Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours:
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068301.html
Name: Tools-mt.708
Ancestors: Tools-mt.707
Fixes the small bug where the code holder and browser omitted to communicate the state of their buttons correctly to the pluggable button.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068302.html
Name: ToolBuilder-Morphic-mt.170
Ancestors: ToolBuilder-Morphic-mt.169
Action buttons being white was surprising and overridden long time ago in PluggableButtonMorphPlus.
Still, retain the hook to build action buttions. Maybe we can distinguish them in the future like tool buttons distinguish from push buttons in other UI frameworks such as Qt.
=============================================
Patrick Rein uploaded a new version of Tests to project The Inbox:
http://source.squeak.org/inbox/Tests-pre.347.mcz
==================== Summary ====================
Name: Tests-pre.347
Author: pre
Time: 23 July 2016, 2:20:08.768962 pm
UUID: 8272055d-5600-c34b-aac4-9c29a9fda5e5
Ancestors: Tests-pre.346
Adds a Compiler test to the list of expected failures as it can not currently work in an image only using the V3PlusClosures encoder. The test is expected to be executable in an image which uses the V3Encoder.
=============== Diff against Tests-pre.346 ===============
Item was changed:
----- Method: ClosureCompilerTest>>expectedFailures (in category 'failures') -----
expectedFailures
+
+ ^#(testDebuggerTempAccess testInjectIntoDecompilations testInjectIntoDecompiledDebugs) ,
+ (({CompiledMethod classPool at: #SecondaryBytecodeSetEncoderClass .
+ CompiledMethod classPool at: #PrimaryBytecodeSetEncoderClass}
+ includes: EncoderForV3)
+ ifFalse: [#(testSourceRangeAccessForBlueBookInjectInto)]
+ ifTrue: [#()])
+ !
- ^#(testDebuggerTempAccess testInjectIntoDecompilations testInjectIntoDecompiledDebugs)!
Item was changed:
----- Method: ClosureCompilerTest>>testSourceRangeAccessForBlueBookInjectInto (in category 'tests') -----
testSourceRangeAccessForBlueBookInjectInto
"Test debugger source range selection for inject:into: for a version compiled with closures"
"self new testSourceRangeAccessForBlueBookInjectInto"
| source method |
+ [source := (Collection sourceCodeAt: #inject:into:) asString.
- source := (Collection sourceCodeAt: #inject:into:) asString.
method := (Parser new
encoderClass: EncoderForV3;
parse: source
class: Collection)
generate: (Collection compiledMethodAt: #inject:into:) trailer.
+ self supportTestSourceRangeAccessForInjectInto: method source: source.]
+ on: Error
+ do: [:e | self fail: e.]!
- self supportTestSourceRangeAccessForInjectInto: method source: source!
Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours:
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068296.html
Name: Graphics-tpr.354
Ancestors: Graphics-mt.353
Extend Color class>>fromString parsing to allow also a string of the forrm 'r,g,b' where each item can be the string representation of an integer between 0 and 255
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068297.html
Name: Files-cmm.160
Ancestors: Files-cmm.159
Enable DirectoryEntry's with directory-tree enumeration and content stream access.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068298.html
Name: System-mt.840
Ancestors: System-mt.839
Fixes a bug when filing out change sets with changed class definitions. The bug was that the concept of "fat class definition", which merges old and new inst vars for the loading process, did not account for existing vars in the super class.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068299.html
Name: System-mt.841
Ancestors: System-mt.840
Make the file-drop-service-menu more descriptive.
=============================================
http://lists.squeakfoundation.org/pipermail/packages/2016-July/068300.html
Name: Morphic-mt.1198
Ancestors: Morphic-mt.1197
Make system progress morph use accessors instead of direct instVar accesses to support smooth feature updates in the future.
Especially #position:label:min:max: is problematic if we add or remove instVars in SystemProgressMorph because that morph is used when loading code. We had a similar issue with HandMorph in the recent past.
Still, bugs can occur when updating from older Squeak images. Hence, we might want to consider "restarting" the update process for each entry in an update map.
=============================================
Marcel Taeumel uploaded a new version of System to project The Trunk:
http://source.squeak.org/trunk/System-mt.841.mcz
==================== Summary ====================
Name: System-mt.841
Author: mt
Time: 22 July 2016, 11:48:11.810945 am
UUID: b1354fd5-23af-774d-a0df-a39882b25093
Ancestors: System-mt.840
Make the file-drop-service-menu more descriptive.
=============== Diff against System-mt.840 ===============
Item was changed:
----- Method: DefaultExternalDropHandler>>chooseServiceFrom: (in category 'private') -----
chooseServiceFrom: aCollection
"private - choose a service from aCollection asking the user if needed"
^aCollection size = 1
ifTrue: [aCollection anyOne]
ifFalse:[UIManager default
chooseFrom: (aCollection collect:[:each| each label])
+ values: aCollection
+ title: 'You dropped a file. Please choose a service:' translated withCRs].!
- values: aCollection].!
Item was changed:
----- Method: ExternalDropHandler class>>chooseServiceFrom: (in category 'private') -----
chooseServiceFrom: aCollection
"private - choose a service from aCollection asking the user if needed"
^aCollection size = 1
ifTrue: [aCollection anyOne]
ifFalse:[UIManager default
chooseFrom: (aCollection collect:[:each| each label])
+ values: aCollection
+ title: 'You dropped a file. Please choose a service:' translated withCRs].!
- values: aCollection].!
This does change the return type in the case of a directory from a
FileDirectory to a DirectoryEntryDirectory.
The intention is to gradually migrate most of FileDirectory's public
api to DirectoryEntryDirectory, and let FileDirectory be concerned
with the platform-specific access issues.
In the meantime, the most compatible thing to do for backward
compatibility with the old #/ is to send #asFileDirectory to the
result.
On Thu, Jul 21, 2016 at 3:44 PM, <commits(a)source.squeak.org> wrote:
> Chris Muller uploaded a new version of Files to project The Trunk:
> http://source.squeak.org/trunk/Files-cmm.159.mcz
>
> ==================== Summary ====================
>
> Name: Files-cmm.159
> Author: cmm
> Time: 21 July 2016, 3:44:04.16263 pm
> UUID: 52077283-7106-4ab6-8368-d5f4b5f6f641
> Ancestors: Files-eem.158
>
> Fix FileDirectory>>#/ to not assume the argument represents another directory; it could be a file.
>
> =============== Diff against Files-eem.158 ===============
>
> Item was changed:
> ----- Method: FileDirectory>>/ (in category 'path access') -----
> / aString
> "Answer a FileDirectory on a subdirectory named aString, of the receiver."
> + ^ FileDirectory directoryEntryFor: (self fullNameFor: aString)!
> - ^ FileDirectory on: (self fullNameFor: aString)!
>
>