Squeak-dev December 2000

squeak-dev@lists.squeakfoundation.org

96 participants
272 discussions

Genetic Algorithms
by Diego Gomez Deck 03 Dec '00

03 Dec '00

This if my very first version of a framework for GA in Squeak... Evaluate "GeneticAlgorithm test" for test. The morph that show the statistics of the run is sized for a 1024x768 display. The size can be changed in GeneticAlgorithmMorph>>createFitnessMorph and in GeneticAlgorithmMorph>>createFitnessMapMorph. TIA, Diego Gomez Deck

1 0

[updates] 3 more
by Dan Ingalls 03 Dec '00

03 Dec '00

Just a couple more updates to tweak the new symbol correction, and bring the Mac sources up to date with the forthcoming 2.9.6 VM. Barring unforseen problems, this is probably a good plateau to bundle on UIUC. - Dan ------------------------- 3060CorrectionTweak2-di -- Dan Ingalls -- 3 December 2000 Loosens the new correction filter a bit so as not to miss good condidates. 3061CorrectionTweak3-di -- Dan Ingalls -- 3 December 2000 Speeds up String numArgs in support of better Symbol correction. Also discontinues the practise of excluding selectors beginning with uppercase letters from evaluation of numArgs. NOTE: This fileOut must be specially ordered, since String initialize must be executed before the new numArgs can be used. 3062MacVMSource2-9-6-JMM -- johnmci(a)smalltalkconsulting.com -- 1 December 2000 Mac VM Source code for 2.9.6, includes sound support, plus embedded image support"

1 0

Re: Double-Clickable Apps
by Bruce ONeel 03 Dec '00

03 Dec '00

Hi, This seems ok so it's up at ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/2.9_alpha/MacVM2.9Alpha6.sit.b in. cheers bruce p.s. I'll upload the new image/changes .zip tonight or tomorrow. Depends on how saturated the pipes are. John M McIntosh <johnmci(a)smalltalkconsulting.com> wrote: > > > > >John McIntosh was idle a few weeks ago, so we pointed him at this > >and, guess what? It works! > > > >On the Mac, at least. Mac VMs numbered 2.9.6 and higher (coming > >soon) include support for a new primitive, #247, entitled > >snapshotEmbeddedPrimitive. > > Yes Yes it's compiling now, and the change set is being reviewed for > correctness, so you'll all have a 2.9.6 VM by mid weekend or so... > > The 2.9.6 VM also has the ability to set/get the left/right sound > volumes, a primitive hook to your favorite OS's underlying noise > interface. Of course we need the linux/windows/bsd/etc VM support for > that feature too... > > > PS If you build embedded mac applications don't forget to set the > application signature to something else other than FAST (the Squeak > signature) Otherwise that shinny new application you just create will > launch next time you double click on a Squeak Image. (not what you > want...) Also dont' forget if you change the signature then you need > to change the BNDL which contains the pretty icons etc, after all you > are building a double clickable Macintosh application and there are a > few rules to follow to ensure you have a well behaved macintosh > application. > > > > > -- > -- > =========================================================================== > John M. McIntosh <johnmci(a)smalltalkconsulting.com> 1-800-477-2659 > Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com > =========================================================================== > Custom Macintosh programming & various Smalltalk dialects > PGP Key: DSS/Diff/46FC3BE6 > Fingerprint=B22F 7D67 92B7 5D52 72D7 E94A EE69 2D21 46FC 3BE6 > ===========================================================================

1 0

Re: 2.9a Squeak VM, apps, and themes
by J Scott Jaderholm 03 Dec '00

03 Dec '00

Bert Freudenberg <bert(a)isg.cs.uni-magdeburg.de> writes: > > > ok, I downloaded the file Squeak-2.9a.pre1-src.tar.bz2. I extract, > > create a build dir, cd build, ../src/unix/configure works fine, but > > make gives me > > > > make: Nothing to be done for `all'. > > You're sure you ran make from the build dir? If it says there's nothing to > do then there must already be a "squeak" executable in that directory. I definitely did run make from the build directory, and there is no squeak executable in that directory. -- "Man's mind, once stretched by a new idea, never regains its original dimensions." Oliver Wendell Holmes (a Lisp programmer at heart)

1 0

Re: 2.9a Squeak VM, apps, and themes
by Bert Freudenberg 03 Dec '00

03 Dec '00

On 2 Dec 2000, J Scott Jaderholm wrote: > Bob Arning <arning(a)charm.net> writes: > > > ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/2.0/SqueakV2.sources.sea.bin > > I can not extract this under linux. Yeah, if you intend to do more serious squeaking, you at least need to get unstuff for linux: http://www.aladdinsys.com/expander/ Mac folks unfortunately keep posting attachments in binhex encoding or loading up .sit archives :( > Are the .sources for 2.9a not available in .zip or .tar.gz ? Of course they are - see the main Unix site at http://www-sor.inria.fr/~piumarta/squeak/ or even the UIUC Archive (ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/2.8/) The sources file is the same for all Squeak 2.x versions. > ok, I downloaded the file Squeak-2.9a.pre1-src.tar.bz2. I extract, > create a build dir, cd build, ../src/unix/configure works fine, but > make gives me > > make: Nothing to be done for `all'. You're sure you ran make from the build dir? If it says there's nothing to do then there must already be a "squeak" executable in that directory. Besides, if you just start with Squeak you should stay with the official releases. You can get binaries from the UIUC archive and somewhere in one of Ian's ftp directories. Although 2.7 is "really" official for Unix, 2.8 is finished, but not blessed yet. The 2.8.5 sources (not only for PPC Linux) are here: http://www-sor.inria.fr/~piumarta/squeak/Jitter-3.1/ppc-linux-gnu/Squeak-2.… -- Bert

1 0

Re: 2.9a Squeak VM, apps, and themes
by Bob Arning 03 Dec '00

03 Dec '00

On 02 Dec 2000 20:00:19 -0700 J Scott Jaderholm <lispbliss(a)sage.cortland.com> wrote: >I'm interested in playing around with Nebraska and 2.9a of Squeak. I >downloaded Squeak2.9a-2988.zip which includes the .image and >..changes. The problem is I don't have a .sources or a VM for 2.9a and >I heard I need a new VM but all I can find is VMs built for MacOS. Looks like .sources is in ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/2.0/SqueakV2.sources.sea.bin I would try following the links on http://minnow.cc.gatech.edu/squeak/388 appropriate to your OS for the VM >Also, can someone direct me where to download the IRC client, the web >browser, and the email client for squeak? I'm interested in fooling >around with these a bit. They are in the image. >Finally, I've seen some cool themes that give the default squeak look >a nice face lift. The site with these says they don't work with like >2.8a or something. Anyone know if I can get themes elsewhere for >2.9a? Try contacting the author or see if they will run anyway. Cheers, Bob

1 0

2.9a Squeak VM, apps, and themes
by J Scott Jaderholm 03 Dec '00

03 Dec '00

Hi, I'm interested in playing around with Nebraska and 2.9a of Squeak. I downloaded Squeak2.9a-2988.zip which includes the .image and .changes. The problem is I don't have a .sources or a VM for 2.9a and I heard I need a new VM but all I can find is VMs built for MacOS. If anyone can point me where to go it would help. I've scavenged the squeak.org site and the squeak swiki (which was VERY sweet. I saw a spelling error, edit, change, done. That is way cool.) with no luck. Also, can someone direct me where to download the IRC client, the web browser, and the email client for squeak? I'm interested in fooling around with these a bit. Finally, I've seen some cool themes that give the default squeak look a nice face lift. The site with these says they don't work with like 2.8a or something. Anyone know if I can get themes elsewhere for 2.9a? Sincerely, jsj -- "Man's mind, once stretched by a new idea, never regains its original dimensions." Oliver Wendell Holmes (a Lisp programmer at heart)

1 0

StandardFileStream>>readOnlyFileNamed:
by Eric Arseneau 03 Dec '00

03 Dec '00

There seems to be a slight performance problem with this method: readOnlyFileNamed: fileName "Open an existing file with the given name for reading." "Changed to open a more usefull popup menu. It now also includes the most likely choices. jaf" | selection dir files choices newName fullName | fullName _ self fullName: fileName. (self isAFileNamed: fullName) ifTrue: [^ self new open: fullName forWrite: false]. ... And now look at: isAFileNamed: fileName "Answer true if a file of the given name exists." | f | f _ self new open: fileName forWrite: false. f ifNil: [^ false]. f close. ^ true Note that there is two new open:forWrite: calls to open a single file. If we change the previous code to: readOnlyFileNamed: fileName "Open an existing file with the given name for reading." "Changed to open a more usefull popup menu. It now also includes the most likely choices. jaf" | selection dir files choices newName fullName | fullName _ self fullName: fileName. f := self new self new open: fullName forWrite: false. f isNil ifFalse: [^f]. ... Would this not be more efficient ? It looks like the latest release could use this same trick. Does anybody see anything wrong with submitting this change as an [enh] ?

2 1

Re: Fw: Block recursion
by Jon K Anderson 02 Dec '00

02 Dec '00

That's intersting. I assume you mean that Dolphin resulted in this output: 0 0 0 0 0 In Dolphin I would expect that the variable #i for this block is bound within the context of the UndefinedObject>>DoIt method (or whatever they use). Since the last value bound to #i is 0, all 0's would be displayed as the recursive calls unwind. You should try this code and see the result: recurseblock:=[ :i | Transcript show: i printString; cr. i > 0 ifTrue: [ recurseblock value: ( i - 1)]. Transcript show: i printString; cr. ]. In my opinion it is hard to call this a bug except maybe that they let you do it at all. Blocks and their arguments are really not so independent but are always a part of the context of its parent method call context. (I hope I said that right.) Of course recursive methods work fine in Squeak. If you define on Object class the following method: recursiveCallTest: i i > 0 ifTrue: [ self recursiveCallTest: ( i - 1)]. Transcript show: i printString; cr. And then execute: Object recursiveCallTest: 4. You will get the expected result in the transcript: 0 1 2 3 4 Gnu Smalltalk is intersting because it accepts recursive blocks in the way you were expecting. I executed this code: Smalltalk at: #RB put: [ :i | i > 0 ifTrue: [ (Smalltalk at: #RB) value: ( i - 1)]. stdout nextPutAll: i printString; nextPut: Character nl ]! (Smalltalk at: #RB) value: 4! And got the resulting transcript listed below. Cheers, Jon --------------------------------------------------------------- GNU Smalltalk Ready st> Smalltalk at: #RB put: [ :i | st> i > 0 ifTrue: [ (Smalltalk at: #RB) value: ( i - 1)]. st> stdout nextPutAll: i printString; nextPut: Character nl ]! Execution begins... 213 byte codes executed which took 0.001 seconds, giving 213000 bytecodes/sec 22 primitives, percent 10.33 self returns 0, inst var returns 2, literal returns 0 41 samples since last cache cleanup, percent 19.25 12 cache hits, 29 misses, 29.27 percent hits returned value is nil st> (Smalltalk at: #RB) value: 4! Execution begins... 0 1 2 3 4 2383 byte codes executed which took 0.001 seconds, giving 2383000 bytecodes/sec 177 primitives, percent 7.43 self returns 20, inst var returns 10, literal returns 10 370 samples since last cache cleanup, percent 15.53 316 cache hits, 54 misses, 85.41 percent hits returned value is nil st>

1 0

Re: Making an inverted ellipse morph?
by Joshua Channing Gargus 02 Dec '00

02 Dec '00

Hi, Couldn't you fill a Form with the color you want the rectangle to be, and then fill an ellipse in it with Color transparent, and then draw this Form on the display? I've got to run right now, but take a look at how the sketch editor eraser works. I'd imagine you could do something very similar. Joshua On Sat, Dec 02, 2000 at 02:37:37PM -0800, Dino wrote: > Actually, the Form>>replaceColor:withColor: thing doesn't work. The > stuff in the drawOn: method uses canvas functions that completely avoid > drawing the object if the color is transparent. Simply replacing the > color with transparency would seem to work, but it doesn't. If you > replace the color with transparency, you get a white background. > > It would be cool if somehow we could have boolean operations with shapes > to get "holes" in shapes... like circle xor: square... > > Seems like the only way to draw a circle cut out of a square is to do > the brute force method which I don't really like doing. > > Dean > > Doug Way wrote: > > > > Dean Pu Mao wrote: > > > > > > I'm curious if it's easily possible to make an inverted ellipse. (It's > > > like an ellipse only that the bounding rectangle is not transparent and > > > the ellipse itself is transparent (and all clicks go through it)). > > > > > > I could subclass EllipseMorph and make it so that the > > > containsPoint: returns an opposite value so that clicks go through, but > > > that still doesn't solve the drawOn: problem. How can I make drawOn: to > > > draw everything around the ellipse keeping the insides transparent? > > > > There are a few different approaches you could try... > > > > The brute force method would be to use the debugger to drill down to where the ellipse drawing code actually happens. If you halt in EllipseMorph>>drawOn:, you'll eventually get down to GrafPort>>fillOval:, which seems to be where the action is. You could try to copy this code (into your subclassed drawOn: method or wherever) and modify it so that it draws outside the ellipse instead of inside. > > > > However, there might be an easier way. The first thought I had was, in your new drawOn: method, to first draw a rectangle (fillRectangle:color:) in the ellipse's color, and then draw an ellipse on top of that in with 'Color transparent'. But that doesn't work because the rectangle shows through behind the transparent ellipse. > > > > Something that would probably work would be to use Form>>replaceColor:withColor:. First you'd draw the rectangle on the Form in one color, then you'd draw the ellipse in a second color, and then do a "aForm replaceColor: secondColor withColor: Color transparent". Then draw this Form on your canvas. (Or, simply get the Form from the canvas in the first place, with something like Canvas>>contentsOfArea:.)

1 0

Jump to page: