Can we* prettyplease do something about the KUrlNavigator?

Thomas L├╝bking thomas.luebking at web.de
Fri Jun 19 23:12:55 BST 2009


Am Friday 19 June 2009 schrieb Matthew Woehlke:
> To be clear, the point is I don't see this as useful unless I am able to
> pre-position the caret with a single click. (And even then, not sure how
> selection would work.)
Ok. The only problem i see on this is to have the desired position in sight.
In other words: The "partial" path feature of the UrlNavigator.
Well, it's feature and you can trigger it off (and actually i have as i'm not 
a fan of siplification or lying when it comes to urls, but that's OT) but it 
would indeed be great to get even this together...

> The ideal would be to be able to switch from breadcrumb to edit mode
> without the text moving...
Because the current mouseposition mismatches the former clicktarget...

Do you know the saying about the prophet and the mountain? =D

One could simply on (modified...!) click:
- calculate the proper text index to move the caret to (no big deal)
- show up the lineedit
- set the caret and...
- warp the cursor to that position

As you get the important thing when warping (all trekkies, su... eat it: the 
flash was a helpout to hide the model change... ;-P hah.) - that is a eye 
catching transition when the lineedit content changes from crumbs to text - 
and the warp will tend to be short - this should not be any irritating.

> but don't know how you would manage that. Or
> alternatively the control would have to switch on mouseover, but even
> that isn't ideal.
esp. as i don't know if Qt posts enter updates if you press a modifier (to 
avoid having the modifier pressed before you entered the widget)

One more thing:
there's currently pretty much emtpy space between the items.
Assuming that you in manymost cases want to paste to the end or in delimiter 
fractions, one could spare 4px or so between the triangle and the crumb for 
the I-beam thus "empty space" (i.e. easy and unmodiefied lineedit) behavior.

Thomas




More information about the kde-core-devel mailing list