323683 - Date picker doesn't handle refresh correctly

Status: RESOLVED FIXED

(Core Graveyard :: XForms, defect)

Description

[aaronr]

inputs bound to dates don't handle refresh correctly.  They don't ever show their initial values and other date pickers bound to the same changed node won't show the changed value.  Testcase to follow.

Comment 1

[aaronr]

There is also another problem.  If @incremental="true", then datepickers bound to the same node should change at the same time, without the user moving focus away from the input field.

Comment 2

[aaronr]

Attachment: testcase

Comment 3

[Doron Rosenberg (IBM)]

(In reply to comment #1)
> There is also another problem.  If @incremental="true", then datepickers bound
> to the same node should change at the same time, without the user moving focus
> away from the input field.
> 

Is this all fixed by removing the refresh method from the xbl?

Comment 4

[aaronr]

(In reply to comment #3)
> (In reply to comment #1)
> > There is also another problem.  If @incremental="true", then datepickers bound
> > to the same node should change at the same time, without the user moving focus
> > away from the input field.
> > 
> 
> Is this all fixed by removing the refresh method from the xbl?
> 

the @incremental bug isn't fixed by removing date picker's "refresh" method.  But removing "refresh" (and thus using input's default refresh) does fix the initialization problem and the update problem.

Comment 5

[Doron Rosenberg (IBM)]

Attachment: patch

First, remove the refresh method that was doing a no-op to fix the instance loading issue.

Secondly, fix a missing bracket bug.

Thirdly, add some _change() calls when the input field changes.

Comment 6

[aaronr]

Comment on attachment 208739 [details] [diff] [review]
patch

In reviewing your patch I also saw that readonly won't work right for date picker.  I'll open another bug for that, though.

Comment 7

[Allan Beaufour]

(In reply to comment #5)
> Created an attachment (id=208739) [edit]
> patch
> 
> First, remove the refresh method that was doing a no-op to fix the instance
> loading issue.

I wanted to check this in, but bug 323829 actually ads functionality to refresh(). What was the issue exactly? The testcase seems to work fine for me with the patch applied (except for the killing of refresh).

Could you check it in Doron, if all is well, etc.?

Comment 8

[Doron Rosenberg (IBM)]

checked into trunk