Re: GObject's property with _ issue

Piotras <pp-VVDi8QVAvoBWk0Htik3J/[email protected]> Mon, 20 Feb 2006 16:49:20 +0100
Newsgroups gmane.comp.web.midgard.devel
Message-ID <[email protected]>
"Jukka Zitting" <[email protected]> wrote:

> Hi,

Hi,
 
> These name mappings could be automatically applied and unapplied
> whenever the API boundaries are crossed. Things like QB and reflection
> capabilities would do this translation also for all property name
> strings.
> 
> I agree that this solution is a bit complex conceptually, but in
> practice it would be totally transparent to applications.

I do not think it's complex , at least we can solve this problem and problem 
with real column names used by property and "reserved" as property's nick. 

The only problem is  is lack of any GObject  API  which can support real property id. 
Also limitation at this moment is the point when you initialize classes after 
you parsed schema. That one requires something like need to redefine internal 
class data.

I wrote today about this because for the first time we had segfault related to 
this issue. It was related to g_hash_table with strings  dependency for property attributes.
But that one can be replaced with any pointers' collection instead of strings. 
( if this is possible for such purpose ).

Piotras