Daily Commit Log

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Daily Commit Log

commits-2
Changes to Trunk (http://source.squeak.org/trunk.html) in the last 24 hours:

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003738.html

Name: Collections-ul.379
Ancestors: Collections-ar.378

- fix: ByteArray should understand #indexOf:startingAt: because StandardFileStream >> #upTo: uses it in binary mode. Also added a general implementation for SequenceableCollection.
- fix: Symbol class >> #allSymbolTablesDo:after: missed some elements, so Alt+q (Cmd+q on Mac) missed some selectors. WeakSet >> #do:after: was also responsible for some misses.
- enh: faster String >> #indexOfAscii:inString:startingAt:

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003739.html

Name: Collections-ul.380
Ancestors: Collections-ul.379

- faster non-primitive String "primitives" by using #basicAt: instead of #at: + #asInteger/#asciiValue
- revert to #to:do: instead of a #whileTrue: in #indexOfAscii:inString:startingAt: , because that's also inlined even if the receiver type is unknown

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003740.html

Name: Collections-ul.381
Ancestors: Collections-ul.380

- simplified SequenceableCollection >> #indexOf:startingAt:

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003741.html

Name: ToolBuilder-MVC-dtl.21
Ancestors: ToolBuilder-MVC-dtl.20

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003742.html

Name: ToolBuilder-Kernel-dtl.37
Ancestors: ToolBuilder-Kernel-eem.36

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003743.html

Name: System-dtl.374
Ancestors: System-ar.373

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003744.html

Name: Tools-dtl.269
Ancestors: Tools-eem.268

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003745.html

Name: ST80-dtl.118
Ancestors: ST80-eem.117

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003746.html

Name: Morphic-dtl.465
Ancestors: Morphic-ar.464

MVC fixes by Florin Mateoc to address debugger opening, including low space interrupt and user interrupt in MVC.

=============================================

http://lists.squeakfoundation.org/pipermail/packages/2010-September/003747.html

Name: Kernel-klc.493
Ancestors: Kernel-nice.492

This makes two significant changes to DateAndTime/TimeStamp either of which fixes a number of TimeStamp tests.

1.  Ensure that when the date is extracted from either DateAndTime or TimeStamp that the date is defined in terms of a DateAndTime.

2.  If a TimeStamp and a DateAndTime represent equivalent instances in time then allow #= to return true when comparing them.

These changes may be controversial.  As mentioned when I discussed this on the #squeak IRC channel there are conceptual differences between TimeStamps and DateAndTimes.  I don't disagree with this notion, however that is not how TimeStamp is implemented currently in Squeak.  I admit I'm taking the easier approach since I would have to give some serious thought and research into a 'proper' redesign of TimeStamp.

For now these changes make TimeStamp/DateAndTime at least work like they seem to be expected to based on the tests.

=============================================