[Mailman-i18n] Recode pipermail templates

[Barry A. Warsaw]

>>>>> "MvL" == Martin von Loewis <loewis@informatik.hu-berlin.de> writes:

    MvL> I made a patch

    MvL> http://sourceforge.net/tracker/index.php?func=detail&aid=634303&group_id=103&atid=300103

    MvL> which recodes the pipermail HTML templates and gettext
    MvL> messages if the charset of the page-under-creation is
    MvL> different from the charset of the templates. It uses HTML
    MvL> character references if the template characters are not
    MvL> supported in the target character set, so it should work fine
    MvL> if the template charset and the target charset are both known
    MvL> to Python. If the template charset and the target charset
    MvL> match, no conversion happens. If the message has no decared
    MvL> charset, but the templates use a non-ASCII charset, a META
    MvL> tag is now inserted into the page.

I posted a follow up to this patch about a simplification that I think
works just as well.  I've now committed all the changes to the
archiver than I think is necessary to get it to properly behave when
the charsets of the content is different than the charsets of the
template.

But this stuff always makes my brain hurt and I'm not 100% sure I've
gotten everything right, or everything that's needed.  Could someone:

- Please double check all the changes that are in the archiver in
  current cvs.  I think what's there will become beta5 unless
  something comes up in the next few hours.

- Are there any other changes to the archiver needed to resolve the
  i18n issues?  If so, please let me know.

- Is patch #634303 still necessary?  I think not, and I'd like to
  close it. 

Thanks,
-Barry