Stéphane Ducasse wrote
> 
> I'm unclear. Does it mean that we have a problem with zip? Package?
> 

I don't really know enough about text encoding to say. It seems that people
have run into things like this before. I collected all the references I
could find and made an issue:
Issue 6160:     Monticello: Zipping Wide Characters
http://code.google.com/p/pharo/issues/detail?id=6160

Pharo2.0a
Latest update: #20160

If one has a method with wide character literals, when Monticello saves the
package, the source.st file inside the mcz file will only be valid if
unzipped inside Squeak/Pharo.

This is because Pharo writes zip files in 4096 byte chunks, which will all
be ByteStrings, except the ones containing wide characters, which will be
WideStrings. Thus, source.st will be a random mashup of such 4096 byte
stretches.

See the following for possibly related info:
*
http://forum.world.st/Monticello-mcz-files-write-their-DataStream-and-a-portion-of-their-chunk-files-as-WideStrings-which--td2294551.html
* https://code.google.com/p/pharo/issues/detail?id=2697
* http://code.google.com/p/pharo/issues/detail?id=6143
* https://code.google.com/p/pharo/issues/detail?id=830

Squeak/Pharo encoding in general
* http://www.visoracle.com/squeak/faq/unicode-utf8byte.html
* http://forum.world.st/squeak-dev-WideString-UTF-8-UTF-32-UCS2-td74243.html
* http://forum.world.st/WideString-performance-td4077404.html

HTH,
Sean

--
View this message in context: 
http://forum.world.st/Problem-with-Monticelo-package-and-corrupted-sourcecode-zip-tp4636458p4636620.html
Sent from the Pharo Smalltalk mailing list archive at Nabble.com.

Reply via email to