Re: [Libevent-users] bug + question

[Jeroen Habraken <vexocide@xxxxxxxxx>]

Hi,

On 5 January 2011 20:16, Nick Mathewson <nickm@xxxxxxxxxxxxx> wrote:
> On Thu, Dec 30, 2010 at 8:21 PM, Mihai Draghicioiu
> <mihai.draghicioiu@xxxxxxxxx> wrote:
>> Hi all. I've found that evhttp_parse_query_str only works if the query
>> string is of the form foo=bar&baz=quux. If we have a query string of
>> the form foo=bar&baz (no value for the last key), it fails. I consider
>> this a bug, because you have all the other valid parameters in there,
>> and it can just set NULL for the value of the baz parameter. I believe
>> the behavior of evhttp_parse_query() is different (correct). I had
>> used evhttp_parse_query() until now, and switched to
>> evhttp_parse_query_str() and discovered this bug.
>
> I'll repost my question from
> https://sourceforge.net/tracker/?func=detail&aid=3037662&group_id=50884&atid=461322
> :
>
> ===
> So, RFC3986 says that the query part of a URL can be any sequence of
> characters in a given set, without providing any instructions on how to
> encode or decode them.
>
> The only place that i can find the key=val&key2=val2 format expressed, on
> the other hand, is in the HTML standard, which doesn't define anything
> about parameters without values.
>
> So I'm inclined to say that this is not-a-bug: If the query parameters are
> not encoded in the HTML-standard way, then absent some other standard, then
> decoding the Request-URI from the Request-Line is up to the library user,
> right?
>
> Is there a standard I'm missing here?
> ===
>
> In other words, according to what standard is foo=bar&baz to be
> interpreted? ÂIf we're considering generic URIs, there is no
> particular interpretation to the query part. ÂIf we're considering GET
> requests generated by HTML forms, I think it's invalid, if I'm reading
> the HTML spec right. ÂSo according to what standard does foo=bar&baz
> have an interpretation?
>

None, unfortunately this has never been standardized. The URI RFC
merely states that the query-string is the part from ? to # as I'm
sure you know, but leaves the interpretation of it to the user.

On Wikipedia <http://en.wikipedia.org/wiki/Query_string> one finds:

* The query string is composed of a series of field-value pairs.
* The field-value pairs are each separated by an equals sign.
* The series of pairs is separated by the ampersand, '&' or semicolon, ';'.

which seems to be the most sane text on the subject (also note the
part on the encoding there).

>
[snip]
>
>
> yrs,
> --
> Nick

Regards,
Jeroen Habraken
***********************************************************************
To unsubscribe, send an e-mail to majordomo@xxxxxxxxxxxxx with
unsubscribe libevent-users    in the body.