Nicolas Cellier uploaded a new version of Monticello to project The Trunk:
http://source.squeak.org/trunk/Monticello-nice.546.mcz==================== Summary ====================
Name: Monticello-nice.546
Author: nice
Time: 31 May 2013, 12:47:59.453 am
UUID: ff2975ee-c296-4d65-8b8a-9b512607a2be
Ancestors: Monticello-kb.545
Let's encode .mcz/snapshot/source.st in UTF8
Let's decode UTF8 in MCStReader.
Previously, it was encoded in latin-1 (iso-8859-L1) for ByteString and UTF-32BE for WideString.
And would always be decoded in latin-1 by MCStReader.
Note that compatibility with legacy code is handled by catching InvalidUTF8 exception.
An alternative would be to use a BOM in new snapshot. However, since the snapshot/source.st is not used by Squeak tools, it's not really worth.
=============== Diff against Monticello-kb.545 ===============
Item was changed:
----- Method: MCMczWriter>>serializeDefinitions: (in category 'serializing') -----
serializeDefinitions: aCollection
+ ^(String streamContents: [:aStream |
+ | writer |
+ writer := self snapshotWriterClass on: aStream.
+ writer writeDefinitions: aCollection])
+ squeakToUtf8!
- | writer s |
- s := RWBinaryOrTextStream on: String new.
- writer := self snapshotWriterClass on: s.
- writer writeDefinitions: aCollection.
- ^ s contents!
Item was changed:
----- Method: MCStReader>>readStream (in category 'evaluating') -----
readStream
+ | contents |
+ contents := stream contents.
+ contents := [contents utf8ToSqueak] on: InvalidUTF8 do: [:exc |
+ "Case of legacy encoding, presumably it is latin-1 and we do not need to do anything
+ But if contents starts with a null character, it might be a case of WideString encoded in UTF-32BE"
+ exc return: (((contents beginsWith: Character null asString) and: [ contents size \\ 4 = 0 ])
+ ifTrue: [WideString fromByteArray: contents asByteArray]
+ ifFalse: [contents])].
^ ('!!!!
+ ', contents) readStream!
- ', stream contents) readStream!
Locked
