Robin Smidsrød wrote:
Andy Wardley wrote:
Does that sound like a good plan to make the least number of people
displeased for as little time as possible?
Sounds good to me.
My only concern is that by using two filters with almost similar names,
but different behaviour might cause people to have weird problems
because of old documentation, and they will revert to the mailinglist
when things aren't working.
Make sure this is documented very clearly in the docs and the FAQ, so we
don't have to answer this question again and again.
Using another name as the "main" filter name for the old behaviour might
help making it more clearer, and maybe only make the url filter name a
synonym which is not used in the documentation. Maybe the filter name
should be 'uri_param' or something similar, to show the intent of the
(broken/old behaviour) filter more clearly.
Here's how javascript does it if that changes anything:
http://xkr.us/articles/javascript/encode-compare/
escape(), encodeURI(), encodeURIComponent()
I wouldn't be surprised if this was the basis for the old method.
encodeURI doesn't encode '/'s.
encodeURIComponent does.
I'd say this probably supports the old behavior of the filter.
I'm still neutral on this issue, though. Andy's solution for uri vs url seems
ok to me. Also the solution of making uri use the old method (for those that
need it) sounds good too.
-- Josh
|
