[Webkit-unassigned] [Bug 22339] Support Indian web sites with EOT by on-the-fly transcoding to Unicode

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Jan 14 21:29:54 PST 2009


https://bugs.webkit.org/show_bug.cgi?id=22339





------- Comment #7 from prunthaban at google.com  2009-01-14 21:29 PDT -------
Hi,

Before submitting a new patch, I would like to get clarification on certain
issues.
The current encoding machinery I use is from Padma and the Padma architecture
is different from the current TextEncoding machinery of Webkit.
(i) Padma uses a generic code base to transcode any font. These generic
routines do bulk of the pre-processing. When you need to add a new font, it is
enough to add two look-up tables and pass the corresponding objects to this
transcoder.
(ii) Padma converts the ASCII characters into an intermediate form before
converting them into unicode characters of a particular language (So there is
an additional level in the transcoding).

Now, if we want to use the TextEncoding architecture of Webkit, (if my
understanding is correct) we need to have one TextEncoding class for each font
we add. But as I stated earlier, Padma transcoders share ~80% code across fonts
& languages and the only way we can achieve this is to make those TextEncoding
objects call the corresponding routines in the Padma implementation. So
effectively we will still have all the Padma architecture hidden behind an
additional level of TextEncoding objects.

If we need to get rid of the Padma code-base, then we need to write encoders
for each font from scratch which is a big effort.

The main classes used in this implementation are,
1) Syllable
2) Parser
3) DynamicParser
4) Converter
5) Encoder

The first 3 classes are the core of the transcoding API. Additionally each font
requires one Encoder and each language requires a Converter. We need to pass
the encoder and converter to the API to perform transcoding on any font (Other
than this there is one Singleton and another static helper class with helper
methods).

So after looking into the TextEncoding API of WebKit, I am not sure how we can
simplify the existing Padma architecture by using TextEncoding.
What are your thoughts on this?

Regards,
Prunthaban


-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.



More information about the webkit-unassigned mailing list