- From: Eric Sedlar <esedlar@us.oracle.com>
- Date: Fri, 3 Mar 2000 10:32:34 -0800
- To: <jamsden@us.ibm.com>, <w3c-dist-auth@w3.org>
One possibility we're toying with is to store a boolean property with a collection resource indicating whether or not its bindings are case-sensitive or case-insensitive. For example, it might be the case that part of your hierarchical namespace is implementing Windows filesystem semantics, which require case-insensitive matching (although they are case-preserving), and another part implementing UNIX filesystem semantics. Otherwise, there will always be some loss of information when moving data in and out of the WebDAV server. --Eric ----- Original Message ----- From: <jamsden@us.ibm.com> To: <w3c-dist-auth@w3.org> Sent: Friday, March 03, 2000 4:09 AM Subject: Re: Case sensitive names and authoring Greg, This is great. I didn't have an HTTP spec in front of me when I wrote the note (need to get one by the bed I guess). We're probably running into the MSFT problem. Maybe this doesn't happen on Windows2000 which seems to leave you file names alone a little better. I'll investigate. On Thu, 2 Mar 2000 jamsden@us.ibm.com wrote: > > HTTP URLs aren't case sensitive, RFC 2616, S3.2.3 states "When comparing two URIs to decide if they match or not, a client SHOULD use a case-sensitive octet-by-octet comparison of the entire URIs, with these exceptions: ..." In other words: they *are* case-sensitive. Apache certainly treats them that way. Try these two URLs: http://www.webdav.org/other/proxy.html http://www.webdav.org/other/Proxy.html You'll find the latter fails. Note that HTTP servers running on an MSFT platform typically treat URLs in a case-insensitive fashion (due to the filesystem being insensitive). > but some resources are. For example, Java > source managed by a WebDAV server must retain the case in the package and > class names or the code won't compile. mod_dav foldes all resource names to > lower case, so .java files on the server won't compile. WHAT?!! mod_dav does NOT do any case-folding. Euh... what made you think it does? > It seems that WebDAV could, like HTTP, be case insensitive while at the > same time require that the server retain the case of resource URLs. That > is, the URL segments in the bindings would retain the case in the target > when it was entered, but access to those resources would be case > insensitive. If this is not the case, WebDAV can't be used to store Java. > That would be a real bummer! > > Would the proposal above be acceptable, and in the spirit of HTTP? No and no. (IMO) I certainly do not want to take the performance hit to start doing case-insensitive work in mod_dav and Apache. I am in great favor of continuing to operate in a case-sensitive fashion, and I believe the relevant RFCs encourage that, too. Cheers, -g -- Greg Stein, http://www.lyra.org/
Received on Friday, 3 March 2000 13:34:26 UTC