KDE/kdevplatform

[Andreas Pakulat]

On 27.11.09 15:34:30, Esben Mose Hansen wrote:
> On Friday 27 November 2009 13:20:26 Andreas Pakulat wrote:
> > > > As I said, the plugin cannot be re-activated without re-activated the
> > > > public API or a major overhaul.
> > >
> > > Can't we just deprecate it? And add a notice that it will be gone by
> > > 4.4.1 or something?
> > 
> > I'm not going to start with public API declared deprecated in the initial
> > release. That simply doesn't make any sense at all.
> 
> I think it is the perfect solution, actually. It is doable, it would harm 
> noone, and it would get a usable kdevelop out of the door. But it is your 
> call, of course. I only strive to help get the best possible software out of 
> the door.
> 
> > > What I don't get: This worked a few weeks ago, last time I updated.
> > > 1052413 seems to still work, but I don't dare upgrade now. I think I am
> > > missing some important piece of the puzzle here: What has happened in
> > > those few days that requires the removal of such an important part of
> > > KDevelop?
> > 
> > Its public API we'd have to maintain, its code that nodoby wants to work on
> > and it has serious design problems that cause crash-bugs when you're trying
> > to configure things. Its simply code thats not up to our standards (IMHO)
> > and thus shouldn't be part of the release. We've removed lots of plugins
> > already for exactly this reason.
> 
> ok, it is the public API part, So I propose my solution above...

If you remove the public API there's no way anymore for anything except the
Menu-Action to do code-formatting. And that is just not important enough to
have in kdevelop, its easily replaced by running astyle in a shell.

> could I beg you to sleep on it before rejecting it? :)

I already slept two nights over the discussion and I'm getting rather tired
of it. I'm re-iterating one last time: Either we ship without
source-formatter, or we don't ship with KDE 4.4 and release on our own.
Vote on that please.

> I know it is a pragmatic solution, but I really cannot see the harm, and
> I can see clear advantage, If KDevelop4.0 enters without code-formatting,
> I fear there might be blood in the reviews :(

Sorry, but reviews are going to take kdevelop4.0 apart anyway. It doesn't
have even half the features of kdev3 or half the features of any other
comparable IDE out there. So I don't care too much about that part.

Andreas

-- 
Long life is in store for you.