some INN bugs from RH bugzilla

[Jeffrey M. Vinocur]

On Dec 2, 2005, at 6:46 AM, Martin Stransky wrote:

> we have some bugs to inn and I'm not completely sure what should I do
> with them:

I'll try to answer, maybe Russ can chime in.


> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=174672

This is a complaint that an application that links against libinn and  
calls server_init() will have TMPDIR set.  The actual scenario in  
which this occurs is not provided, so I'm not sure it's a bug.  But  
we can discuss it...

Poking a little, I see that TMPDIR is set whenever inn.conf is  
parsed.  I'm not sure if that's necessarily correct, but perhaps an  
easier issue to tackle is whether server_init() needs to parse  
inn.conf at all.  Looking through all of lib/clientlib.c, I believe  
that the only thing that ever uses inn.conf is getserverbyfile() --  
so perhaps we could just defer the innconf_read() until then as a  
quick fix?


> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=149831

This is a complaint that XHDR NEWSGROUPS returns a faux,  
reconstructed header based on overview information.

This was fixed in INN 2.4.3.  You can suggest he upgrade.


> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=53948

This is a complaint that INN should not complain about duplicate Cc:  
headers.

As far as I know we're on shaky ground here with regard to having a  
specification to follow, but to my eye, we follow the latest USEFOR  
draft back to RFC 2822.  And RFC 2822 states quite clearly in section  
3.6 that the maximum number of Cc headers is 1.  Therefore I don't  
believe this is a bug, unless someone else feels otherwise?

-- 
Jeffrey M. Vinocur
jeff at litech.org