Squeak-dev April 2000

squeak-dev@lists.squeakfoundation.org

141 participants
506 discussions

RE: World sensitivity (was Re: [updates] 35 more for 2.8alpha)
by Raab, Andreas 30 Apr '00

30 Apr '00

> I not sure what you you mean by "originates" here. There are > quite a few places in the current image that say something like: > > SomeMorph new openInWorld > > What Morph is this originating from? Take the case of the > Compiler doing a fileIn and encountering a syntax error. It > currently opens a SyntaxError appropriate to the Morphic/MVC > environment in which it finds itself. Are you suggesting that > the Compiler be given a reference to the world prior to > starting the fileIn? Actually yes. At least there needs to be some context in which the action is executed. E.g., assume you want to have a remote 'error world' into which errors get logged so that some (remote) user can examine these. This is *not* your current World and while it seems reasonable to open up the error in the current world I can think of all sorts of cases where you really want to open up something in a world which is not current. And that does require you to establish the context and cleans up the default 'currentWorld' use. > Suppose I want my email client to tell > me when mail is received. Where does it perform this > notification? My preference would be that it notify me in > whatever world is currently active, but how could it know > that beforehand? Good point. I think I was more arguing against the use of World for designated receivers (e.g., everything that doesn't quite work in Telemorphic) as well as the 'oh-so-convient' shortcut for determining what GUI is currently primary. E.g., most methods I've seen use World along the lines of World ifNil:[self doSomeMVCStuff]. "We can't do that in MVC" which may be not quite accurate since there is quite a chance of having a Morphic window in MVC open. In particular I don't like those SomeClass>>openInWorld World ifNil:[self openInMVC] ... (or vice versa) methods. > I can imagine Morphs that may want to delete themselves from > the world and then reappear later. Not in their own world, > but in whatever world the user is looking at. Additionally, a > Morph might want to adapt itself in some way to the world > *before* it is opened in that world. Oh but that's okay - it can always do this through #openInWorld: aWorld which should be the message sent once the world has been determined (by whatever means). - Andreas

1 0

RE: RE: World sensitivity (was Re: [updates] 35 more for 2.8alpha)
by Raab, Andreas 30 Apr '00

30 Apr '00

> On Sun, 30 Apr 2000 20:13:23 +0200 Stefan Matthias Aust > <sma(a)3plus4.de> wrote: > >Actually, I think it isn't needed. It's a handy reference but every > >morph can be asked for its world by using #world already. > > Assuming that morph is in the world already. Every Morph originates (one way or another) from some other Morph so there's always some Morph you can ask about it's world. Except from the first one (e.g., when you create a Morphic project or a Morphic window in MVC) in which case the world needs to be explicitly created anyways (and notice that World>>owner == nil, so that myWorld openInWorld: nil should be just fine). Actually I think that *any* Morph with owner == nil should behave as if it defines its own world. > That's a good idea, but it still leaves the question of where > those few methods will find the world. Object>>currentWorld starts out: > > World ifNotNil: [^ World]. Object>>currentWorld is nothing but a convenient shortcut. Those methods should get passed the world explicitly (or the classes should just have an instVar currentWorld if passing the world is to inconvenient otherwise). - Andreas

1 0

RE: World sensitivity (was Re: [updates] 35 more for 2.8alpha)
by Pennell, David 30 Apr '00

30 Apr '00

So what is the global World good for? Do we really need it? -david

2 1

Re: help! hosed precious image: rescuable ?
by squeak＠sysa.barnett.sk 30 Apr '00

30 Apr '00

On Sun, Apr 30, 2000 at 06:25:20PM +0200, Torge Husfeldt wrote: > Hi Michal, > First of all, make a backup of anything left (.image .changes files). > (No real data recovery specialist works with originals). > Have you thougt about reading the image into a InterpreterSimulator and > catching the exception that would surely be raised, to then inspect the > 'lost' image?! I haven't worked much with that thing myself but I know that > much: there are two versions and you have to work with the one matching the > endianness (Smalltalk endianness<Cmd-P>) of your machine (the attached cs > should make the right choice for you). > Then, if you have an intact .changes file (and your notes weren't hidden in > Workspaces) you could use the famous codeBrowser to browse it. > > hope this helps, > Torge > ----- Original Message ----- > From: michal starke <miso.fastlist(a)auf.net> > To: <squeak(a)cs.uiuc.edu> > Sent: Sunday, April 30, 2000 11:50 AM > Subject: help! hosed precious image: rescuable ? I had some like that corrupted image, when i touch my HD cables in phase of "Smalltalk condenseChanges" and machine was frozen. Maybe it is not the correct way, but worked fine for me : I first removed all starting huge noninteresting lines of code up to found interested start point of my code lines and found start signature of ----SNAPSHOT---- . Then I saved file to some new named rescue.st and than i file in that code to found previous operatin image... so try also this with copy of your changes and new image ... Jan Barger New Message Board + Project for collecting web links -> http://www.barnett.sk "The best way to predict the future is to invent it." - Alan Kay "The mind can only see what it is prepared to see. " - Edward de Bono

