[GOODIE] Connectors pre-release version (for 3.2)

Bruce Cohen brucecohen at qwest.net
Sun Mar 24 02:13:25 UTC 2002


Oh, oops, just discovered an oddity while playing with the grouped 
stick figure.  If I
grab the *end* of a NCConnector, and separate it from the circle it 
was attached to, I can't reattach it because it flies off to 
somewhere outside the group.  At a guess, the attachment is in some 
coordinate system very different from what it needs to be to stay 
near where it was before.  This isn't a problem with Connectors, I 
would guess, but may be Group behavior that has to be modified if I 
want to be able to treat a group as if it were a composite figure in 
a PasteupMorph.

Any suggestions on this one before I dive into debugging?

"The joke is over when the head falls off." - Scotts' proverb
=========
Bruce Cohen
brucecohen at qwest.net

At 6:04 PM -0800 3/23/02, Bruce Cohen wrote:
>Ned,
>
>I've gotten the latest Connector version working in 3.2gamma, and 
>it's rather nice.  In particular, you were interested in how we 
>would prefer to interact with the Morphs, so I played around with 
>that some, and learned that the behavior of parts of a connected 
>group of Connector morphs with respect to mouse interaction is 
>especially pleasant.
>
>I built a Connectors flap, using NCMakerButtons containing some 
>basic prototypes: a connector, an ellipse, a label, etc.  That works 
>fine. Using the buttons, I generated some parts and put together a 
>stick figure out of variously-sized circles for the head, hands, 
>feet, and joints, and connectors for the limbs.  Then I selected all 
>the parts of the figure and grouped them.  At that point, if I just 
>grab (left click) one of the circles (NCEllipse), I can move the 
>entire group, but if I try that with a limb (NCConnector), I add a 
>vertex and bend the line.  On the other hand, I can move an 
>individual circle with respect to the group just by left-clicking 
>until I have the halo for that morph alone, then click the move 
>halo.  So I can easily select whether I'm manipulating the 
>individual connector or the group, and the defaults are convenient.
>
>Good work.
>
>"The joke is over when the head falls off." - Scotts' proverb
>=========
>Bruce Cohen
>brucecohen at qwest.net
>>On Wednesday 20 March 2002 08:40 pm, I wrote:
>>
>>>  I've gotten Connectors to a point where I think it's generally useful, and
>>>  would like some testing before officially releasing it.
>>>
>>>  I have finally added a simple programmatic interface for querying
>>>  connections, which should make it easier to use for your own projects.
>>>
>>>  Also, I have changed the wiring to work with both drag'n'drop and
>>>  point'n'click modes.
>>>
>>>  The enclosed change set is intended to be loaded into a 3.2gamma image.
>>>  With 3.2 coming along Real Soon Now, I don't think I need to worry about
>>>  3.0 images.
>>>
>>>  It will attempt to convert existing connectors if you have any in your
>>>  image. It should also convert old connectors in projects when you load
>>>  them. Please tell me if it doesn't.
>>>
>>>  It will open up a couple of sample flaps with connectors and shapes.
>>>  However, the UML shapes there are more a demo than anything else. I would
>>>  hope that you would come up with your own flaps or ButtonBars with your own
>>>  favorite shapes and connectors.
>>>
>>>  The Connectors should work with any Morph.
>>>
>>>  There is still a lingering problem with the bounds of the open-arrowhead
>>>  flavor (in the flap as DirectionalAssociation) not being set right from
>>>  time to time; you will only see this if you try to do a drag selection. If
>>>  anyone has a fix for this please let me know.
>>>
>>>  As always, please don't try this in an image that you don't have backed up.
>>>
>>>  I'm looking forward to some feedback!
>>
>>This version takes care of some problems when loading into images with old
>>Connectors.
>>
>>Also, you can get an updated version of the Connectors demo (as a project) at
>>Bob's Super Swiki; it's called ConnectorsDemo2-nk.001.pr
>>
>>Thanks,
>>--
>>Ned Konz
>>currently: Stanwood, WA
>>email:     ned at bike-nomad.com
>>homepage:  http://bike-nomad.com
>
>
>--


-- 



More information about the Squeak-dev mailing list