gtg-contributors team mailing list archive

[Luca Invernizzi <invernizzi.l@xxxxxxxxx>]

On Fri, Sep 10, 2010 at 12:06:13PM +0200, Lionel Dricot wrote:
> The performance test is mainly affected by the TreeTest class. We should
> either make that TreeTest class seamless or make a test that doesn't use
> it.
> I'm interested in any input regarding performances.
I've made test_speed disable the TreeTest. It does perform way better, as it
takes ~1sec on my machine. I believe we still have a bug, as the time
takes still scales exponentially with respect to the number of nodes.
100 nodes => ~1 sec
150 nodes => ~8 secs
200 nodes => ~40 secs
However, this is not release-critical anymore.

> 
> 
> > - gtk errors causing segfaults (sporadic)
> 
> That's here that anyone could help : find a way to reproduce such errors.
> If we can reproduce it, we could probably write a test and then fix the
> error.
> 
> > - in the tags tree, when adding a tag all the partial tags are
> > displayed(@t, @ta, @tag)
> 
> That should be fixed in rev 892 and there's two tests for that (one in
> liblarch testing, the other in tagstore testing)
It works. Thanks, Lionel!
> 
> 
> Another error that I want to see fixed before implementing drag-n-drop is
> the "child" status in the gtk.TreeView. Namely :
> 
> - Tasks with children doesn't have the + to view children after a refresh
> unless you open the task
Bug opened
> - I didn't find anyway to expand the children by default.
mh, gtk.Treeview().expand_all() doesn't work with liblarch as treemodel. I
wonder if it's a bug.
> 
> 
> Lionel

Attachment: signature.asc
Description: Digital signature