1 0

Porting Java To Smalltalk?
by Conrad Taylor 30 Apr '00

30 Apr '00

Hi, I would like to port Java code to Smalltalk. Thus, is there any references for performing such an action? I quess my main concern would be Java's packages (i.e. their name spaces). Also, most of the classes in Java that I need to convert follow the following pattern: public class className1 extends className0 implements className2 { ... } If anyone has any references and/or examples, please let me know. BTW, I'm using Squeak Smalltalk 2.8a. Thanks, -Conrad __________________________________________________ Do You Yahoo!? Talk to your friends online and get email alerts with Yahoo! Messenger. http://im.yahoo.com/

2 1

[ENH] Method Finder with matching classes marked
by Doug Way 30 Apr '00

30 Apr '00

On Tue, 18 Apr 2000, Michael_Chean wrote: > > Perhaps the original poster _did_ try "reverse<Cmd-Shift-W>" and was > > confused by the fact that "FlippyArray2" does contain a 'reverse' method, > > as do quite a number of classes, sometimes meaning "reverse order" and > > sometimes meaning "opposite colour". > > Yep, actually this is what happened. It was more confusing in that I > was looking for something to do with the class String - so from my > beginners point of view the reversed in OrderedCollection wasn't an > immediate match. Thank goodness that the OrederedCollection actually > had an example of reversed which clued me into the similiarity. > Otherwise the Newbie must try to decipher from the description or code > what is going on. Here is an enhancement to the Method Finder which will hopefully help to narrow down the appropriate class in these situations. (I realize that you were not using Method Finder here, but if you were, this might have helped.) This is an enhancement to the "example mode" of the Method Finder. When an example is given, and one of the found methods is selected from the list, the classes in the class list which match the class (or are a superclass) of the receiver will be specially marked with an asterisk. For example, if the example 'abcd'. 'dcba' is given in the top pane, the method 'reversed' will appear in the method list. When 'reversed' is selected, eight classes will appear in the class list, but now SequenceableCollection>>reversed will be asterisked, since that is the class/method that is actually being used in this example. (because String is a subclass of SequenceableCollection) Generally, the asterisked class is really the one we're looking for... the other seven classes are basically irrelevant to the example, although it's useful to at least leave them in the list. The asterisk appears before the class/method name, so the asterisked classes end up being moved to the top of the list, since they're sorted alphabetically. I think this is handy, anyway. (My original plan was to make the matching classes appear in bold instead of using an asterisk, but that looked like it would be a lot harder to implement.) Here are some other fun examples to try when testing this enhancement: 'abc'. 'ab'. true 'abcd'. $c. 3 1. 2. true #(10 20 30 40). 20. 50. #(10 50 30 40) #(1 2 3) asOrderedCollection. 4. 4 Date. Date today Of course, no classes will be asterisked if you use the "selector fragment" mode. I didn't try to implement it for the "multi-example" mode which appears in MethodFinder>>methodFor:, either. Also, I suppose it might be best to only asterisk the "nearest" superclass and not all superclasses, so that only one class would ever be asterisked. But that looked a little bit hairy to implement, so I kept it simple for now. The attached changeset just makes a couple of changes to SelectorBrowser... mainly, SelectorBrowser>>markMatchingClasses was added, plus selectedClass was tweaked to handle parsing the extra asterisk. (I guess I just missed the last batch of updates for 2.8, if this enhancement is deemed worthy of inclusion... keep it in mind the next time around. :-) ) Regarding the Method Finder, it is true that not very many newbies think to try the Method Finder... it definitely needs to appear in more Squeak tutorials. (Naturally, it isn't going to be in the old Smalltalk books, since it's a new thing.) Also, I think the explanatory text in the lower pane could be improved... the part about the example mode should give a couple of typical scenarios, like a mini-tutorial... I remember I missed the point the first time I read through it. Maybe I'll submit something. - Doug Way EAI/Transom Technologies, Ann Arbor, MI http://www.transom.com dway(a)mat.net, @eai.com

1 1

Re: [BUG] Halo skips SystemWindow in Morphic (2.8alpha)
by Scott Wallace 30 Apr '00

30 Apr '00

At 8:23 PM -0400 4/17/00, Doug Way wrote: >I meant to post about this earlier, as a possible bug in 2.8alpha. > >For some reason, the SystemWindow morph is always being skipped when you >cycle through the various submorph halos. I don't think this was >intended... I'm guessing this is a bug. Actually, it *was* intended. I'm sure that most users will have observed that SystemWindows have an uneasy and often misleading relationship to halos. Unlike other morphs, SystemWindows come armed with their own controls for moving, resizing, labeling, dismissing, and menus -- and these controls are very different from the corresponding halo controls. For example, the way you resize a regular morph and the way you resize a SystemWindow differ radically, and if you try to use the halo to resize a SystemWindow you'll get something very strange. And very few of the standard halo-based menu items are appropriate for SystemWindows. A key goal of the halo rework was to make the standard halo-gesture do what people most of the time want (and to support, by way of a general "escape", a "shifted" halo gesture that allows the halo to be appear on *any* object. When working with "naked morphs" (i.e. morphs not in SystemWindows), it seemed that most often the initial halo should be brought up on the entire outermost naked morph, making it easy, for example, to dismiss a RecordingControlsMorph (formerly this required numerous successive cmd-clicks). But that same philosophy, if applied to SystemWindows, would result in users' frequently getting halos on windows with their first cmd-click. These two worlds of controls -- SystemWindows vs. halos-on-other-morphs -- are so different in flavor that in the halo rework, a conscious decision was made *not* to make it that easy to get a halo for a SystemWindow -- though it is indeed possible, via successive cmd-shift-click gestures. Perhaps this was a wrong choice, but in any case it was made after listening to many suggestions and trying out many alternatives. I still consider the choices settled on to be worthwhile improvements, and I will be interested to hear other opinions on the subject -- especially after the initial shock of unfamiliarity wears off. -- Scott

4 5

Re: [updates] 35 more for 2.8alpha
by R. A. Harmon 30 Apr '00

30 Apr '00

At 01:39 PM 4/29/00 +0200, Stefan Matthias Aust wrote: >At 15:56 27.04.00 -0500, R. A. Harmon wrote: > >>I would suspect not as MVC is slated to be removed from Squeak Central >>release. I'm not sure when. > >I'm not sure whether this statement is correct. Actually, I think, it's >wrong. AFAIK, the default desktop is probably changed from MVC to morphic >but there're no plans to remove MVC. The last time this question came up a couple of months ago I went back and found Dan Ingalls's original post on the future of Squeak and confirmed that he said MVC was being removed. I think I posted an excerpt of it. Unless I missed something since the last discussion, I guess that is still the plan. >I'm also pretty sure that this wouldn't be easy as it's deeply burried in >all parts of Squeak - including Morphic :-) If I remember correctly, during the last discussion Dan explained (something like) that the MVC event mechanism would remain but they still planned to remove the MVC GUI framework. Sorry, I can't give a better explanation. He said anyone could then add the MVC GUI framework to the Squeak Central release, if they wanted it. -- Richard A. Harmon "The only good zombie is a dead zombie" harmonra(a)webname.com E. G. McCarthy Spencer, Iowa

2 1

Re: help! hosed precious image: rescuable ?
by Tim Rowledge 30 Apr '00

30 Apr '00

It's just possible that you can read the image in a text editor of some sort; once there you have a reasonable chance of finding most Strings with your notes by simply doing searches for whatever still seems familiar. If you really stare hard, you might even get to see the patterns of someof your objects. Powerful headache pills may be needed. tim

1 0

Re: Logo in Smalltalk
by Bijan Parsia 30 Apr '00

30 Apr '00

Check out: http://minnow.cc.gatech.edu/squeak/1240 for a description of StarLogoWorld, what seems to be an implementation of StarLogo. StarLogo is exceptionally neat: It allows you to control mulitple turtles in parallel. You can do a lot of very cool simulations very easily with it. (I have a small writeup of the recent Java version of StarLogo at: http://www.monkeyfist.com/articles/333) "Although you are dealing with thousands of turtles you can do neat things. For example we can have the turtles leave scent trails, then we can have the turtles hunt for the best scent intensity. For that example you get clumps of turtles, with turtles occasionally peeling off and going to another scent clump. He will look into the distribution of theis work. This is the type of thing Disney would like to package up and deliver on the network to the education market." I'd love to hear more about the distribution of this work! Cheers, Bijan Parsia. On Fri, 28 Apr 2000, Carl Gundel wrote: > Does anyone know of an example of Logo in Smalltalk? > > -Carl > >

1 0

Jump to page: