"Olaf Alders via RT" <bug-HTML-FormatExternal@rt.cpan.org> writes:
Show quoted text>
> I just had a chat with Dave Rolsky about this and his suggestion was
> that the commented code in the Zen formatter looks to be a good way to
> handle this:
The validate bit? Or convert? One thing I was wary of is I didn't want
to hard-code too much knowledge about what the respective programs could
or couldn't do, as you never know when they might grow etc.
Of course it's not helpful for the module to let bad things happen if
there's a way to do it right. One possibility on the input side would
be to entitize any non-ascii when unsure.
Show quoted text> Was there a particular use case for wanting a different output charset?
Yes, I use it that way. In my rss2leafode I get html from an rss feed
or http fetch in what charset the server gives, and I output utf-8
always for the resulting generated news message. (Could have left the
charset unchanged for the output perhaps, but I also intermingle little
bits of further text.)
I made a start enabling some wide bits. I propose to have wide input
make wide output by default, plus an output_wide=> option to force it.