[opus] Fwd: Re: opusfile 0.4 release

[Ralph Giles]

I don't know. Derf?

 -r


-------- Original Message --------
Subject: 	Re: [opus] opusfile 0.4 release
Date: 	Sat, 24 Aug 2013 14:20:35 -0700
From: 	alpha thinktink



With my old streaming thread I am able to retrieve a unicode file from a
web server by converting the resource name to UTF-8 MBCS then uri
encoding it before sending it to the server with "GET /[resourcename]
HTTP/1.0".  For example, retrieving a file named "Yūji Ōno - The
Psychogun [libopus 1.0.2].opus" the request resource string looks like
this (taken from my local web server's access log):
"Y%C5%ABji%20%C5%8Cno%20-%20The%20Psychogun%20%5Blibopus%201.0.2%5D.opus"

Now when trying to open a unicode file with op_open_url(...) the access
log is not showing an access attempt, the function call returns too fast
and the error is set to -129 (OP_EFAULT).  I am converting the wchar_t
to a MBCS UTF-8 string (using ::WideCharToMultiByte(CP_UTF8,...) like I
do for my old stream thread) before passing it on to op_open_url with
the thinking that it would process it the same way I do.

Am I doing it wrong or is it a limitation in libopusurl?