Squeak-dev April 2001

squeak-dev@lists.squeakfoundation.org

165 participants
614 discussions

[ENH] Hierarchical selection of Projects for 'jump to project'
by JArchibald＠aol.com 03 Apr '01

03 Apr '01

All-- The attached change set modifies the behavior of the menu used to navigate to other projects. The current menu presents (bottom part of menu) a list of Project names in alphabetical order. It seems to be more evocative if this section of the menu presents a hierarchical list instead. This change set provides for this functionality. After using it for a while, I don't know how I survived with the previous behavior -- I guess the truth is that I didn't use it very effectively. Once you start using Projects regularly to segment your work, you soon end up with a LOT of Projects. I think this change help a user see the organization of his Projects when jumping around a lot. Adjustments are made to PopUpMenu to present the same appearance in MVC projects. Enjoy, Jerry. ____________________________ Jerry L. Archibald systemObjectivesIncorporated ____________________________

1 0

Refactoring support in Squeak
by Brian Rice 03 Apr '01

03 Apr '01

Hello all, Before I try to dive through the coding-support classes in Squeak to look for these features, I'd like to know what the status is of porting the Refactoring Browser to Squeak. I've heard that some of the basic refactorings are supported, but apparently they are without user interface hooks. Is someone working on this? If there is only a little left to do, where are the places that need work? Interested in Refactorings, ~

3 2

Re: Dumb question about headless & Nebraska
by Tom 02 Apr '01

02 Apr '01

On Mon, Apr 02, 2001 at 02:32:27PM -0500, Les Tyrrell wrote: > The Java and Windows clients that I've used worked well with the > Linux server. From what I've read, using Unix with X as a server is > perhaps the most optimal choice, other than a special purpose > server which one might build in and for Squeak. Let me emphasize again that Xvnc does not depend on running UNIX as your OS or X11 as your window system. You can think of Xvnc simply as something that takes X11 connections on one end and converts them to VNC connections on the other. You can use it to bring an X11 server to any platform without needing any hardware drivers. I believe people have ported Xvnc to both Windows and MacOS X. Cheers, Thomas.

1 0

Re: [ENH] Better TrueType fonts
by Karl Ramberg 02 Apr '01

02 Apr '01

Ned Konz wrote: > > On Saturday 24 June 2006 22:25, Karl Ramberg wrote: > > Ned Konz wrote: > > > Also, how would you keep the italic characters from getting cut off at > > > the > > > upper right of their bounding boxes (and still have decent looking > > > kerning)? > > > > How does the fonts work on different background colors ? > > I think the form over rule must be modified. The background color of > > the font form must act as transparent. > > It's using Form paint, just like the regular fonts. So you get shades of grey > around the edges. > > This is why I say that a colormap needs to be used: > > 8-bit Font destination > -- > 0 (transparent) transparent > 40 (black) opaque foreground color > other shades of grey foreground color with varying alpha > Ok. But I don't understand why the italic kerning cuts off the fonts... Are font forms placed after eachother based on baseline kerning and assuming the glyphs don't overlap verticaly? And italic does overlap verticaly so the font forms must be copied in front of eachother to not cut the glyphs ? Karl

2 3

[minnow] problems with link and uploading
by Stephan Rudlof 02 Apr '01

02 Apr '01

I have some problems here; link http://minnow.cc.gatech.edu/squeak/1406 ('All Projects' page!) and uploading to http://minnow.cc.gatech.edu/squeak/1810 (so I cannot put a draft here) don't work. Greetings, Stephan -- Stephan Rudlof (sr(a)evolgo.de) "Genius doesn't work on an assembly line basis. You can't simply say, 'Today I will be brilliant.'" -- Kirk, "The Ultimate Computer", stardate 4731.3

1 0

Re: Help wished? [was: Re: Genie on iPaq..comments]
by Stephan Rudlof 02 Apr '01

02 Apr '01

Nathanael, I have anticipated your answer and also prepared my reaction: Nathanael Schärli wrote: > > Stephan, > > Wow, a speedup of factor 5 or more sounds fabulous! I haven't written the > primitive yet, but considering the great speedup we should definitely do it. > The main problem of Genie (especially on PDAs) is the performance and using > such a primitive we could probably solve that. > > I have quite some C/C++ experience but I've never written Squeak primitives > yet and I would appreciate your help! I'm motivated to help here, since your work seems to be very good and many people are interested in it. Fortunately I have some experience with writing Squeak primitives (LargeIntegersPlugin). But since I'm very busy and don't want to take a deeper look onto your Genie work right now (possibly after having BlockClosure semantics ;-) ), I need your help. If you send me a changeset containing - a semantic testing method which tests CRStrokeFeature>>sameClassAbsoluteStrokeDistance:forReference: with a few typical examples; and - a speed testing method, which checks for the speed of just this method *after* the parameter constructing stuff has been done ('self' seen as a parameter, too!); then I'll write you the prim. This means to create some typical instances of CRStrokeFeature first and calling >>sameClassAbsoluteStrokeDistance:forReference: for each from the testing method then:. Semantic testing additionally compares each result with the wished one, speed testing not. Greetings, Stephan > > Nathanael > > > > As far as the delay in the lookup is concerned, I'm really > > thinking of > > > writing a primitive for it. 90% of the time is spent in the method > > > CRStrokeFeature>>sameClassAbsoluteStrokeDistance:forReference: that > > > basically iterates over the vector decompositions of two > > gestures and > > > compares them using integer arithmetic. (I avoided any > > floating point > > > operations because that is even slower on small machines). > > This method has < > > > 100 LOC and it only calls two helper methods (3 LOC each). > > I don't have much > > > experience concerning the performance gain by making a > > primitive out of such > > > a method, but I have a feeling that it could be quite a lot. > > > Does anyone know more? > > > > After taking a look onto it I guess a speedup of factor 5 to > > 10 for this > > method, if written as primitive. This is a conservative estimation. > > > > If you or another person haven't written it so far, I'd like > > to help here. > > > > Greetings, > > > > Stephan -- Stephan Rudlof (sr(a)evolgo.de) "Genius doesn't work on an assembly line basis. You can't simply say, 'Today I will be brilliant.'" -- Kirk, "The Ultimate Computer", stardate 4731.3

