4.2: the binary compatible release

[Kevin Ottens]

Le Monday 20 October 2008, Aaron J. Seigo a écrit :
> On Sunday 19 October 2008, Kevin Ottens wrote:
> > Le Saturday 18 October 2008, Aaron J. Seigo a écrit :
> > >     * Known issues
> > > 	* Service additions
> >
> > Hm, last time I looked at it, it was ok. Did it got much changes I
> > overlooked lately?
>
> no, but i'll be adding hooks to "remote" a Service over Jolie, now that the
> Jolie bits are done. i'm thinking of adding a publish and subscribe set of
> methods so that Plasma::Service can do it all behind the back of the
> Plasma::Service itself by re-routing requests based on the
> publish/subscribe state in Service::startOperationCall

OK, if you want we can take some time on IRC to discuss those ones when you 
reach this task again? I personally envision some gobby foo on some header 
files. :-)

> > > 	* Multiple action runners
> >
> > Not sure I'm qualified to comment on this one. I'm not exactly sure which
> > API got impacted by this feature.
>
> it's Ryan Bitanga's patch, publised on review-board. it has a number of
> issues, however, and simply will not work with the current design of
> KRunner. essentially, it takes AbstractRunner in a completely different
> direction that is wholey incompatible with the current design: it pushes
> the actual action to be trigered out of the AbstractRunner::exec and into
> the Actions instead; those runners would be completely useless depending on
> the UI they are used with and makes it a completely non-starter.

Ow, ouch.

> so i need to spend a few days really soon here to figure out exactly how to
> accomplish the same end without damaging the design.

Okidoki, same offer than for Plasma::Service on my side.
"ervin all your API bitching needs since 1789", yeah babe.

> > > 	* PanelSvg name - apparently people don't like it ;)
> >
> > Seeing the reactions we might find a better name soon. Not a good name
> > (not possible for this class), but at least not a misleading one.
> >
> > I think we could give a try to renaming ConfigXml too. I know it's
> > inheriting KConfigSkeleton, but it's not only about config. I guess this
>
> it isn't about config? what is about?

I've been unclear here, it's that in the plasma API it's not always used for 
config (because of Plasma::Service), otherwise it'd be safe to claim it's 
about config only. It's more of a data scheme loader?

> > one is missing the "loading" metaphor, we've UiLoader, it's doing a
> > similar service for config descriptions.
>
> ConfigLoader? hm..

If we ignore the Plasma::Service use case for a minute, it's IMO a better name 
than ConfigXml already. Teams up nicely with UiLoader, and eludes (works in 
english?) the fact that it's described with some XML (which is kind of an 
implementation detail).

Regards.
-- 
Kévin 'ervin' Ottens, http://ervin.ipsquad.net
"Ni le maître sans disciple, Ni le disciple sans maître,
Ne font reculer l'ignorance."
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
Url : http://mail.kde.org/pipermail/plasma-devel/attachments/20081020/8cd3a82e/attachment.sig