|
Bumping (to be done with other iconv issues)
soichiro ohba added support for ignore. Transliteration appears to be a tougher problem. Marking as 1.x feature since we will not get this done before final 1.0 release (unless someone knows an easy way to add transliteration)
Revisit for 1.1. Maybe not fixed, but we should give it a go.
Unless we can find an outside library which does tranliteration I don't think we can do this since Java does not support it.
Is any additional work needed to make the iconv string at least not blow up? Perhaps we could ignore the translit part or display a warning and continue.
How about the use of ICU4J?
http://www.icu-project.org/index.html Very interesting Koichiro...I did not know this project existed. Would it provide the missing functionality we need? Is there any down side to using it instead of the Java Charset classes?
Yes, ICU4J would provide the translit functionality to iconv on JRuby.
(For further information, see http://www.icu-project.org/apiref/icu4j/com/ibm/icu/text/Transliterator.html But, there is a down side actually. It's too huge. (icu4j.jar<4.4MB>, icu4j-charsets.jar<2.4MB>). I would be satisfied if we can fix it to just ignore the translit and perhaps warn. If anyone complains about translit missing we'll bother them to write us a gem-based Iconv extension with ICU4J.
|
|||||||||||||||||||||||||||||||||||||||||||||||
Finish for 1.0...shouldn't be a huge deal now that we know about it, but any information on the various iconv settings would speed the process along.