1 0

Re: Dumb question about headless & Nebraska
by Kevin Fisher 02 Apr '01

02 Apr '01

Strangely enough, I've done this before...several years back I had a swiki running on Xvnc. You're right, it is slow. :) Still it would be pretty cool just to use Squeak alone, if possible. > > If not, and you have the bandwidth to support it, give VNC a try. VNC used > to be called a frame-buffer protocol, and for things that don't update their > contents frequently it is usable even over relatively slow connections. But I > believe > that Morphic may have enough updates going on in the background that this > approach > would be unusable over a slow connection. It's easy enough to find out, anyway. > > For the more ambitious, try implementing the VNC server protocol in Squeak. > > - les > > > > >

4 3

Re: morphic related question...
by Ned Konz 02 Apr '01

02 Apr '01

On Monday 02 April 2001 04:12, Raymond Tiefenthal wrote: > Btw, theres a method #setTarget: in MorphMenu which allways gives me a > walkback. > The comment says: > "Set the default target object to be used for add item commands, and > re-target all existing items to the new target or the the invoking hand" > The method sends #hand to the argument, which raises "message not > understood". That seems like it expects a MorphicEvent as argument. > Allthough I'm a rookie, that looks like a bug to me. > Anyhow, perhaps I'll investiagte that further and post sepperatly. setTarget: is meant to be called from a menu choice, with the event that initially made the menu come up. It is not intended to be used to change the target of an existing MenuItemMorph programmatically. -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

1 0

Is www.squeak.org down?
by Houser, Robert C 02 Apr '01

02 Apr '01

I can't reach it. I get the following: Error- 404 Requested Information http://www.squeak.org/ is unavailable. Failed to connect to server www.squeak.org (80) reason: hostname unknown ----------------------------------------------- Right to keep and bear Haiku shall not be abridged Or denied. So there.

2 1

Re: Re: Object, or Attribut, or something else...
by GeneCox 02 Apr '01

02 Apr '01

Hans-Martin Mosner， Thank you for your patient. I learn a lot from your words. I want to make something clearer, for I think it be valuable. at 2001-04-02 10:23:00 you wrote： >I think that there is a common misconception that Smalltalk's class tree is >somehow carrying the 'meaning' or 'intention' of Smalltalk programs. It isn't. >Class inheritance is just a mechanism for code sharing, nothing spectacular. Yes... but the design could be influenced much by the way you modeling the world and I was trying to say that take classes as containers holding grouped attribute-blocks might result in better code sharing mechanism and tidier building of class library. >In the current Smalltalks, the relation of objects to their message catalogs >(via their class) is rather fixed, which has the undesirable effect that your >examples of accumulating and refining knowlege about some real-world entity >can not be directly modeled... Though it still would be not too difficult to achieve that even in current system. Performance not concerns, adding a variable #memory and some simple methods like #keep:as/for: to the current root Object( for these attributes will be in common among all) and something more, then a few trikes using blocks might at last make it, but this is *not* what I was seeking for. An attribute-centered ( or some other terms in better English ) Smalltalk system can be very different, and make these kind of things very easy to implement, may we just treat it as side effects. >....... >Things like age, current profession etc. They have nothing to do with class >based inheritance at all. It's very easy to create a kind of 'attribute-holding' >object which can dynamically gain and lose attributes. What is that in object besides its attributes, but still will hold them( those attributes)? With the term attributes I mean information exchanging and action taking "capabilities", or knowledge and behavior accordingly. (Or variables and methods, if it will make it clearer.) Let me put it like this: when looking at codes shared among objects, I see attributes, shared, which can be well grouped as classes while some other people might see inheritance, which, I believe to have made things complicated. And the main difference is, object can be build of attribute blocks across classes(of attributes) with ease if we would like to forget something we put on them as inheritance. I hope it is worth discussing. >Prototype-based languages, such as Self which Jecel mentioned, have taken that >approach to implement full systems. From a practical programming perspective, >I prefer the Smalltalk way of making classes explicit, because I feel that >code is much more tangible in such a system. I agree. Yet classes as attributes grouped together will not sacrifice anything of value in the current system. It will only make the system tidier, simpler, and clearer, surpose we are not to make use of any side effects. Thank you again, and Good luck! GeneCox genecox(a)21cn.com

1 0

Jump to page: