Re: [gnome-db] GdaConnectionEvent and GdaDataModel issues

["Vivien Malerba" <vmalerba gmail com>]

On 3/14/07, Piotr Pokora <piotrek pokora gmail com> wrote:
> "Vivien Malerba" <vmalerba gmail com> wrote:
>
> > > Is there anything I do wrong here?
> >
> > No, it's correct. What each GdaConnectionEvent contains is decided by
> > each provider, so if you don't get what you expected, the providers
> > are the ones to blame, and of course it's possible to improve them and
> > make them set the correct properties in the GdaConnectionEvent.
>
> You mean improve in GdaConnectionEvent implementation?

No, the GdaConnectionEvent is just a GObject in which to store
specific information, it does not do any processing by itself, there
is not much to improve there.

> I just wonder if to make generic SQL query error handling with function connected
> to signal or in functions which are wrappers for all gda execute functions.
> The first one requires additional SQL query string available in user data pointer
> but nicely resolves problem of keeping error handling in one place.
> The second solution is "provider bug free" but... an so on ;)

In the current implementation, each provider is free to create any
GdaConnectionEvent it wants when it wants, and this way of working
will remain. However, most (if not all) the providers create
GdaConnectionEvent objects in one or two methods, so correcting them
is fairly easy.

> > >
> > > Also is bug for gda_data_model_get_column_index_by_name created or reported?
> > > I didn't check libgda issue tracker recently and I have no account there , but this function
> > > simply can not work.
> >
> > Not that I'm aware of. Could you provide some code to reproduce the
> > problem? Also you can fill a bugzilla report (I don't think you need
> > an account to do that).
>
> I mentioned about this in this thread:
> http://mail.gnome.org/archives/gnome-db-list/2007-March/msg00023.html
>
> And provided working code replacement in this mail:
> http://mail.gnome.org/archives/gnome-db-list/2007-March/msg00026.html
>
> I am not sure if this is a bug , but it looks like a bug.
> Anyway , this function can work as expected if you follow Your suggestion from this thread :)

I'm going to have a look at it, thanks.

Cheers

Vivien