Squeak-dev December 2000

squeak-dev@lists.squeakfoundation.org

96 participants
272 discussions

Re: VM Debugging Help, Please
by Lex Spoon 31 Dec '00

31 Dec '00

Tim Rowledge <tim(a)sumeru.stanford.edu> wrote: > Karl Goiser wrote: > > > > Thanks for the Ned. I'm on the mac in Codewarrior, so I don't think I can > > call a function from within the debugger. Somebody else suggested I write > > it into the event code conditional upon a global which I could then turn on > > whenever I wanted to... > consider running the IntepreterSimulator; then you can debug in > Smalltalk! It can be too slow on some machines or under some > circumstances, but you'd be amazed how powerful it can be. > I tried the simulator in a recent 2.9 and it crashed in LargeIntegersPlugin. The method in question was one that checks class membership by name; the relevant method doesn't work in the simulator. -Lex

1 0

Re: Socket and multi-homed hosts
by Lex Spoon 31 Dec '00

31 Dec '00

I think netType determines the arguments to socket(), not to bind(). Currently the options are TCP and UDP, which sounds irrelevant to what you're after. There's currently no way to choose the address that should be listened at, much less the specific interface that should be used. If you want this enough, it seems perfectly reasonable to add something like #bindToAddress:port: to Socket, and then to add the corresponding stuff to SocketPlugin and to sqUnixNetwork.c. The method might look like: bindToAddress: aByteArray port: anInteger "bind this socket to the given address and port. all-0's and 0 may be specified, respectively, if the caller doesn't care" <primitive: 'bindToAddress' module: 'SocketPlugin'> ^nil Then you need to add a platform-independent version of the primitive to SocketPlugin, do "SocketPlugin translate", and add the Linux-specific version to sqUnixNetwork.c. Of course, you need the source files for the Unix VM, too -- Ian Piumart's version 2.8pre5 appears to be the one to go for right now. So, it's doable, if a fair bit of time is available. :| -Lex merlyn(a)stonehenge.com (Randal L. Schwartz) wrote: > > After an admittedly cursory examination of Socket, I see only > the comment of: > > The netType parameter is platform dependent and can be used to > encode both the protocol type (IP, Xerox XNS, etc.) and/or the > physical network interface to use if this host is connected to > multiple networks. A zero netType means to use IP protocols > and the primary (or only) network interface. > > in Socket>>primSocketCreateNetwork:type:receiveBufferSize:sendBufSize:semaIndex: > > Has anyone done any work to document how this works for the various > platforms, or try to make this a little more portable? > > Specifically, I need this to bind to a particular interface on a Linux > box for my ISP. > > It looks like sqUnixSocket.c ignores that parameter, creating > an INADDR_ANY socket every time. > > -- > Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 > <merlyn(a)stonehenge.com> <URL:http://www.stonehenge.com/merlyn/> > Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. > See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

1 0

Re: Morphic
by Lex Spoon 31 Dec '00

31 Dec '00

One quick suggestion is to check out the "pluggable" morphs, if you haven't already. They include text areas, lists, and buttons, which actually seems to be plenty for many UI's. To see them in action -- including the way they deal with menus -- check out FileList class>>openAsMorph. Good luck! Lex "Peter Schuller" <peter.schuller(a)infidyne.com> wrote: > Hello, > > I realize I have just been asking questions since I joined this group. > Hopefully I will eventually be knowledgeable enough to constribute in the > form of *answering* some. But for now, I have yet another question: > > Is there a tutorial/overview/howto for Morphic that covers more than the > extreme basics? I don't have a problem understanding the basics concepts, > but I am still suffering from being "lost" in the API. Probably because I'm > mostly coming from a Swing/Qt background and things are obviously done > quit differently in Morphic. > > The only one I've found that actually introduces you to the API is John > Maloney's "Fun with the Morphic Graphics System" (referenced at > http://minnow.cc.gatech.edu/squeak/30), which is good at what > is does, but is limited in scope and seems to be geared towards those > that aren't familiar with even the Squeak browser yet. > > I'm looking for something to accelerate the initial learning process; i.e. > point out how to do some common stuff so that I can a better "feel" for it. > Like creating windows, managing menus and component layouts. > > Just now I was trying fooling around with menus and the SystemWindow. I was > not able to figoure out how to add menus to a SystemWindow (or if there is > even any special support for such actions). I'm sure I'll get it > straightened out after browsing the API some more, but hence my desire > for a hands-on tutorial just to get a feel for the most common things... > > Thank you, > > -- > / Peter Schuller, InfiDyne Technologies HB > > PGP userID: 0x5584BD98 or 'Peter Schuller <peter.schuller(a)infidyne.com>' > Key retrival: Send an E-Mail to getpgpkey(a)scode.infidyne.com > E-Mail: peter.schuller(a)infidyne.com Web: http://scode.infidyne.com

1 0

Re:speed of 2.9a
by Marcel Weiher 31 Dec '00

31 Dec '00

> From: silence-squeak(a)nilpotent.org (Faried Nawaz) > Also, 2.9 with the latest updates (3144) is a lot less responsive than > 3125. It's barely usable in its current state. I have to click several > times or click-and-hold to select classes in the browser, for example. > Window decorations (close and minimize) are also very slow to react. I'm > not sure which update triggered this, but I can check tomorrow. Yes, I have the same problems with 2.9-2710, which really got to me as I was re-writing the MacOS-X VM. Anyway, I dug a little deeper and found a couple of things that surprised me: 1. MVC vs. Morphic My VM currently has all its input-sensor support #ifdef'ed out and is therefore purely event driven. However, MVC, despite it having to rely on a Squeak-side sensor-simulation of the event-input from the VM, is a whole lot more responsive. Another thing I noticed was that Morphic was synthesizing events from the simulated input-sensor synthesizewd from the VM's events. This changed when I switched to another project and back (it seems the check for wether to get events or simulate them is made at project-switch-time, despite this not really having much to do with the projects, and more with the underlying VM running the show). 2. Event propagation Through some logging, I discovered that events get to the VM quickly, even in Morphic. It is within Morphic's event code that things get sluggish and/or lost. In particular, the top of HandMorph>>processEvents reports events pretty much as they are input from the VM ( still as 8-element arrays / event-buffers ). However, they sometimes arrive much, much later at the top of HandMorph>>handleEvent: (now as real morphic events). Sometimes they don't arrive at all. So something must be killing those events in the meantime. ( I am not talking about mouse-moved events, but clicks). Marcel

1 0

Re: VM Debugging Help, Please
by Ned Konz 31 Dec '00

31 Dec '00

On Thursday 28 December 2000 15:32, Karl Goiser wrote: > My question is: is there a way to log message sends to a file so I can see > what is executing? Otherwise, when in the debugger, is it possible to look > at the source of the methods being executed? Well, there's printCallStack() that you could call from C (you don't say what platform you're running, but using the WinCE remote debugger, I can break program execution, then evaluate printCallStack(); this will print out the call stack on the console (stderr or stdout for regular Win32 or Unix, I imagine). -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

3 2

2.8 vm for unix, speed of 2.9a
by silence-squeak＠nilpotent.org 30 Dec '00

30 Dec '00

Hello, The INRIA site has 2.8pre3 -- is that the latest VM for Unix? I keep seeing references to 2.8p4 and 2.8p5 but they're neither at the INRIA ftp site nor at UIUC. Also, 2.9 with the latest updates (3144) is a lot less responsive than 3125. It's barely usable in its current state. I have to click several times or click-and-hold to select classes in the browser, for example. Window decorations (close and minimize) are also very slow to react. I'm not sure which update triggered this, but I can check tomorrow. Faried.

1 0

Re: Smalltalk Implementation of Linda?
by Stephane Ducasse 30 Dec '00

30 Dec '00

Contact thomas hofmann, he built a distributed space framework on top of distributed Smalltalk. http://www.iam.unibe.ch/~hofmann/ > Resent-date: Fri, 29 Dec 2000 20:10:47 +0000 > Date: Fri, 29 Dec 2000 12:11:30 -0800 (PST) > Resent-from: squeak(a)cs.uiuc.edu > From: Conrad Taylor <conradwt(a)yahoo.com> > Subject: Smalltalk Implementation of Linda? > Resent-sender: squeak-request(a)cs.uiuc.edu > To: squeak(a)cs.uiuc.edu, vwnc(a)cs.uiuc.edu > Resent-cc: recipient list not shown: ; > Resent-message-id: <pnJrN.A.YwF.H_OT6(a)jerry.cs.uiuc.edu> > MIME-version: 1.0 > Content-transfer-encoding: 7BIT > X-Loop: squeak(a)cs.uiuc.edu > X-Mailing-List: <squeak(a)cs.uiuc.edu> archive/latest/27565 > > Hi, does anyone know if there's a Smalltalk > implementation of Linda? BTW, JavaSpaces > is on implementation of Linda but I'm looking > for something in the Smalltalk area. > > Thanks in advance, > > -Conrad > > > > __________________________________________________ > Do You Yahoo!? > Yahoo! Photos - Share your holiday photos online! > http://photos.yahoo.com/ > Stephane DUCASSE (ducasse(a)iam.unibe.ch) http://www.iam.unibe.ch/~ducasse/ "if you knew today was your last day on earth, what would you do different? ... especially if, by doing something different, today might not be your last day on earth" Calvin&Hobbes University of Bern, Institut fuer informatik and Mathematik IAM-SCG, 10 neubruckstrasse, CH-3012 Bern, Switzerland.

1 0

Re: [VM] CompiledMethod header format?
by Bob Arning 30 Dec '00

30 Dec '00

On Fri, 29 Dec 2000 20:22:48 -0500 (EST) "Matthew S. Davis" <mdavis(a)cse.psu.edu> wrote: > I've determined the following header format from the Interpreter >methods: > >bit 0 : constant 1 to indicate SmallInteger >bits 1-9 : primitive index >bits 10-17 : literal count >bit 18 : not used? >bits 19-25 : temporary count >bits 25-28 : argument count >bits 29-31 : primitive index > > Could someone please explain why bit 25 is shared between two >fields? I'm not clear why this is the case. Should the temporary count >really be in bit positions 18-24? If so, it does not look like they are. Matthew, Temporary count is definitely 6 bits in the header. Where did you see an indication that it was 7? And the bit you have marked as unused is the Large/Small context flag. Cheers, Bob

2 1

Re: Squeak Language Tutorial needed...
by Andrew C. Greenberg 30 Dec '00

30 Dec '00

>Hello all, > >In putting together materials for the Squeak CD-ROM (which I intend to >post to the list for comment real soon), I've noticed that there seems >to be a good collection of Squeak tutorial materials, but no real simple >introduction to the language itself with basic programming syntax examples. While not exactly what you are asking for, I am informed that my "Squeak Smalltalk: A Quick Reference," has been found useful for these purposes. It does begin with a basic description of the language and some syntax examples. http://www.mucow.com/squeak-qref.html

2 1

[VM] CompiledMethod header format?
by Matthew S. Davis 30 Dec '00

30 Dec '00

Hi, I'm just starting an attempt to figure out "how the Squeak VM works". I decided to start with the chapters from the Blue Book on the implementation of the VM and to go through my Squeak 2.8 image as I go along. I know that things have changed in this implementation from the Blue Book days, but there is something in the Squeak implementation that is really confusing me. I decided to figure out Squeak's method header format. So I checked out the Interpreter class. Assume that the most significant bit is numbered 31 and the least significant bit is 0 in this question. It seems that bit 25 of the header is associated with two fields: the argument count from Interpreter>>argumentCountOf: and with the temporary count found in Interpreter>>bytecodeActivateNewMethod and Interpreter>>internalBytecodeActivateNewMethod which both implement the retrieval of the tempCount from the header (alas). It also seems that bit 18 is totally ignored. I've determined the following header format from the Interpreter methods: bit 0 : constant 1 to indicate SmallInteger bits 1-9 : primitive index bits 10-17 : literal count bit 18 : not used? bits 19-25 : temporary count bits 25-28 : argument count bits 29-31 : primitive index Could someone please explain why bit 25 is shared between two fields? I'm not clear why this is the case. Should the temporary count really be in bit positions 18-24? If so, it does not look like they are. Thanks! -Matthew

2 1

Jump to page: