[Gambas-devel] Perfomance differences between gb.gtk and gb.qt components

[Benoit Minisini]

On lundi 19 novembre 2007, José Luis Redrejo wrote:
> Hi, I'm doing some tests with a pretty complex application made in gambas.
> It has quite a bunch of classes and about one thousand lines of code.
> I was surprised when I checked the just starting the application and let it
> show the graphical interface, without interacting with it, took a lot of
> cpu use in a quite powerful computer (checked with the gnome-system monitor
> and later with top), but I was much more surprised when changing from using
> gb.gtk to gb.qt the cpu use was as little as it can not be appreciate, so
> it was around 0.
> This can be easily appreciated just creating a clean new project with a
> form, a couple of buttons and a textbox, without adding any code. Running
> it with gb.qt, top only shows gbx2 comsumption at the starting and very
> fast it disappears from the view, running it using gb.gtk the application
> stays on top of the processes all the time, with about a 4% of CPU use.
>
> So, it seems there is some problem with the event loop in gb.gtk component.
> I think this is a very important bug in order to use gambas under Gnome, as
> many users do.
>
> Regards.

Maybe Daniel could give us some details about the gb.gtk event loop.

I noticed that my GTK+ programs were always far slower than my QT programs, 
using Gambas or not, so I didn't notice that specific event loop problem.

Regards,

-- 
Benoit Minisini