kdeutils/kcalc broken from recent kcfg changes

[Albert Astals Cid]

A Dimecres 08 Juliol 2009 03:28:16, Pino Toscano va escriure:
> Hi,
>
> > (If kcfg is not relevant to k-c-d, please drop them with my apologies.)
>
> (the opposite, actually)
>
> > kcalc fails to build due to the additions of _helper methods, since the
> > code relies on variable declarations from previous settings.
> > Unfortunately, things are not as simple as just adding copies of the
> > declarations to each setting, since the code also appears in the ctor.
>
> Yes, I know about it.
>
> > Is there a way to fix this without making many copies of the same
> > variable with different names?
>
> No(t yet).
> Given that adding accessors for the default value, as done originally in a
> recent patch by Albert, can have two down-sides
> a) not being 100% sure <code> snippets are totally independent one from
> each other
> b) avoid people getting "grown" generated cpp/h files for a feature they
> don't use
> I propose, with Albert's approval, the following: explicitely adding in the
> .kcfgc file the option DefaultValueGetters=true for eanbling it (default:
> false). This way, nothing should change in the generated cpp/h files wrt
> KDE 4.3, and only users of that would enable it.
> What do you (all) think about it?

It's ok for me, actually it's something Parker already mentioned on the 
reviewboard stage.

>
> > (Also, who do I yell at for changing the code generation w/o checking
> > that it didn't break existing users? ;-) </friendly ribbing>)
>
> Given that the feature was committed by Albert, who left for akademy the
> day after, I fixed it on my own making it working (modulo <code> snippets
> for default values depending on other ones); so I guess I could (but I
> won't) take your yell.
> Although, given that akademy people had a crappy (at most) internet working
> (so had difficulties in reaching him), and that after all trunk is open
> since few days (and the breakage is there since not even half a week), what
> about being a bit more patient, instead to "have to yell" at first?
> </friendly ribbing>

Actually i think your use of <code> abuses the system, but given the new 
option Pino plans to introduce it'd be "fixed".

Albert