Re: [RFC] Non-normalizing Unicode Composition Awareness

Thomas Ã…kesson <thomas_at_akesson.cc> writes:

> If you, or someone else with WC insight, could provide some details on
> when/how conversions in the opposite direction is performed (e.g. svn
> stat and most commands taking path arguments), that would be
> incredibly useful to me. I would like to explore the option to somehow
> work around the "irreversible problem".

The Subversion libraries generally deal with UTF-8 paths exclusively.
When the use runs

   $ svn stat foo

the path 'foo' is typed in some local encoding. The program converts the
path from the local encoding to UTF-8 before passing it to the client or
working copy libraries. These library then treats the path as UTF-8 in
almost most cases. When making a system call such as stat() they pass
the UTF-8 path to the Subversion low-level IO functions in libsvn_subr.
These low-level functions convert the UTF-8 to the local encoding before
making the system calls.

-- 
uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com