Squeak-dev November 1999

squeak-dev@lists.squeakfoundation.org

164 participants
545 discussions

[FIX] FreeCell bug after update 1619
by Pennell, David 29 Nov '99

29 Nov '99

------_=_NextPart_000_01BF3A8C.E0E37A20 Content-Type: text/plain; charset="iso-8859-1" Replaces contains: with anySatisfy: to be consistent for update 1619. This problem is exposed if you run FreeCell in a SystemWindow and activate the window. -david ------_=_NextPart_000_01BF3A8C.E0E37A20 Content-Type: text/plain; name="FreeCell-satisfaction.cs" ; x-mac-type="65417070" ; x-mac-creator="43534F6D" Content-Disposition: attachment; filename="FreeCell-satisfaction.cs" Content-Transfer-Encoding: imap_stub 0,1771,2,488,0, ------_=_NextPart_000_01BF3A8C.E0E37A20--

1 0

Re: Intervals
by Bob Arning 29 Nov '99

29 Nov '99

On Mon, 29 Nov 1999 08:00:28 -0800 Dan Ingalls <Dan.Ingalls(a)disney.com> wrote: >Here is a place that I wish we had a type inference system >working already, because then we could examine all the place where #includes: >was being sent to an interval, and decide which meaning was intended. Short of that, an empirical approach may provide useful (though not necessarily complete) information. The idea would be to watch for any invocation of Interval>>includes: and record that fact. Later examination of the data collected (if any) could point to places that need fixing. Enclosed is a snippet to do that (it could be easily modified to watch for just about anything). Cheers, Bob ======== 'From Squeak2.7alpha of 9 November 1999 [latest update: #1660] on 29 November 1999 at 11:41:18 am'! Object subclass: #FlagWatcher instanceVariableNames: '' classVariableNames: 'Gotchas ' poolDictionaries: '' category: 'Kernel-Objects'! !Object methodsFor: 'flagging' stamp: 'RAA 11/29/1999 11:41'! flag: aSymbol "Send this message, with a relevant symbol as argument, to flag a message for subsequent retrieval. For example, you might put the following line in a number of messages: self flag: #returnHereUrgently Then, to retrieve all such messages, browse all senders of #returnHereUrgently." aSymbol == #rangeIncludes ifTrue: [ FlagWatcher gotcha: thisContext shortStack. ].! ! !FlagWatcher commentStamp: 'RAA 11/29/1999 11:40' prior: 0! I am used to collect interesting things for later examination. To add something to me, do FlagWatcher gotcha: 'anything you like'. To find out what has been seen so far (and to reset the current collection) do FlagWatcher show.! !FlagWatcher class methodsFor: 'as yet unclassified' stamp: 'RAA 11/29/1999 11:37'! gotcha: something (Gotchas ifNil: [Gotchas _ Bag new]) add: something! ! !FlagWatcher class methodsFor: 'as yet unclassified' stamp: 'RAA 11/29/1999 11:38'! show | temp | Gotchas ifNil: [^1 beep]. temp _ Gotchas. Gotchas _ nil. temp sortedCounts inspect. ! ! !Interval methodsFor: 'accessing' stamp: 'RAA 11/29/1999 11:35'! includes: aNumber self flag: #rangeIncludes. ^ aNumber between: self first and: self last! !

1 0

Intervals
by Andrew P. Black 29 Nov '99

29 Nov '99

--============_-1268253414==_============ Content-Type: text/plain; charset="us-ascii" I hesitate to raise this one, since the original code is stamped "di", but the include method on class Interval is manifestly broken. There are some other subtle problems with Intervals, particularly empty intervals and thos with negative steps, but the attached fileIn does at least fix the most obvious problem. As released, include: works only for intervals with a step size of 1. I was wondering about the pros and cons of re-implementing intervals as two subclasses, one for +ve steps and one for -ve steps. Many of the methods have two cases in their code. Would this be a very minor gain in speed for a significant amount of code bloat. The problem is for example: (1 to: 20 by: 3) asArray prints as (1 4 7 10 13 16 19 ) and so (1 to: 20 by: 3) asArray includes: 8 prints as false but (1 to: 20 by: 3) includes: 8 prints as true Nevertheless, (1 to: 20 by: 3) occurrencesOf: 8 prints as 0! --============_-1268253414==_============ Content-Type: text/plain; name="Interval-includes.st" ; x-mac-type="54455854" ; x-mac-creator="522A6368" Content-Disposition: attachment; filename="Interval-includes.st" Content-Transfer-Encoding: imap_stub 0,1767,2,737,0, --============_-1268253414==_============--

1 0

Let's Get Bizzy
by Dan Ingalls 29 Nov '99

29 Nov '99

Folks - I awoke this morning with a scene from the movie "Look who's Talking" in my mind (I've never really seen the whole movie, but I did pass by the tube when my family was watching this scene). A 5- or 6-year old girl is having a fantasy of playing basketball (which she loves) with Michael Jordan (I think). She utters the subject exhortation, and then procedes to go out and tear up the court. I hope some of you recognize the scene. At any rate, I have this feeling about Squeak right now. It's a time when there are great challenges in every direction, but I also think we can get through them and be on a completely new plateau in a month or two. It's also a time when we could use some help, so I figured it might be worth a message to the list. My top priority My sole focus, and I will be working with Ted and consulting with Andreas about this, is to bring together ImageSegment swapping Project export/import A simple name space architecture Our goal is before the end of the year, to be able to start up a small kernel Squeak image, and then browse around to active content (remote projects) on the web, importing any that look interesting. The new content would be swappable off the disk thereafter so it is quick to access, and yet doesn't make the image too big for small machines. There are two ways outsiders can help with this -- first is to critique the design, and second is to test the result. I greatly appreciate the feedback I got last month. OOPSLA caused me to suspend the work, but I'm back at it again. I plan to send out another message on this in the next day or two. I think you can see that if this all works at all well, it will give us the effect of everyone in the world having access to the best Squeak demos and active tutorials. Remember that we have a working eventRecorder with built-in compressed voice-over. AND you can now launch Squeak from within Netscape. Tee Hee. You Can Help The rest of the gruop is also pretty heavily loaded too, with John, Alan, and Scott focussed on end-user syntax and scripting issues, and Andreas up to his ears in polygons and JIT-Blt. There are at least two small items that would help us, or at least help me, just to keep the pace going in this busy time. World Within a World Window. One of the tasks on my list is to implement a world within a world window for Morphic. I would love to hand this off to someone else, and I know there are several Squeakers out there who could do the job. Note that we already have the ability to have a morphic world in an MVC window. What we need is a morphic world in a morphic window. I got this partly working a month ago, and would be happy to pass along that code. One tricky thing to deal with is: is the inner world always active, or only when the window is active? The latter would be simplest and would be adequate for now. Another thing to deal with is global flaps. Currently global flaps are single instances that get moved around from project to project. The analogous thing to do for an inner world would be that when you activate the inner window, flaps disappear from the outer project and appear in the inner window. An alternative would be not to show global flaps in nested worlds. Collecting Recent Submissions to 2.7alpha It would also save me a couple of days if someone could run through Squeak mail since the last time we were caught up with various enhancements and bug fixes. It would be ideal if that person would assemble the submissions, help me pick out any that might be questionable, and make sure the rest are happy in 2.7apha. I could then very easily turn the result around into the update stream. That way we can all keep going forward while I'm working on other stuff. Volunteers for either of the above? Squeak Socket Primitives I have held back from participating in this discussion because it is not an area in which I feel competent to make judgements. However I think we're converging on an approach that makes the best of the existing primitive interface, and also takes advantage of Craig's work on Correspondents. I'm going to stay out of details here, since if you get Squeak mail you have already been immersed in the whole discussion. At any rate, we'll do what we can to bring this to closure in the next little while. Jitter/J3/TGV/Squam, but not "Late for Lunch" Ian seems to have his project well in hand, but I thought it deserves some mention in this list of semi-coordinated activity. Ian has made great progress, as those of you who were at OOPSLA will be aware. I'm hoping we can pick a reasonable plateau in the next few weeks and formally distribute a J3 VM, so that people can start enjoying the fruits of this very cool work. - Dan, for SqC

