[tex-hyphen] hyphenation patterns with apostrophe

Mojca Miklavec mojca.miklavec.lists at gmail.com
Mon Jun 6 16:34:56 CEST 2011

On Mon, Jun 6, 2011 at 15:59, Arthur Reutenauer wrote:
>> It would be nice and it would be time to solve the ignored issue with
>> apostrophe. At the moment XeTeX uses U+0027 only and LuaTeX uses
>> U+2019 only.
>  Can you please remind us what the problem was exactly?

a) When users write "l'arcobaleno" with
without turning the auto-replace-U+0027-with-U+2019 and if patterns
include <U+2019> as apostrophe (which is the case in LuaTeX), this
won't be hyphenated properly.

b) On the other hand if users turn the autoreplacement on (which is
the proper way to do it anyway) and patterns include <U+0027> as
apostrophe (which is the case in XeTeX), it won't be hyphenated
properly again.

Now, honestly, the change that I proposed would only effect the case
(a) which is a bad typesetting practice anyway. Maybe we don't want to
promote bad typesetting habits, so it might be perfectly legitimate to
say: "please use proper apostrophe if you want proper hyphenation".

The case (b) which is a much bigger problem at the moment could also
be solved by replacing U+0027 with U+2019 in all sources and then
convert U+2019 into U+0027 on the fly, like we do for other UTF-8

Alternatively we could auto-generate files with duplicated patterns
for all apostrophes and load that file in addition to the main one.
(We would have to write to XeTeX mailing list about that as well.)


More information about the tex-hyphen mailing list