Patch for compilation errors with INN 2.3.2

Alex Kiernan alexk at demon.net
Fri Jul 20 08:30:31 UTC 2001



Roman Gollent <roman at pobox.com> writes:

> I came to the mistaken conclusion that nobody used forte C
> to compile inn since there was a problem getting configure
> to properly detect and link in the appropriate libraries
> when using it and the --with-openssl option (it would seem
> that Forte C is a bit pickier about the order in which the
> libraries are linked in).

I'd never tried with OpenSSL, I just don't have a need for it.

> The reason I thought/think that the compiler might be to blame is
> that I tried to reinitialising the history* files (as well as the
> cycbuffs used for CNFS) several times, through several recompiles
> (with the same configure options) using Forte C and never was able
> to get grephistory to give me the correct tokens.

If you can reproduce it simply & get me some history files, I'd love
to take a look. Certainly I'm not seeing it here as grephistory was
one of those things I did major surgery on to move it to the history
API.

> As soon as I recompiled (with exactly the same options) using gcc,
> the problem went away and everything is working as one would expect.
> I think we might be running an unpatched version of Forte C, when I
> get to work tomorrow I can find out if this is the case.

http://access1.sun.com/forte is the place to find them all in one
place (apologies if this is a comment from the egg sucking dept).

> It's always possible that the problem is somewhere between the
> keyboard and the chair, but I've taken a lot of time and made an
> effort to try to rule it out in this case.  :-)
> 

:-)

> I can set up another news box with the forte C compiled
> version of INN to reproduce the problem and then I will mail
> you the resulting history*.  How does that sound?
> 

Fantastic.

-- 
Alex Kiernan, Principal Engineer, Development, Thus PLC


More information about the inn-bugs mailing list