Jan Vrany <jan.vrany@labware.com> [Thu, 12 Nov 2020 12:02:05 +0000] rev 25417
Add `CharacterArray >> #emphasisAtPoint:on:` for compatibility
Jan Vrany <jan.vrany@labware.com> [Mon, 31 Aug 2020 12:01:25 +0100] rev 25416
Cherry-picked changes from eXept, part 2
...from commit a6b6dda4caff:
* CharacterArray.st
* String.st
* SequenceableCollection.st
Jan Vrany <jan.vrany@labware.com> [Mon, 31 Aug 2020 11:59:30 +0100] rev 25415
Cherry-picked changes from eXept, part 1
...from commit a6b6dda4caff:
* CompiledCode.st
* ExecutableFunction.st
* Method.st
* RecursionLock.st
Jan Vrany <jan.vrany@labware.com> [Thu, 13 May 2021 11:26:34 +0100] rev 25414
Add more compatibility methods to `UninterpretedBytes`
Jan Vrany <jan.vrany@labware.com> [Wed, 05 May 2021 10:40:40 +0100] rev 25413
Honour formatter settings in `SmalltalkLanguage >> #methodDefinitionTemplate...`
Jan Vrany <jan.vrany@labware.com> [Mon, 29 Mar 2021 21:22:40 +0100] rev 25412
Fix `Collection >> #sortedBy:` w.r.t unordered collections
...such as `Set`s
Jan Vrany <jan.vrany@labware.com> [Thu, 25 Mar 2021 20:30:03 +0000] rev 25411
Fix unlikely but possible race in `WeakValueDictionary`
It may happen that value in `valueArray` could have been already
collected by the GC but #clearDeadSlots have not yet been called.
When this happened, `#at:ifAbsentPut:` returned tombstone rather
than updating the dictionary with value from block.
This commit fixes this by checking whether `valueArray` contain
the tombstone and if so, clearing up the dead slots and restarting
the operation. HTH.
Jan Vrany <jan.vrany@labware.com> [Wed, 03 Feb 2021 11:52:30 +0000] rev 25410
Reworked `SimulatedClassPool >> #associationAt:ifAbsent:` to throw an error
...if there's no such class variable. This makes it behave more like
dictionary / pool.
Jan Vrany <jan.vrany@labware.com> [Wed, 03 Feb 2021 06:58:24 +0000] rev 25409
Add (missing) `CharacterArray >> #isBinary`
`#isKeyword` and `#isUnary` were already there so having `#isBinary`
seems right.
Jan Vrany <jan.vrany@labware.com> [Tue, 02 Feb 2021 09:15:34 +0000] rev 25408
Fix `PositionableStream >> #peekForAll:` w.r.t end of stream
`#peekForAll:` raised an error when trying to peek past the end
of the stream, whereas it should (it is more logical) to just
return `false`). This is how it behaves in other dialects, too.