Re: Help! Perl unicode & Anomy? (was: Question on failures on testall)

From: Paul Wallingford (
Date: Mon 19 May 2003 - 10:41:26 GMT

  • Next message: "failure notice"

    > Hi all!
    > This is a request for help - it seems that perl 5.8.x
    > breaks Anomy in certain cases when handling 8bit data,
    > because of the Unicode support.
    > I would very much appreciate it if someone out there
    > with 5.8.x and reasonable perl experience could let me know
    > if:
    > a) the test cases fail and
    > b) if it is there is some simple way to
    > tell perl 5.8 to revert to 5.6
    > Unicode
    > behavior?
    > Any feedback would be most appreciated.
    > Ron - this appears to be what is causing your problem, I
    > don't have a solution at the moment, hopefully someone on
    > the list can help?
    > On 2003-05-05, 10:36:29 (-0400), Ron
    > wrote:
    > [snip]

    Try the following:

    1) It sounds like Ron is using RedHat. Other Linux / Unix distros may
    suffer from this also. Check the /etc/sysconfig/i18n file. It was
    changed in RedHat 8 and higher and broke a bunch of stuff. This policy
    of breaking existing software or dumping support for older software in
    favor of beta or alpha code is why I no longer use RedHat (see mod_perl
    for Apache 2 which is not even done yet).

    The i18n file includes a new utf8 LANG definition. This wreaks havoc
    with a lot of Perl scripts. I changed mine to LANG="en_US", rebooted
    and everything started working again.

    2) Since you will probably have to support utf8 mail, you need to be
    careful to not exclude it. Careful use of the "use bytes;" and "use
    utf8;" pragmas can alleviate some of the problems described above.

    hosted by