Squeak-dev May 2005

squeak-dev@lists.squeakfoundation.org

153 participants
243 discussions

Deriving TcpService.
by François THIMON 04 May '05

04 May '05

Hi, as I've been told here, I tried to derive the TcpService class of the KomServices package but I still get message errors. It seems that the pvtBindings methods always return null, I don't know why. The only method I had to redefine was serve:, I did this but it doesn't work. (TcpService on:1024) start. does work but replacing TcpService by my server class's name messes it all up. Can someone help please? By the way, do you know how i can close all the port that got open each time I tried? Closing squaek can do nothing to help but for now I have to increment the port number each time and this is quite pain in the neck. Thanks in advance.

1 0

Harvesting Party Sunday May 8th 05:00 GMT
by Ken Causey 04 May '05

04 May '05

Just a reminder that the next Harvesting Party will take place between 05:00 (5 AM) and 09:00 (9 AM) GMT Sunday May 8th, 2005. As usual we will meet in #squeak on irc.freenode.net. More information about the #squeak IRC channel can be found at http://people.squeakfoundation.org/article/7.html Information about the Harvesting Party idea can be found at http://people.squeakfoundation.org/article/29.html If you have the time please attend and work toward the future of Squeak. The next seven parties are scheduled for 09:00 (09:00 AM) GMT Thursday May 12, 2005 13:00 (01:00 PM) GMT Monday May 16, 2005 17:00 (05:00 PM) GMT Friday May 20, 2005 21:00 (09:00 PM) GMT Tuesday May 24, 2005 01:00 (01:00 AM) GMT Sunday May 29, 2005 05:00 (05:00 AM) GMT Thursday June 02, 2005 09:00 (09:00 AM) GMT Monday June 06, 2005

1 0

FloatArray>primitiveDivFloatArray
by Tim Rowledge 04 May '05

04 May '05

Minor problem with this when 64bitting it. There is a loop that tries to quickly check for any zero values by extracting the floats as plain ints and comparing with (int)0. So far as I can work out from webpages purporting to explain the IEEE float standard this misses out half the possible 0 values since the S bit can be 0 or 1. Is there anybody out there with a good grip on the IEEE stuff that can confirm, deny, explain this? The code in question is (the C code generated, that is):- rcvrPtr = ((float *) (interpreterProxy->firstIndexableField(rcvr))); /* Check if any of the argument's values is zero */ argPtr = ((float *) (interpreterProxy->firstIndexableField(arg))); for (i = 0; i <= (length - 1); i += 1) { if ((longAt(argPtr + i)) == 0) { return interpreterProxy->primitiveFail(); } } tim -- Tim Rowledge, tim(a)sumeru.stanford.edu, http://sumeru.stanford.edu/tim World Ends at Ten! Pictures at 11 on Fox News!

2 2

Re: Question about flushing Display on Canvas
by Lex Spoon 03 May '05

03 May '05

> > It would be great if there were a way that was as easy to draw in > > Morphic as Milan is seeking. If you draw onto the Display directly, > > Squeak is smart enough to flush things immediately as you draw. It's > > not fast, but it's easy to use and understand. On the other hand, if > > you play by Morphic's rules, then what can you do? > > Implement a drawOn: method ;-) Yes, that's what people should eventually learn, but think way back to before you were comfortable programming GUI's... Many people have a hard time thinking about this idea that a widget is passive and waits to be asked to draw itself. Many people would like to write code along the lines of: 1. compute a new point to draw. 2. draw it. 3. go to 1. Or maybe: Whenever a building is destroyed, draw an explosion on top of it. Or: Every 10 milliseconds, draw a new square at a random place. These are awkward for a beginner in Squeak right now. Before you can implement them, you have to code up some extra infrastructure to manage a form behind the scenes. On further thought, though, there actually is a way to do this in Squeak right now: using addMorph:. Instead of drawing an ellipse, add an EllipseMorph to the right location. There are (or should be) morphs for all of the canvas commands.... -Lex

1 0

Re: Monticello and image bloat
by Chris Muller 03 May '05

03 May '05

flushAllCaches took me down to 44MB. I did not know about that. Thanks! Chris --- Cees De Groot <cdegroot(a)gmail.com> wrote: > Hi Chris, > > It's a known feature/problem/bug/wart of MC that it eats up a lot of > space. Usually, doing "MCFileBasedRepository flushAllCaches; Smalltalk > garbageCollect" takes care of that. > > Regards, > > Cees > > On 5/3/05, Chris Muller <chris(a)funkyobjects.org> wrote: > > I've been experiencing gradual image-growth as I develop code, and now my > > 100-meg image has become unwieldly. SpaceTally reports 59MB used on String > > instances! > >

1 0

Monticello and image bloat
by Chris Muller 03 May '05

03 May '05

I've been experiencing gradual image-growth as I develop code, and now my 100-meg image has become unwieldly. SpaceTally reports 59MB used on String instances! The first few entries of String allInstances asBag sortedCounts are: a SortedCollection(48161->'cmm' 39710->'' 6481->'building SAR component' 3582->'New prereqs.' 1818->'empty log message' 1543->'Magma client-private' 1432->'Ma time objects-cmm.1' 1263->'<historical>' 1008->'trying Monticello' ...) These appear to be Strings of Monticello's domain. But why so many duplicates? 6481 copies of 'building SAR component'? I remember saving a *few* (maybe 10) with that description, but why are there so many of it now? Is there any way to share the same String instance instead? Using the 'explore pointers' function, I traced one of them to MCCacheRepository. By niling out its 'default' class-instance variable and immediately saving the image chopped off TWENTY megabytes! Ok, making progress. But my String counts are still extremely high: a SortedCollection(30377->'cmm' 27930->'' 3804->'building SAR component' 2237->'New prereqs.' 1010->'empty log message' 957->'<historical>' 847->'Magma client-private' 801->'cmm 2/17/2005 10:47' 784->'Ma time objects-cmm.1' ...) How can I clean out more of these? Obviously, 'cmm' is my developer initials. Is it feasible to share this String? What Monticello action(s) should I avoid to keep this from happening? I found one; looking at the historial entries of a package seems to cause rapid image-growth.. Are there others? Thank you, Chris

3 2

Re[4]: Dorado bytecodes per second
by Andres Valloud 03 May '05

03 May '05

Hello Jecel, Monday, May 2, 2005, 1:18:22 PM, you wrote: JAJ> I want to write polymorphic code and let the compiler generate case JAJ> statements (PICs - polymorphic inline caches). JAJ> I want to write well factored code and let the compiler generate big, JAJ> efficient blobs (with inlining). JAJ> I want to play with blocks and let the compiler extract control flow JAJ> strctures wherever it can. JAJ> In short, I want to program in Smalltalk and let the compiler generate JAJ> the Fortran that the hardware seems to crave. Totally, I agree. Smalltalk is for us. And when hardware changes, we'll just change the VM and be done with it. -- Best regards, Andres mailto:sqrmax@optonline.net

1 0

REMINDER: Harvesting Party Wednesday May 4th 01:00 GMT
by Ken Causey 03 May '05

03 May '05

Just a reminder that the next Harvesting Party will take place between 01:00 (1 AM) and 05:00 (5 AM) GMT Wednesday May 4th, 2005. As usual we will meet in #squeak on irc.freenode.net. More information about the #squeak IRC channel can be found at http://people.squeakfoundation.org/article/7.html Information about the Harvesting Party idea can be found at http://people.squeakfoundation.org/article/29.html If you have the time please attend and work toward the future of Squeak. The next seven parties are scheduled for 05:00 (05:00 AM) GMT Sunday May 08, 2005 09:00 (09:00 AM) GMT Thursday May 12, 2005 13:00 (01:00 PM) GMT Monday May 16, 2005 17:00 (05:00 PM) GMT Friday May 20, 2005 21:00 (09:00 PM) GMT Tuesday May 24, 2005 01:00 (01:00 AM) GMT Sunday May 29, 2005 05:00 (05:00 AM) GMT Thursday June 02, 2005

1 0

[ANN] Kedama version 2.0
by Yoshiki Ohshima 03 May '05

03 May '05

Hello, I've been working on a parallel particle programming system called Kedama. The last version wasn't quite good enough to announce, but I feel this version is a bit better. Kedama is deeply inspired by StarLogo and StarSqueak. In Kedama, you can control thousands of objects with eToy-like tile scripts. Please take a look at: http://www.is.titech.ac.jp/~ohshima/squeak/kedama2/ and give it a try. Thank you! -- Yoshiki

7 6

Re: Dorado bytecodes per second
by Andreas Raab 02 May '05

02 May '05

> For the case of manipulating bitmaps we can always have a coprocessor > help out to fix the problem. While it would be great if some hotspot > could be found and eliminated from Morphic it is far more likely that > the overhead is distributed all over the place. That is most likely correct. Have a look at the current Compiler and ParseNodes which were written in the ugly old days and notice how they painfully avoid sending messages, and how they use pools for stuff that's considered really bad style today. What has changed since the old days is the general attitude - we happily pay the "price" of the extra message if it gives us a bit more flexibility, if it makes the code a little more easy to read. This simply didn't used to be the case "back then" and MVC is one of the areas that is full of "good old code" - I bet you a "modern" MVC would not perform any better or worse than Morphic. But adding a graphics processor is a good idea and will win you *big* on the small machines - most of which are bound by transfer speed over the bus anyway so having graphics go fast makes a much bigger difference than on the faster machines. Cheers, - Andreas

7 9

Jump to page: