Found Bug But Not Sure How to Proceed To Fix

Attached is a hack to keep rotated rectangles (well, most anything rotated, but rectangles are were it really shows) all aligning as one might expect them. It basically assumes that the bounds have been expanded by 1 and scoots the drawing back to the top left when rotated by 90, 18 or 270. See if you like it.

On 3/2/18 8:59 PM, Jeremy Landry wrote:

So far, with a little testing, everything lines up much better with .1 instead of removing it.  removing that offset still sees to make it offset, but .1 lines everything up visually more...still not perfect, but much less flagrant.  It still has a clipping glitch (you can see it with a rectangle when rotating: the border gets clipped off) but it does visually appear more 'solid' and expected when aligning rectangular shapes together.

I'm going to keep experimenting to find something that keeps the rotated bitmap in tact while still aligning properly.  I feel like I probably glanced at a method that needs an edit over and over but haven't realized which one it is yet. I still need to look into the cursor not being properly redrawn as well.  I might be stretching or trying to make a connection that isn't there, but something 'feels' related to this and the cursor, like some common reference point is being shifted so that the cursor's area for drawing doesn't reach the left and bottom edges just like the origin doesn't reach the left and bottom edges correctly for the this rotation issue.

On Fri, Mar 2, 2018 at 5:37 PM, David T. Lewis <[hidden email]> wrote:

Does anything bad happen if the "expandBy: 1" is removed? I tried removing it
in my Squeak trunk image, and "CompassDialMorph new openInWorld" (which is one
user of TransformationMorph) seems to work nicely.

And if the "expandBy: 1" is removed in an Etoys image, does anything bad happen?
Does it fix the issue that Jeremy is seeing?

This would be an easy thing to fix in Squeak trunk, which hopefully will become
part of future Etoys releases, knock wood.

Dave


On Fri, Mar 02, 2018 at 03:19:54PM -0500, Bob Arning wrote:
> TransformationMorph>>computeBounds
> ?????? self hasSubmorphs ifTrue:
> ?????? ?????? [bounds _ (transform localBoundsToGlobal:
> ?????? ?????? ?????? ?????? ?????? (Rectangle merging:
> ?????? ?????? ?????? ?????? ?????? ?????? (self submorphs collect: [:m | m
> fullBounds]))) truncated
> ?????? ?????? ?????? ?????? expandBy: 1].
> ?????? fullBounds _ bounds.
>
> Note the expandBy: 1
>
>
> On 3/2/18 1:59 PM, Jeremy Landry wrote:
> >Actually to add a little more insight if possible, you can create bug
> >with the following actions:
> >
> >make a new playfield
> >make 2 new rectangles and place them into the playfield.
> >Rotate one of them 90 degrees.
> >Turn gridding on in the playfield.
> >
> >The rotated and unscaled rectangle will not align with the non-rotated
> >one with gridding.?? Something about making a bitmap copy is offsetting
> >the values, my guess is there's a rounding problem that is popping up
> >on these objects, thus scale and rotation get offset for some reason,
> >and specifically changes the origin by 1@-1 somewhere...it causes
> >graphical glitches leaving 'garbage' with playfield indicators when
> >the object that has the indicator is moved and also shows up with
> >rotated ellipses.?? What is the common thread through them all??? What
> >method are they all accessing?
> >
> >On Fri, Mar 2, 2018 at 10:45 AM, Jeremy Landry <[hidden email]
> ><mailto:[hidden email]>> wrote:
> >
> >    You will also notice if you rotate any shape, it becomes a bit
> >    map.?? Scale it and you will see that problem re-occur even with
> >    built-in objects.?? Could it be the masking offset, too??? that
> >    causes this??? I'm still investigating, but I'm not sure how to
> >    step through 'already working' things specifically related to what
> >    I'm trying to look at.
> >
> >    On Fri, Mar 2, 2018 at 10:42 AM, Jeremy Landry <[hidden email]
> >    <mailto:[hidden email]>> wrote:
> >
> >        Hi, it only appears when scaling a morph.?? Here is
> >        example...it is my belief (how does one step through code with
> >        debugger if there's no bug happening?);
> >
> >        Whatever code is common that scales morphs is being offset
> >        incorrectly. The first image shows the 'garbage' left behind
> >        by playfield indicator if that option is set.?? It seems it was
> >        only tested in 'holders' and thus likely not picked up during
> >        'intended' use.
> >
> >        Next, two pictures show scaling as where the misalignment
> >        starts.?? This might be why you could not recreate bug;
> >        resizing rectangle changes rectangle morph extent, whereas
> >        scaling is a differnet piece of code and therefore different
> >        outcome.
> >
> >        Image included.?? This works with *anything* scaled.
> >        https://imgur.com/a/6zEWq
> >
> >        On Fri, Mar 2, 2018 at 10:20 AM, Bob Arning
> >        <[hidden email] <mailto:[hidden email]>> wrote:
> >
> >            I tried to recreate your example and could not. How did
> >            you create the two morphs in your example? Are you certain
> >            the read/gray morph does not have some transparent pixels
> >            along the top and left?
> >
> >
> >            On 3/2/18 12:50 PM, Jeremy Landry wrote:
> >>            I did a little more digging, just for the record...and it
> >>            turns out it's scaling error.?? No matter what, scaling
> >>            offsets morphs by 1@-1.?? I'm still looking for where this
> >>            occurs and why.?? So far, injecting intentionally bad
> >>            values into bordered morph and sketch classes hasn't
> >>            produced any change.?? It 'feels' like most of that code
> >>            isn't even used, but then again, I might not have
> >>            genuinely created a new object with that code and system
> >>            might have been giving me what was already in system...
> >>
> >>
> >>            On Fri, Mar 2, 2018 at 7:15 AM, Bert Freudenberg
> >>            <[hidden email] <mailto:[hidden email]>> wrote:
> >>
> >>                That bug tracker is pretty much dead, since all
> >>                ongoing development moved to squeak.org
> >>                <http://squeak.org>.
> >>
> >>                The plan was to make a new Etoys release based on the
> >>                latest Squeak source base, but we have not have
> >>                enough developer time to make that actually happen.
> >>
> >>                - Bert -
> >>
> >>                On 1 March 2018 at 06:03, Nicco Kunzmann (rambler)
> >>                <[hidden email]

> >>                <mailto:[hidden email]>> wrote:
> >>
> >>                    Hi,
> >>
> >>                    I found a bug tracker on the squeakland site:
> >>                    http://tracker.squeakland.org/secure/Dashboard.jspa
> >>                    <http://tracker.squeakland.org/secure/Dashboard.jspa>
> >>
> >>                    I think, you can describe the problem there as an
> >>                    issue and hand in the fix.
> >>                    They may also know how to guide you further if
> >>                    the code base is somewhere else.
> >>                    If you have problems there (login, no response),
> >>                    I guess, this mailing list is one place to ask
> >>                    for help.
> >>
> >>                    Best,
> >>                    Nicco
> >>
> >>
> >>                    On 03/01/2018 03:34 AM, Jeremy Landry wrote:
> >>>                    Hi, while working on a project needing precise
> >>>                    alignment of sketches/bitmaps and rectangles
> >>>                    using playfield 'use gridding', it seems that
> >>>                    there's a difference of detected origin.
> >>>
> >>>                    Here's a screenshot with magnifier showing the
> >>>                    discrepency.
> >>>                    https://imgur.com/a/tcaRu
> >>>
> >>>                    I'm quite certain this bug is the one that
> >>>                    causes bad redrawing, especially when showing a
> >>>                    cursor move when 'indicate cursor' is activated
> >>>                    in a playfield.
> >>>
> >>>                    If I correct this bug inside of a project, will
> >>>                    it load into an image where it has not been
> >>>                    fixed and be fixed?
> >>>
> >>>                    I haven't actually fixed the bug yet, but didn't
> >>>                    want to waste time on it if the fix will not
> >>>                    travel with the project.
> >>>
> >>>                    Thanks in advance for any input.
> >>>
> >>>
> >>>                    _______________________________________________
> >>>                    squeakland mailing list
> >>>                    [hidden email]

> >>>                    <mailto:[hidden email]>
> >>>                    http://lists.squeakland.org/mailman/listinfo/squeakland
> >>>                    <http://lists.squeakland.org/mailman/listinfo/squeakland>
> >>
> >>
> >>                    _______________________________________________
> >>                    squeakland mailing list
> >>                    [hidden email]
> >>                    <mailto:[hidden email]>
> >>                    http://lists.squeakland.org/mailman/listinfo/squeakland
> >>                    <http://lists.squeakland.org/mailman/listinfo/squeakland>
> >>
> >>
> >>
> >>                _______________________________________________
> >>                squeakland mailing list
> >>                [hidden email]
> >>                <mailto:[hidden email]>
> >>                http://lists.squeakland.org/mailman/listinfo/squeakland
> >>                <http://lists.squeakland.org/mailman/listinfo/squeakland>
> >>
> >>
> >>
> >>
> >>            _______________________________________________
> >>            squeakland mailing list
> >>            [hidden email]
> >>            <mailto:[hidden email]>
> >>            http://lists.squeakland.org/mailman/listinfo/squeakland
> >>            <http://lists.squeakland.org/mailman/listinfo/squeakland>
> >
> >
> >            _______________________________________________
> >            squeakland mailing list
> >            [hidden email]
> >            <mailto:[hidden email]>
> >            http://lists.squeakland.org/mailman/listinfo/squeakland

> >            <http://lists.squeakland.org/mailman/listinfo/squeakland>
> >
> >
> >
> >
>

> _______________________________________________
> squeakland mailing list
> [hidden email]
> http://lists.squeakland.org/mailman/listinfo/squeakland

_______________________________________________
squeakland mailing list
[hidden email]
http://lists.squeakland.org/mailman/listinfo/squeakland

_______________________________________________
squeakland mailing list
[hidden email]
http://lists.squeakland.org/mailman/listinfo/squeakland

_______________________________________________
squeakland mailing list
[hidden email]
http://lists.squeakland.org/mailman/listinfo/squeakland

Which option is it that makes this appear?

On 3/2/18 1:42 PM, Jeremy Landry wrote:

Whatever code is common that scales morphs is being offset incorrectly. The first image shows the 'garbage' left behind by playfield indicator if that option is set.  It seems it was only tested in 'holders' and thus likely not picked up during 'intended' use.