[XeTeX] Polytonic greek and XeLaTeX

[Ross Moore]

Hi Peter,

On 03/01/2008, at 1:38 AM, Peter Dyballa wrote:

>
> Am 02.01.2008 um 05:09 schrieb Ross Moore:
>
>> When both methods are supported, the lines beginning 'decomposed'
>> show the result of "Omega + combining comma" (i.e.  U+039F;U+0313;)
  oops, I meant Omicron (not Omega) here

>> appearing three times on the line, with the 4th instance being
>> 'Omega with psili'  (i.e. U+1F48; ). The difference is usually
    and here.

>> visible.
>
>
> The correct way to produce Ὀ (GREEK CAPITAL LETTER OMICRON WITH
> PSILI, U+1F48) would be to use Ο  (GREEK CAPITAL LETTER OMICRON, U
> +039F) with ᾿ (COMBINING COMMA ABOVE, U+0313), which is faked here
> with GREEK PSILI at U+1FBF. The combining effect already failed for
> Johannes (he is referencing a scheme in X11 to compose one character
> from more than one key press event, which seems to have failed), so
> we see the two characters. And since COMBINING COMMA ABOVE, U+0313,
> is a dead character, it should never be printed but lead to sorts of
> failure. To see a psili character the use of GREEK PSILI at U+1FBF is
> needed. To see COMBINING COMMA ABOVE, U+0313, as stand-alone psili
> character, it must be combined with a space character like Latin ^,
> ´, `, ~, or ¨.
>
> I'm not sure whether his original post seems to suggest that a method
> exists in at least some fonts that combines a letter with a combining
> letter (via liga table?) ...

My testing shows that some (but not all) of the GFS fonts have this.
when "Script=Greek" is specified using the commands in  fontspec.sty .
(GFS Theocritis seems to have it automatically, but the actual glyphs
there for he diacritics are very small, almost invisible.)

> so it looks most promising to create a
> "greek-tex-text" mapping to pass to XeTeX the already composed
> characters – in case Johannes assumes that by writing down two (or
> more) characters XeTeX should produce only one. I think he did not
> realise that the composition in his editor failed.

JK's comment suggests that it should not be necessary for the editor
to produce a single character, but ensure that the 2 characters
entered remain adjacent. So his editor did the right thing, but
he was let down by the font not doing the combination correctly.
(TeXshop's editor gets this wrong by inserting a space between
them, under some circumstances.)

Since there are quite a few fonts that do not do it right, then
yes, a  "greek-tex-text"  mapping would be an appropriate thing
to have available when using those fonts. It would be an optional
extra that an author could specify to make XeTeX compensate for
the lack of a standard font feature.

>
> --
> Greetings
>
>    Pete


Cheers,

	Ross

------------------------------------------------------------------------
Ross Moore                                         ross at maths.mq.edu.au
Mathematics Department                             office: E7A-419
Macquarie University                               tel: +61 +2 9850 8955
Sydney, Australia  2109                            fax: +61 +2 9850 8114
------------------------------------------------------------------------