Re: Representative glyphs for combining kannada signs

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Fri Mar 17 2006 - 15:21:39 CST

  • Next message: Sandra Bostian: "Re: New symbol for Russian rouble?"

    From: "Peter Constable" <petercon@microsoft.com>
    >> From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org] On
    >> Behalf Of Antoine Leca
    >
    >> "Most up-to-date" (available to the general public) would be Vista's March
    >> CTP, correct?
    >
    > February CTP. Another refresh is coming. (Sorry, can't do an upgrade from Feb CTP.)
    >
    >> Since it is a beta, perhaps you can consider a bug report, then.
    >
    > What's the bug? Probably Philippe was trying to display these on XP without having enabled support for complex scripts, in which case that's an expected outcome and not a bug.

    No! Asian and Middle-East Complex scripts ARE enabled in Windows XP Control Panel; additionally, all the extra code pages are installed, to install also the large core fonts needed for Han and GB18030). So this is not an expected outcome, because most Indic scripts still have rendering bugs (except a few, notably Devanagari, Tamil and Kannada which are working properly in IE6, IE7Beta, Firefox, MS Office, and OpenOffice). I have also no problem with Arabic, Hebrew, Thaana, Syriac, Ethiopian.

    But I still have problems with UCAS in Internet Explorer only, and with a few scripts for which it is impossible to select the appropriate OpenType font despite it is installed (and working in Notepad, Word, and Firefox): Myanmar, Oriya.

    And non properly working Indian scripts include:
    * Bengali, Oriya, Malayalam (left matras displayed on the right),
    * Telugu (minor: only one matra is subjoined, others are not subjoined or upper joined but always displayed after the consonnant)

    Thai is working but due to its special visual encoding order, so reordering is not a problem.

    I am still experimenting with Lao, and I'm still not sure that Tibetan is correct with the Tibetan font I have installed. The purpose of these tests is to set a configuration guide for users, and to detect documents whichare not properly encoded and use various tweaks when trying to work around common rendering bugs. From the data I have, there are legions of documents with such tweaks, so this is now a serious interoperability problem.

    This archive was generated by hypermail 2.1.5 : Fri Mar 17 2006 - 15:27:01 CST