[UPDATES] 7 for 3.6alpha
Bruce ONeel
edoneel at sdf.lonestar.org
Mon Jun 2 13:43:36 UTC 2003
Thanks, these are now all packed up in:
ftp://st.cs.uiuc.edu/pub/Smalltalk/Squeak/3.6alpha/Squeak3.6a-5247.zip
cheers
bruce
Doug Way <dway at riskmetrics.com> wrote:
>
> These are the 4 items which were [approved] last week, plus one which
> Daniel approved about a month ago that I somehow missed. (The subject
> line for that one was munged a bit, though... I'll use that as an
> excuse. ;-) )
>
> - Doug Way
>
>
> -------------------------------
>
> 5241instrPrinter-md -- Marcus Denker -- 8 April 2003
> A refactoring for InstructionPrinter,
> AbstractInstructionPrinter and InstVarRefLocator:
> AbstractInstructionPrinter:
> moved the instVar and a method into InstVarRefLocator, now this class
> can
> be used as a Abstract Superclass for
> both InstructionPrinter and InstVarRefLocator
> Renamed to InstructionClient (c.f. VisualWorks).
> InstructionPrinter: refactored to be a
> subclass of InstructionClient"
>
> 5242HashChanges1-SqR -- Andres Valloud, Stephan Rudlof -- 13 August 2002
> Goals:
> Implement an improved hash for large integers.
> Refactor/rename some hash methods to make them more intention revealing.
> Pass 1 - Implement refactoring so it can be bridged later.
> sr: I have changed #byteAt: to #basicAt: to make it Slang compilable
> (and changed some var names)."
>
> 5243HashChanges2-SqR -- Andres Valloud -- 13 August 2002
> Pass 2. Implement LargePositiveInteger>>hash.
> Rehash sets as appropriate."
>
> 5244HashChanges3-SqR -- Andres Valloud -- 13 August 2002
> Pass 3. Make String and ByteArray use the
> refactoring instead of the original
> implementation."
>
> 5245MorphicClsComments-efc -- Eddie Cottongim -- 7 March 2003
> Class Comments for Morph, ImageMorph, SimpleButtonMorph and
> StringMorph. These were subject
> to review and editing on the swiki, and thus should be fairly
> comprehensive."
>
> 5246Sibling-dup-tk -- Ted Kaehler -- 11 March 2003
> Whenever you ask for a sibling instance of a Player in an EToy, the
> Players of the new object is an instance of the original Player class,
> instead of having a new class. When you duplicate a morph the holds a
> set of siblings, there was bug that made the new objects not be
> siblings. This fixes the bug. Alan wants this fixed for the cells of
> a Tic-Tac-Toe game.
> To test, make a playfield with a blob in it. Make a sibling of the
> blob. Tell the playfield to 'make sibling instance'. All four blobs
> will remain siblings.
> In addition, the production of a sibling instance is now cleaner.
> (Updated to fix a conflict with the 5240 MCP update. -dew)"
>
> 5247VeryDeepCopyUsing-tk -- Ted Kaehler -- 13 May 2003
> Object>>veryDeepCopyUsing: is a utility method. When you have a
> complex data structure, involving several custom classes, veryDeepCopy
> on a subpart may not do what you want. When you veryDeepCopy one of
> your objects, it does not know which instance variable are actually
> pointers back to a 'global' object, which should not be duplicated.
> The 'parent' links in a tree are an example. These can be handled by
> creating veryDeepInner: and veryDeepFixupWith:. However, it may be
> easier to make a custom veryDeepCopy method. In it, you create a
> DeepCopier, and store into its references dictionary with (globalObject
> -> globalObject). Then you want to start a veryDeepCopy using that
> DeepCopier, and calling veryDeepCopyUsing: is the way to do that."
More information about the Squeak-dev
mailing list
|