I have an XSLT script which will convert our XML into a PE ready version. It converts into a , which has, in the past/for years, allowed the composition to honor that to be the preferred word breaking point. At some point this stopped working, however, but I'm not sure when.
Is this the preferred method to get discretionary hyphens to output using PE?
Firstly, U+200B is zero-width space and not discretionary hyphen. For discretionary hyphen (soft hyphen) you should use character U+00AD ().
Secondly, recent Arbortext releases now ship with two formatting engines. The original FOSI engine is used when selected in Styler, for XSL-FO, or for FOSI stylesheets. The new APP engine is used by default in Styler, or for .3F stylesheets.
I tested in the APP engine and it worked as I expected it to. I didn't get around to testing FOSI yet.