soft hyphens in docx conversion output

[quinta@ebf.cz]

Soft hyphens marks (characters U+00AD, or entitities #173 or shy), originally existing in html, are exported to docx (again) as shy characters (code 00AD).

Which is not quite desired behaviour, cause MS Word implements optional word breaks differently, and characters 00AD itself are simply displayed (visually simillary as standard hyphens).

Exported docx document containing shy characters can be repaired by searching shy characters (using symbol ^0173), and replacing them: either by Word "optional word break" (^-), or (mostly in my case) just deleting them by replacing by nothing...

Anyway: Is such export behaviour intentional? Or - mayby - is for some reason inevitable? Is there any way how to achieve replacing shy characters to MS Word "optional word break" as part of conversion?

[kovidgoyal]

Well just use the search and replace feature in the conversion dialog to replace the soft hyphen character with whatever you like, (IIRC the zero width non-joiner is what word uses for optional spaces).

[kovidgoyal]

And note that in DOCX softhyphens are represented as a special tag not as a unicode character. From the next release calibre will convert soft hyphens to that tag automatically. https://github.com/kovidgoyal/calibr...c9948658be0db8

[quinta@ebf.cz]

Originally Posted by kovidgoyal

And note that in DOCX softhyphens are represented as a special tag not as a unicode character.

Yes, docx softhyphens (optional word brakes), seem to look more like objects then characters. They do not react to unicode-revealing shortcut Alt+X, their "ascii value" is 31 (?), their XML representation is element <w:softHyphen/> (not character)...

My attempts to create optional word brakes using suggested Calibre "search and replace" export feature was (yet) not succesfull. Well, all I tried was replacing using expression \u200C (which is unicode value of suggested "zero width non-joiner"), and using expression \u001F (hexa value of 31)... Excuse my naive approach. : )

Possible good reason for converting soft hyphens to OWB as default Calibre export behaviour: MS Word itself is behaving that way. OWB are converted to SHY when exported to HTML, and vice versa (just tested in Word 2010).

Originally Posted by kovidgoyal

From the next release calibre will convert soft hyphens to that tag automatically. https://github.com/kovidgoyal/calibr...c9948658be0db8

Ooh, I think thats better message than expected. Thank you.