Squeak-dev September 2000

squeak-dev@lists.squeakfoundation.org

115 participants
411 discussions

Re: [BUG] #hash broken
by Andres Valloud 29 Sep '00

29 Sep '00

Hi. > I just noticed that after the recent changes to #hash the following code > will break: > > (Array with: SmallInteger maxVal with: SmallInteger maxVal) hash. > > Since I'm not sure how to fix it I'd appreciate if somebody could help out. Use the hash fix I sent to the list some time ago. Andres.

2 1

[BUG] Southeast Gradient
by Bert Freudenberg 29 Sep '00

29 Sep '00

Something strange happens with a gradient going down to the left. From a certain line on, all is filled with the second color: Morph new extent: 200@200; fillStyle: ((GradientFillStyle ramp: { 0.0 -> Color red. 1.0 -> Color blue}) origin: 100@100; direction: -50@50; normal: 50@50); openInWorld Or is it just me? -- Bert

1 0

New user
by David G. Durand 29 Sep '00

29 Sep '00

I'm not an experienced Smalltalker by any stretch of the imagination, though I've played with Smalltalk systems a number of times. However, spurred by an interest in playing with the Udanax.com release of the Xanadu code, I've become re-interested in Smalltalk and I'm now thinking of doing some side projects in it. My problem is that I decided that I really want the Antialiased fonts, but I can't get the released version to work properly on Squeak 2.8. I fixed a couple of small breakages (the removal of the QuickPrint Object and its replacement by a method in DisplayScanner, for instance), but it just doesn't work properly. I've tried to trace the rendering flow of control, and I determined that adding a displayString:on:from:to:at:kern: method to AntialiasedFont seems to fix text display in the MVC world, but characters in TextMorphs would always come out as black blobs. At this point, my grasp of the fine points of BitBlt and display depths, etc. fails me. It seems significant to me that I can hack MVC to work and not morphic, but I can't tell why. Has anyone faced and conquered this problem? Are the anti-aliased fonts something that might make their way into the mainline Squeak development at some point? I'm happy to send my changes so far, if they'd help someone else who's interested in this problem and competent to solve it. -- David -- ------------------------------- David Durand david(a)dynamicDiagrams.com Chief Technical Officer Dynamic Diagrams http://www.dynamicDiagrams.com/

2 1

Re: TGen in 2.9a?
by Ned Konz 29 Sep '00

29 Sep '00

Rob Withers wrote: > Now that I think about it, TGen4.1a was a file I ended up with trying to > (unsuccessfully) merge the main file with your fixes. That must be why > it's 4.1a. I'll take a peek this weekend, to try and consolidate them. > Are all of your fixes represented by TGen.4.cs and the > ConcreteStreamFixes file, Ned? I don't remember. I'll poke around and see if I have an ICFP image somewhere. -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

1 0

Re: TGen in 2.9a?
by Rob Withers 29 Sep '00

29 Sep '00

Helge Horch wrote: > At 11:18 28.09.2000 -0400, Mark Guzdial wrote: > >Has anyone got TGen to work in 2.9a? I was trying to load up the > >ray-tracing language GML into 2.9a (using the tgen at > >http://calliso.swiki.net), > > I seem to recall it was <URL:http://callisto.swiki.net> (with an extra t). > > >but I kept getting some really severe crashes [...] I backed down to a 2.8 > >image and had no problem. > > Hmm, it seems to have been developed with 2.7 (1762/1782) due to ICFP > contest constraints. I tried it at the time in a fully updated 2.9a2558 > under Morphic with no problems at all. > > After diff'ing T-Gen.4.cs against T-Gen.4.1a.cs, I decided in a hurry on > this installation procedure: > Hi Helge, I am trying to remember what TGen.4.1a.cs is for. I've attached a diff. There are conversion methods for lf to cr line endings in the TranslatorGenerator method.. There also looks to be some file name manipulation changes. I use TGen.4.cs, then, since I use Linux, I load the FixConcreteStream changeset. I am still using the SourceForge VM, and a base 2.8 image so I'm sorry I can't help. 0.) T-Gen.4.cs (just out of sheer paranoia) 1.) FixConcreteStream-TGen-nk.1.cs 2.) MorphicInstallTGenButton.cs 3.) GML-RayTracer.3.cs is what I do... Somebody sent me a fixes pdf userGuide (thanks to whomever did that!) which I will post to the Callisto site as well (and go for another rev of the tgen tgz file) > > 0.) T-Gen.4.cs (just out of sheer paranoia) > 1.) T-Gen.4.1a.cs (which should be okay on its own, but heck) > 2.) FixConcreteStream-TGen-nk.1.cs > 3.) MorphicInstallTGenButton.cs > 4.) GML-RayTracer.3.cs > > Subsequently I browsed the class side of GMLObject and picked examples from > the method comments at random. Fun! I'm glad you like it. Unfortunately we didn't win, and I think that we were the "three days to render an image and so we eliminated them" entry (see http://www.cs.cornell.edu/icfp/contest_results.htm). I tried to render a Go game image and I quit after 24 hours. > > Anyone know where to get the GML source for the Aspirin(tm) and Adrenalin > examples that are rendered on the Swiki page? > I have them at home so I'll post them later tonight. I thought someone had asked about them before; it must have been you. I was broken for a few days after the contest was over. ;-) cheers, Rob > > >maybe I have a black thumb when it comes to Squeak these days... > > Time to crank up your karma. Submit a couple of class comments. ;) > > HTH, > Helge 1c1 < 'From Squeak2.7 of 5 January 2000 [latest update: #1762] on 27 August 2000 at 5:54:01 am'! --- > 'From Squeak2.7 of 5 January 2000 [latest update: #1762] on 27 August 2000 at 7:38:26 pm'! 6399c6399 < !TGenUI methodsFor: 'initialization' stamp: 'rww 8/27/2000 05:33'! --- > !TGenUI methodsFor: 'initialization' stamp: 'rww 8/27/2000 15:13'! 6407,6408c6407,6408 < do: [:each | specName := specName, each]. < baseName _ baseName directory fullNameFor: specName. --- > do: [:each | specName := specName, each, '.']. > baseName _ baseName directory fullNameFor: (specName copyFrom: 1 to: (specName size - 1)). 7193c7193 < !LL1Parser methodsFor: 'parsing' stamp: 'tfei 10/29/1999 21:01'! --- > !LL1Parser methodsFor: 'parsing' stamp: 'rww 8/27/2000 16:57'! 9498c9498 < !TranslatorGenerator methodsFor: 'public access' stamp: 'tfei 11/12/1999 16:33'! --- > !TranslatorGenerator methodsFor: 'public access' stamp: 'nk 8/27/2000 19:12'! 9509a9510 > tokenSpecification replaceAll: Character lf with: Character cr. 9519a9521 > grammarSpecification replaceAll: Character lf with: Character cr. 9526a9529 > inputText replaceAll: Character lf with: Character cr.

2 1

RE: [ANN] stencili pre-announcement
by Raab, Andreas 29 Sep '00

29 Sep '00

> If you're really interested, I'll upload the image/changes. Yes I am ;-) - Andreas

2 1

[BUG] in 2.9a
by Raymond Asselin 29 Sep '00

29 Sep '00

I noted that the BobUI is now include in the 2.9a release...after trying to load it a couple of time without success :) Fine. But...as I understand... a method to initialize a class variable (CloseForm) in the class UIWindowMorph is lacking (UIWindowMorph initialize). So I tried to add it copying that same method from the CS I have at hand, but the CS #2630 (curiously named BadHashFix) :) don't like that...there is a new method lacking 'hashMultiply' which seems to give hard time to others Squeakers. "The crash is because the change set redefines SequencableCollection>>hash in a way incompatible with the current definition. " as recalled Bob Arning to Mark Guzdial about TGen. Sorry.. no fix.

1 0

Re: TGen in 2.9a?
by Bob Arning 29 Sep '00

29 Sep '00

On Thu, 28 Sep 2000 11:18:19 -0400 Mark Guzdial <guzdial(a)cc.gatech.edu> wrote: >Has anyone got TGen to work in 2.9a? I was trying to load up the >ray-tracing language GML into 2.9a (using the tgen at >http://calliso.swiki.net), but I kept getting some really severe >crashes ("Hit any key to restart"). Then I took a fresh 2.9a image, >loaded all updates, then tried again from MVC. I kept getting that >Object did not understand variableSubclass... to create classes like >Grammar. I backed down to a 2.8 image and had no problem. Mark & others, The crash is because the change set redefines SequencableCollection>>hash in a way incompatible with the current definition. I manually edited that method out of the change set and it filed in OK. Whether or not it still works, I cannot say. Cheers, Bob

1 0

[ANN] stencili pre-announcement
by Bolot Kerimbaev 29 Sep '00

29 Sep '00

I'll need some time to package it, but here's a teaser: http://gilligan.cc.gt.atl.ga.us/~bolot/squeak/stencili contains a few screen shots. Stencili is an end-user vector-graphics support package for Squeak. Currently, it offers some of EU functionality of Xara and Illustrator. If you're really interested, I'll upload the image/changes. More information will be available shortly. Bolot

1 0

[CCC] Re: TGen in 2.9a?
by Helge Horch 29 Sep '00

29 Sep '00

>At 11:18 28.09.2000 -0400, Mark Guzdial wrote: > >but I kept getting some really severe crashes [...] I backed down to a 2.8 > >image and had no problem. By golly, Mark is right, of course. Filing in TGen-4.cs or 4.1a.cs under Squeak 2.9a2710 results in a strange MNU/unable-to-open-debugger crash. The crash seems to occur around line 5492 of said changesets (coincides with progressbar going to ca. 40%). The last bits from the .changes file mention [--- snip ---] !SequenceableCollection methodsFor: 'comparing' stamp: 'tfei 10/31/1999 10:49' prior: 67574585! hash "Answer an integer hash value for the receiver such that, -- the hash value of an unchanged object is constant over time, and -- two equal objects have equal hash values." ^self size! ]style[(205)f1! ! ----STARTUP---- [--- snip ---] However, there is *no* style information in the original changeset for that #hash method. Or any method. In fact, the string ']style[' does not occur at all in the cs. I'm stumped as to where and why things go wrong, though. What is it since update 2558 that is logging generated style information? Ouch -- bad karma here too, it seems. I'd better include a class comment quick: [---- snip ---] 'From Squeak2.9alpha of 23 June 2000 [latest update: #2710] on 29 September 2000 at 2:07:53 am'! !PinMorph commentStamp: 'hh 9/29/2000 02:07' prior: 0! I am the visual representation of a pin on a Component. I display as an image indicating whether I am an input, output (triangle) or bidirectional (diamond) pin. I maintain a collection of wires (and through them, the connecting pins) that I am connected to. instance variables: - component: a ComponentLikeModel - pinSpec: a PinSpec - pinForm: a Form - wires: an OrderedCollection of WireMorphs! [---- snip ---] Gee, I hope this helps, I dearly do. Cheers, Helge

1 0

Jump to page: