This is a nasty first-cut patch to add proper character set support to CGI::Wiki, as per my anguished rambling in #openguides today. It's nasty. But then, frankly, CGI::Wiki could do with a bit of database abstraction...
good things about it: I don't think it'll change anything if you don't use it. The default is iso-8895-1, which is the effective default if you don't say anything else. Search::InvertedIndex is broken and won't index stuff with the utf-8 flag set, so I index the raw bytes, again, this means that current live data shouldn't be affected. Not that I've tried this.
bad things about it: apart from the ugliness? It requires perl 5.8. This is a very bad thing, but there are ways of getting around it. New version soon. There's a work-around in the patch for HTML::Parser's uselessness as well, this can go away when http://rt.cpan.org/NoAuth/Bug.html?id=7014 is landed.
I have an example living at http://movieos.org/wiki/wiki.cgi - this is a patched CGI::Wiki and a CGI::Wiki::Kwiki patched to use it. You'll need a decent font. :-)
tom
On Fri 03 Sep 2004, Tom Insam tom@jerakeen.org wrote:
This is a nasty first-cut patch to add proper character set support to CGI::Wiki, as per my anguished rambling in #openguides today.
OK - nothing is going to get done about this until I get back from holiday in a week, so if you want to tidy it up/add anything else/abstract it a bit more, you have until then. *Mail me again no sooner than a week on Monday*.
bad things about it: apart from the ugliness? It requires perl 5.8.
Yes, you need to fix that :)
Kake
On Sep 3, 2004, at 18:44, Kake L Pugh wrote:
On Fri 03 Sep 2004, Tom Insam tom@jerakeen.org wrote:
This is a nasty first-cut patch to add proper character set support to CGI::Wiki, as per my anguished rambling in #openguides today.
OK - nothing is going to get done about this until I get back from holiday in a week, so if you want to tidy it up/add anything else/abstract it a bit more, you have until then. *Mail me again no sooner than a week on Monday*.
bad things about it: apart from the ugliness? It requires perl 5.8.
Yes, you need to fix that :)
And I've fixed it. Attached is a patch to add charset support to CGI::Wiki that falls back to the old behaviour if you are running under 5.6.
The current CVS CGI::Wiki::Kwiki takes advantage of this, and default new C::W::K installs will be full utf-8 wikis. Imports from CGI::Kwiki wikis that have used nasty utf8 chars seem to work, and you can put high-bit stuff into pages.
tom
On Mon 13 Sep 2004, Tom Insam tom@jerakeen.org wrote:
And I've fixed it. Attached is a patch to add charset support to CGI::Wiki that falls back to the old behaviour if you are running under 5.6.
Applied, released as 0.56. Note that this charset support is completely untested - tests welcomed.
The current CVS CGI::Wiki::Kwiki takes advantage of this, and default new C::W::K installs will be full utf-8 wikis. Imports from CGI::Kwiki wikis that have used nasty utf8 chars seem to work, and you can put high-bit stuff into pages.
Hurrah, release it!
Also new in 0.56 is a Plucene search backend. Please try it out - it promises to be much better than the Search::InvertedIndex one.
Kake