Doug Way <dway at riskmetrics.com> wrote:
> On Saturday, August 9, 2003, at 09:34 PM, Daniel Vainsencher wrote:
> > Doh. Difference between reviewing and testing. Deleted the width: line
> > in my image, and all is good. The update effect is prettier.
> Not sure what you mean by the "update effect".  (Just curious)

What I meant is that when you select an item, between mouse down and
mouse up, the item gets a border. Before the border used to be two
horizontal lines, since the morph didn't end in sight.

> Also, another bonus is that the StringMorph halo doesn't extend all the 
> way to the right side of the screen, which is really annoying... you'd 
> have to drag your mouse way over to debug/inspect the morph.  Given 
> that, I'd say the halo behavior is vastly improved overall with these 
> two changesets.  (I would often need to inspect StringMorphs when 
> working on Whisker... oh wait, those were IndentingListItemMorphs. :-) )
I have nothing to do with that in particular, that's Eddie's idea and
implementation (though I'm including it in my changesets). You get to
see the point of my changes only if you either patch MessageSet (or some
other list browser) to use a factory, or read the code changes. It has
no effect on default behavior.

This should also allow trivial implementation of many other gui tricks,
like icons in lists, stuff like the CDT package does, and so forth.

> > Anyone played with the 2nd version? any comments?
> Seems to work fine so far.
That's something :-)

> Now we just need to make a similar fix for IndentingListItemMorphs, 
> which are used in the SimpleHierarchicalListMorphs in the file list, 
> object explorer, etc. :-)  (see 
> SimpleHierarchicalListMorph>>adjustSubmorphPositions)
I'm staying away from SHLMs for now, I want to finish the closer PLM
subclasses first.

Daniel