Squeak-dev January 2000

squeak-dev@lists.squeakfoundation.org

131 participants
468 discussions

[FIX] explainNumber
by Andrew P. Black 06 Jan '00

06 Jan '00

I just noticed that selecting -> (the message selector that creates an association) in a browser and asking it to "explain" it, results in a walkback. The problem is in the logic for deciding is the selected string is a number, which of course it isn't. This is not a big issue, but it is one that a new user might easily fall over. An apparent fix is attached. Andrew

1 0

[BUG?] Decompiler can't decompile TestCClass1>>method9
by Michel Tilman 06 Jan '00

06 Jan '00

[Squeak 2.6 / Win98] Not important to me, just thought I would let you know ... michel

3 2

RE: [Q] Exact WarpBlt (etc.) transformation to 1/3 width?
by Raab, Andreas 05 Jan '00

05 Jan '00

Henrik, > And Andreas, btw, regarding using PostScript (cubic) font > outlines, you wrote that they would be slow to convert to > squared ones. I probably didn't write this exactly (or if I did then it wasn't accurate). Converting cubic into squared curves isn't a big problem (though it'll be inaccurate of course). What I was (probably) referring to is an extension of the Balloon engine to handle cubic curves directly. > But you only do this once, ie. on import, right? Hm ... not necessarily. It's usually not a good idea to throw away accurate information ;-) So I guess I'd rather keep the cubic curves and convert them on demand (and who knows, some future Balloon version might be able to handle cubics directly). > So is there any other problem? There has never be a 'problem' ;-) It's just work. A.

2 1

[FIX] World's viewbox obsolete
by Bert Freudenberg 05 Jan '00

05 Jan '00

"Change Set: wiwfix-bf Date: 5 January 2000 Author: Bert Freudenberg Removes numerous references to the world's viewBox origin. A not so long time ago the viewBox origin determined an offset that had to be used for low level stuff, while nowadays all coords are global, even in inner worlds. One bug example is changing the border width of a morph in an inner world. Also fixes Morph>>openInWorld so it is positioned at the world's viewbox's origin. ToDo: ObjectExplorerWindow should be removed - it contains copies of old code. Check PasteUpMorph>>displayWorld* for non-deferred updating." -Bert-

2 1

[FIX] Indented text and selection
by Hans-Martin Mosner 05 Jan '00

05 Jan '00

This changeset fixes a bug with selection to the left of indented text (the virtual tabs used for indenting would cause a crossedX condition, which leads to wrong cursor positioning) Hans-Martin 'From Squeak2.7alpha of 11 October 1999 [latest update: #1716] on 5 January 2000 at 5:40:55 pm'! "Change Set: Indentation-hmm Date: 5 January 2000 Author: Hans-Martin Mosner Fixes a bug with selecting to the left of text with a TextIndent attribute"! !CharacterBlockScanner methodsFor: 'scanning' stamp: 'hmm 1/5/2000 07:13'! characterBlockAtPoint: aPoint index: index in: textLine | runLength lineStop done stopCondition | line _ textLine. characterIndex _ index. " == nil means scanning for point" characterPoint _ aPoint. (characterPoint == nil or: [characterPoint y > line bottom]) ifTrue: [characterPoint _ line bottomRight]. (text isEmpty or: [(characterPoint y < line top or: [characterPoint x < line left]) or: [characterIndex ~~ nil and: [characterIndex < line first]]]) ifTrue: [^ (CharacterBlock new stringIndex: line first text: text topLeft: line leftMargin@line top extent: 0 @ textStyle lineGrid) textLine: line]. rightMargin _ line rightMargin. destX _ leftMargin _ line leftMarginForAlignment: textStyle alignment. destY _ line top. lastIndex _ line first. self setStopConditions. "also sets font" self adjustForIndentationLevel. runLength _ text runLengthFor: line first. characterIndex ~~ nil ifTrue: [lineStop _ characterIndex "scanning for index"] ifFalse: [lineStop _ line last "scanning for point"]. runStopIndex _ lastIndex + (runLength - 1) min: lineStop. lastCharacterExtent _ 0 @ line lineHeight. spaceCount _ 0. done _ false. [done] whileFalse: [stopCondition _ self scanCharactersFrom: lastIndex to: runStopIndex in: text string rightX: characterPoint x stopConditions: stopConditions kern: kern displaying: false. "see setStopConditions for stopping conditions for character block operations." self lastCharacterExtentSetX: (specialWidth == nil ifTrue: [font widthOf: (text at: lastIndex)] ifFalse: [specialWidth]). (self perform: stopCondition) ifTrue: [^ (CharacterBlock new stringIndex: (characterIndex==nil ifTrue: [lastIndex] ifFalse: [characterIndex]) text: text topLeft: characterPoint extent: lastCharacterExtent) textLine: line]]! ! !CharacterBlockScanner methodsFor: 'private' stamp: 'hmm 1/5/2000 07:05'! adjustForIndentationLevel "Advance destX to the indented position without executing the tab action. IndentationLevel is cleared to suppress its effect later. This is needed to fix a bug when a selection is made to the left of indented text. - hmm" self indentationLevel timesRepeat: [ destX _ textStyle nextTabXFrom: destX leftMargin: leftMargin rightMargin: rightMargin]. indentationLevel _ 0! ! !CharacterBlockScanner methodsFor: 'private' stamp: 'hmm 1/2/2000 20:56'! buildCharacterBlockIn: para | lineIndex runLength lineStop done stopCondition | "handle nullText" (para numberOfLines = 0 or: [text size = 0]) ifTrue: [^ CharacterBlock new stringIndex: 1 "like being off end of string" text: para text topLeft: (para leftMarginForDisplayForLine: 1) @ para compositionRectangle top extent: 0 @ textStyle lineGrid]. "find the line" lineIndex _ para lineIndexOfTop: characterPoint y. destY _ para topAtLineIndex: lineIndex. line _ para lines at: lineIndex. rightMargin _ para rightMarginForDisplay. (lineIndex = para numberOfLines and: [(destY + line lineHeight) < characterPoint y]) ifTrue: ["if beyond lastLine, force search to last character" self characterPointSetX: rightMargin] ifFalse: [characterPoint y < (para compositionRectangle) top ifTrue: ["force search to first line" characterPoint _ (para compositionRectangle) topLeft]. characterPoint x > rightMargin ifTrue: [self characterPointSetX: rightMargin]]. destX _ leftMargin _ para leftMarginForDisplayForLine: lineIndex. nextLeftMargin_ para leftMarginForDisplayForLine: lineIndex+1. lastIndex _ line first. self setStopConditions. "also sets font" self adjustForIndentationLevel. runLength _ (text runLengthFor: line first). characterIndex ~~ nil ifTrue: [lineStop _ characterIndex "scanning for index"] ifFalse: [lineStop _ line last]. (runStopIndex _ lastIndex + (runLength - 1)) > lineStop ifTrue: [runStopIndex _ lineStop]. lastCharacterExtent _ 0 @ line lineHeight. spaceCount _ 0. done _ false. [done] whileFalse: [stopCondition _ self scanCharactersFrom: lastIndex to: runStopIndex in: text string rightX: characterPoint x stopConditions: stopConditions kern: kern displaying: false. "see setStopConditions for stopping conditions for character block operations." self lastCharacterExtentSetX: (font widthOf: (text at: lastIndex)). (self perform: stopCondition) ifTrue: [^ CharacterBlock new stringIndex: lastIndex text: text topLeft: characterPoint extent: lastCharacterExtent]]! !

1 0

Re: MajorShrink & SystemTracer
by Yoshiki Ohshima 05 Jan '00

05 Jan '00

Hi, Probably someone else will give us much more better explanations:-) > Could you tell me which classes are the "obsolete" > classes? (I did not know that term). There is (should be) no obsolete class in the vanilla image. But every class about to be removed from system is once marked as "obsolete". Usually, the class object is unreachable from anywhere and reclaimed by the garbage collector sometime later. However, sometimes the "removed" class object is still reachable from somewhere (existing instances, method literals, etc.) and remains in the image. Such class is called obsolete class. See the implementors and senders of #obsolete. Though I'm not 100% confident it is the problem (sorry), but I guess it has a relation to the problem. -- Yoshiki

2 1

Re: testing smalltalk applications.
by Stephane Ducasse 05 Jan '00

05 Jan '00

To go look at they have Sunit testing frmaework for Squeak http://www.xprogramming.com/software.htm > Resent-date: Mon, 03 Jan 2000 16:05:09 +0000 > Date: Mon, 03 Jan 2000 08:03:19 -0800 > Resent-from: squeak(a)cs.uiuc.edu > From: TECHIN A KANG <TECHIN.A.KANG(a)mail.sprint.com> > Subject: testing smalltalk applications. > Resent-sender: squeak-request(a)cs.uiuc.edu > To: squeak(a)cs.uiuc.edu > Resent-cc: recipient list not shown: ; > Resent-message-id: <3aBXrB.A.xNG.1iMc4(a)jerry.cs.uiuc.edu> > MIME-version: 1.0 > X-Loop: squeak(a)cs.uiuc.edu > X-OpenMail-Hops: 1 > X-Mailing-List: <squeak(a)cs.uiuc.edu> archive/latest/13973 > > Hi, > > Is there any package in squeak community to "test" smalltalk > application? > > In general, what are the methods to test smalltalk application? Module > or Control, or View individually, or combination of M-V-C. Manual test, > or automatic test? > > Testing to my concern is validating a feature or features that are up to > expectation. For example, how do I know at specific area the color is > red, not green; not exactly #CC0000 vs #-00CC00, but relatively. > > Regards, > Techin Alex Kang > Stephane DUCASSE (ducasse(a)iam.unibe.ch) http://www.iam.unibe.ch/~ducasse/ "if you knew today was your last day on earth, what would you do different? ... especially if, by doing something different, today might not be your last day on earth" Calvin&Hobbes University of Bern, Institut fuer informatik and Mathematik IAM-SCG, 10 neubruckstrasse, CH-3012 Bern, Switzerland.

1 0

[UPDATES] Closing on 2.7 -- really
by Dan Ingalls 05 Jan '00

05 Jan '00

Folks - Here are 12 more updates. I hope to get 2.7 out in the next day or so now. Any loose ends will henceforth become part of the 2.8alpha project. - Dan ---------------------- 1749LZ77 -- Andreas Raab -- 2 January 2000 Adds LZ77 based compression (e.g. PKZip/GZip) to Squeak. GZip compression is available from the FileList ('compress file') but will take a *long* time without the plugin support. The quality and speed of compression is similar to that of other compressors from the LZ77 family. Note: I have run a regression test on 83,210 files (11,583,048,046 bytes; every accessible file on my hard disk) to make sure compression works as expected. "Remove some old classes we don't need anymore" Smalltalk removeClassNamed:#InflatePlugin. Smalltalk removeClassNamed:#GZipStream. Smalltalk removeClassNamed:#ZLibStream. "Add two new pools" Smalltalk at: #ZipConstants ifAbsentPut:[Dictionary new]. Smalltalk at: #GZipConstants ifAbsentPut:[Dictionary new]. 1750CompressionTweaks-ar -- Andreas Raab -- 2 January 2000 Tweaks for compressing non-binary contents (e.g., using Strings instead of ByteArrays)." 1751MoreCompressionTweaks -- Andreas Raab -- 2 January 2000 Fixes problems introduced with the last change set (seems like I wasn't really recovered from the party yet...)" 1752env6-tkML 1753filePath-tkMQ -- Ted Kaehler -- 2 January 2000 Recent changes to FileDirectories made the top level of the FileList not work. If selected [] to choose your hard disk or a server, and chose the hard disk, it broke. In that case, we now look up HardDisk: instead of just HardDisk (with suitable delimiter for each platform). And don't call checkName:fixErrors: when the local name is empty (i.e. the top directory)." 1754preDebug-tkMR -- Ted Kaehler -- 3 January 2000 Allow morphic preDebug windows to use the full width for the message. Widen the window to 450 pixels because it now says MessageNotUnderstood: instead of the previous shorter text." 1755TTFontReaderPatch-sma -- Stefan Matthias Aust (sma(a)3plus4.de) -- 1 January 2000 This change set adds support for cmap format 4 and 6 encoding to the TTFontReader class (format 4 is actually recommended by Microsoft). This allows to read in most existing true type fonts without errors - actually, of some 30 freeware and shareware fonts I tested all worked. It also corrects the format 0 encoding and fixes a problem, that the reader was preferring an missing mac encoding over an existing windows encoding. Furthermore, a bug in TTSampleFontMorph's smoothing menu has fixed. I also reformatted a few methods." 1756ScanCharacterFailFix -- Dan Ingalls -- 3 January 2000 Fixes a strange bug originally reported by Bob Arning, wherein if the text scanCharacters primitive failed due to, eg, translucency mode, its destX value had been changed by partial operation of the primitive up to when the BB failure takes effect. This resulted in improper positioning when the Squeak-level retry was attempted. Now when such a failure occurs, no store-back occurs, and the retry takes place normally. 1757DirectoryEntryFix -- Dan ingalls -- 4 January 2000 Fixes a problem pointed out by Ted Keahler, and clarified by Dragos-Anton Manolescu. When the 6th token of a ServerDirectory entry is a time, Squeak had assumed the year was the current year, whereas it should be the year that this month must be, given that it is within the past 6 months. I have now fixed this. 1758ScanCharsFailFix2 -- Dan Ingalls -- 4 January 2000 Supplants the change of #1756 with an earlier check for failure. While the other fix did improve display, several situations remained with glitches. Those are now fixed. 1759DecompileFix -- Dan Ingalls -- 4 January 2000 Fixes a decompile bug introduced as part of the new brace compilation/decompilation. 1760TwoWayScrollPane -- Bob Arning -- 4 January 2000 Since it was in Undeclared I(di) took the liberty of including the version in FSM. We probably need to clean this up some, but at least there's code there now.

1 0

vm & persistence
by Ryan Daum 05 Jan '00

05 Jan '00

some squeakers may be interested in the documents at this url: http://field.medicine.adelaide.edu.au/~colin/coldpub/ and also at http://coldstore.linuxbox.com/ -- -------------------------------------------------------------------- "The only truth that gets through will be what we force through: the victory of reason will be the victory of people who are prepared to reason, nothing else." -- Bertolt Brecht, "Galileo" -----------------------------------------------------ryan(a)techno.ca-

1 0

Thank you
by Maurice Rabb 04 Jan '00

04 Jan '00

As we tick down to the last few hours of the year, now is a good time to publicly acknowledge some of the great work that many in the Squeak community have been quietly doing. Though I have been quiet for most of this past year, we here at Stono have been very busy. We are doing almost 100% our development work in Squeak. Last month we completed rev.2 of an application that we developed to do animated presentations for promotions at large scale sponsored events (e.g. at stadiums and conferences). We would not have been able to complete our work without the help of many great Squeakers. In particular, I want to thank Bob Arning and Doug Way who did some contract work for us. Doug Way did a _fabulous_ job building the new user interface for our application. Bob Arning, as most of you know, wrote the Morphic UI tool kit on which the interface to our application is based. Bob was extremely helpful in write code in a pinch, and helping me understand various Squeak matters. I am extremely grateful for his patience. I also want to thank Jim Benson, Bob Hartwig, and Mark Schwenk. Jim built a really cool QuickTime plug-in that we experimented with for a while. Bob and Mark have helped me out from time to time, and run the CSTUG (Chicago Smalltalk Users' Group) which has put on some great sessions. (Any of you in SqC coming to Chicago anytime soon? ;-) ) Finally, a want to thank SqC, and the entire Squeak list for providing such great technical support, thoughtful discussion, and technical comraderie. Have a great New Year! Sincerely, --Maurice --------------------------------------------------------------------------- Maurice Rabb 773.281.6003 Stono Technologies, LLC Chicago, USA

2 3

Jump to page: