Re: tibetan rendering support in ICU

Guntupalli Karunakar <[email protected]> Thu, 4 Nov 2004 18:05:04 +0530
Newsgroups gmane.comp.lib.icu.general
Organization Netcore Solutions
Message-ID <[email protected]>
On Tue, 02 Nov 2004 09:43:23 -0800
Eric Mader <[email protected]> wrote:

> Hello Karunakar,
> 
> Jens' analysis looks correct to me. The featureTags variable is a 
> pointer to a pointer to a tag, i.e. a pointer to an array of tags,
> so you need the "&" operator.
> 
 there was a bigger mistake I was doing in 
  for (le_int32 idx=offset; idx < count+offset ; idx++ ) {

changing to this made it work ;)
  for (le_int32 idx=0; idx < count ; idx++ ) {

 I actually simplified it, just assigning all tags to each char. (Will
have to test well how good this approach is, but it would get us some
working stuff to try in Openoffice)

        for (le_int32 idx=0; idx < count ; idx++ ) {
                featureTags[idx]=tagArray;
        }

> After a quick glance at your code, I have a few other comments:
> 
> * The adjustGlyphPositions() and glyphProcessing() methods look like
> 
> they're the same as the ones which you would inherit from 
> OpenTypeLayoutEngine. You can probably just inherit.
>
> * If you have the option, I suggest that you work w/ a newer version
> of ICU - there have been many improvements and bug fixes made to the
> 
> LayoutEngine since ICU 2.2. (We're about to release ICU 3.2!)

Basic aim was to get it working in OOO, 1.1.3 was using ICU 2.2 , will
nyways try with 3.2 also.

Regards,
Karunakar