|
From: Ralf S. <ral...@we...> - 2006-09-26 07:40:00
|
On Mon, Sep 25, 2006 at 20:41 -0400, Peter Baker wrote:
>
> How embarrassing that the "liag" error is in the current release.
> Obviously it's long since been corrected in the CVS version. I'll try
> disconnecting the small cap ligatures and moving the " RQD" feature to
> the top of the list. The result should be in the CVS tree tomorrow
> morning, in case anyone wants to test.
BTW, why do you use 'liga' for thinhs like 'Hcircumflex <= H
circumflexnosp'? IMHO something like 'ccmp' would be more appropriate
here. And why do define it only for some characters with diacrtical
marks?
One other thing: AFAIK one has to add features to all script/language
combinations for which they should be active. For example, right now for
latn{ISL} you only define the required feature (why not 'locl') that
replaces [Tt]horn and eth. If I understand it correctly this means that
thngs like ligatures or small caps are no longer available when
latn{ISL} is selected. I don't know why this is the case, though.
I have just tested this with XeTeX: Here the result is indeed that
ligatures or small caps no longer work when latn{ISL} is selected. In
addition, the required feature ' RQD' does not work as expected, ie, I
don't get the nordic thorn. I am not sure if this is a font specifc
errot (maybe feature order), a fontforge error (the way this
non-standard feature is implemented) or a XeTeX/ICU error.
cheerio
ralf
|