2014-10-01 15:19 GMT+02:00 Bertrand Delacretaz <[email protected]>:

> Hi,
>
> On Wed, Oct 1, 2014 at 3:13 PM, Carsten Ziegeler <[email protected]>
> wrote:
> > 2014-10-01 14:22 GMT+02:00 Bertrand Delacretaz <[email protected]>:
> >> ...Should have both local (to the feature) and global variables?...
>
> > Variables are currently local to a feature - do we need global variables?
> > When e.g. overriding a variable from a system property it could use
> > "feature.VARNAME" as the key...
>
> That should work, I agree, but in your example is "feature" the ID of
> a feature? If yes where is it defined?
>
> It's the feature name from the first line: [feature name=REQUIRED]

>> What is "sling.options", does it already have a defined meaning?...
> > This is a defined framework property for the sling run mode handling. It
> > tells Sling's launchpad that one of the two run modes will be active, but
> > exactly one...
>
> Ah, ok - the name does not say anything about that, should rather be
> "sling.runmode.options"? We might accept both names for now and
> deprecate the old "sling.options" name?
>
> Sure we could do that - I agree that sling.runmode.options is a little bit
nicer. I don't have a strange urge to do the deprecation though :)


> > ...I think having the attributes of a section (run mode, start level) in
> the
> > same line is more readable and understandable....
>
> Ok, that works for me as long as run mode and start level as treated
> in the same way, as they are orthogonal.
>

Yes, I agree.

Cool, it seems we have a first agreement. I'll update the implementation
with the points discussed.

Thanks

Carsten


>
> -Bertrand
>



-- 
Carsten Ziegeler
Adobe Research Switzerland
[email protected]

Reply via email to