2 5

Re: Several beginner questions
by Ian Trudel 29 Nov '99

29 Nov '99

-----Message d'origine----- De : Mike Thomas <miketh(a)ptmnet.com> ¿ : squeak(a)cs.uiuc.edu <squeak(a)cs.uiuc.edu> Date : mercredi 24 novembre 1999 19:30 Objet : Several beginner questions >Hi all.. Heya there. >2. The Swiki site has loads of customisation pictures, but no descriptions >of how to do that customisation. For example, how do I change my >environment background (or any other window for that matter) to an image >instead of garish green? You can use the FileList's "open image as background" feature to do so. It supports Squeak's common file format (GIF, JPG, Form stoteOn: (run coded) and BMP). This feature works in MVC world as well. To proceed, you need to open a FileList. Search for your favorite wallpaper and then select it with the yellow button. If you chose a valid supported file, you should see "open image as background". Select it and voila. That's what I wrote on Swiki, http://minnow.cc.gatech.edu/squeak/772 ::) Regards, Ian

2 1

RE: Why i use Squeak ???
by Alejandro F. Reimondo 29 Nov '99

29 Nov '99

Hi Piero!, sorry, but I consider this topic interesting to be copied to Squeak & SUGAR's list... [please be patient with this little introduction] If you can read Spanish, please go to http://www.sugarWeb.com You will find there a lot of interesting material about Smalltalk (in Spanish) and an active mail list about Smalltalk and Object Tecnology for Spanish spoken people. We (Xavier Alvarez and I) founded SUGAR (the Smalltalk User Group of ARgentina) 3 (or 4?) years ago because we saw that smalltalkers does not talk about their works. One of the things I saw is that the Smalltalker didn't talk because they didn't have "big problems" working with objects... [they don't know something too important to talk about. big problems are many "small" problems...] The wrong thing is that people didn't know all the thing that was been done in Smalltalk. We study this side effect of smalltalking (all people see their objects and not see/talk each other) and founded SUGAR. Now SUGAR is a group that exceeds Argentina limits; but for us is the definition of the smalltalk community in our country. But why I said it to you? because we also learn that Smalltalk is something more than a language+tools... >> Smalltalk si not a language. Smalltalk is a place. >interesting sentence.....:; One of the most difficult things to teach (when you teach Smalltalk and Object Technology) is the notion that there is a REAL Technology change. When, we, mere smalltalkers speak about Smalltalk to "conventional" ;^) people; immediately the people try to match our sentences to their well known world of languajes and tools. I introduce Object Technology principles (Smalltalk principles) to people writing the sentences: Object Technology (O.T.) or Object Environment (OE) and Object Oriented Languages (O.O.L) then I make the people read each world to know what it says and what does NOT say... O.T. is the TECHNOLOGY to work with Objects. Object Environment is a place with things(called objects) , Smalltalk is an OE, then is a place. O.O.L are Languages (not a technology change), and ORIENTED to objects not really WITH objects. In OOL there is a pointing to OT, but only as an orientation, not as reality. IN an object environment, we are embedded in a virtual world of objects, creating, manipulating, touching them. Our activities are not declarative, but gestural (toucing, repairing, building them) When you teach OT in that way, people can see that there are something more that languages and tools to program computers. Using conventional technology, we humans, talk to the computer as if it is a person too. Computers was a mirror reflecting the working power of the human. (you can see old movies where "the machine" attacks humans and having human-like roles) In this technology the power of the language is very important. We can do all the actions that can be expressed by language. Now in the network era, we have learned that we have virtual objects; and each object has a language. Now we talk to objects, and the machine disappears... Now the machine is a place, an environment. >> our brains, but also because IT is important for our "mind contents"... >eh eh....this is a very phylosophical question....let's talk about it Now, the point is: Do whe use computers to do our work? or as a medium where information can migrate and be more active. (can migrate prior to the dead of the host) I am currently studing evolution of Smalltalk environments (at slow speed, because I must work in commercial development too, always using Smalltalk from 1987). I have placed some references at sugar's web site, and many are announced at the sugar's mail list. You can find an interesting document (in Spanish) at http://www.sugarWeb.com/Theory/StGen.htm > You can read "Behavior of Information" > http://www.truxton.com/~trux/etc/boi/index.html [...snip...] >i have read it a little...in the first chapter he has written about >tha fact that information lives...it is surprising for me !! However i >haven't figured out what is the purpose of that book...and you ? I think the purpose is to transmit new points of view to understand the nature of information as is (not as we, humans want to see it). There, you will find a very interesting material explaining abstraction and classification techniques as a well known mechanism to propagate information. Applying this kind of thinking to our Object Environments can be used to know what "we" really are doing when we build virtual (non-material) spaces. Other interesting pointers to related material http://www.lycaeum.org/~sputnik/Memetics/index.html http://www.physics.wisc.edu/~shalizi/Dawkins/viruses-of-the-mind.html http://www.tufts.edu/as/cogstud/papers/time&obs.htm http://www.lycaeum.org/~sputnik/Memetics/index.html http://www.cpm.mmu.ac.uk/jom-emit/ Now, we can talk about the Smalltalk "language"... ;^) We know that Smalltalk is a place (full of objects). But actually we can't talk to the objects in "natural" language. We cant touch them with our hands and talk them in a friendly manner... Our object machines (actually computers) are very rudimentary; we use rigid syntactic languages and ony a finger (mouse, click, doubleclick, drag&drop) to comunicate with virtual objects. Unfortunately the computer language to communicate with objects is called Smalltalk too! Now, the language of communication between objects is also called Smalltalk. Imagine a (near) future when we can take an object (aCollection) in our hands and pick an object from it. This action is the same of sending the message #at: to aCollection; But the message is not said explicitly (the language disappears, only the "motion" (spreading) of information is done). We can change the syntactic rules (Language) of an Smalltalk Environment and it will continue been an Smalltalk environment. We can change it's syntax or does not send messages explicitly, But it will be an Smalltalk. Then Smalltalk is not a language; but currently HAVE a language (also called Smalltalk) and tools like browsers & inspectors. The future of a language is to be replaced by another. The future of an environment is an evolved environment. Now a reflection for fun: what is the future of Java? ;^) Algol / Fortran / Pascal / Prolog / Modula / Ada / C / C++ / Java / Java++ ;^) They all have in common that are languages, and the human talks to the computer... (now object oriented) Java, C or Fortran are specimens of Language species. Your Smalltalk, as my Smalltalk are specimens of Smalltalk species. The REAL technological change is to HAVE objects; to live with them, and evolve with them in an stable environment/system. If the objects are all virtual, real or a mix; does not care. [now you can stop reading] :^) - - - - - cut here - - - - - A "virtual space" of objects is not only some communicating objects (sending messages). They define an environment, with relations and physical properties. Our objects are the atoms of our virtual worlds. All things evolve in parallel in Nature, and we humans study Nature from things of size similar to us, to the greatest (universe) and smallest particles (subatomic particles). Provably because God creates the Universe and place us saying we must do things as he did. But, in real world all things has been created (by God), and human can't create anything? We can only mix the things and try to understand them. But where can we create? We can create in our minds, using our imagination power. With imagination we can create new worlds and do the noblest thing. Now with Smalltalk we can place our creations outside our minds and do not kill them (to fileOut/printOut a class is like placing a dead butterfly in a book; it can look fine, but it only can fly in our mind; [here you can think again in Object Environment vs. Object Oriented Languages]). Now we can create our virtual/active worlds. But we have ony one processor by world and we have only atoms (objects). :^( The construction of molecules (frameworks) has begun, but only recognizing that there is a physics of our system, we can start to think in a chemistry and biology of a virtual world. - - - - - cut here - - - - - bye, Ale. Ah, remember to visit http://www.sugarWeb.com if you can read Spanish... ---------- De: Piero Campanelli[SMTP:pc@skylink.it] Enviado el: Viernes 26 de Noviembre de 1999 19:38 Para: Alejandro F. Reimondo Asunto: RE: Why i use Squeak ??? Hi Alejandro !! > Smalltalk si not a language. Smalltalk is a place. interesting sentence.....:; > There, we can create our creatures and help them to evolve. > I use Smalltalk(Squeak) because I can HAVE the power of creation. this is very poetic....probably you're a very creative guy ...;))) Squeak is a tool for a developer as the paper is a tool for a poet...but what do you do with your creativity and Squeak...??? (sorry..i think to be a poor creative :) > our minds an move (using Internet) at speed of light. > I think that we use Smalltalk (object environment) because > it is important to us to export our mind contents outside > our brains, but also because IT is important for our "mind contents"... eh eh....this is a very phylosophical question....let's talk about it !!!! > You can read "Behavior of Information" > By Luc Claeys > > http://www.truxton.com/~trux/etc/boi/index.html i have read it a little...in the first chapter he has written about tha fact that information lives...it is surprising for me !! However i haven't figured out what is the purpose of that book...and you ? Thank you ... Piero

1 0

Re: Some questions about Squeak
by Stephane Ducasse 28 Nov '99

28 Nov '99

> No! Don't do it that way! > It is definetely not recommended to implement : > > new > ^super new initialize > > in a class and its subclass, because initialize will be sent > twice (several times) to the same object. The super call > makes no difference here because the item returned by > super new does not know about it anymore. Thanks I already know. > So it is pretty easy to solve your problem without patching > the VM. It suffices to implement: Defining something into Object class is not good. The class behavior is inherited from Behavior This is Behavior that new should be redefine to call initialize > Object class>>new > "Just do it here so my subclasses need not do it" > ^super new initialize > > Wait! Don't forget to implement first: > > Object>>inititalize > "Do nothing" > ^self. > > >

1 0

[BUG][FIX] FileList>>addNewFile and Dir
by Steve Elkins 28 Nov '99

28 Nov '99

The change set included below fixes the same bug in FileList>>addNewFile and FileList>>addNewDirectory. The methods were so similar I felt obliged to parameterize them into one. --------8<-------->8-------- 'From Squeak2.7alpha of 13 November 1999 [latest update: #1660] on 28 November 1999 at 9:13:02 am'! "Change Set: FileListBugFixes Date: 28 November 1999 Author: Steve Elkins Parameterizes 2 very similar FileList methods, #addNewDirectory and #addNewFile, as well as removing the same bug from each, that pushing the cancel button in the FillInTheBlank dialog gave a walkback."! !FileList methodsFor: 'file list menu' stamp: 'sge 11/28/1999 09:03'! addNew: aString byEvaluating: aBlock "A parameterization of earlier versions of #addNewDirectory and #addNewFile. Fixes the bug in each that pushing the cancel button in the FillInTheBlank dialog gave a walkback." | response newName index ending | self okToChange ifFalse: [^ self]. (response _ FillInTheBlank request: 'New ',aString,' Name?' initialAnswer: aString,'Name') isEmpty ifTrue: [^ self]. newName _ response asFileName. Cursor wait showWhile: [ aBlock value: newName]. self updateFileList. index _ list indexOf: newName. index = 0 ifTrue: [ending _ ') ',newName. index _ list findFirst: [:line | line endsWith: ending]]. self fileListIndex: index. ! ! !FileList methodsFor: 'file list menu' stamp: 'sge 11/28/1999 09:04'! addNewDirectory self addNew: 'Directory' byEvaluating: [:newName | directory createDirectory: newName] ! ! !FileList methodsFor: 'file list menu' stamp: 'sge 11/28/1999 09:04'! addNewFile self addNew: 'File' byEvaluating: [:newName | (directory newFileNamed: newName) close] ! !

