The encoding that it's set to use is utf-8. Do people have browsers
out there that would be sending data encoded in something other than
utf-8?
This value is still changeable in the config file. It just seemed that
the current default of not decoding the input is not a desirable
default.
Kevin
--
Kevin Dangoor
Author of the Zesty News RSS newsreader
email: k...@blazingthings.com
company: http://www.BlazingThings.com
blog: http://www.BlueSkyOnMars.com
> The encoding that it's set to use is utf-8. Do people have browsers
> out there that would be sending data encoded in something other than
> utf-8?
IIRC, the encoding used to send data should be the same that is used to
receive data. So, if someone configures it's code to use, e.g., iso-8859-1
then the browser will be expecting iso-8859-1 data and will be sending
iso-8859-1 as well. The same might happen for more esoteric encodings such as
the ones used by Windows (cp*).
> This value is still changeable in the config file. It just seemed that
> the current default of not decoding the input is not a desirable
> default.
Agreed. I'd go a step further and configure it to use the same value as
kid.encoding = "utf-8" in app.cfg. This ties the configuration and makes it
consistent between TG and Kid.
--
Jorge Godoy <jgo...@gmail.com>
I think you're right about the browsers working that way. I've changed
it to use kid.encoding.
Kevin
I think we can go one _more_ step further and set default content-type,
namely:
if decoding_filter is on and tg.content_type not set, set default
content-type to "text/html; charset=" kid.encoding.
That's a good idea. Would you mind opening a ticket to do that so that
we're sure it gets done?
Kevin
...which would work great if the server always called the client first.
;) There's quite a bit of good research baked into CP's decodingfilter,
to fall back to required/reasonable guesses when the charset is not
specified in the Content-Type header.
Robert Brewer
System Architect
Amor Ministries
fuma...@amor.org