[HarfBuzz] Fwd: 'rakar' for cluster സ്പ്ര is not reordered when there is full base conjunct

Rajeesh K Nambiar rajeeshknambiar at gmail.com
Mon Mar 4 11:18:17 PST 2013

On Mon, Mar 4, 2013 at 10:43 PM, Cibu Johny (സിബു) <cibu at google.com> wrote:

> This happens in hb 0.9.13
> and h
> appens in Uniscribe as well.
This (in general 'Ra') is one failure case we have with 'old-spec' (mlym).
The issue is also part of language reformation - the oldlipi had glyphs for
all consonants+Virama+Ra rendering 'Ra' as "Post-Base", but the reformed
lipi cut the Virama+Ra part into "Pre-Base" to be rendered in front of the

In this specific example "സ്പ്ര" (Sa+Virama+Pa+Virama+Ra) ideally, the
following in the sequence should happen (for mlym, at least) to define

1. If there is a glyph defined for "സ്പ്ര" (Sa+Virama+Pa+Virama+Ra) in the
font, use that.
2. If there is a glyph defined for "സ്പ" (Sa+Virama+Pa), use that glyph and
render the "്ര" (Virama+Ra sign) as pre-base to this glyph. Here 'Ra'kar is
render pre-base to cluster of consonants (Sa+Virama+Pa).
3. If there is a glyph defined for "പ്ര", then render it as "സ്‌‌പ്ര"
(Sa+Virama+ <Glyph for Pa+Virama+Ra).
4. Default case to be "സ്്രപ" (Sa+Virama+"്ര" (Virama+Ra sign)+Pa). Here
'Ra'kar is joined with last consonant Pa.

Santhosh, please correct/add if I missed something.

> Font is Rachana from
> ttf-indic-fonts-core
>  (/usr/share/fonts/truetype/ttf-indic-fonts-core/Rachana_04.ttf) in a
> Debian
> /Ubuntu
> dist.
Latest version of Rachana is here:

> _______________________________________________
> HarfBuzz mailing list
> HarfBuzz at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/harfbuzz

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/harfbuzz/attachments/20130305/dd097ca4/attachment.html>

More information about the HarfBuzz mailing list