[squeak-dev] The Trunk: MonticelloConfigurations-edc.60.mcz

Edgar J. De Cleene edgardec2001 at yahoo.com.ar
Fri Aug 21 11:00:38 UTC 2009


This was the last three, I think people wish know what's  going on.

Very, very good job Andreas for move Squeak forward.
Still I wish see number change and name the old thing Squeak 3.11.
So we should now have Squeak 3.11.alpha.nnnn and all squeakers work on this.

Edgar



-------------- next part --------------

========== Collections-ar.114 ==========

ToolBuilder cleanup for Transcript opening.

========== EToys-ar.32 ==========

ToolBuilder cleanup with Lexicon and InstanceBrowser.

========== Graphics-ar.64 ==========

Avoid using class Line from MVC in floodFill: debug printing.

========== Kernel-nice.212 ==========

Patch for http://bugs.squeak.org/view.php?id=7373
-2.0 isInf -> true 

>>> Kernel-ar.211 <<<
Allow CompiledMethod to provide a suitable drag label. Used by DnD in browsers.

>>> Kernel-ar.210 <<<
Closure fallout: With closures, all object enumeration must have a definitive end marker since enumeration will almost always create new contexts from block activations.

========== Morphic-ar.155 ==========

Fix references to PreferencesPanel.

>>> Morphic-ar.154 <<<
Provide a stub implementation of SmalltalkEditor>>select which is sent by Parser's variable correction when the new editors are used.

>>> Morphic-jmv.153 <<<
Lots of small issues in new Editor hierarchy.

>>> Morphic-ar.151 <<<
ToolBuilder cleanup. Fix various (now obsolete) protocols used in the world menu etc.

========== MorphicExtras-ar.38 ==========

Closure fallout: With closures, all object enumeration must have a definitive end marker since enumeration will almost always create new contexts from block activations.

========== PreferenceBrowser-ar.35 ==========

Provide a compatibility method for creating preferences for clients which might use the older interface using view registries directly.

>>> PreferenceBrowser-ar.34 <<<
Include a halo type in view registry for halo themes.

>>> PreferenceBrowser-ar.33 <<<
Moves PreferenceView/PreferenceViewRegistry into PreferenceBrowser since they are no longer central to the preference mechanism but rather a part of how the particular preference browser organizes access to preferences.

========== ST80-ar.49 ==========

There is currently no preference browser in MVC. Inform the user about it.

>>> ST80-ar.48 <<<
Push optional buttons and annotation pane down to CodeHolder. This avoids ugly appearance of users of (StringHolder new contents: '') openLabel: 'Title'.

>>> ST80-ar.47 <<<
StringHolder needs to implement contents:notifying: to support the standard code pane.

>>> ST80-ar.46 <<<
Better layout for optional buttons and annotation panes. As an experiment, annotation panes are now at the bottom (kinda like status lines).

>>> ST80-ar.45 <<<
Tools should be using a dynamic #labelString not a static one when constructed.

>>> ST80-ar.44 <<<
Refactoring of ToolBuilder construction methods. Breaks them out into separate methods and makes it easier to mix and match the different parts.

>>> ST80-ar.43 <<<
ToolBuilder cleanup in StringHolder. Avoid referencing either Morphic or MVC wherever possible.

========== Services-Base-ar.35 ==========

Fix usage of outdated preferences protcols.

========== System-ar.131 ==========

More preferences rework: Moved PreferenceView, PreferenceViewRegistry etc. into PreferenceBrowser since they should not live in the System package. The only remaining classes are Preferences, Preference, PragmaPreference and PreferenceExample. For the time being, this is probably as small as we can get and removes another set of dependencies (both Morphic and MVC) from core packages.

>>> System-ar.130 <<<
ToolBuilder cleanup. Removal of a few unused (and difficult to convert) methods and fixes of others.

>>> System-ar.129 <<<
Closure fallout: With closures, all object enumeration must have a definitive end marker since enumeration will almost always create new contexts from block activations.

========== ToolBuilder-Morphic-ar.32 ==========

Morphic ToolBuilder tweaks: Leave buttons with a white background to have proper mouse over highlight. Update enabled selector *after* updating color to get disabled state in buttons right.

>>> ToolBuilder-Morphic-ar.31 <<<
Don't remember text widgets as updatable panes - they are not supported and create useless warnings.

>>> ToolBuilder-Morphic-ar.30 <<<
Changes the default color of pluggable text to match current morphic style.

>>> ToolBuilder-Morphic-ar.29 <<<
Update button colors more often; otherwise inheritance button gets wrong color.

>>> ToolBuilder-Morphic-ar.28 <<<
A few subtle tweaks for pluggable panels in Morphic. Makes button rows and other composite panels look nicer.

>>> ToolBuilder-Morphic-ar.27 <<<
Use the preferred code font for pluggable texts in tools.

>>> ToolBuilder-Morphic-ar.26 <<<
Morphic ToolBuilder tweaks: buttons need to update more often, and their appearance should match current looks.

========== Tools-ar.102 ==========

Push optional buttons and annotation pane down to CodeHolder. This avoids ugly appearance of users of (StringHolder new contents: '') openLabel: 'Title'.

>>> Tools-ar.101 <<<
Fix coloring of inheritance button. Remove extra switches in hierarchy browser.

>>> Tools-ar.100 <<<
Annotation panes are back. As an experiment, they are now at the bottom and also show the one-line preview of the class comment when a class is selected.

>>> Tools-ar.99 <<<
ToolBuilder refactoring. Better structured construction of the various pieces. Also puts back DnD support in the browser and removes another bunch of Morphic idiosynchracies that were introduced for DnD support.

>>> Tools-ar.98 <<<
Put two methods back that were still in use and cause no harm.

>>> Tools-ar.97 <<<
A *serious* round of toolbuildering the tools. This removes a load of methods for constructing the interfaces and replaces them by ToolBuilder specs and interface construction. Does not only help to reduce the dependencies between tools and particular ui frameworks but also helps to make tools available for MVC that previously werent (like Lexicon etc). Some fixups in other packages still coming up...

>>> Tools-ar.96 <<<
Closure fallout: With closures, all object enumeration must have a definitive end marker since enumeration will almost always create new contexts from block activations.

========== MonticelloConfigurations-ar.59 ==========

Faster updates. MCMcmUpdater now remembers the last update map seen in each repository and only considers newer ones when updating. This avoids downloading old update maps that are long past.

========== Collections-nice.116 ==========

Patch for
http://bugs.squeak.org/view.php?id=7269
OrderedCollection streamContents broken
http://bugs.squeak.org/view.php?id=2390
ReadWriteStream with an OrderedCollection bugs when growing  

>>> Collections-nice.115 <<<
Patch for
http://bugs.squeak.org/view.php?id=7296
forceTo:paddingWith: doesn't work with OrderedCollections 

========== CollectionsTests-nice.89 ==========

Test for
http://bugs.squeak.org/view.php?id=7269
OrderedCollection streamContents broken
http://bugs.squeak.org/view.php?id=2390
ReadWriteStream with an OrderedCollection bugs when growing  

>>> CollectionsTests-nice.88 <<<
Test for
http://bugs.squeak.org/view.php?id=7296
forceTo:paddingWith: doesn't work with OrderedCollections 

========== Graphics-ar.66 ==========

Several font display fixes:
- BitBlt primDisplayString needs to consult the glyph map in its non-primitive code path
- CharacterScanner will no longer interpret characters 256/257 as stop conditions when occuring in a string
- The "stops" argument is used consistently instead of the ivar stopConditions
- StrikeFont m17n display should not pass on the entire remaining string to its fallback font
- The default fallbackFont substitution font *must* use the font it is backing since the fallback font will not be installed independently from the original.

>>> Graphics-ar.65 <<<
Fixes StrikeFont>>displayMultiString: to do proper two-pass rendering for fonts requiring two-pass rendering.

========== Kernel-ar.217 ==========

Styling hooks:
- Add ParagraphEditor>>blinkParen and fix issue in ParagraphEditor>>clearParens for syntax highlighting compatibility.

>>> Kernel-ar.216 <<<
Another round of reducing MVC/Morphic dependencies:
- Introduce the #isWindowForModel: protocol which avoids the use of patterns like "(d isKindOf: SystemWindow orOf: StandardSystemView) and:[d model == self]"
- Simply Object>>currentWorld to neither look at MVC embedded worlds nor create new MVC-based morphic worlds

>>> Kernel-mha.215 <<<
removed obsolete method from CompiledMethod

>>> Kernel-mha.214 <<<
fixed problem in MethodProperties>>#analogousCodeTo: that would cause an error in TraitTest

>>> Kernel-mha.213 <<<
implemented isMethodContext to avoid an error in MethodContextTest

========== KernelTests-mha.84 ==========

deleted obsolete BlockClosureTest

>>> KernelTests-mha.83 <<<
fixed BlockContextTest and MethodContextTest (both sent isBlockClosure instead of isClosure)

>>> KernelTests-mha.82 <<<
fixed MethodContextTest and BlockContextTest (both sent isBlockClosure instead of isClosure)

========== Monticello-bf.318 ==========

- store and read preambles/postscripts in textual source, too
- show preambles/postscripts when browsing a snapshot


>>> Monticello-bf.317 <<<
- merged bp.316 from inbox
- deal with nil user/password in MCHttpRepository

>>> Monticello-bf.316 <<<
- sort class vars and pools (merging Eliot's and my fixes)

========== Morphic-ar.159 ==========

Styling hooks:
- Refactor the TextEditor>>changeEmphasis: logic  bit
- Provide shout support methods in TextEditor
- Add SmalltalkEditor>>styler ivar.

>>> Morphic-laza.158 <<<
Add check for owner chain without any of them having a model -> nil


>>> Morphic-laza.157 <<<
I guess not too many use ProportionalSplitters to change
the layout of windows these days, but they are broken in the
latest trunk image. This is because the Method
containingWindow in Morph has changed. The containing
Window is now searched based on the model of the
component, but ProportionalSplitters do have no model.
A quick fix is to look up the owner chain of a
ProportionalSplitter to find an owner that has a model. Most of
the time (always?) the owner will be the containing window, so
I don't know if the ifFalse: Branch is needed.

>>> Morphic-ar.156 <<<
Another round of reducing MVC/Morphic dependencies:
- Remove the old openInMVC methods
- Provide SystemWindow>>isWindowForModel:
- Remove an indirect MVC dependency via window colors

========== MorphicExtras-ar.42 ==========

Another round of reducing MVC/Morphic dependencies:
- Provide MVCWiWPasteUpMorph>>isWindowForModel:
- Remove convertToMVCWiWPasteUpMorph

>>> MorphicExtras-sm.41 <<<
A class comment for ProjectNavigationMorph, and a fix for the parent's.

>>> MorphicExtras-ar.39 <<<
Retire PreferencesPanel. It has served its purpose well, but now it's time to get some rest.

========== Multilingual-ar.40 ==========

Merge Multilingual-sn.38:
Added missed UTF32InputInterpreter class.
Corrected MacUnicodeInterpreter for working with v.4.x Mac VMs 

>>> Multilingual-bf.39 <<<
- remove null characters from LanguageEnvironment's class comment

>>> Multilingual-sn.38 <<<
Added missed UTF32InputInterpreter class.
Corrected MacUnicodeInterpreter for working with v.4.x Mac VMs 

>>> Multilingual-ar.38 <<<
Several font display fixes:
- CharacterScanner will no longer interpret characters 256/257 as stop conditions when occuring in a string
- The "stops" argument is used consistently instead of the ivar stopConditions
- The default fallbackFont substitution font *must* use the font it is backing since the fallback font will not be installed independently from the original.

========== PreferenceBrowser-ar.36 ==========

Clean out obsolete preference view registries.

========== ST80-ar.50 ==========

Another round of reducing MVC/Morphic dependencies:
- Remove the durable menus which created a cyclic dependency between MVC, Utilities, and Morphic
- Provide StandardSystemView>>isWindowForModel:

========== Services-Base-ar.36 ==========

Put BlockClosure>>valueWithRequestor: back for services.

========== System-ar.132 ==========

Another round of reducing MVC/Morphic dependencies:
- Remove various Utilities methods we no longer need

========== ToolBuilder-Kernel-ar.19 ==========

Styling hooks:
- Provide PluggableCodePaneSpec which is a sub-type of PluggableTextSpec and assumed to be used specifically for code editing tools.

========== ToolBuilder-Morphic-ar.33 ==========

Styling hooks:
- Provide styling support in PtmPlus
- Set up the default styler for a pluggableCodePane


========== Tools-bf.106 ==========

Fix Message Finder problem (bug #7380)

>>> Tools-ar.105 <<<
Styling hooks:
- Provide #aboutToStyle: interface throughout the tools (similar to #shoutAboutToStyle: but passes styler instead of view)
- Use pluggableCodePaneSpec for the tools that should support syntax highlighting.

>>> Tools-ar.104 <<<
Fixes SelectorBrowser. It had its morphicWindow method nuked but not been given the equivalent ToolBuilder build method.

>>> Tools-bf.103 <<<
- fix UIManager typo in Browser>>recent
-------------- next part --------------

========== Morphic-ar.135 ==========

Adds a nice speedup to window activation by removing a bit of code intended as optimization but actually causing significnt slowdowns when opening new windows. Speeds up the overall browser benchmark by 20-25%.

========== MorphicExtras-edc.36 ==========

I just add package MorphicExtras-Flaps for having same as in http://www.squeaksource.com/Ladrillos.html.


-------------- next part --------------

========== MonticelloConfigurations-ar.58 ==========

Fixes an interesting issue in the MCM updater. Apparently, #allAncestorsDo: can loop indefinitely. Use ancestorsDoWhileTrue: instead and keep track of the versions that we've seen already to avoid this problem.

>>> MonticelloConfigurations-ar.57 <<<
UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> MonticelloConfigurations-ar.56 <<<
Optimizes updateFromRepositories so that we do not need to download new versions if the this version is already loaded in the image in question. This greatly simplifies the update process.

>>> MonticelloConfigurations-ar.55 <<<
Progress while loading configurations.

========== Collections-jmv.113 ==========

Added #withoutJustTrailingDigits, that is like #withoutTrailingDigits, but does not trim at the first space.

>>> Collections-ar.112 <<<
Text class fromString: no longer creates a text attribute referring to the current default font index. This caused extremely strange effects when such text is displayed using any non-default text style.

>>> Collections-nice.111 <<<
Fix for
http://bugs.squeak.org/view.php?id=7172
CharacterSetComplement byteArrayMap optimization is slow 

Add an inst var to cache this information.
Rationale: #byteArrayMap is a method dedicated to optimization of speed, there is no reason why we should spoil the game with introduction of CharacterSetComplement in 3.10.

>>> Collections-nice.110 <<<
Fix for
http://bugs.squeak.org/view.php?id=6777
CharacterSet copy is shallow... 

"If we change the copy, do we change the original?
We'd better not"
| cs1 cs2 |
cs1 := CharacterSet separators.
cs2 := cs1 copy.
cs2 remove: Character tab.
^self assert: (cs1 includes: Character tab). 

>>> Collections-nice.109 <<<
Test for
http://bugs.squeak.org/view.php?id=6482
Interval storeOn: incorrect for some ScaledDecimal

Some ScaledDecimal have more bits than printed.
Example: 0.5s1 squared will print 0.2s1
However 0.5s1 squared * 4 = 1

Thus (0 to: 1 by: 0.5s1 squared) size = 5
(0 0.25 0.5 0.75 1)

If you storeString and reinterpret, you get size 6
(0 0.2 0.4 0.6 0.8 1.0)

Interval SHOULD use storeOn: and not rely on Number>>printOn:

This is some sort of violation of encapsulation.
And it is defeated by introduction of a new species of Number (ScaledDecimal)...


>>> Collections-nice.108 <<<
Fix for
http://bugs.squeak.org/view.php?id=7175
CharacterSetComplement cannot printString 

>>> Collections-nice.107 <<<
Fix for
http://bugs.squeak.org/view.php?id=6782
1.0 / (FloatArray with: 2.0)

Note: the fix also handle case of gradual underflow like:
self assert: (1.0e-39 / (FloatArray with: 1.0e-39)) first < 2.

>>> Collections-nice.106 <<<
Fix for 
http://bugs.squeak.org/view.php?id=6998
WideString>>at:put: does not return the put-object

>>> Collections-nice.105 <<<
Fix for 
http://bugs.squeak.org/view.php?id=7121
add:beforeIndex: and add:afterIndex: do not work as advertised 

>>> Collections-nice.104 <<<
Fix for
http://bugs.squeak.org/view.php?id=7136
LinkedList #add:after: bug

>>> Collections-nice.103 <<<
Fix for
http://bugs.squeak.org/view.php?id=6994
#(1 2 3) asByteArray as: Interval.

Note: this does not fix #(nil) as: Interval.

>>> Collections-ar.102 <<<
Conversion functions for base64. Just as we have 

	'Hello World' zipped unzipped = 'Hello World'

we can now use the similarly useful

	'Hello World' base64Encoded base64Decoded = 'Hello World'

or more practically useful since this is what often happens in transmissions of lengthy data over text channels:

	'Hello World' zipped base64Encoded 
		base64Decoded unzipped = 'Hello World'.


>>> Collections-ar.101 <<<
Allow Text>>runLengthFor: to return 0 for an empty text so that client code does not need to guard specifically against the situation. This is similar to other methods that behave identically in this situation (#emphasisAt:, #fontAt:) .

>>> Collections-nice.100 <<<
Fix for Interval>>#includes: and #indexOf:
http://bugs.squeak.org/view.php?id=1602
http://bugs.squeak.org/view.php?id=1603
http://bugs.squeak.org/view.php?id=6455
http://bugs.squeak.org/view.php?id=7180

Note: repeat after me:
Interval of Floats are evil.

Though the noble fuzzy inclusions tests (corrected by this patch), some interval will fail in finding their own elements:

| eps i |
eps := (1.0 timesTwoPower: -52) * 1.25.
i := (1 to: 1+(99*eps) by: eps).
^(i includes: (i at: 2))

========== CollectionsTests-nice.87 ==========

Test for
http://bugs.squeak.org/view.php?id=6777
CharacterSet copy is shallow... 

"If we change the copy, do we change the original?
We'd better not"
| cs1 cs2 |
cs1 := CharacterSet separators.
cs2 := cs1 copy.
cs2 remove: Character tab.
^self assert: (cs1 includes: Character tab). 

>>> CollectionsTests-nice.86 <<<
Test for
http://bugs.squeak.org/view.php?id=6482
Interval storeOn: incorrect for some ScaledDecimal

Some ScaledDecimal have more bits than printed.
Example: 0.5s1 squared will print 0.2s1
However 0.5s1 squared * 4 = 1

Thus (0 to: 1 by: 0.5s1 squared) size = 5
(0 0.25 0.5 0.75 1)

If you storeString and reinterpret, you get size 6
(0 0.2 0.4 0.6 0.8 1.0)

Interval SHOULD use storeOn: and not rely on Number>>printOn:

This is some sort of violation of encapsulation.
And it is defeated by introduction of a new species of Number (ScaledDecimal)...


>>> CollectionsTests-nice.85 <<<
Test for
http://bugs.squeak.org/view.php?id=7175
CharacterSetComplement cannot printString 

>>> CollectionsTests-nice.84 <<<
Test for
http://bugs.squeak.org/view.php?id=6782
1.0 / (FloatArray with: 2.0)

>>> CollectionsTests-nice.83 <<<
Test for 
http://bugs.squeak.org/view.php?id=6998
WideString>>at:put: does not return the put-object

>>> CollectionsTests-nice.82 <<<
Tests for 
http://bugs.squeak.org/view.php?id=7121
add:beforeIndex: and add:afterIndex: do not work as advertised 

>>> CollectionsTests-nice.81 <<<
Test for
http://bugs.squeak.org/view.php?id=7136
LinkedList #add:after: bug

>>> CollectionsTests-nice.80 <<<
Test for
http://bugs.squeak.org/view.php?id=6994
#(1 2 3) asByteArray as: Interval.

>>> CollectionsTests-ar.79 <<<
Simple tests for base64Encoded/base64Decoded.

>>> CollectionsTests-nice.77 <<<
Test for Interval>>#includes: and #indexOf:
http://bugs.squeak.org/view.php?id=1602
http://bugs.squeak.org/view.php?id=1603
http://bugs.squeak.org/view.php?id=6455
http://bugs.squeak.org/view.php?id=7180

These tests are only non regression test about previous fuzzy inclusion errors.

REMINDER:
Interval of Floats are evil.
Fuzzy inclusion does not make them behave that better.

For example, I did not include this funny test, known to fail due to fuzzyfication efforts:

| eps i |
eps := (1.0 timesTwoPower: -52) * 1.25.
i := (1 to: 1+(99*eps) by: eps).
^(i includes: (i at: 2))

========== EToys-ar.31 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Files-ar.26 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> Files-mha.25 <<<
make tests more disciplined: remove files created during testing

>>> Files-mha.24 <<<
make tests more disciplined: remove files created during testing

========== Graphics-ar.63 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> Graphics-jmv.62 <<<
And do run the install again!

>>> Graphics-jmv.61 <<<
When setting new fonts to use the underscore and caret glyphs, don't mess with older existing StrikeFonts.

>>> Graphics-jmv.60 <<<
Another round of small fixes for AA StrikeFonts.

>>> Graphics-jmv.59 <<<
Moves dual pass for AA StrikeFonts back to BitBlt, without needing any new instvar (that would break WarpBlt). Enhanced a bit the spacing between glyphs.

>>> Graphics-ar.58 <<<
Adds CompositionScanner>>computeDefaultLineHeight to compute the height of an empty string correctly from its attributes.

>>> Graphics-ar.57 <<<
Removes the extra iVars from BitBlt again (pushed down into GrafPort) since they would cause WarpBlt primitives to fail. Sigh.

>>> Graphics-jmv.56 <<<
Moves the double pass for AA StrikeFonts from GrafPort up to BitBlt.

>>> Graphics-jmv.55 <<<
Fixes Display alpha channel for AA StrikeFonts.

>>> Graphics-jmv.54 <<<
Additional DejaVu StrikeFonts, sizes 7 and 12.

>>> Graphics-jmv.53 <<<
Part of StrikeFont enhancements to support glyphs up to 32bpp, with anti aliasing.

>>> Graphics-ar.52 <<<
Add subfamily name 'Negrata' to the list of known styles of Truetype fonts.

>>> Graphics-ar.51 <<<
Integrate a fix from Multilingual-Support-FromEToys.

>>> Graphics-ar.50 <<<
Allow non-portable fonts in Morphic font selection menu when TTFileDescription>>offerNonPortableFonts is enabled.

>>> Graphics-ar.49 <<<
Addresses an interesting bug in [Multi]DisplayScanner. All stopConditions exept #endOfRun have lastIndex past the last character displayed. #endOfRun sets it *on* the character. If we display up until lastIndex then we will also display invisible characters like CR and tab. This problem should be fixed in the scanner (i.e., position lastIndex consistently) but I don't want to deal with the fallout right now so we keep the fix minimally invasive.

>>> Graphics-ar.48 <<<
Addresses an interesting bug in [Multi]DisplayScanner. All stopConditions exept #endOfRun have lastIndex past the last character displayed. #endOfRun sets it *on* the character. If we display up until lastIndex then we will also display invisible characters like CR and tab. This problem should be fixed in the scanner (i.e., position lastIndex consistently) but I don't want to deal with the fallout right now so we keep the fix minimally invasive.

========== Kernel-ar.209 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> Kernel-ar.208 <<<
Fixes debugging of FFI methods.

>>> Kernel-nice.207 <<<
Fix for
http://bugs.squeak.org/view.php?id=6983
Float nan isLiteral. It should not. 

>>> Kernel-nice.204 <<<
Fix for
http://bugs.squeak.org/view.php?id=4378
ScaledDecimal storeString incorrectly use printOn: and loose exactness 

>>> Kernel-nice.203 <<<
Speed up number parsing by refactoring SqNumberParser
http://bugs.squeak.org/view.php?id=6976
http://bugs.squeak.org/view.php?id=7174
+ some refactoring applied in Pharo

Main speed up come from:
- not using Character>>#digitValue
- avoid creating ifFail: [] blocks, but rather return nil
- compose and conquer algorithm for parsing large integers (so as to make as many computations with SmallInteger as possible, in the spirit of divide and conquer printing)

>>> Kernel-nice.202 <<<
Fix for
http://bugs.squeak.org/view.php?id=6779
ScaledDecimal readFrom: '5.3'.
did not answer a ScaledDecimal 

>>> Kernel-nice.201 <<<
Fix for 
http://bugs.squeak.org/view.php?id=7169
self assert: (SqNumberParser parse: '0.50s2') = (1/2).

>>> Kernel-nice.200 <<<
Faster printString for LargeInteger
http://bugs.squeak.org/view.php?id=6887

Basic idea is to use divide and conquer algorithm:
split the LargeInteger in two halves and recursively so as to fallback to faster SmallInteger arithmetic as soon as possible.

This change improves
[:n | [n printString] bench] value: 100 factorial.
by roughly a factor 4

>>> Kernel-nice.199 <<<
Fix for http://bugs.squeak.org/view.php?id=6874
(2 raisedTo: 31) negated bitAnd: (2 raisedTo: 31) negated - 1.

>>> Kernel-nice.198 <<<
Fix for http://bugs.squeak.org/view.php?id=7114
SmallInteger minVal highBitOfMagnitude.

========== KernelTests-mha.81 ==========

fixed FloatTest >> #testNaN5

This test would not accept an IEEE-754 compliant NaN converted to a string. It now accepts both quiet and signalled NaNs.

>>> KernelTests-nice.80 <<<
Test for
http://bugs.squeak.org/view.php?id=6983
Float nan isLiteral. It should not. 

>>> KernelTests-nice.79 <<<
Test for
http://bugs.squeak.org/view.php?id=4378
ScaledDecimal storeString incorrectly use printOn: and loose exactness 

>>> KernelTests-nice.78 <<<
Add a non regression test against number parsing speed up bug introduced by early implementation of:
http://bugs.squeak.org/view.php?id=6976

>>> KernelTests-nice.77 <<<
Test for
http://bugs.squeak.org/view.php?id=6779
ScaledDecimal readFrom: '5.3'.
would not answer a ScaledDecimal 

>>> KernelTests-nice.76 <<<
Test for 
http://bugs.squeak.org/view.php?id=7169
self assert: (SqNumberParser parse: '0.50s2') = (1/2).

>>> KernelTests-nice.75 <<<
Additional tests for LargeInteger printString
This is to ensure http://bugs.squeak.org/view.php?id=6887 is OK 

>>> KernelTests-nice.74 <<<
Test for http://bugs.squeak.org/view.php?id=6874
(2 raisedTo: 31) negated bitAnd: (2 raisedTo: 31) negated - 1.

>>> KernelTests-nice.73 <<<
Test for http://bugs.squeak.org/view.php?id=7114
SmallInteger minVal highBitOfMagnitude.

========== Monticello-ar.315 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Morphic-ar.150 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> Morphic-ar.149 <<<
Integrate Cuis text editors. The new classes are Editor and TextEditor (in Morphic-Text Support). TextMorph now honors the preference #userNewEditors to determine whether to use the (old) TextMorphEditor or the (new) TextEditor (shameless plug: pragma preferences rock; check out TextMorph class>>useNewEditors). There is still a ways to go before this will work but it gives us the starting point to knock down the remaining issues with the editors (mostly keyboard shortcut issues).

>>> Morphic-jmv.148 <<<
Another round of small fixes for AA StrikeFonts.

>>> Morphic-jmv.147 <<<
Moves dual pass for AA StrikeFonts back to BitBlt, without needing any new instvar (that would break WarpBlt). Enhanced a bit the spacing between glyphs.

>>> Morphic-ar.146 <<<
Fix typo in MenuMorph>>confirm:orCancel:.

>>> Morphic-ar.145 <<<
Use the correct default line height for empty text.

>>> Morphic-ar.144 <<<
With our new great fonts we also need a decent cursor to go along with it. Use a classic blinking text caret instead of a green invisible something.

>>> Morphic-ar.143 <<<
Provide a set of interfaces to MenuMorph which allow us to vector UIManager requests directly to it instead of having to indirect through PopUpMenu. A necessary step to decouple MVC and Morphic. Also adds NewParagraph>>replaceFrom:to:with: for easier compatibility with Cuis' editors.

>>> Morphic-ar.142 <<<
Choosing Help>>Preferences from the World menu should open a PreferencesBrowser not the ancient PreferencesPanel.

>>> Morphic-ar.141 <<<
Push down the additional state from BitBlt to GrafPort since BitBlt cannot deal with the added state. Also rename the state vars internally from "secondPassRule / secondPassColorMap" to "extraPassRule / extraPassColorMap" to avoid an exercise in Monticello madness.

>>> Morphic-jmv.140 <<<
Moves the double pass for AA StrikeFonts from GrafPort up to BitBlt.

>>> Morphic-jmv.139 <<<
Fixes Display alpha channel for AA StrikeFonts.

>>> Morphic-jmv.138 <<<
Part of StrikeFont enhancements to support glyphs up to 32bpp, with anti aliasing.

>>> Morphic-mha.137 <<<
some World menu convenience...
* refactored World menu construction a bit
* there is now a method makeConvenient that inserts some items at the top of the World menu
* for now, there's four popular entries that make common tools available with one less click ;-)

>>> Morphic-ar.136 <<<
Proper round rect drawing. Instead of just corner rounding we can now draw round rects just like any other rectangle with an optional corner radius, for example:
	Display getCanvas frameAndFillRoundRect: (100 at 100 extent: 200 at 50) radius: 10 fillStyle: Color white borderWidth: 2 borderColor: Color black.


========== MorphicExtras-ar.37 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Multilingual-ar.37 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> Multilingual-bf.36 <<<
- use UTF8ClipboardInterpreter for all platforms, not just Windows

>>> Multilingual-ar.35 <<<
Integrate Multilingual-Support-FromEToys.

>>> Multilingual-ar.34 <<<
Addresses an interesting bug in [Multi]DisplayScanner. All stopConditions exept #endOfRun have lastIndex past the last character displayed. #endOfRun sets it *on* the character. If we display up until lastIndex then we will also display invisible characters like CR and tab. This problem should be fixed in the scanner (i.e., position lastIndex consistently) but I don't want to deal with the fallout right now so we keep the fix minimally invasive.

>>> Multilingual-ar.33 <<<
Allow TTCFont to return nil from formOf: and handle it gracefully by using #hasGlyphOf: where required. This paves the way to allowing on-disk TrueType fonts where the actual font file may be missing after an image move and consquently no glyph can be computed for the font.

========== Nebraska-ar.17 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Network-ar.34 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== ST80-ar.42 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> ST80-jmv.40 <<<
Another round of small fixes for AA StrikeFonts.

========== ScriptLoader-ar.326 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Services-Base-ar.34 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== Sound-ar.7 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== System-ar.128 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> System-jmv.127 <<<
Remove the old entry in TextConstants!

>>> System-jmv.126 <<<
Run again the install of the DejaVu fonts

>>> System-jmv.125 <<<
Another round of small fixes for AA StrikeFonts.

>>> System-jmv.124 <<<
Do Run the postscript to install new fonts!

>>> System-jmv.123 <<<
Moves dual pass for AA StrikeFonts back to BitBlt, without needing any new instvar (that would break WarpBlt). Enhanced a bit the spacing between glyphs.

>>> System-ar.122 <<<
Introducing pragma preferences. Preferences no longer have to litter the Preferences class but can be defined as class variables local to the place where they are needed. Preferences are discovered by using the <preference:category:description:type:> pragma, for example:

samplePref
	<preference: 'Sample Preference'
		category: 'Examples'
		description: 'A Sample Preference'
		type: #Boolean>
	^SamplePref

Four types of preferences are currently supported: #Boolean, #Number, #String, and #Color. When the preference is changed, the according setter method is being called.

>>> System-jmv.121 <<<
Fixes Display alpha channel for AA StrikeFonts.

>>> System-jmv.120 <<<
Additional DejaVu StrikeFonts, sizes 7 and 12.
Run the install.

>>> System-jmv.119 <<<
Additional DejaVu StrikeFonts, sizes 7 and 12.

>>> System-jmv.118 <<<
StrikeFont useUnderscore

>>> System-jmv.117 <<<
Part of StrikeFont enhancements to support glyphs up to 32bpp, with anti aliasing.

>>> System-nice.116 <<<
Tet and fix for
http://bugs.squeak.org/view.php?id=7351

The code to extract image name without .image extension gets fooled if there is another dot in the name like in
Squeak3.10.2-7179-basic.image

========== Tests-mha.36 ==========

* make tests more disciplined: remove files created during testing
* moreover, several methods have a changed revision number (not changed by me)

========== ToolBuilder-Morphic-ar.25 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

>>> ToolBuilder-Morphic-ar.24 <<<
Morphic UI Manager requests are now vectored directly into FillInTheBlankMorph and MenuMorph instead of their MVC equivalents (FillInTheBlank and PopUpMenu). This helps decoupling MVC and Morphic a little more.

>>> ToolBuilder-Morphic-ar.23 <<<
Add separators to Morphic ToolBuilder menus.

>>> ToolBuilder-Morphic-ar.22 <<<
Fix an old bug in MorphicToolBuilder which would ignore the default size of the model being opened.

========== Tools-ar.95 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.

========== TrueType-ar.7 ==========

Integrate Mac Truetype fixes from Bert.

>>> TrueType-ar.6 <<<
Add support for Truetype files from disk. TTFileDescription is polymorphic with TTFontDescription but does not read the file into memory but works directly from disk. TTFileDescription has a series of font search paths that can be platform specific (see TTFileDescription>>fontPathsDo:) and use those to locate font files. In order to install a family of fonts execute, e.g.,

	TTFileDescription installFamilyNamed: 'Arial'.

The (non-portable) list of available platform families is also visible in the regular TextMorph font menu (Alt-k). To enable/disable having those fonts visible, execute:

	TTFileDescription offerNonPortableFonts: true/false.

If an image gets moved to a platform that does not have the font file it uses the fallback font (which currently results in mostly question marks). 

========== Universes-ar.40 ==========

UIManagerization. Replaces all the trivial references to PopUpMenu, SelectionMenu, CustomMenu, and FillInTheBlank.


More information about the Squeak-dev mailing list