Full-width English letters are ignored by XLIFFWriter

[Xinqi Su]

Hello,

Full-width English characters in segment source/target content are ignored by XLIFFWriter when writeUnit() is called. Same happens to half-width East Asian language characters.

We have such use cases

• both full-width and half-width English characters co-exist in source content

• both half-width English and Japanese characters co-exist in source content

Part of the content in above use cases is missing after converted to XLIFF because Okapi XLIFFWriter ignores them. Would it be possible for Okapi to handle those cases? That will be very helpful. Thank you.

Please find the relevant Okapi source code at https://tinyurl.com/y9k5bg64

[YvesS]

Fixed bug preventing some Unicode characters to be output. This resolves issue #21.

→ <<cset ce1615ca8e23>>

[YvesS]

Well that was a major bug. Thanks for catching and reporting it.

The Maven artifact for the updated 1.7-SNAPSHOT should be here: http://repository-okapi.forge.cloudbees.com/snapshot/ And we should release 1.7 before the end of the year.