Hola Andres!
-----Original Message----- From: Andres Valloud [mailto:sqrmax@prodigy.net]
[snip]
What do you do when developing?
Ok, and you load anything and everything but the SIFConfig, which depends on the Module/SemanticModel being around. Thanks for the pointer to the Grooming page.
The image is tight! It has lost the VMConstruction stuff as well as Balloon3D and much other stuff.
The idea behind this is that the VM construction stuff, and/or VMMaker, would be a module loadable upon request. Same with Balloon3D or anything else.
yep, and by extension to what other folks have been doing, the Networking, Morphic, MVC, etc. would all be external modules/components too.
[Bad] Missing SemanticModel. We like it, we love it, we want more of it! :)
Well... we'll have to ask Paul or Joseph :).
I have and they are working on it. :)
[Ugly] The UI is now outdated.
I guess that you could improve it. But so that you know, the fonts don't have any license (as opposed to Squeak), and the XML GUI builder should work both on Morphic and MVC.
Yep. The XMLUIBuilder is cool technology. However, a lot of work has occurred since the initial 3.1a image came out. The 3.1a environment looks great!!!!!
Do you have scripts that will allow us to take the latest and greatest image and strip it down to it's fur?
I don't think it's a good idea, anyway. I think it's much better to do it once (and only once), do it really well (no undeclareds, all source code can be recompiled without problems, no obsolete classes, etc), and from then on just put modules into the repository when you want new stuff.
I would disagree, although your next paragraph puts it in a different light. The fact is that there are 4 - 5 other people doing this same exercise right now. All it is is an evolving stripper script, including code to address Undeclared, Obsolete and so on. I'm sure you *want* people kicking the tires and checking the engine. This includes prepping a release image...
Incidentally, those scripts never existed. We had to do it manually for the most part, and it was a lot of work!
yes. Isn't a large part of that work, 'modularizing' the base classes to remove methods that are used by a stripped framework?
Over the next week, I will produce a version of my code that runs on your new events. I don't know the story with exceptions - sorry, dude. :)
later, Rob
Hola Rob :))).
Ok, and you load anything and everything but the SIFConfig, which depends on the Module/SemanticModel being around. Thanks for the pointer to the Grooming page.
Oh?... well, actually I tend to load just what I need. Leaving the rest out means I don't have to stumble on it while I am dealing with something else.
yep, and by extension to what other folks have been doing, the Networking, Morphic, MVC, etc. would all be external modules/components too.
Exactly.
The new environment looks great!!!!!!!
While it can be more enjoyable, I'd prefer more beautifully crafted code. Squeak has a long history of being a sugar coated turd, if you excuse my Spanish :).
I would disagree, although your next paragraph puts it in a different light. The fact is that there are 4 - 5 other people doing this same exercise right now. All it is is an evolving stripper script, including code to address Undeclared, Obsolete and so on. I'm sure you *want* people kicking the tires and checking the engine. This includes prepping a release image...
What can I tell you... I just enjoy Tim's messages about how frustrated he gets. For starters, he could have saved himself a lot of trouble by getting the SWT image and loading new stuff into it. But oh well :)))... and if it was hard to do with a 2.8 image, I don't want to know what happened after 1000+ updates!
yes. Isn't a large part of that work, 'modularizing' the base classes to remove methods that are used by a stripped framework?
Not only that... just removing the horribly tight references is just a mess. Believe me: it is not worth doing again. But oh well.
Andres.
RE: [ANN] SWT 0.6 releasedI've been playing with embedding morphs in a TextMorph. No matter how I embed the morph, by pasting or programmatically I experience some strange behavior when wrapping. To see this evaluate the following:
morphAsText := Text string: '*' attribute: (TextAnchor new anchoredMorph: (TextFieldMorph new contents: 'data here'; yourself)). text := (TextStream on: Text new) nextPutAll: 'This is the text before a morph: '; nextPutAll: morphAsText; nextPutAll: ' here is some text following the morph.'; contents. (PluggableTextMorph on: nil text: nil accept: nil) setText: text; openInWindow
Resize so everything is on one line then slowly narrow the window until the embedded TextFieldMorph just fits on the first line. So far all should be okay. Narrow the window so the embedded morph just doesn't fit on the first line. Notice how TextFieldMorph wraps but then indents itself by its own width!
To get a hint of what may be going on type a character just prior to the TextFieldMorph making sure there is no space between them and repeat the resizing exercise. Seems if the $* which carries the TextAnchor attributes is at a line wrap point the layout mechanisms are getting confused.
My image is #4282 with the HTML Table changes set 0.6 installed.
Joerg
squeak-dev@lists.squeakfoundation.org