Re: [Planner Dev] Fix for patch for the Link button patch against 0.11 CVS.

["lincoln phipps openmutual net" <lincoln phipps openmutual net>]

Attached is a new patch - this maintains the visual
orientation by using the ... mrp_task_get_position ()
to make sure that the order is maintained.

Can this be checked ? It works OK on my local copy in
that I now correctly see the links being applied
top-down.

I didn't know about mrp_task_get_position(task) but now
I do ;) Quite obvious now that I look at it !.

Rgds,
Lincoln.


lincoln phipps openmutual net wrote:

> I did pkg-config --modversion gtk+  and got 1.2.10
> Is that the right shell command to use ?.
>
> Yes it looks like what I build ends up sorted
> differently from what you build from CVS.
>
> I get a clean compile and nothing is squeeking as
> to wrong version so this anomaly needs to be corrected
> in code.
>
> Certainly Task1 -> Task2 ->Task3 is what we want by
> default. We dont have a row-id ordinal of some kind to
> verify the sort order and then choose to run up or down
> the list i.e. some number that runs 1,2,3,4,5,6 based
> on the position that the gantt is displayed but can we
> compare GtkTreeIter.stamp ?
>
> Rgds,
> Lincoln
>
> Richard Hult wrote:
>
> > mån 2004-02-16 klockan 15.21 skrev lincoln phipps openmutual net:
> >
> > > Richard,
> > >     I just checked out the latest CVS copy (from anon so
> > > I had to wait a bit as the CVS server was playing up and out
> > > of sync), and the link button is around the wrong way from
> > > how my original patch did it i.e. the expected link isn't
> > > top-down visually but bottom up visually.
> > >
> > > Attached is a patch to the current CVS to fix this orientation ;)
> >
> >
> >
> > Hm? Here, CVS does the following:
> >
> > Task 1
> > Task 2
> > Task 3
> >
> > select all and link:
> >
> > Task 1 -> Task 2 -> Task 3
> >
> > Which should be the expected behavior I'm sure?
> >
> > If that's not what you are seeing, there might be a difference in how
> > the selection is returned in different versions of GTK+ (I'm using
> > 2.3.x) , and we might need to make sure that the list is sorted
> > according to the visual order.
> >
> > /Richard
> _______________________________________________
> Planner-dev mailing list
> Planner-dev lists imendio com
> http://lists.imendio.com/mailman/listinfo/planner-dev
RCS file: /cvs/gnome/planner/src/planner-task-tree.c,v
retrieving revision 1.10
diff -u -b -B -p -r1.10 planner-task-tree.c
--- src/planner-task-tree.c	14 Feb 2004 02:21:21 -0000	1.10
+++ src/planner-task-tree.c	16 Feb 2004 23:30:38 -0000
@@ -2098,6 +2098,7 @@ planner_task_tree_link_tasks (PlannerTas
 
 		task = l->data;
 
+		if (mrp_task_get_position(target_task) > mrp_task_get_position(task)) {
 		if (!mrp_task_add_predecessor (target_task,
 					       task,
 					       relationship,
@@ -2112,6 +2113,23 @@ planner_task_tree_link_tasks (PlannerTas
 			gtk_widget_destroy (dialog);
 			g_error_free (error);
 		}
+		}
+		else { /* The other orientation: Different Gtk sort selections different */
+		if (!mrp_task_add_predecessor (task,
+						target_task,
+					       relationship,
+					       0,
+					       &error)) {
+			dialog = gtk_message_dialog_new (NULL,
+							 GTK_DIALOG_DESTROY_WITH_PARENT,
+							 GTK_MESSAGE_ERROR,
+							 GTK_BUTTONS_OK,
+							 "%s", error->message);
+			gtk_dialog_run (GTK_DIALOG (dialog));
+			gtk_widget_destroy (dialog);
+			g_error_free (error);
+			}
+		}
 		
 		target_task = task;
 	}