I've incorporated your suggested language.
Thanks,
Marc.
On Jul 13, 2008, at 7:41 PM, Manger, James H wrote:
> I was wrong a fortnight ago when I though any sensible @Path value
> with encode=false could be re-written with encode=true.
> A @Path value does need encode=false when it includes:
> * a %-escaped <reserved> character (eg a ‘/’ that is NOT a path
> segment separator); or
> * a %-escaped ‘{‘ or ‘}’ (so it is not misinterpreted as a {name}
> placeholder).
>
> Example: @Path(value=”/math/pi=22%2F7”, encode=false)
>
>
> When encode=true I assume ‘?’, ‘#’ and ‘%’ characters in a @Path
> value are encoded (as %3F, %23 and %25) as they are not allowed
> (alone) in a path, despite being “valid URI characters”. Instead of
> talking about “valid URI characters”, the Path#encode javadoc could
> say:
> “If true, any percent characters or characters that are not valid
> in a URI path will be automatically encoded.”
>
>
> James Manger
>
---
Marc Hadley <marc.hadley at sun.com>
CTO Office, Sun Microsystems.