PDF rendering (and SVG-in-PDF rendering) of Arabic fonts not working correctly with ligatures for many fonts

[Robert Gurol <ro...@signavio.com>]

Hi all,

for rendering PDF containing both Arabic text and SVG containing arabic
text, I recently experimented with different fonts.

Except for the Tahoma font (which I'd rather not use for its license
restrictions), Arabic text is rendered without ligatures, i.e. the letters
of words are displayed as if they were written alone, while Arabic demands
they be rendered differently inside of words and at the beginning and end
of words, so they apper connected like hand-written text.

That even though most fonts I tried seem to work perfectly well with PNG
transcoding (when registered with the GraphicsEnvironment beforehand),
generating Arabic text as expected.

The fonts I tried so far are Lateef, Scheherazade and (Wikipedia-used)
Amiri.
Amiri (Regular) contains almost all characters of the unicode character
sets Arabic, Arabic Supplement, Arabic Presentation Forms A and Arabic
Presentation Forms B.

As far as I know, ArabicTextHandler tries to replace adjacent letters with
the respective connected forms and annotates the characters (not: code
points, alas) with the respective in-word-position information for the
letter, and it is called in both PNG transcoding and SVG-in-PDF transcoding.
However, PNG looks much better in all cases, the ligatures seem to work
only there.

Is this a known issue?

If not, I will gladly provide more information and test resources.

Regards,

Robert

Re: PDF rendering (and SVG-in-PDF rendering) of Arabic fonts not working correctly with ligatures for many fonts

[Glenn Adams <gl...@skynav.com>]

Arabic is not presently supported in SVG when using FOP. For direct FOP use
(not SVG), then if you are using any version of FOP prior to FOP 1.1,
Arabic is not supported. If you are using 1.1 or current dev build, then
you are doing something wrong.

You need to provide more details.


On Wed, May 14, 2014 at 11:08 AM, Robert Gurol <ro...@signavio.com>wrote:

> Hi all,
>
> for rendering PDF containing both Arabic text and SVG containing arabic
> text, I recently experimented with different fonts.
>
> Except for the Tahoma font (which I'd rather not use for its license
> restrictions), Arabic text is rendered without ligatures, i.e. the letters
> of words are displayed as if they were written alone, while Arabic demands
> they be rendered differently inside of words and at the beginning and end
> of words, so they apper connected like hand-written text.
>
> That even though most fonts I tried seem to work perfectly well with PNG
> transcoding (when registered with the GraphicsEnvironment beforehand),
> generating Arabic text as expected.
>
>  The fonts I tried so far are Lateef, Scheherazade and (Wikipedia-used)
> Amiri.
> Amiri (Regular) contains almost all characters of the unicode character
> sets Arabic, Arabic Supplement, Arabic Presentation Forms A and Arabic
> Presentation Forms B.
>
> As far as I know, ArabicTextHandler tries to replace adjacent letters with
> the respective connected forms and annotates the characters (not: code
> points, alas) with the respective in-word-position information for the
> letter, and it is called in both PNG transcoding and SVG-in-PDF transcoding.
> However, PNG looks much better in all cases, the ligatures seem to work
> only there.
>
> Is this a known issue?
>
> If not, I will gladly provide more information and test resources.
>
> Regards,
>
> Robert
>