1 0

RE: Some questions about Squeak
by Raab, Andreas 28 Nov '99

28 Nov '99

Stephane, > **I would like to know if there is a way to propose new improvements > for the smalltalk kernel. You're using the right way: Just post it to the list. > **I tried several times to get automatic update from the server > but I NEVER succeeded. Is there somethign wrong? Unlikely. But looking onto your email address it might be better to use a server in Europe. I *think* that the updates are mirrored in Germany (Bert - is this correct?!) so you might want to patch your Squeak to go to a different server. > **The mailing list is quite buzy. Is there no plans to have a newgroup? > I know that some of you think that Squeak is not ready for the mainstream > people but this would help people that are interested to communicate more > easily. I don't see a difference between having a mailing list and a newsgroup *except* that having a mailing list avoids a lot of spam (both, to the newsgroup and to the people posting on it) and that it is a somewhat larger download. Since I don't mind downloading a bit more stuff and I really hate spam my personal preference is clearly that of having a mailing list. > ** The uiuc server is often pretty difficult to obtain. Yesterday for > the first time I was able to play with the music (just because I never > thought before that I could download another midi file). It could be > fun to have a small set of midi file on the Squeak ftp site too. Try going to ftp://ftp.cs.uni-magdeburg.de/pub/Smalltalk/Smalltalk - the entire contents of the uiuc Smalltalk directory is mirrored there. And it is right around the corner so you should be able to get better access to it. BTW, another nice place to look at is www.leo.org or ftp.leo.org - they have quite a selection of music related stuff there and again, it's even closer to you (in Munich). Andreas

4 3

[UPDATES] Order emerging
by Dan Ingalls 28 Nov '99

28 Nov '99

Here below is the latest news. Many thanks to all the test pilots that have crashed and the mechanics that have made repairs (esp Bob). By the way, if you have tuned in late, don't forget... After updates 1648-1650 any Morphic projects in MVC must be converted to a new form. To do this, execute in the 2.7alpha root project (an MVC project) MorphWorldView convertToMVCWiWPasteUpMorph You may then re-collapse the converted windows and all should be fine. Enjoy - Dan --------------------------- 1655ProtoObjectTweak-di -- Dan Ingalls -- 26 November 1999 A few further tweaks to the ProtoObject change, to use Object where appropriate in traversing the entire Object hirearchy. Fixed duplicate entries in browse senders, implementors. 1656PreferencesFix-bf Fixed a problem with the MVC preferences window 1657WiWTweaks2 -- Dan Ingalls -- 27 November 1999 Attempts to catch the last few places where Morphic worlds in MVC require MVCWiWPasteUpMorph in place of PasteUpMorph. 1658ColorAndProjectFix -- Dan Ingalls -- 27 November 1999 Makes Choose Desktop Color work in inner worlds. Preempts attempt to change projects from inner worlds. 1659WiWTweaks3-RAA -- Bob Arning -- 27 November 1999 A number of 'last inch' changes to support Windows in Windows... 1. If the user (in an inner world) changes the display depth or otherwise causes Display to get a new bitmap, the outer world will need to know about this to avoid leaving parts unrefreshed. 2. The ColorPickerMorph used to change the desktop color sometimes worked and sometimes didn't, depending on world position. 3. If the inner world is active and you click in the outer world, you didn't always get a world menu. PasteUpMorph>>restartWorldCycleWithEvent: should now deal with this. 4. If you followed Marcus Denker's report, you had an inactive inner world trying to run it's own #displayWorld and sometimes this would generate a walkback (like the hand was not yet initialized). 1660BrownDragFix-RAA -- Bob Arning -- 27 November 1999 Fixes a problem reported by Tansel Ersavas that constrained dragging was broken (can't drag in a flap, brown-drag in any pasteUpMorph doesn't work).

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
55
Older →

Jump to page: