Squeak-dev

squeak-dev@lists.squeakfoundation.org

15 participants
79907 discussions

Re: Squeak crash upon full GC
by sqrmax＠cvtci.com.ar 20 Jan '98

20 Jan '98

Hi. >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 already have a CRC32 checker coded entirely in Smalltalk... big problem with it is that it crunched 18kb per second in Visual Smalltalk. Definitely too slow to check the image when saving or so. At that time I thought it was so slow because it was using LargeIntegers (that crc is 32 bits long) instead of SmallIntegers, wasting a lot of time. I also have the assembler version (quite portable, less that a dozen lines), which can be embebbed in a DLL or primitive in C. At that time, the assembler version was able to crunch about 10mb of data per second... but I'm not so sure about that figure. I'm sure it was able to crunch 6mb of data per second, but I also remember that I optimized it some later. There below is the code of the CRC32 class appended. There are two details about it. First, it was thought to be fed with BitStreams, a bit oriented counterpart of Stream. However, it can be easily adapted for Streams (the #nextByte selector should be replaced with #next). The second insignificant detail is that it uses the selector #shr:, which can be translated to #>>. Overall, only two selectors have to be changed and those two methods appear only in one selector, #codeFor:. Andres. Content-Type: application/octet-stream; name="Crc32.st" Content-Disposition: attachment; Filename="Crc32.st" Attachment converted: Cog:Crc32.st (????/----) (0000AB53)

1 0

Re: Sound problem in 1.30
by Ian Piumarta 19 Jan '98

19 Jan '98

> Now, sound seems to be "disabled" for Unix by having all primitives in > sqUnixSound.c fail. I thought they all "silently succeed" instead. E.g: #define FAIL(X) { return X; } int snd_AvailableSpace(void) FAIL(8192) Maybe there's something in the imagfe that's repeatedly calling the sound primitives because they appear to be succeeding instantly? I'll take a closer look sometime. (I guess this problem never surfaced before because the "page turning" sound in the BookMorph is a new feature.) > Is this problem related to the recently-fixed problem of > improperly-handled primitive failures? No. That fix (and the problem is cures) only relates to user-defined primitives. Ian

1 0

Re: Problem in 1.30
by Andreas Raab 19 Jan '98

19 Jan '98

Bob, > An annoying problem has surfaced in 1.30, on my machine at least. I'm > running on a machine without sound support, and after bringing up the > 'Play With Me - 4' example in 1.30 I find I get a walkback every time I > try to move from page to page using the controls in the upper left part > of the screen. Bringing up a debugger shows the following message > sends: 'Play With Me - 4' uses sound when turning the pages. Since you seem to have no sound card installed that's a bit of a problem. I don't think that returning true on a missing hardware is a good idea. Rather than this I'd like to add a flag 'disableSounds' to Preferences and modifying some method appropriately (see below). Hope this helps, Andreas 'From Squeak 1.3 of Jan 16, 1998 on 19 January 1998 at 8:34:06 pm'! !Preferences class methodsFor: 'initialization' stamp: 'ar 1/19/98 20:31'! chooseInitialSettings "Restore the default choices for Preferences." "Preferences chooseInitialSettings" self setPreference: #reverseWindowStagger toValue: true. self setPreference: #thoroughSenders toValue: true. self setPreference: #uniformWindowColors toValue: false. self setPreference: #warnIfNoChangesFile toValue: true. self setPreference: #warnIfNoSourcesFile toValue: true. self setPreference: #disableSounds toValue: false. ! ! !SoundPlayer class methodsFor: 'playing' stamp: 'ar 1/19/98 20:32'! playSound: aSound "Reset and start playing the given sound from its beginning." Preferences disableSounds ifFalse:[ aSound reset. self resumePlaying: aSound. ]! ! Preferences chooseInitialSettings.! -- 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

Re: API Calls for Squeak 1.3
by R.L.J.M.W. van Loon 19 Jan '98

19 Jan '98

Hi Ale, I see you fixed the bug in getCurrent@KernelDLL class (we experienced it this morning after building everything from scratch). KernelDLL class methodsFor: 'opening' stamp: 'afr 1/14/98 21:52'! getCurrent "Private - get (open) an instance of the receiver." self current: (self fromInteger: self new getHandle). ^self current! Why not change the code to: KernelDLL class methodsFor: 'opening' stamp: 'afr 1/14/98 21:52'! getCurrent "Private - get (open) an instance of the receiver." ^self fromInteger: self new getHandle. Remember, getCurrent should answer something which becomes current. The setting of current is handled up in the hierarchy. Now, getCurrent also sets the current. We've got things compiled and running. When evaluating the getVersion example it prints 16r4 (good!) but when I evaluate the WinExec 'calc.exe' the Windows Calculator comes up and then Squeak crashes in Win32Alloc (vMemIsReservedPage...Query) etc. After testing 3 or more examples leading to the same result we stopped. Any suggestions as to fix things or what we might be doing wrong? Anyway, well done! Reinier. P.S. I've sent it to the mailing list to. Might be interesting for others. > Evolution makes Smalltalk gurus. > Mutation makes Java fans. Nice, I like this.

1 0

Re: Squeak 1.3 for Unix
by Ian Piumarta 19 Jan '98

19 Jan '98

Todd, > I am having problems generating a new (1.3) VM with your > user primitive change set. The resulting executable dies > on start up with a Segmentation Fault in the BitBlit code. I've placed a new version (1.3-compatible) of the userPrims in ftp://alix.inria.fr/pub/squeak/unix/goodies/userprims and I've checked that it works with the example UDPs in that same directory. (I also fixed the bug to make the UDPs fail correctly.) The reason you were seeing the SEGV is that Dan changed the BitBlt code to use a dispatch table instead of a series of nested "if"s and/or a switch statement. By filing-in the old UDP stuff you were removing the initialisation of this dispatch table from the interpreter's startup code. Regards, Ian

1 0

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

Jump to page: