Daniel Shahaf wrote on Thu, Apr 26, 2012 at 18:36:54 +0300: > Without reading the context: Philip Martin posted just this week with > a detailed explanation of when non-ASCII passwords would (and wouldn't) > work over http.
(The bottom line was that the input on the client must result in the same bytes the server expects to see --- but the input on the client depends on locale. (Also, he didn't touch on that issue there, but I'm quite sure that the infamous NFC/NFD issue applies here too; namely, é has two encodings in UTF-8 and the client and server must agree on which one to use...)) > > Nico Kadel-Garcia wrote on Thu, Apr 26, 2012 at 08:12:27 -0400: > > On Thu, Apr 26, 2012 at 7:36 AM, Cooke, Mark <mark.co...@siemens.com> wrote: > > > > > Hi Steve, > > > > > > Glad to have been useful (and to know it is not just me)! > > > > > > Unfortunately I have not managed to get to the bottom of how or why > > > subversion <-> httpd is mangling the password and what (if anything) > > > subversion (or neon or serf?) could change to fix this... > > > > > > ~ mark c > > > > > As I remember, not much. Passing non-ASCII characters through any kind of > > password handling tool is a problem, for a lot of reasons.
