Re: UnicodeData.txt is NOT invalid, flawed, broken, corrupt OR wrong

From: Philippe Verdy (
Date: Mon Jun 13 2005 - 09:48:23 CDT

  • Next message: Erik van der Poel: "Re: Arabic letters separated by markup"

    From: "Theodore H. Smith" <>
    >I think I found the bit I'm looking for:
    > OK. So then the UnicodeData.txt is not invalid, flawed, broken, corrupt
    > and wrong :)
    > Well I did say right from the start that I could be wrong and I'd like to
    > get the official answer, which I did not get last time I raised this
    > question.

    Whatever you said in your email, the first thing readers have seen is your
    affligeant subject line (that I need to correct here).
    I have detailed you in further private emails all the relevant info and
    links that you had missed.

    It would have been much more polite to use a more appropriate subject line
    like: "normalized compositions cause problem with my application", and then
    asking for help.

    It would have been obvious to test your implementation and compare its
    results with other existing and conforming implementations. It was obvious
    to use the normalization test data which is in the UCD: so now, download
    this big file and make sure that your corrected implementation passes this
    documented test. If it still does not, reread the specs, or ask for help if
    there's things you still don't understand in those specs.

    This archive was generated by hypermail 2.1.5 : Mon Jun 13 2005 - 09:51:16 CDT