Squeak-dev May 2002

squeak-dev@lists.squeakfoundation.org

160 participants
453 discussions

[FIX] Interval class>>newFrom:
by Richard A. O'Keefe 31 May '02

31 May '02

In Squeak 3.0.1, try Interval newFrom: (1 to: 1) You expect a fresh copy of (1 to: 1). You GET a division by 0. This is because the implementation starts | newInterval | newInterval := self from: aCollection first to: aCollection last by: (aCollection last - aCollection first) // (aCollection size - 1). If aCollection size = 0, this errs instead of returning an empty Interval. If aCollection size = 1, this gets a divide-by-zero. Instead, it should read | newInterval n | (n := aCollection size) <= 1 ifTrue: [ n = 0 ifTrue: [^self from: 1 to: 0]. ^self from: aCollection first to: aCollection last]. newInterval := self from: aCollection first to: aCollection last by: (aCollection last - aCollection first) // (n - 1). Has this already been fixed in 3.2?

1 0

[FIX] raok-elementPrinting
by ok＠cs.otago.ac.nz 31 May '02

31 May '02

from preamble: "Change Set: raok-elementPrinting Date: 30 May 2002 Author: Richard A. O'Keefe Collection>>printElementsOn: used to use the #skip: method of streams. This limits the range of streams that #printOn: can be used with, as some quite reasonable streams (such as XML generation) may not support #skip:. The existing code in Collection>>storeOn: does not need #skip:. This change-set makes #printElementsOn: use the same technique as #storeOn:, so that #skip: isn't needed any more."!

1 0

Sensible dynamic menus?
by danielv＠netvision.net.il 30 May '02

30 May '02

I recently post a change to make TheWorldMenu "open" submenu dynamic. What does that mean? just that between the "web browser" entry, which is still hardwired into TWM, and the "mvc project" entry (likewise), it puts up the contents of a dynamic list that Celeste and IRCChat register into. It cuts the link from morphic to the applications, which is nice, but the fact that IRC and Celeste are both internet apps is incidental. If we start moving the other entries into the registry, we'll get a mess - browser (b) mail reader file list web browser .. you get the idea. This simply depends on the load order, which is simple, but butt-ugly and inconsistent to the user. BTW, this is a problem we'll have to solve for many instances - the browser list menus, the huge and evil ParagraphEditor menu, the menus for change sets (which still have the mail to list even without Celeste, but what if there's no sockets?) And those also want a reasonable interaction between dynamic menus and menus that split with "more..." Daniel

3 2

[FIX][BC] Minor BC fixes
by Anthony Hannan 30 May '02

30 May '02

from preamble: Change Set: MinorBCFixes-ajh Date: 30 May 2002 Author: Anthony Hannan Minor fixes to CompiledMethod2 #= and MethodContext2 #isDead that was causing problems for Ken Dickey. Thanks to Ken for finding the bugs. Ken, this will fix your latest problem. Cheers, Anthony 'From Squeak3.2gamma[BC] of 15 January 2002 [latest update: #4743] on 30 May 2002 at 12:33:41 pm'! "Change Set: MinorBCFixes-ajh Date: 30 May 2002 Author: Anthony Hannan Minor fixes to CompiledMethod2 #= and MethodContext2 #isDead that was causing problems for Ken Dickey. Thanks to Ken for finding the bugs."! !CompiledMethod2 methodsFor: 'testing' stamp: 'ajh 5/30/2002 12:22'! = method "Answer whether the receiver implements the same code as method." (method isKindOf: CompiledMethod2) ifFalse: [^ false]. self size = method size ifFalse: [^ false]. header = method header ifFalse: [^ false]. bytecodes = method bytecodes ifFalse: [^ false]. 1 to: self size do: [:i | (self at: i) = (method at: i) ifFalse: [^ false]]. ^ true! ! !CompiledMethod2 methodsFor: 'testing' stamp: 'ajh 5/30/2002 12:23'! hash ^ bytecodes hash! ! !MethodContext2 methodsFor: 'testing' stamp: 'ajh 5/24/2002 11:00'! isDead "A frame becomes dead once execution returns from it (once it is popped)" stack size < (frameIndex + FrameActiveOffset) ifTrue: [^ true]. (stack at: frameIndex + FrameActiveOffset) == self ifFalse: [^ true]. "very rare but the above could be true even if self is dead, if it just happened to be put on the stack by another frame in the same place where it used to reside." (stack at: frameIndex + FrameMethodOffset) class == CompiledMethod2 ifFalse: [^ true]. "The last two conditions combined are very very rare if self is dead, so we can assume self is not dead" ^ false! !

1 0

SystemTracer2 (was: [updates] Nearly final 3.2)
by Anthony Hannan 30 May '02

30 May '02

Dan Ingalls <Dan(a)SqueakLand.org> wrote: > Does anyone know fixes to the simulator or tracer that should go in? At http://minnow.cc.gatech.edu/squeak/uploads/2119/NewSystemTracer-ajh.1.cs you will find a new SystemTracer called SystemTracer2. I use if for converting standard images into BC images. The changeset only adds new classes and new messages to the image so it is safe to file in. The old SystemTracer is left untouched. SystemTracer2 is a total refactoring making it simpler (I think) to convert/trace images. Just execute "SystemTracer2 new writeImageFile: 'clone.image'" to trace your image, creating a new image and changes file. To convert objects while tracing, subclass SystemTracer2 and override methods such as initSwapMapPreShutdown. swapMap is a dictionary of existing objects to new converted object that you create. This requires more space, but makes the tracing/converting simpler since all the objects are converted ahead of time. There is also a hook to convert while tracing if you want. Cheeers, Anthony

1 0

[ENH] DupHandleBehavior
by schwa＠cc.gatech.edu 30 May '02

30 May '02

from preamble: "Change Set: DupHandleBehavior Date: 30 May 2002 Author: Joshua Gargus Brings the behavior of the halo duplication (green) handle in line with that of the grab (black) handle. More specifically, the grab handle offers two ways to end a grab interaction, depending on whether the grab was initiated by clicking or dragging the grab handle. Another click is required to drop a morph picked up with a click, but morphs picked up with a drag are dropped at the end of the drag. Such was not the case for morphs created via the dup handle; they have always required another click to drop them. This has been a niggly ever since the 'click' behavior replaced the 'drag' behavior. Now both work!! " !

1 0

Mouse event problems.
by Caitlin Rene Bovee 30 May '02

30 May '02

I am trying to write a mouseDown method and a mouseUp method which will return to me the coordinates of the mouse at those points. However, I can't ever seem to get a value. This is the only piece of code I have: startCoord :=self position. Do I need to use super? The way I am testing this is printing out the values to the transcript when I use the function which manipulates my sketchPad PluggableButtonMorph. I also have other buttons to set various tools for the sketch pad and the tool variable is set correctly when the button is pushed. I have no idea what I am missing, but any help would be appreciated. Thanks, Caitlin Bovee Email: gte049l(a)prism.gatech.edu stasid(a)cc.gatech.edu

2 1

[BUG] ESC behavior is different with regard to menus for text panes and worlds
by Craig Latta 30 May '02

30 May '02

Hi-- In 3.3a.4881: - Open a World menu by pressing escape. - Press escape again; the menu closes. - Open a text pane menu by pressing escape after selecting a text pane. - Press escape again; the menu doesn't close. As far as I can tell, one must close that menu via its "close" halo. (I'd prefer that this last escape-key event closed the menu. FWIW, I think menu-manipulation via the escape key is useful.) thanks, -C -- Craig Latta improvisational musical informaticist craig.latta(a)netjam.org www.netjam.org Smalltalkers do: [:it | All with: Class, (And love: it)]

2 1

[GOODIE] Jabber protocol
by Duane Maxwell 30 May '02

30 May '02

This is the Jabber low level protocol excised from the exobox project. The exobox client code is substantially tied to the rest of the project, so it's not very extractable. The code was originally written by me, but had substantial improvements by Andres Valloud and Luciano Notarfrancesco, among others. It was designed to work in a 2.8 image with the exobox XML parser. It has not been loaded or tested in any way with any other images or parser, but is submitted "as is" in the hope that someone may want to pick it up and bring it up to date. It also references a Jabber server that no longer exists - someone will have to set one up, or utilize the jabber.org servers. -- Duane

4 4

Interval>>copyFrom:to:
by Richard A. O'Keefe 30 May '02

30 May '02

In Squeak 3.0, if you do (1 to: 12) allButFirst you get an array, not the Interval I was expecting. There are a whole lot of "slicing" methods in SequenceableCollection which end up calling #copyFrom:to:. The default implementation builds an instance of "self species new". It's pretty obvious how I could fix this to work the way I'd expect: Interval>>"in category 'copying'" copyFrom: firstIndex to: lastIndex self assert: [firstIndex >= 1 and: [lastIndex between: firstIndex-1 and: self size]]. ^self class from: (firstIndex - 1) * step + start to: (lastIndex - 1) * step + start by: step My question is whether (1 to: 12) allButFirst is SUPPOSED to return an Array for some reason, or whether this was an inintended consequence of an incomplete edit? Has this been changed later? (I keep waiting for the announcement of 3.2final...)

1 0

← Newer
1
2
3
4
5
6
7
...
46
Older →

Jump to page: