Squeak-dev October 2000

squeak-dev@lists.squeakfoundation.org

112 participants
424 discussions

Re: More Information on Squeak Textbook
by Ivan Brusic 04 Oct '00

04 Oct '00

Jeff Szuhay wrote: > > >Currently ranked in sales at 38,828! > > Wow! and this is scheduled to ship until December?! But wait! The current sales rank is 7,610. Mark was kidding about the "with a bullet" part, but it really is true. Congrats Mark! Ivan P.S. I will be buying the book, but not from Amazon. ;) -- Ivan Brusic | "What about tone deaf people, such as myself, E-mail: ibrusic(a)wsicorp.com | who can't tell Em from a cat being shoved Phone: (978) 262-0758 | in a blender?" - overheard on alt.guitar

1 0

Re: [updates] 14 for 2.9alpha
by Adam Bloom 04 Oct '00

04 Oct '00

>From: Dan Ingalls <Dan.Ingalls(a)disney.com> >Reply-To: squeak(a)cs.uiuc.edu >To: "Leandro Caniglia" <caniglia(a)dm.uba.ar> >CC: squeak(a)cs.uiuc.edu >Subject: Re: [updates] 14 for 2.9alpha >Date: Tue, 3 Oct 2000 21:36:52 -0700 I'm new to Squeak so this may sound like a stupid question, what does this program do? Adam > >Remember, you asked for it... > >There's really not much to Part I. > >Simply define... >!PolygonMorph methodsFor: 'drawing' stamp: 'di 10/3/2000 09:50'! >drawArrowOn: aCanvas at: endPoint from: priorPoint > "Draw a triangle oriented along the line from priorPoint to endPoint." > | pts | > pts _ self arrowBoundsAt: endPoint from: priorPoint. > (self valueOfProperty: #arrowSpec ifAbsent: [5@4]) y >= 0 > ifTrue: [aCanvas drawPolygon: pts fillStyle: borderColor] > ifFalse: [aCanvas drawPolygon: pts fillStyle: Color transparent > borderWidth: borderWidth+1//2 borderColor: borderColor]! ! > >This actually gives you transparent arrowheads, because... > >When you invoke 'customize arrows', it will record an arrowSpec that has x >and y values that only make sense in a positive dierection. The negative x >is already used to say that you want concave bases on your arrows. The >above solitary change says that if you go negative in y, then it means the >same about width (which already takes the abs), but that the center should >be transparent. Voila! > >There is a method, > computeArrowFormAt: endPoint from: priorPoint >which I thought I would have to change as well, but in fact it is not used >for drawing, but mainly for bounds and hit testing (conatinsPoint:), in >which it is fine for it to continue to think that the arrows are opaque. > >Now, here is what remains to be done to do the job right (Part II) in my >opinion... >If you look at the results with the above change, you will see two main >problems: > > 1. The arrow points look yukky because they are made with wide lines, and > > 2. The shaft of the line intrudes into the transparent heart of the >arrow. > >(2) is relatively simple to address. You will find (copied in two places >-- boo!) code which shortens the last line segment so that it does not >interfere with the arrow tips at present. > >The methods are drawBorderOn: and drawDashedBorderOn:, in each of which you >will see I have inserted the following ugly logic: > > arrows == #none ifFalse: > ["Shorten line ends so as not to interfere with tip of arrow." > ((arrows == #back or: [arrows == #both]) and: [p1 = vertices first]) > ifTrue: [v _ p1 - p2. > d _ Point r: v r - (self borderWidth*3) degrees: v degrees. > p1i _ (p2 + d) asIntegerPoint]. > ((arrows == #forward or: [arrows == #both]) and: [p2 = vertices last]) > ifTrue: [v _ p2 - p1. > d _ Point r: v r - (self borderWidth*3) degrees: v degrees. > p2i _ (p1 + d) asIntegerPoint]]. > >Your first assignment is to factor these two methods, or at least the above >code, so that it isn't quadruplicated (it's on my to do list, but you don't >want to go there...). > >Having done that, you need to look in the code for arrowBoundsAt: endPoint >from: priorPoint, in which it will be revealed exactly where the "base" of >the arrow is (namely "wingBase"). So the above code needs to be refined so >that the length of the line ends will be shortened, not just to r - (self >borderWidth*3), but to end at the windBase. Of course you will have to >max: with 0, since, on a curve, the last segment may be quite short, so >lots of this logic may fail. If you make it work for a sharply curved line >(which will involve interfering with more than the last segment), you get a >silver star. > >OK, so much for the shaft intruding into the heart of the arrow. > Now how about that ugly shape made with a thick border? >Here's what I would do... > >I would add a special method to the protocol of canvas for arrowheads with >some sort of flag for transparent centers. In fact you should be able to >use existing methods to accomplish it, but what has to happen is that you >need to supply not only the points on the outside of the arrow (already >available from arrowBoundsAt: endPoint from: priorPoint), but also the >points on the inside of the arrow (you could probably just scale them all >about a center), and cause the resulting border to be painted as a shape. >This will give you sharp exterior points as well as sharp interior points. >When you're done, you can crow to the Squeak list, and Ned can tackle his >UML editor. [You don't get to really crow until you verify that it does >the right thing in Postscript as well as on the screen.] > >Have fun! > - Dan > > _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. Share information about yourself, create your own public profile at http://profiles.msn.com.

2 1

[ENH] preliminary AttachmentPointMorph stuff
by Ned Konz 04 Oct '00

04 Oct '00

Here's where I am so far. The behavior of the AttachmentPointMorphs is not yet pluggable. See what you think, though... This looks better with the new improvements to the arrowheads that Dan sent out. -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

1 2

More Information on Squeak Textbook
by Mark Guzdial 04 Oct '00

04 Oct '00

I've had several questions about my Squeak textbook, with suggestions that I post the answers to the list. Again, the book is at: http://www.amazon.com/exec/obidos/ASIN/0130280283/o/qid=970306184/ (Currently ranked in sales at 38,828! (With a bullet? :-)) The publisher (which I can't believe Amazon doesn't list!) is Prentice-Hall. My editor is Alan Apt. Both the Table of Contents and Back cover text for the book are WRONG. <sigh> (I've alerted P-H, and they're working on it.) Below is the actual back cover contents and the manuscript table of contents (which is changing some -- copyeditors are pretty amazing :-). The cover for the textbook and the edited volume (http://coweb.cc.gatech.edu/squeakbook) will be the same image, with altered color schemes. Maybe I shouldn't spoil the surprise about the cover images -- let's just say that they're wonderful and entirely appropriate for Squeak. Rick Zaccone of Bucknell, Rik Fischer Smoody of Portland State, and Cullen O'Neill of U. Michigan all test-drove the book this last year. Let me know if you have any other questions about it. Mark Squeak: Object-Oriented Design with Multimedia Applications Mark Guzdial "In Squeak, you have your hands in one of the most late bound, yet practical, programming systems ever created. It is also an artifact which is wide, broad, and deep enough to permit real scientific study, creation of new theories, mathematics, engineering constructions, and even its own successor! Please use this book to learn how to do "all of the above." - Alan Kay, V-P Research, Disney Imagineering "The text presents OO concepts in the context of Smalltalk (Squeak), which is by far the best environment and language in which to learn the concepts. The author is clearly enthusiastic about the subject matterauthoritative in explaining the basics of Smalltalk. - Keith R. Pierce, St. Paul Companies This innovative text teaches object-oriented programming using Squeak - a powerful, freely available language that is highly effective for teaching object-oriented programming. Using a focus on the basics and compelling case studies, the text covers the whole process of object-oriented development - from object-oriented analysis and design, though user interface design and application evaluation. FEATURES/BENEFITS Free Squeak CD with book ** CD includes version 2.7 of Squeak compatible with Windows 95/98/NT, WindowsCE, Macintosh, Linux, Solaris, SGI, Acorn, BeOS. ALSO includes source code for examples, Squeak extras, and lecture notes in PowerPoint and Squeak-native formats. Focus on Squeak ** Gives both professionals and students great flexibility in working in an exciting environment that supports the latest in multimedia, including 3-D computer graphics, MIDI audio, and Flash animations. It is the first printed documentation for Squeak's advanced features, Morphic user interface, and support for multimedia. Rich description of object-oriented process. ** Introduces readers to an object-oriented design process, using UML standard class diagrams. A user interface design and evaluation process. ** Leads users through constructing user interfaces, then designing and evaluating user interfaces. Case studies present the design and implementation of exciting multimedia applications, then follow-up with real evaluation results from interaction with users. ABOUT THE AUTHOR Mark Guzdial is an Associate Professor in the College of Computing at Georgia Institute of Technology. His research focuses on learning sciences and technology, specifically, "collaborative Dynabooks," which are tools that support learning through collaborative multimedia construction. He has been teaching object-oriented analysis, design, and programming in Smalltalk for over five years. His Ph.D. is in Education and Computer Science from the University of Michigan. Georgia Tech praised his teaching with an award for "Outstanding Use of Educational Technology" (1997). He has received a prestigious National Science Foundation CAREER Award. He is on the editorial board of IEEE Multimedia, Journal of the Learning Sciences, Journal of Interactive Learning Research, and Interactive Learning Environments Journal. TABLE OF CONTENTS FOREWORD SOFTWARE: ART, ENGINEERING, MATHEMATICS, OR SCIENCE? BY ALAN KAY 10 PREFACE 1 1 APPROACH OF THE BOOK 1 1.1 START FROM WHERE THE STUDENTS ARE 2 1.2 LEARNING INVOLVES TESTING AND FAILURE 4 1.3 GENERATION AND INQUIRY, NOT TRANSMISSION 5 2 CONTENT OF THE BOOK 5 3 ACKNOWLEDGEMENTS 6 PART 1 FOUNDATIONS OF OBJECT-ORIENTED PROGRAMMING AND SQUEAK 8 CHAPTER 1 OBJECTS, SMALLTALK, DYNABOOKS, AND SQUEAK: WHERE THE OBJECTS COME FROM 9 1 OBJECT-ORIENTED PROGRAMMING 9 2 BIRTH OF OBJECTS 10 3 "A PERSONAL COMPUTER FOR CHILDREN OF ALL AGES" 13 4 BACK TO THE FUTURE 18 5 COMMON ANCESTRY OF OTHER OBJECT-ORIENTED LANGUAGES 21 CHAPTER 2 A TOUR OF SQUEAK 24 1 BASIC RULES OF SMALLTALK 24 2 DOING "NORMAL" THINGS IN SQUEAK 27 2.1 VARIABLES AND STATEMENTS 28 2.2 CONTROL STRUCTURES 30 2.3 LITERALS, NUMBERS, AND OPERATIONS 33 2.4 STRINGS AND ARRAYS 34 2.5 FILES 38 3 DOING "OBJECT" THINGS IN SQUEAK 38 3.1 BLOCKS 39 3.2 VARIABLES AND MEMORY 40 3.3 CREATING CLASSES, INSTANCES, AND METHODS 41 3.4 THE SQUEAK MODEL OF EXECUTION 43 4 USING SQUEAK 45 4.1 STARTING A NEW PROJECT 48 4.2 EXTENDED EXAMPLE: MUPPETS IN SQUEAK 50 5 TOOLS AND STRATEGIES: FINDING MORE IN SQUEAK 57 5.1 FINDING CLASSES: THERE HAS GOT TO BE A WINDOW OR A TEXTFIELD CLASS AROUND HERE SOMEWHERE. WHERE IS IT? 57 5.2 EXPLORING OBJECTS: I'M EXPLORING A COMPLICATED OBJECT, AND NOW I'VE GOT A BAZILLION INSPECTORS ALL OVER THE SCREEN. IS THERE SOME OTHER WAY TO EXPLORE AN OBJECT? 59 5.3 FINDING METHODS: I REMEMBER THERE'S A WAY OF GETTING AN ELEMENT, 'AT'-SOMETHING, BUT I CAN'T FIND IT. 60 5.4 FINDING IMPLEMENTORS: I SEE A REFERENCE TO A METHOD NAMED FINDTOKENS: (OR WHATEVER). WHERE IS IT? WHAT DOES IT DO? 62 5.5 FINDING SENDERS: THAT'S WHAT FINDTOKENS: DOES. WHO USES IT? 63 5.6 DEBUGGING: I CAN'T FIGURE OUT WHAT MY CODE IS DOING! I'VE GOT AN ERROR, BUT IT MAKES NO SENSE TO ME. 64 5.7 LEARNING TO USE A CLASS: WHAT ALL DOES STRING OR OTHER CLASSES UNDERSTAND? 66 5.8 THERE SEEM TO BE A HUGE NUMBER OF COMMAND KEYS. IS THERE A LISTING SOMEWHERE OF ALL OF THEM? 67 6 HOW DO YOU MAKE AN APPLICATION IN SQUEAK? 68 CHAPTER 3 YOUR FIRST PROGRAM: JOE THE BOX 71 1 ADELE GOLDBERG'S JOE THE BOX 71 2 TOOLS AND STRATEGIES: FILING IN NEW CODE 71 3 PLAYING WITH BOXES 73 3.1 JOE AND JILL AS EXAMPLE OBJECTS 76 3.2 ADDING A NEW KIND OF BOX 77 4 CREATING THE BOX CLASS AND BOX INSTANCES 79 5 BASICS OF DRAWING 81 5.1 CREATING FORMS 81 5.2 TEACHING BOXES TO DRAW 84 5.3 GETTING INPUT FROM THE USER 86 6 EXTENDING BOX TO CREATE NAMEDBOX 87 7 TOOLS AND STRATEGIES: HOW TO GO FROM "SAMPLE CODE" TO "REUSE" 90 8 IMPROVING BOXES: EFFICIENCY, ANIMATION, AND DESIGN 93 8.1 DRAWING BOXES BETTER 93 8.2 ANIMATING BOXES 95 8.3 DESIGNING BOXES 96 CHAPTER 4 DESIGNING OBJECT SYSTEMS 99 1 THE OBJECT-ORIENTED DESIGN PROCESS 99 1.1 OBJECT-ORIENTED ANALYSIS STAGE 100 1.1.1 Brainstorming Candidate Classes 100 1.1.2 Class Responsibility Collaboration (CRC) Cards 102 1.2 OBJECT-ORIENTED DESIGN STAGE 104 1.2.1 UML Tools 105 1.2.2 Relationships between Classes 105 2 YOUR FIRST DESIGN: A CLOCK 107 2.1 DOING IT QUICKLYAND WRONG 107 2.2 OBJECT-ORIENTED ANALYSIS OF THE CLOCK 108 2.2.1 Brainstorming a Clock 109 2.2.2 CRC Cards for a Clock 109 2.3 OBJECT-ORIENTED DESIGN OF A CLOCK 113 2.3.1 Considering An Alternative Design 115 2.4 OBJECT-ORIENTED PROGRAMMING FOR THE CLOCK 116 2.4.1 Implementing SecondsTicker 117 2.4.2 Side Trip on Processes 118 2.4.3 Implementing the Clock Class 119 3 SPECIALIZING CLOCK AS AN ALARMCLOCK 122 3.1 OOA FOR ALARM CLOCK 122 3.2 OOD FOR ALARM CLOCK 123 3.3 OOP FOR THE ALARMCLOCK 124 4 REUSING THE CLOCK AND ALARMCLOCK 126 4.1 REUSE IN A VCR 126 4.2 REUSE IN AN APPOINTMENTBOOK 129 4.2.1 Programming the Appointment Book 131 4.2.2 Implementing the AppointmentBook with Collections 131 4.2.3 Implementing the Appointments 133 5 IMPLEMENTING MODELS 134 6 RULES OF THUMB FOR GOOD OBJECT-ORIENTED DESIGNS 136 6.1 RECONSIDERING JOE 137 7 TOOLS AND STRATEGIES: PROGRAMMING IN GROUPS 137 7.1 CREATING PROJECTS AND CHANGE SETS 138 7.2 WORKING WITH SOMEONE ELSE'S CHANGES 139 7.3 RECOVERING FROM SOMEONE ELSE'S CHANGES 140 CHAPTER 5 BUILDING USER INTERFACES IN SQUEAK 143 1 ISSUES IN BUILDING A USER INTERFACE 143 2 DEVELOPING MODEL-VIEW-CONTROLLER 144 2.1 ROUND 1: NO SEPARATION AT ALL 145 2.2 ROUND 2: SEPARATE WINDOWS AND BUTTONS 150 2.3 ROUND 3: SEPARATING EVEN THE TEXT 157 2.4 STRENGTHS AND WEAKNESSES OF MODEL-VIEW-CONTROLLER 163 3 BUILDING PLUGGABLE USER INTERFACES IN SQUEAK 165 3.1 CREATING PLUGGABLE INTERFACES IN MVC 167 3.2 CREATING PLUGGABLE INTERFACES IN MORPHIC 171 3.2.1 Introducing Morphic 171 3.2.2 Doing Pluggable Components in Morphic 176 3.3 MENUS AND DIALOGS IN A PLUGGABLE WORLD 181 4 BUILDING MORPHIC USER INTERFACES 183 4.1 PROGRAMMING MORPHS FROM THE VIEWER FRAMEWORK 184 4.1.1 Adding an Instance Variable 186 4.1.2 Making our Ellipse Fall 188 4.1.3 Building the Kicker 190 4.2 PROGRAMMING BASIC VARIABLES AND EVENTS OF MORPHS 192 4.2.1 Instance Variables and Properties 193 4.2.2 Morphic Events 194 4.2.3 Animation 196 4.2.4 Custom menus 196 4.2.5 Structure of Morphic 197 4.3 PROGRAMMING A MORPHIC FALLING OBJECT 198 4.3.1 Implementing the Falling Object 200 4.3.2 Implementing the Kicker 201 4.3.3 Running the Text Falling Simulation 202 4.3.4 Changing the Gravitational Constant 202 5 TOOLS AND STRATEGIES: USING MORPHS THAT YOU HAVEN'T MET YET 203 CHAPTER 6 DESIGNING USER INTERFACES IN SQUEAK 206 1 KNOW THY USERS FOR THEY ARE NOT YOU 206 1.1 HOW DO YOU DECIDE BETWEEN USER INTERFACE ALTERNATIVES? 207 2 UNDERSTANDING THE USER 208 2.1 UNDERSTANDING THE TASK 209 3 MATCHING USERS TO INTERFACE: AVOIDING USER ERROR 210 4 A USER INTERFACE DESIGN PROCESS 212 5 CRITIQUING OUR CLOCK INTERFACE 214 6 EVALUATION OF USER INTERFACES 215 6.1 EVALUATION BEFORE USER INVOLVEMENT 215 6.2 EVALUATION WITH USERS 216 6.3 EVALUATING GROUPWARE 219 CHAPTER 7 MULTIMEDIA NUTS-AND-BOLTS 221 1 TEXT 221 1.1 EXPLORING SQUEAK'S TEXT 221 1.2 PROGRAMMING SQUEAK'S TEXT 223 2 GRAPHICS AND ANIMATION 225 2.1 SIMPLE GRAPHICS TOOLS 225 2.2 PROGRAMMING SIMPLE GRAPHICS 226 2.3 3-D GRAPHICS FOR END-USERS 228 2.3.1 Programming Wonderland without the Script Editor 232 2.4 FLASH MOVIES 233 3 SOUND 235 3.1 RECORDING, VIEWING, AND EDITING SOUND 235 3.2 SOUND CLASSES 238 3.3 MIDI SUPPORT 240 3.3.1 MIDI Support Classes 243 3.4 SPEECH IN SQUEAK 244 4 NEW MEDIA IN SQUEAK 245 4.1 SQUEAKMOVIES 245 4.2 BOOKMORPHS 248 5 MAKING THE DYNABOOK IN SQUEAK 250 PART 2 CASE STUDIES 252 CHAPTER 8 CASE STUDY: AUDIO NOTES 253 1 MOTIVATION FOR AUDIO NOTES 253 2 USING AUDIO NOTES 253 3 DETAILS ON USE 254 4 OBJECT-ORIENTED DESIGN AND PROGRAMMING OF AUDIO NOTES 255 4.1 BUILDING THE USER INTERFACE 256 4.2 HANDLING THE NOTES LIST 258 4.3 IMPLEMENTING THE LIST UI 259 4.4 SETTING UI OPTIONS 260 4.5 RECORDING AND PLAYING SOUND 260 5 EVALUATING THE DESIGN AND INTERFACE 264 5.1 HEURISTIC UI EVALUATION 265 CHAPTER 9 CASE STUDY: PLUGGABLE WEB SERVER AND SWIKI 268 1 NETWORKING SUPPORT IN SQUEAK 268 1.1 USER INTERFACES FOR NETWORKING 268 1.2 PROGRAMMING NETWORK ACCESS 270 1.3 WEB SERVING WITH PWS 271 1.3.1 Programming the PWS 272 2 SWIKI AS A PWS APPLICATION 275 2.1.1 Uses of the Swiki 277 2.2 OBJECT DESIGN OF SWIKI 277 2.2.1 Responsibilities of Swiki Classes 278 2.2.2 Critiquing Swiki's Object Design 281 2.2.3 Evaluating Swiki's Usability 281 CHAPTER 10 CASE STUDY: MAT (MULTIMEDIA AUTHORING TOOL) 284 1 MOTIVATION FOR AND USE OF MAT 284 2 OBJECT DESIGN AND PROGRAMMING OF MAT 287 2.1 IMPLEMENTING MAT 289 2.1.1 Drag-and-Drop into MAT 290 2.1.2 Handling the Menu in MAT 291 2.1.3 Generating HTML in the ParagraphEditor 291 3 USER EVALUATION OF MAT 298 CHAPTER 11 CASE STUDY: PROTOTYPING A PLAY-WRITING WORKBENCH 302 1 MOTIVATION AND SETTING THE GOAL 303 2 OBJECT ANALYSIS OF THE WORKBENCH 304 3 IMPLEMENTATION OF THE PLAY WORKBENCH 306 3.1 PROTOTYPE SCRIPTINGBENCH AND SCRIPTINGCONTROLS 306 3.2 IMPLEMENTING THE ACTORS 308 3.3 IMPLEMENTING PLAYING THE SCRIPT 310 4 ITERATING ON THE PLAYBENCH 312 5 EVALUATING THE PLAYBENCH: EXPERT EVALUATION AND OBSERVATION 314 5.1.1 Summarizing the Evaluation 316 APPENDIX A QUICK START GUIDE OF SQUEAK EXAMPLES 319 1 CONSTANTS 319 2 CONTROL STRUCTURE EXPRESSIONS INVOLVING BLOCKS 319 3 TESTING OBJECTS 320 4 BASIC OBJECT BEHAVIOR 320 5 BOOLEAN OPERATIONS 320 6 MAGNITUDE AND NUMBER OPERATIONS 321 7 CHARACTER OPERATIONS 322 8 COLLECTION OPERATIONS 322 9 STRING OPERATIONS 323 INDEX 325 -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 Associate Professor - Learning Sciences & Technologies. Collaborative Software Lab - http://coweb.cc.gatech.edu/csl/ (404) 894-5618 : Fax (404) 894-0673 : guzdial(a)cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html

1 0

Paper on Using Squeak to teach UI
by Mark Guzdial 04 Oct '00

04 Oct '00

I wrote a paper on how I use Squeak to teach UI (with a little on how we use the CoWeb), including some empirical evidence on how the new Squeak-based approach has improved student performance on MVC. Basically, the idea is to build a UI toolkit from scratch, polling Sensor and writing to Display -- an approach which is near impossible in most other environments, and which seems to result in better student learning. The paper was submitted to ACM's SIGCSE 2000 conference, and this version will be a GVU TR, #GIT-GVU-00-17. FYI, the paper is available at: http://guzdial.cc.gatech.edu/papers/sigcse00/sigcse-squeak.pdf http://guzdial.cc.gatech.edu/papers/sigcse00/sigcse-squeak.ps Mark -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 Associate Professor - Learning Sciences & Technologies. Collaborative Software Lab - http://coweb.cc.gatech.edu/csl/ (404) 894-5618 : Fax (404) 894-0673 : guzdial(a)cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html

1 0

[FIX][ENH] MemoryHogs revisited
by Georg Gollmann 04 Oct '00

04 Oct '00

Somewhere in the 2.9a upgrade stream the memory hogs mechanism was disabled (but not fully removed). This fileIn reintroduces it in a slightly robuster way. Georg -- ---- Dipl.Ing. Georg Gollmann TU-Wien, Zentraler Informatikdienst Wiedner Hauptstr. 8-10 phon:(+43-1) 58801 - 42022 A-1040 Wien fax: (+43-1) 58801 - 42099 mail:gollmann@zid.tuwien.ac.at http://macos.tuwien.ac.at/Gollmann.html

1 0

Re: VM source stucture
by Tim Rowledge 04 Oct '00

04 Oct '00

Edward P Luwish <eluwish(a)uswest.com> is widely believed to have written: > The class InterpreterSupportCode currently contains only the Mac-specific > files. I suggest that we create subclasses for each platform and maintain > the port-specific code and build instructions as part of the Squeak sources > (or at least, allow for platform-specific fileins). I kindof agree, but it would be some pretty big literal strings to bung about. I think there is a reasonable argument for actually getting rid of the Mac stuff and having a minimalistic exemplar pseudo-platform. I realised while trolling through Ian's latest makefile stuff that it ought to be possible to (ab)use his idea of (from the README):- > The directory structure is as follows: > > [A-Z]*/ - hand-written platform-independent plugin source > generated/ - automatically generated VM source > generated/*Plugin.c - automatically generated plugin source > generated/LargeIntegers.c - automatically generated plugin source > unix/ - hand-written Unix-specific VM source > unix/[A-Z]*/ - hand-written Unix-specific plugin source > > Note that hand-written plugin source files must be in directories > whose names begin with a capital letter, and generated plugins in > files called "generated/*Plugin.c", in order for configure to > recognise them. (With the exception of "generated/LargeIntegers.c", > since there always has to be someone who likes to make life difficult > by being different.) ... such that FilePlugin (for example) could be actually one of those [A-Z]* directories despite the fact that it is not handwritten nor entirely platform independant. Indeed, most plugins are a mix of generated, portable, hand-written etc. It may even be possible to understand enough of the magical auto-make/config/whatever to make it work with a sensible structure. tim -- Tim Rowledge, tim(a)sumeru.stanford.edu, http://sumeru.stanford.edu/tim Useful random insult:- Wise as the world is flat.

2 2

Detecting Circular Lists
by Andrew P. Black 04 Oct '00

04 Oct '00

A friend recently sent me the following email: >Andrew, > >I remember a really cute algorithm that detect when a linked list >traversal has gone circular (due to a bad entry on the list). It involved >using a trailing pointer and allowing the trailing pointer to fall further >behind as the list was traversed. > >Do you happen to have a reference to that algorithm? Well, I had never heard of this algorithm, but given the general idea, it took me only a couple of minutes to reconstruct it on the whiteboard. My variant moves the tailing pointer forward one link every time the traversal pointer moves forward two links. Tim Sheard, a colleague here at OGI, told me that he thought that the algorithm was originally due to Dijkstra, but that he thought that he had seen it in CommonLisp, in the form of a predicate cyclic, defined something like overlaps a b = if (eq a b) then true else overlaps (next a) (next (next b)) cyclic a = overlaps a (next a) Needless to say, there is no such function in the CommonLisp manual. Just for grins, I implemented this test in Squeak as part of a LinkedList method #do:ifCyclic: (ChangeSet attached). Not that Squeak makes much use of linked lists ... but if I need this someday, I want it written down! I would still be interested in an original reference. Andrew

1 0

[FIX] Trivial fix to Link class
by Andrew P. Black 04 Oct '00

04 Oct '00

The Link class method #nextLink:, which should answer a new instance of Link, instead answered nil. The fix is the obvious addition of 'yourself'. The one place that should have used this method -- SystemDictionary>>useUpMemory -- carefully programmed around this bug, so I changed that code so that it now uses #nextLink: in the natural way. Andrew

1 0

[BUG] PluggableTest class open
by Andrew P. Black 03 Oct '00

03 Oct '00

While playing with PluggableLists, I came across PlugableTest class, which contains an example of how to use PluggableViews. So, I, tried the example "PlugableTest open". SAVE YOUR IMAGE BEFORE YOPU TRY THIS. In Morphic, it hangs the image. Command-period brings up a notifier, but I can't get a debugger, or anything else. In MVC, it works just fine. This is in a vanilla 2.8 image and 2.8 VM. Andrew

1 0

← Newer
1
...
36
37
38
39
40
41
42
43
Older →

Jump to page: