Simon Willison wrote:
> I don't see any harm in META using unicode strings, whereas
> if it were to use bytestrings our documentation ends up being that
> little bit more confusing (we can't just claim everything is a
> unicode string).
We can't do it anyway. First example is "raw_post_data
On 19 Jul 2006, at 15:30, Ivan Sagalaev wrote:
> I just thought that may be they shouldn't. If META is a reflection of
> CGI's environment that is derived from HTTP environment that is
> essentially in byte strings then I think META being unicode is may be
> useless and misleading.
>
> Instead t
Simon Willison wrote:
> request.META should contain unicode strings that directly reflect the
> underlying raw bytestrings
I just thought that may be they shouldn't. If META is a reflection of
CGI's environment that is derived from HTTP environment that is
essentially in byte strings then I t
On 19 Jul 2006, at 14:37, Gábor Farkas wrote:
> 1. request.META should contain raw bytestrings (like it's currently)
> 2. request.META should contain unicode strings. for QUERY_STRING, we
> should convert it to unicode using the 'ascii' charset. and we should
> not url-decode it.
request.META s
Simon Willison wrote:
>
> On 19 Jul 2006, at 13:19, Ivan Sagalaev wrote:
>
>> Talking about QUERY_STRING... While the string itself is in ASCII
>> it has
>> urlencoded data and there the encoding matters. As fas as I can see in
>> practice browsers tend to encode those data in the same encodin
Simon Willison wrote:
> We shouldn't be decoding QUERY_STRING in request.META at all - we
> should leave it as urlencoded ASCII. request.META is meant to give
> you access to the 'raw data' from the browser.
>
> We do however need to take charset stuff in to account when creating
> the requ
Ivan Sagalaev wrote:
> Simon Willison wrote:
>> In the absence of anything better than that, I think it's safe to
>> assume that CGI environment variables should always be ASCII encoded.
>
> Talking about QUERY_STRING... While the string itself is in ASCII it has
> urlencoded data and there th
On 19 Jul 2006, at 13:19, Ivan Sagalaev wrote:
> Talking about QUERY_STRING... While the string itself is in ASCII
> it has
> urlencoded data and there the encoding matters. As fas as I can see in
> practice browsers tend to encode those data in the same encoding as
> the
> page from where t
Simon Willison wrote:
> In the absence of anything better than that, I think it's safe to
> assume that CGI environment variables should always be ASCII encoded.
Talking about QUERY_STRING... While the string itself is in ASCII it has
urlencoded data and there the encoding matters. As fas as I
Bill de hÓra wrote:
> gabor wrote:
>
>> questions:
>> 1. django publishes the whole environ dictionary as request.META. the
>> environ dictionary is a normal byte-string dictionary. so, should we
>> convert it to unicode so that the request.META dictionary only contains
>> unicode strings?
>>
gabor wrote:
> questions:
> 1. django publishes the whole environ dictionary as request.META. the
> environ dictionary is a normal byte-string dictionary. so, should we
> convert it to unicode so that the request.META dictionary only contains
> unicode strings?
>
> 1.a: if yes, how? some thin
On 18 Jul 2006, at 23:30, gabor wrote:
> 1. django publishes the whole environ dictionary as request.META. the
> environ dictionary is a normal byte-string dictionary. so, should we
> convert it to unicode so that the request.META dictionary only
> contains
> unicode strings?
>
> 1.a: if yes,
On Wed, 2006-07-19 at 00:30 +0200, gabor wrote:
> questions:
> 1. django publishes the whole environ dictionary as request.META. the
> environ dictionary is a normal byte-string dictionary. so, should we
> convert it to unicode so that the request.META dictionary only contains
> unicode strings
13 matches
Mail list logo
