Forwarding it to the list so people don't bother to debug themselves.
It's tested with cifs and a local change which enables upas/fs to do
IMAP with NTLM auth.
---------- Forwarded message ----------
From: Yaroslav <yari...@gmail.com>
Date: 2012/3/30
Subject: Re: [9fans] SSHv2
To: erik quanstrom <quans...@quanstro.net>
2012/3/30 erik quanstrom <quans...@quanstro.net>:
> On Fri Mar 30 09:48:50 EDT 2012, yari...@gmail.com wrote:
>> 2012/3/30 erik quanstrom <quans...@quanstro.net>:
>> > what's the basis for this claim? it might be broken, since we don't use it
>> > much, but it's not missing.
>>
>> Yes, sorry, I've meant it's broken; details follow.
>
> try pulling a new copy.
>
> - erik
There are phases mismatch between factotum's chapconv() and libauth's
auth_respond(). I've checked plan9port - it has identical problem.
Patch which fixes it follows:
term% diff -c /n/sources/contrib/quanstro/root/sys/src/cmd/auth/factotum/chap.c
/sys/src/cmd/auth/factotum/chap.c
/n/sources/contrib/quanstro/root/sys/src/cmd/auth/factotum/chap.c:120,125
- /sys/src/cmd/auth/factotum/chap.c:120,126
ret = -1;
chal = nil;
+ res = nil;
k = nil;
attr = c->attr;
/n/sources/contrib/quanstro/root/sys/src/cmd/auth/factotum/chap.c:153,166
- /sys/src/cmd/auth/factotum/chap.c:154,163
c->state = "read challenge";
if((nchal = convreadm(c, (char**)(void*)&chal)) < 0)
goto out;
- if(astype == AuthMSchap && nchal != MSchallen)
- c->state = "write user";
if((user = strfindattr(k->attr, "user")) == nil){
werrstr("key has no user (cannot happen?)");
goto out;
}
- if(convprint(c, "%s", user) < 0)
- goto out;
c->state = "write response";
if((pw = strfindattr(k->privattr, "!password")) == nil){
--
- Yaroslav
