Squeak-dev April 2001

squeak-dev@lists.squeakfoundation.org

165 participants
614 discussions

Re: Smalltalk & Squeak featured on Slashdot
by Stephane Ducasse 30 Apr '01

30 Apr '01

on 4/26/01 2:45 PM, Henrik Gedenryd at Henrik.Gedenryd(a)lucs.lu.se wrote: > Bijan Parsia wrote: > >> Now, testing for beginsWith: 'Morphic' gets us: >> >> 3379693 >> >> Bloat at last!!! >> >> But gee, to echo your point, Morphic is a *lot* of stuff. There's a PDA, a >> bunch of games, a EPS outputter, a set of interface widgets (a lot of >> which are in use in my system, since I'm running in morphic and have a >> gazillion windows open, etc.). > > Right, a lot of stuff, and I've said this before: most of this is not > Morphic per se, but applications of it. The PDA and the games are typical > goodies, and so on. Calling these "Morphic" is as logical as if every use of > MVC would become a part of MVC itself. > > And the whole eToy system has nothing to do with Morphic per se. In fact, a > lot of newcomers are confused (and I sure was) since there is no clear line > between Morphic and eToy. So when there are really just a few things to > learn, you have no way of knowing what that is. Etoy is clearly in the > domain of SqC's own experimentation, and I've said before that it should be > kept separate. In Stable Squeak I'm pretty sure this will happen. I think > this alone would cut the number of methods in Morph to less than half of > what it is now. And this would be great to have classes that have less than 200 methods ;). In the same idea, I strongly think that string does not have to know asURL as.... asblabla Because then everything is linked together. >

2 1

Headless squeak on linux 2.7.
by olczyk＠interaccess.com 30 Apr '01

30 Apr '01

How do I run a headless squeak on linux? I try squeak -headless and get a message: /usr/bin/squeakvmh missing ( find / -name squeakvmh finds nothing either ). If I need the file squeakvmh, where can I get it? I am using 2.7.

3 2

window's file handling
by Projet DeepLow 30 Apr '01

30 Apr '01

Hi everyone, first of all we would like to thank all those who helped us on our previous problems (Squeak VW compatibility), and then we would like to submit another one. as part of our masters project we are trying to make a windows explorer in squeak, and we were wondering if anyone could help us with file handling. Does anyone know how the file browser works (the pink window), and which methods it invokes ? there are many classes containing the word "file", which ones should we focus on to make get the file tree of a computer, to open a file, to execute a file from squeak(if it is possible) and to copy/paste and delete a file. Thanks in advance for your help. Projet DeepLow Maîtrise Informatique Université de Savoie (France) Explorateur 3D en SmallTalk Serna Audrey (madame_dodue(a)hotmail.com) Fleutot Pierre (foreveroux(a)yahoo.fr) Josserand Samuel (samjosserands(a)yahoo.fr) Loiseau Mathieu (MateOstl96(a)aol.com)

2 1

Re: Squeak for 3-year-olds
by Karl Ramberg 30 Apr '01

30 Apr '01

Cees de Groot wrote: > > Alan Kay <Alan.Kay(a)disney.com> said: > >There > >are a number of examples on Squeakland.org, and there will be many > >more over the next few weeks and months. > > > Squeakland has a download for MacOS only. I've been toying with the > plugin on Linux, but I don't think I've got the right image to go with the > samples on Squeakland. Would it be possible to provide just the image for > people that already have the plugin installed? Or is the image available > elsewhere (I tried both the original plugin image - 2.8 - and a stock > 3.0 image amended with the plugin source fileIns, but both complain that > they can't find certain methods when downloading the projects). I think you can use the 3.1 image. Check the preferences so it will not complain that sources and changes are not present. There will probably be some issues if the plugin is 2.8 and the image is 3.1. Karl

1 0

RE: runing squeak at sun Solaris 8.0
by Pennell, David 30 Apr '01

30 Apr '01

Bert sent this on the VNC list quite a while ago. It works well for me. -----Original Message----- From: Bert Freudenberg [mailto:bert@isgnw.CS.Uni-Magdeburg.De] Sent: Friday, March 17, 2000 3:57 AM To: Pennell, David Subject: Re: Display depth on Solaris w/ VNC (I'm CCing to PWS list because these guys might be interested in VNC, too. Look at the bottom, Swikiers) On Thu, 16 Mar 2000, David Pennell wrote: > If I start up VNC on Solaris 2.7 with the default vncserver script > (depth=8), I get a message from the Squeak VM complaining about not > being able to support "class TrueColor" [nb - this is an X class, not > a Squeak class] on a display depth of 8. I changed the vncserver to > depth 16, one of the two depths that are supported for TrueColor in > sqXWindow.c and all works, but still rather slowly (100 Mbps between a > 300 MHz NT box and a 4xx MHz Ultra 60). > > Squeak's "Display depth" indicates 8 bits. So - if the image is using > 8 bits, where did the TrueColor come from in the VM and how can I get > everybody back to 8 bits and make things a little snappier? VNC uses a rather unusual 8-bit TrueColor visual by default, while the Unix VM expects an PseudoColor visual in 8 bit. This is a known issue (http://www.uk.research.att.com/vnc/xvnc.html): Option -cc n Sets the colour Visual class used by the server. Some X applications don't cope too well with the TrueColor visual normally used by an 8-bit-deep Xvnc. You can make the server use a PseudoColor visual by specifying -cc 3. I use the following for 8/15 bit Squeak servers: vncserver -geometry 800x600 -depth 8 -cc 3 -dontdisconnect vncserver -geometry 800x600 -depth 15 -pixelformat RGB555 -dontdisconnect 15 bits is faster than 16 because no bit-shifting is necessary (Squeak actually uses 15 bits even if it claims to use 16). Look for stHasSameRGBMask16 in sqXWindow.c ... Also, with a VNC server at :1 start Squeak with "-xshm -display unix:1" options to make it faster. Using all this I even can squeak remotely over an ISDN line with SSH encryption. That way I can fully operate my Swiki servers from home :-) All in all, the Squeak/VNC combination is so powerful someone should implement the RFB protocol in Squeak directly - that way Mac/Win Squeakers could use it to full advantage, too (Mac/Win VNC servers only provide the local desktop remotely, while XVNC creates a real virtual desk which is independend of the local display). -Bert- -----Original Message----- From: Jörn Eyrich [mailto:Joern.Eyrich@brokat.com] Sent: Monday, April 30, 2001 6:38 AM To: squeak(a)cs.uiuc.edu Subject: Re: runing squeak at sun Solaris 8.0 TECHIN.A.KANG(a)mail.sprint.com wrote: > 2) running squeak from webpage connect to server using VNC, got the > following error: > Visual class TrueColor is not supported at depth 8 have you tried to start the vnc server with a higher display depth? use "-depth 32" (or 16, or 24) I don't know what this does to performance, however Joern

1 0

Re: Debugger-step into.
by Ned Konz 30 Apr '01

30 Apr '01

On Sunday 29 April 2001 04:08, phiho.hoang wrote: > Henrik wrote: > > There is only one thing I miss. The restart command starts executing the > > current method from the beginning even if you've gone halfway thru it. I > > just wish one had the option of restarting with the debugger, so one > > could debug into a message that one has passed over by accident. > > I agree, this would be very convenient. I wonder if that feature is > within the power of Smalltalk. I posted a version of this to the list earlier. I haven't used it yet; it'll be interesting to see if anyone finds it useful (besides PhiHo and Henrik, of course <g>). Some more things that should be pretty easy to add: * run to cursor: step until PC is at (or just past?) where you put your cursor * reset PC to cursor (backwards or forwards?). Could skip code in this way. Any thoughts on these? I'd really like to have a copy of Terry Raymond's _Smalltalk Professional Debug Package_ (http://www.craftedsmalltalk.com) for Squeak, but I doubt that Terry sees much of a market here. It works great with VisualWorks, though. -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

2 1

Debugger shows message from class that isn't the receiver's
by Andrew P. Black 30 Apr '01

30 Apr '01

I'm playing around with BookMorphs -- actually, I'm trying to get my ECOOP presentation into a book, just to see if I can. On this journey, I have encounterd an error that I just plain don't understand. That is: I don't understand the message that the debugger is sending me! After reading in (in my case, right now, zero pages), the image reading code writes out the bookMorph. I get the same problem if I just write out a bookMorph from the menu. I see a progress dialog 'please wait while the objects are counted', and then get a MNU. Here is the (top of the) walkback: 29 April 2001 6:18:04 pm MessageNotUnderstood: xxxInstVarAt: VM: Mac OS - nil Image: Squeak3.1alpha [latest update: #'Squeak3.1alpha' of 12 April 2001 update 3910] PasteUpMorph(Object)>>doesNotUnderstand: PasteUpMorph>>DoItwith: ReferenceStream(DataStream)>>writeInstance: ReferenceStream(DataStream)>>nextPut: [] in ReferenceStream(DataStream)>>nextPutAll: Array(SequenceableCollection)>>do: ... What's weird here is that writeInstance -- code shown here --- writeInstance: anObject "PRIVATE -- Write the contents of an arbitrary instance." ^ anObject storeDataOn: self is sending "storeDataOn: self" to anObject, and anObject is "a PasteUpMorph<page>(2386)", the pasteUpMorph that is the top page in my book. So, far, so good. But the method that shows up in the next frame of the stack trace is "DoItWith". When I browse that frame with the debugger, what I see is the following method: storeDataOn: aDataStream "Store myself on a DataStream. See also objectToStoreOnDataStream. must send 'aDataStream beginInstance:size:'" | cntInstVars | cntInstVars _ self class instSize. "cntIndexedVars _ self basicSize." aDataStream beginInstance: self xxxClass size: cntInstVars "+ cntIndexedVars". 1 to: cntInstVars do: [:i | aDataStream nextPut: (self xxxInstVarAt: i)]. " 1 to: cntIndexedVars do: [:i | aDataStream nextPut: (self basicAt: i)] " This is not a method from PasteUpMorph at all. It's a method from ObjectOut, which is why it is reasonable for it to be sending xxxInstVarAt: to self. But self is my PastUpMorph, hence the MNU. So, what's going on here? Perhaps my PasteUpMorph has become an ObjectOut and then become a PasteUpMorph again? If so, why is it trying to store itself on the disk? Was it reconstituted a bit too soon? Here be deep magic! All that I wanted to do was load some images ... Here is the whole of SqueakDebug.log: 29 April 2001 6:18:04 pm MessageNotUnderstood: xxxInstVarAt: VM: Mac OS - nil Image: Squeak3.1alpha [latest update: #'Squeak3.1alpha' of 12 April 2001 update 3910] PasteUpMorph(Object)>>doesNotUnderstand: PasteUpMorph>>DoItwith: ReferenceStream(DataStream)>>writeInstance: ReferenceStream(DataStream)>>nextPut: [] in ReferenceStream(DataStream)>>nextPutAll: Array(SequenceableCollection)>>do: ReferenceStream(DataStream)>>nextPutAll: ReferenceStream(DataStream)>>writeArray: ReferenceStream(DataStream)>>nextPut: SmartRefStream>>instVarInfo: [] in SmartRefStream>>nextPut: ProgressInitiationException>>defaultAction ProgressInitiationException(Exception)>>pass [] in BlockContext>>valueUninterruptably ProgressInitiationException(Exception)>>handlerAction ProgressInitiationException(Exception)>>signal ProgressInitiationException>>display:at:from:to:during: ProgressInitiationException class>>display:at:from:to:during: String>>displayProgressAt:from:to:during: SmartRefStream>>nextPut: StandardFileStream(ReadWriteStream)>>fileOutClass:andObject: [] in BookMorph>>saveIndexOnURL BlockContext>>on:do: [] in BlockContext>>valueUninterruptably BlockContext>>on:do: [] in BlockContext>>valueUninterruptably BlockContext>>on:do: BlockContext>>valueUninterruptably BlockContext>>ensure: CursorWithMask(Cursor)>>showWhile: BookMorph>>saveIndexOnURL BookMorph>>saveAsNumberedURLs BookMorph>>savePagesOnURL BookMorph>>loadImagesIntoBook [] in MenuItemMorph>>invokeWithEvent: BlockContext>>on:do: [] in BlockContext>>valueUninterruptably BlockContext>>on:do: [] in BlockContext>>valueUninterruptably BlockContext>>on:do: BlockContext>>valueUninterruptably BlockContext>>ensure: CursorWithMask(Cursor)>>showWhile: MenuItemMorph>>invokeWithEvent: MenuItemMorph>>mouseUp: MenuItemMorph>>handleMouseUp: MouseButtonEvent>>sentTo: MenuItemMorph(Morph)>>handleEvent: MorphicEventDispatcher>>dispatchDefault:with: MorphicEventDispatcher>>dispatchEvent:with: MenuItemMorph(Morph)>>processEvent:using: MorphicEventDispatcher>>dispatchDefault:with: MorphicEventDispatcher>>dispatchEvent:with: MenuMorph(Morph)>>processEvent:using: MenuMorph(Morph)>>processEvent: MenuMorph>>handleFocusEvent: [] in HandMorph>>sendFocusEvent:to:clear: PasteUpMorph>>becomeActiveDuring: HandMorph>>sendFocusEvent:to:clear: HandMorph>>sendEvent:focus:clear: HandMorph>>sendMouseEvent: HandMorph>>handleEvent: HandMorph>>processEvents [] in WorldState>>doOneCycleNowFor: Array(SequenceableCollection)>>do: WorldState>>handsDo: WorldState>>doOneCycleNowFor: WorldState>>doOneCycleFor: PasteUpMorph>>doOneCycle [] in Project class>>spawnNewProcess [] in BlockContext>>newProcess

