Re: Handling of Surrogates

From: Asmus Freytag (
Date: Fri Apr 17 2009 - 00:17:07 CDT

  • Next message: Philippe Verdy: "RE: Handling of Surrogates"

    On 4/16/2009 9:32 PM, Mark Davis wrote:
    > It only implies it if it was spec'ed to imply it.
    All of us can agree with that. I understood the original question to be
    "how should it be spec'ed" ?
    > Which it doesn't, at least in cases I'm familiar with.
    And (assuming this was a question about how to define the spec) that
    leaves us with the question of whether the existing loose practice is
    preferable for SQL.

    If escape codes are only used when entered (usually manually) into
    program source code, that's one type of use. If they are used instead to
    represent field data whenever the field character set is not Unicode,
    that would be a different type of use.

    Only in the latter would you run into scenarios where data could be
    transferred to/from a server and its semantics in UTF-32 could change.

    I thought that the original post contemplated uses like the second case.
    If, instead, all that was meant is source text for SQL statements, that
    could be less critical, I guess.

    So it depends - does anyone know?

    > Mark

    This archive was generated by hypermail 2.1.5 : Fri Apr 17 2009 - 00:19:15 CDT