|Anonymous | Login||03-29-2020 22:34 UTC|
|Main | My View | View Issues | Change Log | Docs|
|Viewing Issue Simple Details [ Jump to Notes ]||[ View Advanced ] [ Issue History ] [ Print ]|
|ID||Category||Severity||Reproducibility||Date Submitted||Last Update|
|0003760||[Squeak] Graphics||major||always||06-02-06 03:41||07-10-06 02:12|
|Summary||0003760: In 7033 (and 6665) Thick Ellipse Borders incorrectly drawn when ellipse has gradient fill|
For a dramatic view of this one
From object get an ellipse using the red menu.
Make the border with very large.
Make the ellipse large enough you can see the center of it.
(Notice that the border and the ellipse are both elliptical)
Request a gradient fill.
Notice the change in the shape of the ellipse.
I have uploaded
which contains snapshots of an ellipse
with and without thick borders and gradient fills.
The problem lies in the border drawing routines and probably somewhere in the assumptions that balloon engine is making.
This is related to 0003736 and 0000564
4xSameCircle.gif [^] (13,111 bytes) 06-02-06 03:41
EllipseBorders1.gif [^] (23,669 bytes) 06-11-06 02:58
(0005218 - 966 - 1050 - 1050 - 1050 - 1050 - 1050)
And more analysis.
From overlaying various combinations of solid and gradient filled ellipses with thick translucent or transparent borders here is what I've deduced is going on:
With solid filled ellipses, half the boarder width is within the bounds and half is outside the bounds. The inner outer and bounds are ellipses similar in shape and concentric with the solid fill ellipse.
With a gradient fill ellipse the border has the same outer bounds. It looks like it is drawn by using the original ellipse as a penpoint and moving it around a rectange inset from the outer bounds by half the border width.
With thick borders, the intersection of the four borders is where the original ellipse shows thru.
Transparent borders do not get drawn and show the morph as it would be without a border.
There is probably a difference with outer bounds when a transparent border is used rather than a zero width border.
(0005220 - 635 - 689 - 689 - 801 - 801 - 801)
This bug is also related to the previously reported 0000858 and are both probably caused by the same problem.
In Sq3.8-6665Full with has the vmMaker stuff in it I think I tracked things down to BalloonEnginePlugin>>primAddOval.
plus a truncation problem in the use of #loadPoint:from:
this will sometime clip the left and bottom edge in the wrong direction giving rise to the trailing glitches and artifacts.
It would be good to use an alternate #loadFlooredPoint:from: routine at the very least.
Or better write the routines to deal with a few more decimal (binimal? ) places of accuracy. This, I admit, would be a lot of work
(0005659 - 119 - 119 - 119 - 119 - 119 - 119)
|Any proposals for how to fix this before 3.9 goes out? Or should we just confirm the bug and leave for some other time?|
(0005700 - 365 - 407 - 407 - 407 - 407 - 407)
This one is deep in places I haven't probe yet.
I believe the actual fix will have to be done in the Balloon Plugin.
I haven't fiqured out a workaround for 3.9 either .
So if you want to acknowledge it, thats cool.
Also if you have any clues as to how to fix this that would be good add here. Something might spark a fix.
Yours in service, --Jerome Peace
|06-02-06 03:41||wiz||New Issue|
|06-02-06 03:41||wiz||Status||new => assigned|
|06-02-06 03:41||wiz||Assigned To||=> andreas|
|06-02-06 03:41||wiz||File Added: 4xSameCircle.gif|
|06-11-06 02:58||wiz||File Added: EllipseBorders1.gif|
|06-11-06 03:15||wiz||Note Added: 0005218|
|06-11-06 05:33||wiz||Note Added: 0005220|
|07-09-06 06:54||andreas||Note Added: 0005659|
|07-09-06 06:54||andreas||Status||assigned => feedback|
|07-10-06 02:12||wiz||Note Added: 0005700|
| Mantis 1.0.8[^]
Copyright © 2000 - 2007 Mantis Group
54 total queries executed.|
38 unique queries executed.