Squeak-dev February 2000

squeak-dev@lists.squeakfoundation.org

163 participants
655 discussions

Re: [UPDATES] 26 more for 2.8a
by merlyn＠stonehenge.com 29 Feb '00

29 Feb '00

>>>>> "Dan" == Dan Ingalls <Dan.Ingalls(a)disney.com> writes: Dan> 1876Explain-th -- Torge Husfeldt -- 2 February 2000 Dan> Explain in code pain was broken [...] A freudian slip (I mean, it hurts to code sometimes, but... )? :-) -- Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 <merlyn(a)stonehenge.com> <URL:http://www.stonehenge.com/merlyn/> Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

1 0

Re: [UPDATES] 26 more for 2.8a
by John Clonts 29 Feb '00

29 Feb '00

Sorry, but I'm missing something here. Are you referring to a URL other than the one supplied below? Or, what do you mean by this?. Thanks, John Mark Guzdial wrote: > > A Word of THANKS to Bruce! It's *SOOO* Nice to be able to just > double-click his URL to peek at the latest-and-greatest without even > dealing with updating code from server. THANK YOU, BRUCE! > > Mark > > >Hi, > > Now ftp://st.cs.uiuc.edu//Smalltalk/Squeak/2.7/files/Squeak2.8a-beo.zip > >is updated with these changes. > > > >cheers > > > >bruce

1 0

RE: [CCC] Object class comment
by Norton, Chris 29 Feb '00

29 Feb '00

[Thank you Ranjan for replying to Andrew for me...] On 2/28/2000, Ranjan Bagchi [ranjan_bagchi(a)yahoo.com] wrote: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "Proxies don't subclass UndefinedObject, they subclass nil, much like how Object subclasses nil." ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I believe this is correct. On 2/28/2000, Andrew P. Black [black(a)cse.ogi.edu] wrote: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ "It does achieve one _undesired_ effect -- making all the Proxy objects print as nil!" ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Although I do not have a ton of recent experience with proxies, I think that the proxy class implements #doesNotUnderstand: to get around these problems. What you do is you define a "realSubject" method that returns the class you are masking with your proxy. Any desirable protocol that the realSubject has that you want your proxy to have you can redirect to its realSubject. Of course, if the protocol is commonly used, you should probably avoid the #doesNotUnderstand: and instead implement the necessary protocol in the proxy. hopefully that wasn't too confusing... :-) ---==> Chris

1 0

Re: Change the Squeak UI Look
by Dino 29 Feb '00

29 Feb '00

Where can I get a changeset for this "skins" thing? Dino "Stephan B. Wessels" wrote: > > I wrote some code to do this which is avaiable via the swiki at > http://minnow.cc.gatech.edu/squeak/683. Take a look at the bottom of the web > page. This code works with V2.8a. > > Maybe that will help. > > - Steve > > montgomery farrar tidwell wrote: > > > Howdy, > > > > Torsten.Bergmann(a)phaidros.com wrote: > > > > > > ... > > > - open a file list, select a JPEG or GIF and select 'open image as > > > background' > > > > when i do this, the image is tiled. is there anyway to get it to center > > and stretch? > > -- > > > > \\//_

3 2

[ENH] Stream input in words/lines/paragraphs as string/object
by Stephen Travis Pope 29 Feb '00

29 Feb '00

(This time *with* the promised attachment...) Adds a number of string-reading methods such as nextLineString and nextWordObject (evaluates the next word) to PositionableStream. -- stp Stephen Travis Pope stp(a)create.ucsb.edu -- http://www.create.ucsb.edu/~stp "Change Set: streamIO.cs Date: 23 Feb 2000 Author: Stephen Travis Pope Adds a number of string-reading methods such as nextLineString and nextWordObject to PositionableStream." 'From Squeak2.8alpha of 13 January 2000 [latest update: #1852] on 19 February 2000 at 10:35:55 pm'! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:40'! nextLineObject "Deliver the next line string as an object." | string | string := self nextLineString. ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. ^Compiler evaluate: string! ! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:23'! nextLineString "Deliver the next line (up to CR, possibly empty) as a String." self skipSeparators. ^self upTo: Character cr ! ! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:43'! nextParagraphObject "Deliver the next whole paragraph as an object." | string | string := self nextParagraphString. ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. ^Compiler evaluate: string! ! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/19/2000 21:24'! nextParagraphString "Deliver the next text block (up to CR/CR, possibly empty) as a String." self skipSeparators. ^self upToAll: (String with: Character cr with: Character cr)! ! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:41'! nextWordObject "Deliver the next ascii word as an object." | string | string := self nextWordString. ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. ^Compiler evaluate: string! ! !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/19/2000 21:24'! nextWordString "Deliver the next 'word' (up to and separator, possibly empty) as a String." | newStream element | self skipSeparators. newStream _ WriteStream on: (collection species new: 100). [self atEnd or: [(element _ self next) isSeparator]] whileFalse: [newStream nextPut: element]. ^newStream contents! !

1 0

Re: [ENH] Stream input in words/lines/paragraphs as string/object
by Lex Spoon 29 Feb '00

29 Feb '00

By the way, squeak already has some of this stuff. For instance, nextLine and nextWord. Lex Stephen Travis Pope <stp(a)create.ucsb.edu> wrote: > This is a multi-part message in MIME format. > --------------3C7A918CACB687B1CA8BBCA7 > Content-Type: text/plain; charset=us-ascii > Content-Transfer-Encoding: 7bit > > > (This time *with* the promised attachment...) > > Adds a number of string-reading methods such as nextLineString and > nextWordObject (evaluates the next word) to PositionableStream. > > -- > > stp > Stephen Travis Pope > stp(a)create.ucsb.edu -- http://www.create.ucsb.edu/~stp > --------------3C7A918CACB687B1CA8BBCA7 > Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; x-mac-creator="522A6368"; > name="streamIO.cs" > Content-Transfer-Encoding: 7bit > Content-Description: Unknown Document > Content-Disposition: inline; > filename="streamIO.cs" > > "Change Set: streamIO.cs > Date: 23 Feb 2000 > Author: Stephen Travis Pope > > Adds a number of string-reading methods such as nextLineString and nextWordObject to PositionableStream." > > 'From Squeak2.8alpha of 13 January 2000 [latest update: #1852] on 19 February 2000 at 10:35:55 pm'! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:40'! > nextLineObject > "Deliver the next line string as an object." > > | string | > string := self nextLineString. > ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. > ^Compiler evaluate: string! ! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:23'! > nextLineString > "Deliver the next line (up to CR, possibly empty) as a String." > > self skipSeparators. > ^self upTo: Character cr > ! ! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:43'! > nextParagraphObject > "Deliver the next whole paragraph as an object." > > | string | > string := self nextParagraphString. > ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. > ^Compiler evaluate: string! ! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/19/2000 21:24'! > nextParagraphString > "Deliver the next text block (up to CR/CR, possibly empty) as a String." > > self skipSeparators. > ^self upToAll: (String with: Character cr with: Character cr)! ! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/18/2000 16:41'! > nextWordObject > "Deliver the next ascii word as an object." > > | string | > string := self nextWordString. > ((string == nil) or: [string isEmpty]) ifTrue: [^nil]. > ^Compiler evaluate: string! ! > > !PositionableStream methodsFor: 'fileIn/Out' stamp: 'stp 02/19/2000 21:24'! > nextWordString > "Deliver the next 'word' (up to and separator, possibly empty) as a String." > > | newStream element | > self skipSeparators. > newStream _ WriteStream on: (collection species new: 100). > [self atEnd or: [(element _ self next) isSeparator]] > whileFalse: [newStream nextPut: element]. > ^newStream contents! ! > > > > --------------3C7A918CACB687B1CA8BBCA7--

1 0

RE: [newbie] clipped method display in system browser
by Lex Spoon 29 Feb '00

29 Feb '00

Doug Way <dway(a)mat.net> wrote: > So, I think we're back to hoping for the self-installing Squeak to work > sometime soon... :-) Sigh. > Well the original point was, this really is unclear. What is a self-installing Squeak? Is it an exe you can click on and have it run? If so, it's something you have to do on a platform-by-platform basis. And in fact, it's basically done (in spirit) for certain distributions of Linux.... I'm sure there are utilities to make a Squeak-package for Windows, and perhaps there are some for Macs as well. In fact, it's a tad surprising that no one has released a one-EXE distribution of Squeak for Windows. Oh well, I happen to actually think ZIP and TAR files aren't too bad, so take it all with a grain of salt. :) Lex

1 0

Re: RE: [newbie] clipped method display in system browser
by Lex Spoon 29 Feb '00

29 Feb '00

Bob Arning <arning(a)charm.net> wrote: > On Sun, 27 Feb 2000 20:28:20 -0500 "Lex Spoon" <lex(a)cc.gatech.edu> wrote: > >It's just the Mac distribution which has this particular gotcha. Is > >there really no way to work around this? Has anyone tried setting the > >app-type-thingy to something other than TEXT and seeing if StuffIt will > >still do it's auto-conversion? > > Umm... I think it's more than just Mac, at least in some cases. I have seen Linux ftp clients do a conversion from CR to LF. > True, but it's much worse on Macs due to the "smartness". If you ftp with ascii mode, then probably *all* of your files will be corrupted, and so you'll notice it pretty quickly. StuffIt only corrupts one of the files, and so the error goes undetected for a while. Well anyway, perhaps changing the type to "STdb" will fix things. By the way, there are replacements for the stock ftp client, which default to binary mode instead of ascii mode. I find this to be much nicer. "ncftp" is one such client, and it comes installed on most Linux systems. Lex

1 0

[UPDATES] 26 more for 2.8a
by Dan Ingalls 29 Feb '00

29 Feb '00

Folks - These incorporate many of the submissions from SqueakVille in the last month. Once again, much credit goes to Stefan Matthias Aust for taking the trouble to cull through them and check for various conflicts. Enjoy - Dan --------------------- 1874cleanup-tkNE -- Ted Kaehler -- 23 February 2000 Defines (SmartRefStream cleanUpCategories) to remove obsolete conversion methods. Includes 29 removals or currently obsolete conversion methods." 1875projectEnv-tkNJ 1876Explain-th -- Torge Husfeldt -- 2 February 2000 Explain in code pain was broken because doitFlag in Parser was sometimes not initialized correctly. I think I found the right place to fix it." 1877FontDisplayEnh-acg -- Andrew C. Greenberg -- 2 February 2000 A small enhancement to display all characters of a strikefont." 1878SysWindowReframe-dns -- David N Smith -- 2 February 2000 A small update to SystemWindow which provides an option to suppress reframe handles of the window. This can be desirable when an application wants control over its own layout and wants no one (like an 8-year old) to resize the window. The default allows reframing; the option can be set and reset during execution however one wishes. Please note the postscript." 1879FillInTheBlankFix-sma -- Stefan Matthias Aust -- 10 February 2000 Stewart MacLean noticed a bug in the MVC version FillInTheBlankController." 1880GraphicalMenuBug-dns -- David N Smith -- 4 February 2000 Small bug in the comment of GraphicalMenu>>initializeFor:... No big deal." 1881PrefFix-ssa -- Sam S. Adams -- 9 February 2000 Spelling error in preferences fixed." 1882LastUpdString-sma -- Stefan Matthias Aust -- 6 February 2000 I tried a ''Smalltalk majorShrink; abandonSources; lastRemoval'' on a 2.8a/1789 image and noticed that here's an error in SystemDictionary>>lastUpdateString which is called from condenseChanges during abandonSources. When this method tries to determine the patch level by calling ''ChangeSorter highestNumberedChangeSet'' the ChangeSorter class has already been stripped. This changeset works around this problem." 1883ColorPrintOn-dns -- David N Smith -- 6 February 2000 Print named colors with names instead of (Color r: 0.833 g: 0.588 b: 0.524)" 1884BlobEnh-dns -- David N Smith -- 9 February 2000 Blobs stay upfront now and you can click through them (use Cmd-click to get them). They also don't show handles when opened from the ''new morph'' menu (sma)." 1885DecompileUnknown-th -- Torge Husfeldt & SMA -- 7 February 2000 A smaller (and enhanced) version of Torge's submission to decompile unbound methods. Torge wrote: 'Here's a little snippet I wrote to decompile CMs that are no longer in any method dictionary (such as DoIts that are about to be deleted, but there are other examples). Why shouldn't we want do do this sometimes.' " 1886ProtoObjComment-di -- Dan Ingalls -- 8 February 2000 Please note the postscript! Dan's comment template the ProtoObject's comment which started up a long discussion on the mailing list. There were better solutions provided by Dan and Vassili Bykov but this is a first step --sma." 1887UncatchableHalt-sma -- Stefan Matthias Aust -- 12 February 2000 This is my proposal for an uncatchable halt, that is [self halt] on: Error do: [] will not silently do nothing but still open a debug notifier. On the other hand, [self halt] on: Halt do: [] will catch the halt." 1888BrowseIt-slr -- Steve Rees -- 11 February 2000 Minor fix to browseIt on ParagraphEditor. Previously ParagraphEditor assumed that a selection beginning with an uppercase letter represented a Class name and that a selection beginning with a lowercase letter was a method name. Whilst the second assumption is valid, the first is not, since although conventionally method names begin with a lowercase letter they can also begin with an uppercase letter. This changeset modifies the assumptions in ParagraphEditor>>browseIt so that if the selected name begins with an uppercase letter, but no entry is found in Smalltalk a search is done for implementing methods with that name." 1889HtmlFmtFix-jcc -- John Clonts -- 11 February 2000 A small correction in the example comment. But while I was testing the code, I noticed a problem in DisplayScanner using MVC and fixed it --sma" 1890CmdKeyDesc-sma -- Stefan Matthias Aust -- 12 February 2000 Updated command key description as well as the disablement of the enablement of Ctrl+( enclosement keys for Windows because this disabled this function on Macs. Okay? Unfortunately, the different platforms are quite different in what key codes they generate. I tried to work around this problem but failed. I'd suggest to double check the VM code for all platforms and make sure that they generate identical codes. This simplify the paragraph editor code and also enables a few more keyboard shortcuts for Non-Macs. Frankly, I'm a bit annoyed that often platform-independence for Squeak actually means Mac-dependence. Added back Cmd-Shift+I for exploreIt." 1891ClassComments-dwh -- Dwight Hughs and Göran Hultgreen -- 26 February 2000 Added method categories to the PenRecorder --sma" 1892TTStringMorph-sma -- Stefan Matthias Aust -- 26 February 2000 Brian Mc Enery noticed: If multiple TTSampleString's are being copied (where each instance uses the same font) then each copy will automatically have a new version of the TTFontDescription including the glyphs and the mapping table. This occurs if a TTSampleString is created out of a parts bin, or if it is duplicated using the halo menu." 1893EnhTranscript-sma -- Stefan Matthias Aust -- 26 February 2000 Now, show: can show any object, not just Strings. Also added method categories to the Transcripter class." 1894FillInFixes-ssa -- Sam S. Adams -- 14 February 2000 Here's a fix for FillInTheBlank so it complies with Preference settings for scroll bars and rounded corners." 1895EmptySymbol-sma -- Stefan Matthias Aust -- 26 February 2000 Richard A. Harmon noticed that the empty symbol ('' asSymbol) didn't work." 1896RmEmtyCatFix-sma -- Stefan Matthias Aust -- 27 February 2000 The browser was missing the function #removeEmptyCategories. Added." 1897NameOfClass-sma -- Stefan Matthias Aust -- 27 February 2000 It seems that somebody accidentally created a 'NameOfClass' class. Removed." 1898PPBlockTemps-sma -- Stefan Matthias Aust -- 27 February 2000 Pretty printing of block temporaries didn't work. Fixed." 1899segBug-tkNL -- Ted Kaehler -- 28 February 2000 Update 1860 added debugging class variables to ImageSegmentRootStub. They allow the user to see the stack of what caused a segment fault. Very useful. However, the new class variables were never defined in the fileOut. Nor were they initialized. Any fault caused an error. We fix it here."

1 0

RE: [CCC] Object class comment
by Ranjan Bagchi 29 Feb '00

29 Feb '00

Andrew, Proxies don't subclass UndefinedObject, they subclass nil, much like how Object subclasses nil. i.e. nil subclass: #Object [...] nil subclass: #ObjectTracer [...] nil subclass: #Proxy [...] Ranjan --- "Andrew P. Black" <black(a)cse.ogi.edu> wrote: > At 14:22 -0500 2000.2.28, Norton, Chris wrote: > >In addition to the other fine responses to your > question (ProtoObject is the > >superclass of Object, et. al), I believe that > people also define Proxy > >objects by subclassing nil (UndefinedObject), > rather than Object. Proxy > >objects are common ... > > > Unfortunately, since UndefinedObject is itself a > subclass of Object, > making Proxy a subclass of UndefinedObject does not > achieve the > desired effect. (It does achieve one _undesired_ > effect -- making > all the Proxy objects print as nil! > > I believe that the reasons for the creation of > ProtoObject include > the need to have a place to put Proxy objects that > is not a subclass > of Object. ObjectOut and ImageSegmentRootStub are > two proxy objects > in the image that are subclsses of ProxyObject. > Marcel's Trampoline > code should probably be modified so that Trampoline > is a subclass of > ProtoObject rather than UndefinedObject. > > > > __________________________________________________ Do You Yahoo!? Talk to your friends online with Yahoo! Messenger. http://im.yahoo.com

1 0

Jump to page: