Thank you, Arthur (and all others who have helped to identify the
cause). As an interim work-around, I had added :
\def ͵{\char `\͵ \penalty 1 \relax}
\def ʹ{\penalty 1 \char `\ʹ \relax}
which worked but which /may/ have had undesirable side-effects (none
noticed so
With ancientgreek selected, XeTeX hyphenates :
Τῆ Εʹ τῆς ΚΑʹ ἑβδοsuppliedμάδος
as
Τῆ Ε-ʹ τῆς ΚΑʹ ἑβδοsuppliedμάδος
i.e., between the Ε-majuscule and the number sign ʹ.
It sounds like you have set up the numeral sign for hyphenation (i. e.,
given it a non-zero \lccode).
Are you sure that that is the UNIcode code point of the numeral sign? To
me it appaears as an apostrophe. In any case is a strange situations.
Examine the hyphenation files for ancient greek; thay are prepared by
Apostolos Syropoulos, who is very knowledeable in this matter.
I am not even
Are you sure that that is the UNIcode code point of the numeral
sign?
Yes, this is U+0374 GREEK NUMERAL SIGN. Depending on the font, it may
well be displayed with a curl like an apostrophe, but it clearly
shouldn't be treated as a letter.
Arthur
On 6/11/14 12:33, Arthur Reutenauer wrote:
Are you sure that that is the UNIcode code point of the numeral
sign?
Yes, this is U+0374 GREEK NUMERAL SIGN. Depending on the font, it may
well be displayed with a curl like an apostrophe, but it clearly
shouldn't be treated as a letter.
Its
Its Unicode general category is Lm (Letter Modifier), which means
it'll be assigned \catcode = 11 and \lccode = itself by default. So
you may want an extra pattern entry in the Greek files:
8ʹ
to suppress a possible hyphenation position when it occurs after a
Greek letter.
Good
No, no changes to any \lccodes or \uccodes at all, Arthur.
Understood. But as is implied in the rest of the thread, XeTeX does
that for you in the format (via unicode-letters.tex). For the time
being, just reset the character's \lccode to zero and you'll be fine.
I'll be discussing a