API would be subject to change. This change was further discussed
on kde-core.

See http://www.kde.org/announcements/announce-1.89.html

It states:
"the following areas major changes are foreseen between now and KDE 2.0: 

Interface and launch method of the IO-Slaves (kio / kioslaves) 
File selection dialog (kfile rewrite) 
Parts embedding (kparts redesign) 
Advanced features of HTML widget (khtml) 
Embedding within kicker (kapplet) 
Sound system"
(kfile should have been kfiledialog I think).

It did take me by surprise and kfiledialog was completely broken
last time I updated, which was time consuming to work around for me.

> Please explain me, why KRASH has been released with a "stabilized API" and
> two  weeks later (28/12/1999) all my ported applications do not compile
> anymore?!? What the fuck are you doing there? That's exactly the point I meant
> with "loss of time"!!!!!
> 
> Can you tell me (and others) what changed and why, Why, WHY?
> 
> Would some KDE-developers please be so nice to answer me? No, Waldo, everyone
> but you, please! Excuse me, but I would prefer an answer from someone who thinks
> before writing...

I do not maintain the code that is the source of you confusion. However I am in
a similar position to you in that I want to develop applications rather than
waste time dealing with changes in the APIs, (this is why I respond to show
that you are not alone, I hope that is some condolence).

I spent many weeks finding out why an application I work on would
spontaneously core dump in often non-reproduceable ways after the libs had
changed underneath it. This is a pretty frustrating experience, the cause of
most of my problems were changes in QT and for a while I was sending a bug
report every other day. (The Trolls were extremely responsive to my reports and
fixed the problems promptly, so at least these things don't affect you).

So like you changes in the libraries result in much "wasted time" for me. All I
can do is have faith that the changes made are necessary and beneficial, and
trust that the core developers realise the extent their changes inconvenience
application developers and retard progress in the application development
sphere. (They should have some idea about this as they have been through the
same thing relying on MICO and QT, and many core developers are also application
developers and their changes have broken some of the applications they have
created).

Basically this is the cost of rapid non-backwards compatible progress. The cost
of bugs in the APIs grows exponentially with time so it's probably I good idea
to fix them ASAP, and this was version 1 of the 'product' so it's kind of
expected that the API wouldn't be perfect.

That's about all I have time for right now.

BFN,
Don.

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