Need step for removing spaces in Asian output

[Former user]

Original issue 133 created by @ysavourel on 2010-03-31T19:57:12.000Z:

White spaces need to be removed in between segments.
It would be nice to have a step for this.

[Former user]

Comment [1.](https://code.google.com/p/okapi/issues/detail?id=133#c1) originally posted by @ysavourel on 2010-03-31T20:30:30.000Z:

Fredrik: This is from the ENG meeting. Not urgent. I may do it at some point. The backend is more important.

[Former user]

Comment 2. originally posted by @ysavourel on 2013-02-04T17:28:18.000Z:

[Former user]

Comment 3. originally posted by ting... on 2014-01-16T22:31:33.000Z:

I've been messing around with a stupidly simple step to solve this problem. I was keeping it separate for development since I haven't really thought out things like options. I can add it to the core when it's more fleshed out.

https://github.com/tingley/okapi-strip-whitespace-step

[Former user]

Comment 4. originally posted by ting... on 2014-01-29T19:00:44.000Z:

It may not be possible to fix this in a single step for all formats, because of an unfortunate hack in the merge code for Ontram's XINI format.

For all non-XINI formats, it's possible for a simple step to handle each TextUnit, check the targets for a ZH/JA locale, and trim whatever whitespace is needed.

However, XINI passes the translation as the *source*. (The merger code sets a `useSource` flag to indicate this internally.) In this case, the target locales must be injected as a step parameter and the source must be trimmed if a ZH/JA target locale is found.

So far, I haven't found a way for the step to detect which mode to use. Can I get merge information from the START_DOCUMENT event? I'll have to check. That might work.

[Chase Tingley]

Fixed by @DenisKonovalyenko with the AddWhitespaceAfterKutenStep and RemoveWhitespaceAfterKutenStep classes.