Re: fo:character and zero width space

On Fri, Feb 12, 2016 at 12:08 AM, Carlos Villegas <[email protected]> wrote:

> Thanks, I'll try your suggestions.
>
> The open issues in Thai glyph composition are major, basically the output
> is simply not acceptable.
>

Understood.


> Correct positioning of vowels, diacritics and tonemarks is not a esthetic
> thing only, sometimes when they are in the wrong position they print on top
> of each other or on top of the wrong base character.   So I wouldn't say
> FOP supports Thai!
>

Actually, I read and speak Thai, so I would agree. It will be useful to
resolve these last bugs on the Thai script support so folks can start using
FOP for Thai content.


>
> I have Java code that I can donate that does glyph substitution for Thai,
> but this is based on the fact that Thai fonts also encode alternate glyphs
> into the Unicode user range.
>

This approach is not acceptable in FOP due to the reliance upon
non-standard coding and font dependent usage. FOP will only use of
GSUB/GPOS font tables, so we need to make this work properly for Thai.


> There are basically two encodings: Microsoft's and Mac's, and my code
> works with Microsoft's fonts. However, I think, the right way to do this is
> to read the font's own substitution tables, this information must be there
> inside the fonts, and use that. But if the font doesn't really have that
> information, my code could be use as a fallback if you at least know that
> it is a Microsoft encoded font.
>
> Carlos
>
>
> On 2/12/16 15:11, Glenn Adams wrote:
>
>
>
> On Thu, Feb 11, 2016 at 9:23 PM, Carlos Villegas < <[email protected]>
> [email protected]> wrote:
>
>> I'm trying to use FOP 2.0 to render Thai language. FOP doesn't really
>> support Thai script
>
>
> FOP does support Thai, at least glyph composition, though there appears to
> be a few open issues (FOP-2092
> <https://issues.apache.org/jira/browse/FOP-2092>, FOP-2355
> <https://issues.apache.org/jira/browse/FOP-2355>). It is also true that
> FOP is not presently performing line breaking in phrase internal positions (
> FOP-2066 <https://issues.apache.org/jira/browse/FOP-2266>), which
> proposes a solution to employ ICU and use dictionary based LB.
>
>
>> , so I'm using an external program (libthai) to do glyph composition and
>> word breaking.
>> Thai doesn't use spaces to separate words, so I was just inserting zero
>> width spaces to separate words. This works fine.
>> However, for justified text, these zero width spaces don't stretch, so
>> nothing happens, I get no justified text.
>> So, I tried to use fo:character with character set to zero width space,
>> and setting the letter-spacing.maximum to allow some stretching.
>>
>
> Zero width space (U+200B) is not treated as a word separator space by
> default. Have you tried something like:
>
> <fo:character treat-as-word-space='true'>&#x200B</fo:character>
>
> If that doesn't work, then keep in mind that letter spacing has no meaning
> when applied to a single character, but applies to a sequence of
> characters. Accordingly, I would try something like:
>
> <fo:inline letter-spacing.minimum="0em"
> letter-spacing.maximum="0.1em">&#x200B;&#x200B;</fo:inline>
>
> or
>
> <fo:inline letter-spacing.minimum="0em"
> letter-spacing.maximum="0.1em">&#x200B;<fo:character
> treat-as-word-space='true'>&#x200B;</fo:character></fo:inline>
>
> If neither approach works, it may require some time with a visual
> debugger, e.g., Eclipse, to step through the code and find out where things
> go wrong.
>
>
>
>> But now FOP doesn't seem to treat the zero width space as such this way,
>> it doesn't break anymore at all!
>>
>> Note that Thai does make use of regular spaces to separate sentences and
>> for other purposes but there's no guarantee that you have spaces in a given
>> paragraph or in case of long sentences.
>>
>> Is there a way to insert a breakable space with minimum and optimum to
>> zero width and some small value for maximum?
>>
>> Carlos
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [email protected]
>> For additional commands, e-mail: <[email protected]>
>> [email protected]
>>
>>
>
>