Squeak-dev October 2005

squeak-dev@lists.squeakfoundation.org

180 participants
296 discussions

veryDeepCopy, ImageSegments, and actionMaps
by Ned Konz 17 Aug '23

17 Aug '23

Any Object can have dependents whether or not it has an instance variable to keep track of them (these are kept in the DependentsFields class variable of Object). The only class that keeps its own dependents is Model. Any Object can have an actionMap. These are supplied by EventManager class by default, and are stored in the ActionMaps class variable of EventManager. The only class that keeps its own action maps is Morph. When we veryDeepCopy an object we also copy the global dependents (see DeepCopier>>fixDependents). However, we don't copy the actionMap in those cases where it's held by the EventManager class. When we file out an object that uses DependentsFields out into an ImageSegment, we also file out the dependents (see ImageSegment>>dependentsSave:) if it's in a project. We do this by adding a project parameter called #GlobalDependentsInProject. However, we don't file out the actionMap in those cases where it's held by the EventManager class. My question is this: should we fix these behaviors (copying and serialization) to include these "global action maps" just like we already include the "global dependents"? Or should we discourage the use of the "global action maps"? -- Ned Konz http://bike-nomad.com GPG key ID: BEEA7EFE

3 2

Trying to build SerialPlugin for IrDA under MPW (MacOS 8.6)
by Hans-Martin Mosner 24 May '12

24 May '12

Hi, I recently got a serial IrDA adapter and wanted to play around with it under Squeak. This adapter has a baudrate switching mechanism which depends on toggling the DTR and RTS lines in the correct sequence. To support this, I wanted to build a new SerialPlugin with an additional primitive. However, the plugin as built with MPW does not seem to be accepted by the VM, which still uses the internal plugin in all cases. I've tried placing the plugin in a 'Plugins' subfolder of the VM's folder, and then directly into the VM folder, without any success. I have considered placing debug outputs in appropriate places in the VM, but building a complete VM with MPW is not easy - I would have to create an MPW makefile for the whole shebang, which is more work than I can invest at the moment. Does any of you Mac aficionados have an idea where the problem could be? I have appended the plugin and corresponding source files in a ZIP file (it's not big) so if you want to try it out you could do it right now :-) BTW, would it make sense to incorporate support for switching serial port parameters of an open port into the standard plugin? Closing and opening a port just to change the baud rate seems a little overkill... Cheers, Hans-Martin

6 14

hash method
by Houssam Fakih 31 Jan '07

31 Jan '07

Hi, When I redefine the = operation for a class I should redefine its hash method. My question is about the definition of the hash method. I read that the method must return a SmallInteger whose value is related to the receiver. I would like to know if there is a general pattern adopted in Squeak for that. I get a look in the Squeak classes but it doesnt help. Thanks, Houssam ____ "Le paradis touche les pieds des mères..." >>> http://csl.ensm-douai.fr/fakih

4 4

RE: Progrmaming in Bytecode?
by Swan, Dean 24 Nov '06

24 Nov '06

>> It's not uncommon to find x86 code padded with a lot of NOPs >> that weren't in the original source. > >Cowardly assembler! ;-) > >> [NOPs...] Sometimes they're there to pad the execution pipeline, and >> sometimes it's just because the code generator allocates a whole large >> enough for the largest form of an instruction to simplify the branch >> target calculations. > >We must be talking about M$oft assemblers here. ;^p) Yes. This has been pretty characteristic of MS assemblers as far back as I can remember (i.e. about MASM 2.0 - I remember being quite excited when MASM 4.0 came out - '186 and '286 code!). >The real problem is that the solution in general might require N >iterations for a piece of code with N branches -- if promoting one of them >from short to long causes the next (or, much more problematically, a >preceding) one to become long too. You're such an optimist. O(n) is close to best case for this problem. >This is a weak form of what I suppose one might call an "oscillating >constraints" problem. (Hands up anyone who has ever seen [La]TeX get Hence, stochastic optimization. Even MS uses stochastic optimizers now, even in vb. (didn't want to say that too loud here. ;) ) >The usual cause for "padding" in (capable) assemblers (other than for >prefetch alignment, as you point out) is due to dynamic linking where the >size of call sites isn't known until link time. (On modern systems the Not only prefetch alignment. I have heard, although I don't know the details that sometimes it's desirable to force a dead cycle in the execution pipeline. I can only guess that this would be to influence the scheduling of the execution units for complex instructions, but I consider myself to be only a "casual observer" with respect to this kind of thing. >In Smalltalk you've also got a bigger problem: for very long methods the >range of even unconditional branches might not be enough, at which point >I know this has nothing whatsoever to do with Smalltalk...) Well, to kind of tie this back into something relevant to Squeak - I know a few people have talked about "new" VM instruction sets. It seems that it might help to think about a VM where all instructions are the same size. This could simplify code generation (i.e. compiling methods), and possibly improve performance and *reduce* code size (at least in the case of long methods). The anticipated performance improvement would come from forcing aligned fetches more often and potentially simpler instruction decoding. It's really nice to work with an architecture where all instructions take up one word of memory, take one cyle to execute, and sometimes can cause five or six different useful things to occur during that one cycle. (ok, so I write a lot of DSP code...) The single cycle thing isn't necessarily relevant for Smalltalk, but those "extended" bytecodes are really distasteful to me. -Dean

10 15

[ANN] Audio Video Library
by Martin Kuball 07 Oct '06

07 Oct '06

Hi I would like to annonce my project to build a Library for processing audio and video material with squeak. I put a 1st version on the SqueakSource at kilana.unibe.ch. It's not much yet. But you can load and analyze avi container with mpeg 4 video and mp2 and mp3 audio content. I did this because I needed a tool to repair some broken avi files. Other tools I tried just failed badly. Try this: RIFFArchive openOn: 'test.avi' "use explore it to get a nice tree view of the chunks in the avi" "open a transcript and execute the following to get some information about the contents" (RIFFArchive openOn: 'test.avi') analyzeVideoStream "To extract the streams inside the avi container do:" (RIFFArchive openOn: 'some.avi') extractAllStreamsToFile: 'base_fileName'. "If you have raw streams you can create an avi this way:" | avi | avi := AVIFile with: (M4VStream on: 'video.m4v') with: (MPAStream on: 'audio.mp3'). avi runtime: 100. avi frameRate: 24000 / 1001. avi createIndex. "not working yet" avi writeTo: (StandardFileStream newFileNamed: 'test.avi') Here is my tentative todo list. - write more documentation and tests - cleanup and speedup - add more container formats like mp4, matroska - add mpeg2 handling (-> dvd authoring???) - connect with the other multimedia bits in squeak (sound, video playing, ...) The project is of course open for everyone to join. Enjoy, Martin

5 6

[ANN] SqueakLight status
by Lic. Edgar J. De Cleene 07 Dec '05

07 Dec '05

People who could have interest on Small Images I actualize http://minnow.cc.gatech.edu/squeak/3935 So now the gizmo is smaller (4.8Mb) and powerful what never before, having some of MorphicWrappers built inside. I wish thanks special this time Francisco Garau last port of original work by Luciano Notarfrancesco and Gerardo Richarte. Edgar ___________________________________________________________ 1GB gratis, Antivirus y Antispam Correo Yahoo!, el mejor correo web del mundo http://correo.yahoo.com.ar

3 5

[BUG] SM2 Burps after loading Breakpoint support or ColorDevTools
by Thomas A Petersen 05 Dec '05

05 Dec '05

Herman, Goran, et. al, I don't think your Jacaranda script is the problem. I tried the same exercise from a clean 3.7 5566 image and loading SM2 Loader first thing. I get the same crash with Breakpoint Support and ColorDevTools. Checking the change sorter shows that things are installed. There appears to be something going on as SM2 finishes things. Goran, Oh Great Sage of SM.... Can you enlihton us? Thanks much, tap P.S. Here's the debug log. Could this be a side effect of the translation stuff?? ----------------------------------------------------- MessageNotUnderstood: Character>>first 1 December 2003 3:22:52 pm VM: Win32 - Squeak3.5 of '11 April 2003' [latest update: #5180] Image: Squeak3.7alpha [latest update: #5566] Character(Object)>>doesNotUnderstand: #first Receiver: $5 Arguments and temporary variables: aMessage: a Message with selector: #first and arguments: #() Receiver's instance variables: value: 53 SMSqueakMap>>installedVersionOf: Receiver: a SMSqueakMap Arguments and temporary variables: aPackage: SMPackage[Breakpoint Support] Receiver's instance variables: packages: a Dictionary(an UUID ('010894c3-0b8a-47fc-8deb-7c08996bfa71')->SMPackage[OSProcessPlugin] an UUID('017b1882-9376-4bf7-b...etc... accounts: a Dictionary(an UUID ('034ab782-a939-de49-976c-b009701b4406')->SMAccount[Todd Blanchard] an UUID ('035d2bda-c301-6049-b...etc... objects: a Dictionary(an UUID ('007e8c1c-02fa-45fc-97be-7ea4d2f67134')->SMCategory: Package group an UUID ('008c9940-4caf-ac49-a...etc... categories: nil dir: 'sm' transactionCounter: 3011 firstTransactionNumber: 2929 daysBacklog: 182 adminPassword: nil installedPackages: a Dictionary(an UUID ('07469ced-289a-4b6f-bd17-27ec567e681a')->'1.25' an UUID ('1200100f-1a95-...etc... installCounter: nil fileCache: a SMFileCache users: nil --- The full stack --- Character(Object)>>doesNotUnderstand: #first -----------------------------------------------------

5 4

[ANN] Bootstrapped kernel image
by Pavel Krivanek 26 Nov '05

26 Nov '05

Hi all, I have prepared an experimental kernel image based on Squeak 3.7. Properties: - size: 1.3 MB - no Morphic - no MVC - no Sound - no Network - no Browser, Debugger etc. - no SUnit - no Projects support (there’s only one default project) - only necessary processes - no ObsoleteClasses (consistent references in sources) - single font - simple console with clipboard support - transcript output (with file log) - transparent creation mechanism Download: - image with changes: ftp://comtalk.net/pub/bootstrap/kernel.zip - bootstrapping environment: ftp://comtalk.net/pub/bootstrap/Bootstrap.zip - please, can somebody place it to a new location with better connectivity? How to prepare it: - use QuiteSmall image (http://www.swerlingphoto.com/squeak/QuiteSmall-3.7.zip) - load Mirrors.1.cs - load Bootstrap.st - evaluate "Bootstrap start" - in console type "Bootstrap revert" - wait a while - condense changes, clear transcript, save image How it works: - for all classes in system "mirrors" are created. Mirrors are classes with prefix "Xxx" that contain only class definition and methods. They contain no valid references to class variables. - mirrors are modified so they contain whole class hierarchy for new system. There are original classes used in code of methods. - initial cleanup is executed and system processes are recreated - for all original classes, methods of mirrors are assigned. Methods that aren’t in mirrors are removed. Then the new definition of class is used. - unnecessary original classes are removed - mirrors are removed - whole sources are recompiled - additional cleanup is executed The console is very simple. Now it cannot simulate FillInTheBlank, confirmations, user interrupt etc. Feel free to clean, improve and fix this image. I have to say that I’m VERY tired of creation of this image now and I need a pause :-) -- Pavel

7 14

Henrik Gedenryd has passed away
by goran＠krampe.se 23 Nov '05

23 Nov '05

Dear Squeakers, I don't recall why I started looking around the Internet for information about Henrik the latest week - I have always had a strange feeling about him so abruptly leaving the Squeak community but thought he had just moved on to other frontiers. I have now verified (99.9%, it is a very uncommon name in Sweden) with the Swedish IRS that Henrik Gedenryd in fact passed away in november 2002 at the age of 32. I recall he was working in England at the time and in an email he sent me just a month before he said he was moving back home to Sweden in november. I have no idea what happened. Henrik started participating in the Squeak community back in 1999, I think his first post is from september that year. IIRC he first "fixed a few things" like for example getting beautiful fonts and a "nice look" before he duyg into other things, like Modules. He helped Dan and SqC at the time to create Squeak 3.3 - with support for Modules being the main feature. He put in a great effort but due to various factors 3.3 never came to be a success - but I am sure it was not due to lack of quality (I have read his code and it is very good), rather the fact it was too "different" from 3.2 and people simply felt uneasy - so they didn't "move into the new house". And a Squeak version that noone "lives in" just dies. We exchanged quite a bit of emails and the original intention with SqueakMap was actually to act as a complementary structure on top of Henrik's Modules system. Henrik and I met in person in 2001 when we visited OOPSLA together and shared a hotel room. I found it very interesting to hear his insights and I realized that, hey, this guy is probably a bit ahead of most of us. One thing that clued me in to this is his description of his working prototype of "Universal Composition" which he also wrote a draft 35-page paper about (please *do* read it): http://www.lucs.lu.se/People/Henrik.Gedenryd/Squeak/UC/UCpaper.pdf ...with a corresponding image including code: http://www.lucs.lu.se/People/Henrik.Gedenryd/Squeak/aspectblt.zip (the above image was recently mentioned because of the fonts and the look, but that is definitely not the reason it is published, as the name shows) I also recall a very nice dinner with him, me and Ian Piumarta where he and Ian drifted off talking about weaving and software composition and I realized I was out of my depth. :) At OOPSLA Henrik wanted to show his Unversal Composition prototype to Dan Ingalls and especially how he had rewritten BitBlt using it in a very beatiful way (see paper for details) but I am not sure they really got time to look at it properly. Another thing I remember from that trip is how we played around with the idea of using a transactional object memory to implement a reversible debugger. Henrik's Ph.D thesis on "How Designers Work" has evidently also created a bit of attention: http://www.lucs.lu.se/People/Henrik.Gedenryd/HowDesignersWork/index.htm l ...which among others Chris Rust (Professor of Design and Head of Art & Design Research Centre and Head of Product & Knowledge Research Centre, Sheffield Hallam University) refers to in very impressive words on his page: http://www.chrisrust.pwp.blueyonder.co.uk/academic/research.htm Henrik was a really nice guy, even though I know he upset at least one or two Squeakers over the years, but who hasn't managed to do that? :) I think it was a great loss to our community and the least thing we can do to honour his memory is to read his very intriguing paper on Universal Composition - especially today when we are working with getting Traits into Squeak. regards, Göran

20 26

Test VMMaker branch for 38b4 on SVN server
by Tim Rowledge 18 Nov '05

18 Nov '05

As a small experiment I've just made a branch on the SVN server that _should_ contain a tree of platform files suitable for working with a 3. 8 image and the VMMaker38b4 on SM. It ought to come your way in response to svn co http://squeak.hpl.hp.com/svn/squeak/branches/VMM38b4 This does NOT include the changes towards 64 bit cleanliness and thus does not need the small changes that ought to be in the update stream. If this seems to work, we'll try making a new branch for each future version of VMMaker so that the head can be left as the dangerous place. tim -- Tim Rowledge, tim(a)rowledge.org, http://www.rowledge.org/tim Oxymorons: Living dead

4 12

Jump to page: