Squeak-dev

squeak-dev@lists.squeakfoundation.org

18 participants
79922 discussions

Squeak crash upon full GC
by Andreas Raab 19 Jan '98

19 Jan '98

Paul, The crash upon GC looks like being solved. > > Strange thing. I just downloaded it again, extracted it and _could_ save > > it without any problems. > > Sounds like a problem on my end or transmission error then. It looks very much like this. I've run a binary diff on your image and the one in the ftp directory and it turned out that one bit differed. > Gee, it would be nice if Squeak did a CRC on images. I wonder how I > would add that - maybe a check at startup? Maybe adding that should be > my punishment if it is just a transmission error. Perhaps it could be an > optional method - but the CRC value would have to be computed on image > save - which would slow things down. I don't think that this would slow down things too much. Would be great if somebody would add some CRC checking to Squeak. I'm not sure how many images are hanging around with the one or other bit toggled. In your case we did find only because it was in the object header ... yes I think doing some CRC checking would be a _very_ good idea. > I'm going to download the zip file again and see if it improves. That > really should have been the first thing I did before mentioning it. > Sorry. Don't worry. These things happen all the time. And it also shows the need for CRC checking. In this sense it may turn out as a good starting point for future development. Andreas -- Linear algebra is your friend - Trigonometry is your enemy. +===== Andreas Raab ============= (raab(a)isg.cs.uni-magdeburg.de) =====+ I Department of Simulation and Graphics Phone: +49 391 671 8065 I I University of Magdeburg, Germany Fax: +49 391 671 1164 I +=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+

3 3

Speed hack dirty?
by Christopher Oliver 19 Jan '98

19 Jan '98

In my cryptology program, I need to rotate a 32 bit value by a multiple of eight bits. I started with: bitRot: word32 by: bitCount "Rotation of long words. This really really belong with the integers." ^ ((word32 bitShift: bitCount) + (word32 bitShift: bitCount - 32)) bitAnd: 16rFFFFFFFF However this is VERY slow. I rewrote this as three explicit methods using digitAt: and digitAt:put: which speeds up the rotation by about an order of magnitude. The real question is just how forbidden is this sort of thing? -- Christopher Oliver Traverse Communications Systems Coordinator 223 Grandview Pkwy, Suite 108 oliver(a)traverse.com Traverse City, Michigan, 49684 (define magic (lambda (f) (lambda (x) x))) (define (more-magic n) (lambda (f) (lambda (x) (f ((n f) x)))))

1 0

API Calls for Squeak 1.3
by Alejandro F. Reimondo 19 Jan '98

19 Jan '98

I have updated my extension for API calls for Squeak. You can download it (213Kb) from http://www.sugarweb.com/Miniatures/Files/SqApi13.Zip Comments will be greatly apreciated. Ale. Alejandro F. Reimondo Feel free to visit Smalltalk User Group of Argentina (SUGAR) at http://www.sugarWeb.com Evolution make Smalltalk gurus. Mutation make Java fans.

1 0

Coordinate order in Array2D
by Christopher Oliver 19 Jan '98

19 Jan '98

Hey folk! Why does Array2D at:at: and at:at:put: use a coordinate order reversed from normal mathematical conventions? The first coordinate is traditionally the matrix row. I just got bitten while transliterating some cryptology code using matrices. This sort of difference can cause severe brain hemorrhage. -- Christopher Oliver Traverse Communications Systems Coordinator 223 Grandview Pkwy, Suite 108 oliver(a)traverse.com Traverse City, Michigan, 49684 (define magic (lambda (f) (lambda (x) x))) (define (more-magic n) (lambda (f) (lambda (x) (f ((n f) x)))))

1 0

GIFReadWriter in 1.3 corrupt
by R.L.J.M.W. van Loon 18 Jan '98

18 Jan '98

Was it meant to be that GIFReadWriter is in the image? Anyway my image only shows one class method and the menu option in the file list window does not work. Any comments? Reinier. ! Smalltalk = PowerTalk ! rvloon(a)inter.nl.net

2 1

Re: Squeak 1.3 for Unix
by Ian Piumarta 18 Jan '98

18 Jan '98

Todd, > Am I missing a step here? I don't think so. Assuming the change set in the goodies directory was actually reliable. I simply copied it from 1.23, and if something obscure changed in the Interpreter's primitive dispatch since 1.23 then there's little hope of it working. > And, are you planning to compile and ship your 1.3 port with user > primitive support? Sure, if it's useful to somebody. I'll do the necessary stuff later tonight. Ian

1 0

Re: Squeak 1.3 for Windows
by R.L.J.M.W. van Loon 18 Jan '98

18 Jan '98

It works. And it's beautiful. (Windows 95 OSR2 and WINDOWS NT Server 4.0 sp1). Reinier.

1 0

RE: Squeak 1.3 for Window Full screen bug on Win95
by Serg Koren 18 Jan '98

18 Jan '98

Hi Andreas, Thanks for the response. On the recursion error, it was my mistake...I re-filed in my code and it works fine. Cheers, Serg -----Original Message----- From: Andreas Raab [SMTP:raab@isgnw.cs.Uni-Magdeburg.DE] Sent: Sunday, January 18, 1998 8:42 AM To: squeak(a)cs.uiuc.edu Subject: Re: Squeak 1.3 for Window Full screen bug on Win95 Serg, > I'm running a Matrox Mystique at 640x480 Hi Color. Have my Windows task = > bar and Office97 startup bar docked to the bottom. Going full screen in = > 1.3 produces a white stripe where the bars are docked as if the update = > rectangle is incorrect and forcing a redraw doesn't solve it. I'd say that there are two possibilities. Either Office97 does something strange to the Explorer task bar (which would not surprise me since the MS Office always used some of the more undocumented Windows) or it's a problem with your Matrox graphics driver (which would also be no surprise - I had trouble with Matrox cards before). I'll give it a look. > Also, it looks like I'm getting a recursion error when I run the simple > animated TowerOfHanoi sample from "A Taste of Smalltalk". I haven't > looked into this yet, but this code worked fine under 1.2. The error > throws up a modal dialog and then quits the 1.3 VM. I don't have the book. Can you email me the sample? Andreas -- Linear algebra is your friend - Trigonometry is your enemy. +===== Andreas Raab ============= (raab(a)isg.cs.uni-magdeburg.de) =====+ I Department of Simulation and Graphics Phone: +49 391 671 8065 I I University of Magdeburg, Germany Fax: +49 391 671 1164 I +=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+

1 0

Squeak mutlitple versions..."official/blessed" release questions...
by Serg Koren 18 Jan '98

18 Jan '98

I guess this leads me to ask... Since there seem to be many people making modifications to various images of Squeak; and some of these changes may affect portability, etc. is there a mechanism in place to consolidate these changes from image to image? It sounds from my short time on this group that a lot of neat stuff is being done by various people, but that these are pretty independant efforts. And if someone releases an image (such as I did for the Mac) how does it become "official" or "blessed"? I guess this is one reason I released my Mac version as 1.24SK (appending my initials) both to tell people it's not really an "official/blessed" version--if there is such a thing, and to avoid a versioning conflict with anyone else creating/working on a 1.24 version. If it helps and there's enough interest I'd be willing to keep and maintain a central site of 'fileOut's by OS type as well as links to people who have Squeak code to distribute. This could be used by people to pick/choose options they may want to load into their images that doesn't fall within an "official" release. There are a lot of sites out there that have good stuff but there doesn't seem to be a "central" site that links to them or organizes what is out there. Personally, I think the cross-platform portability of the code and base interepreter VMs is a key benefit of Squeak over other languages, and I'd be upset to see anyone compromise that. It's ok if someone creates a VM that's specific to an OS, but such modifications I think should fall under the platform specific code files within Interpreter support code. On the other hand any changes to the base interpreter code or non-hardware specific classes should be published so all images on all OSes can incorporate it. I guess one thing I'm disappointed with is that the Windows version of Squeak (I haven't played with the Unix ones) doesn't allow you to create all of the hardware specific code. That is, the WIndows version (at least the 1.3 release) still contains the primitive code in the goodies directory) under 1.23 and am having problems generating a new (1.3) VM with your user primitive change set. The interp.c you ship doesn't seem to have the user primitive support functions (I can't link my primitives against the current interpreter --- unresolved refs to the functions declared in sqUserPrims.h). So, I filed in the 1.23 user primitive change set (from goodies) into the 1.3 Squeak and generated a new interpreter with: Interpreter translate: 'interp.c' doInlines: true I then recompile the VM with the new interp.c and my user primitives. The resulting executable dies on start up with a Segmentation Fault in the BitBlit code. I haven't tried to debug it yet... But.... Am I missing a step here? And, are you planning to compile and ship your 1.3 port with user primitive support? -- Todd Coram (Maroc Ddot) tcoram(a)BLaCKSMITH.com | tcoram(a)pobox.com Sardonic slaughterer of sacred cows.

1 0

Squeak 1.3 for Unix
by Ian Piumarta 18 Jan '98

18 Jan '98

Unix Squeakers, The 1.3 version of Unix Squeak is now available at: ftp://alix.inria.fr/pub/squeak/unix/Squeak-1.3 It includes precompiled binaries for the following platforms: Intel/Linux Sparc/SunOS Sparc/Solaris Alpha/Digital Unix There are numerous improvements to the Unix support code, approximately as follows. support for X Shared Memory extensions rolled into the main sources + command-line option "-xshm" + env var "SQUEAK_XSHM" accompanied by support for asynchronous screen updates + command-line option "-xasync" + env var "SQUEAK_XASYNC" (effective only when xshm is enabled) support for asking the WM not to decorate window + command-line option "-notitle" + env var "SQUEAK_NOTITLE" which works by marking the Squeak window as a "transient for" the root window. I know, it's ugly. But it works fixed a race condition that sometimes caused Squeak and X to disagree about the size of the Squeak window at startup. Changed the default screen offset to 0@0 to cure a problem with initial sizing when user-placement is disabled enabled use of backing store to avoid repainting areas of the Pixmap on exposure, at the expense of server memory (this should be an option, but I'm too lazy [and I've got sooo much memory that I don't care ;)] -- it's easy to identify and disable the relevant lines in sqXWindow.c if you dislike it intensely) changed class/resource name to "Squeak" (no '!' at the end). WM configs (e.g. mine specifies NoTitle and NoBorder for "Squeak") may need to be updated as a result fixed timezone handling on (hopefully) all platforms, and for all locations (whether east or west of Greenwich meridian) -- many thanks to Richie Bielak for finally getting to the bottom of this! init code now switches off unaligned access reporting for DEC Alpha, so there's no more need to issue "uac p 0" from the shell before starting Squeak Unix and MacOS now share common versions of all header files (sq.h, sqConfig.h, and sqMachDep.h) "gnuify" script simplified to remove stuff that is irrelevant with the new Jitter (the original script is still there as "gnuify.trad", and the correct one will be chosen automatically if you're using GNU make) GNUmakefiles totally reworked, with a new top-level GNUmakefile that supports per-platform build directories. there are certainly problems that I've not yet spotted, but please bear with me: the new GNUmakefiles are a huge improvement over the old ones (especially for frequent rebuilds of VMs for multiple platforms ;o) Please let me know of any problems ASAP. I have unfortunately not had chance to test these binaries (1Mb of X bitmap takes a tad too long over a 22.6kbps modem link), but I have every reason to expect them to work. (I built VMs from essentially the same sources on Friday, and experienced no problems whatsoever.) Enjoy! Ian

2 1

Jump to page: