Re: UnicodeData.txt problem

From: Mike Ayers (
Date: Fri Dec 09 2005 - 14:36:27 CST

  • Next message: Tom Emerson: "Re: UnicodeData.txt problem"

    Tom Emerson wrote:

    >>There are known parsers of UnicodeData.txt that currently can't handle
    >>additional syntax. There are "unknown unknown" parsers that may not
    >>handle it either...
    > So we break them. We make them add a line that ignores any line that
    > starts with a '#'. For crying out loud, the advnatages of including
    > the version number in thee files far outweights the minimal
    > inconvenience it would cause to people parsing them.

            For you, perhaps, assuming that you are correct in your implicit belief
    that you do not constantly rely on one of these breakable parsers. Me,
    I'd rather not risk finding out.

    > Telling people to compare file dates, file sizes, or whatever,
    > rediculous.

            If, by "rediculous", you mean "common engineering practice", then yes,
    it's quite rediculous. Best to normalize line endings, then use file
    size + CRC to compare - the odds of this failing are sub-negligible,
    whereas file timestamps are notoriously unreliable.

            Better question - what would be the benefit of... no, no, I know better
    than to ask...


    This archive was generated by hypermail 2.1.5 : Fri Dec 09 2005 - 14:37:41 CST