|
|
Posted by Michael Winter on 09/26/06 16:49
Jonathan N. Little wrote:
[snip]
> URLs are not *always* case-sensitive
Unless explicitly defined otherwise by the scheme, most URI components
are assumed to case-sensitive. The scheme and host components are a
special case. See section 6.2.2.1 in RFC 3986.
> and case-sensitivity *is* platform dependent.
Whether the mapping of URIs to the file system is case-sensitive or not
is platform-dependent, yes, but that's irrelevant to whether URIs
themselves are case-sensitive.
[snip]
> I have yet to come across a IIS that won't parse:
>
> http://www.example.com/somepath/somefile.ext
> http://www.example.com/SOMEPATH/SOMEFILE.EXT
> http://www.example.com/somepath/somefile.EXT
> http://www.example.com/SOMEpath/somefile.ext
> http://www.example.com/sOmEpAtH/sOmEfIlE.eXt
>
> as equivalent so I would argue that case-sensitivity is server
> platform dependent.
Check your browser cache. You should find separate entries for each one
(I do).
It doesn't matter that the server returns precisely the same response:
even after case normalisation (which usually affects only the scheme,
host, and percent-encoding), the URIs will be considered different when
a comparison is made because they /are/ different (certainly so under HTTP).
> I would say building a Linux box would be preferable because it would
> force you to be case-aware ...
I would recommend a much simpler approach: make sure that all resources
use lower-case names, and write lower-case URIs.
Mike
Navigation:
[Reply to this message]
|