[BUG] Bounds of rotated Morphs are incorrect inside GeeMailMorphs

Bert Freudenberg bert at isg.cs.uni-magdeburg.de
Tue Mar 6 08:48:19 UTC 2001


On Mon, 5 Mar 2001, Ned Konz wrote:

> This isn't just a halo problem; if you ask the rectangle its bounds or 
> center, these will also be out of place.

I think this is the same bug I reported a while ago
(http://swiki.gsug.org/sqbugs/96.html). 

The real problem IMHO is the use of absolute coordinates in Morphic. The
transforms in a scroll pane violate this (stupid) principle by changing
the coordinate system, and some Morphic code can't deal with the
situation. For example, when a morph accepts a dropped morph the
coordinates would need to be transformed into the accepting morph's
coordinate space. Actually all querying of the hand position should apply
the morph's transform, but this is only done at a few places - which works
fine only as long as the morph is not transformed. 

Even more wierd stuff happens if you want to build a real magnifier - just
put a system window inside a transform that scales everything up by two.
It basically works, but I managed to crash the VM this way :(

-- Bert





More information about the Squeak-dev mailing list