2 1

Re: [OT] is morphic too simple?
by Andrew P. Black 30 Apr '01

30 Apr '01

At 17:34 +0100 2001.04.21, John Hinsley wrote: > >I can't believe that Sun had this (in Self) and decided to go with Java. Excuse me; it's late, I'm sick, and I've been thinking too much ... But it seems to me that the state of our "science" today is very much like a primitive tribe. When someone breaks out in hives, or has a stomach ache, they first talk to their friends, but if they don't get help, they go and see the witch doctor. They are told to cross a bunch of sticks, dance around them, and sacrifice a sparrow. What happens? Well, quite often, the hives go away or the stomach ache gets better. But they have no idea why. Of course, the witch doctor likes it that way, or else people would not give him any beads. When some software starts doing mysterious things -- like zapping my filestore or my computer refusing to boot -- I first talk to my friends, and if that doesn't help, I call tech support. (I've had some pretty desperate problems lately ;-). The tech support Wizard walks me through a process of muttering incantations, reinstalling something for the third time ... and, quite often, things start working again. I have no idea why. I wonder if the computer companies like it that way? I said that I've been thinking. I'm trying to figure out _why_ I like to spend time in Squeak. After designing and studying programming languages for over 30 years, why am I back with Smalltalk-80? I think that there are several answers, but John is nosing around at least one of them, with the email to which this is a reply. He said "Morphic is too simple". That prompted the zeroth reason that I spend time here. (0) Squeak is simple. Maybe too simple. As a language, it's like XP: it's the simplest thing that could possibly work. No protection, no syntax (above the method level), no initialization, no types, no packages, no modules, no to most of the things that I (as a language designer) thought were important. (1) No Magic. If I want to know why true & 3 answers 3, I can go and read the & method for true. If I want to know why I'm asked for my initials when the moon is blue, I can go and read SystemDictionary>>openSourceFiles. It may not always be easy to figure something out, but I know that it's always possible. (2) It's fully reflective. When I want to figure something out, I don't have to switch to another environment (browsing a source tree with a text editor). I can stay in Squeak and find what I need. (That's introspection.) Not only that, I can change what I want, and the system behavior changes. I'm not meddling with an image of the world, I'm meddling with the world itself. (That's reification.) About the only place where this is not true is in the process of building a VM, which may be why I haven't explored doing that. (3) The development tools are not "special". Anything that I can do in a browser I can do with a doIt. If the tools don't provide a button for what I wan to do, I can fake it. A while ago I was wondering how many methods were understood by String. (I was trying to convince a student that it would be fun to implement tree-structured strings.) So I brought up a protocol browser, got a Morphic halo on the PluggableListMorph, inspected it and found the answer. Everything is an object, duh ... such a simple idea, yet I'm still being surprised by its profundity. (4) Squeak isn't a language, it's a place. This idea was startling to me when I first read it -- go back to Alejandro F. Reimondo's mail to this list on 28th November 1999 -- a very thought-provoking posting. Despite all the power of human languages, they are an obstacle to communication as much as they are an aid. That's why we have to meet each other face to face, rather than sending email and writing documents. That's why XP encourages developers and customers to talk together daily, rather than sharing requirements specs. Can you imagine doing a woodworking project by talking to the tools and telling them what to cut? It might work quite well for cutting boards to size. But at some point I would need to feel the tools in my hands ... the feel of the plane biting into the wood and the way that I react to it are far too subtle to be put into words. I thought at first that Morphic scripting letting us build applications without programming was just a toy for children. But I was the child; it finally dawned on me that this was the way that all applications will be built in a hundred years. We are all pioneers here, I think ... (5) Squeak is a community. A community in which I can risk saying stuff like this and not get laughed at. A bit like the world used to be in the early days, before computing became big business. So, thanks for listening, and sorry if this has rambled a bit. Andrew

2 1

Re: Squeak for 3-year-olds
by Steve Wart 30 Apr '01

30 Apr '01

I agree that building (and destroying) things is important for young kids. At this age they seem to be desperately trying to develop a casual facility with everything around them, whether it's sand toys, crayons or electronic gadgets. And the average home is chock-a-block with electronic gadgets. When Daddy spends most of his day on the computer, kids are really serious about figuring out what that is all about. I don't want to create the impression that I'm trying to "teach my son about computers" -- for me, it's just a way to show him some of the stuff I do, and maybe teach him some rules about my work space. There is a mutual fascination when the two of us play with this stuff. He definitely wants to be here -- when he wants to go play with cars or puzzles or anything else, he is clear about that too. I get a kick out of it because I get to see him go from bashing on the keys to making connections with some pretty complex concepts. It's really just a game and the only rule is that it has to be fun for both of us. The connections I get to make surface in questions like, "what do I need to teach my kids and why is it important?". If I can figure that out I think the "how" will become much easier. Cheers, Steve Alan Kay <Alan.Kay(a)disney.com> wrote: > > For those of you who have very young children, you are much better > off having them start to learn to build things with their hands. Most > computers are not equipped with enough tactile feedback to be great > learning environments for very young children. > > A great book to read for those who are interested is "Towards a > Theory of Instruction" by Jerome Bruner. It doesn't talk about > computers but it is one of the best books about designing good > educational environments for children of various ages. Last year I > wrote a chapter for a book about kindergarten that discusses the > suitability of computers for 3 and 4 year olds. This will also go > online sometime this summer.

2 1

[ENH] ClockMorph gets 24 hr display
by Andrew P. Black 30 Apr '01

30 Apr '01

From the preamble: Date: 29 April 2001 Author: Andrew P. Black Added an instance variable and associated logic and menu item to control whether the display of the clock should be in 24 hr mode or am/pm mode. The default is still am/pm mode. Maybe there should be a preference for this? Also did some method categorization, and added a class comment.

1 0

Jump to page: