patches (was: Re: flexfax: hfaxd exits)

[Matthias.Apitz@SOFTCON.de]

Nico Garcia wrote:
    
    > Anyway, my problem is that on my new RedHat 5.0 system hfaxd
    > exists as soon as I start it. In /var/log/messages I find
    > the following message:
    
    
    I've got a stack of patches for RedHat 5.0 Linux: Matthew has been
    graciously putting them up at the FTP site, and the originals are at
    http://cirl.meei.harvard.edu/~raoul/patches/

	....

Am I Matthew? :-)) Don't know but anyway....

Just to clarify the above mesaage a bit and also some other notes
about the ftp-servers and the patch policy at all:

At the moment I don't host the above mentioned patches in a
public area. The reason for this is that about some of
the patches Nico and I (and perhaps some others too) disagree;
and two or three of these patches will be in a future pl3 or must
be still verified on other systems at least.

In generally my policy is to keep *one* official HylaFAX version
and patches between these official versions should only be placed
in a public area if there are emergency problems to be fixed (like
the UCP problem). All other patches should be collected, tested and
made public in the next official version. The only reason for that
policy is to make it possible to track down problems that
others have. Imagine one sends a message to the list: "I applied
patch1, patch3 and patch7 of someone else and now it is
broken -- please help me." This isn't the way a good support
can be work.

Since pl2 we have now also ftp sites here in Germany
for HylaFAX:

ftp://ftp.leo.org/pub/comp/os/unix/networking/fax/hylafax/
ftp://ftp.sisis.de/pub/hylafax/

The latter one (ftp.sisis.de) will only host the sources and
does not host binary distributions (because of the bandwidth
we have at the moment).
    
    The published RPM for this had a number of things I agree with (such
    as renaming xferstats to xferfaxstats to avoid conflicts with the
    wu-ftpd configuration). It also had some incorrect or fairly careless
    things, such as naming the boot scripts "S80fax" instead of something
    like "S80hylafax", and requiring ghostscript fonts of at least
    version 5.10 when those are *identical* to the version 4.xx fonts.

Yes, this is also a good example for the reason of my policy
I explained above.
    
    So I suggest recompiling with the latest source, version 4.0pl2, and the
    patches I've set up to start with.
    
    > Feb 20 16:02:28 plum HylaFAX[11471]: getpeername: Socket operation on non-socket
    > 
    > strace produces the following:
    > 
    > getpeername(0, 0x8077d4c, [16])         = -1 ENOTSOCK (Socket operation on non-socket)
    > 
    > I'm using hylafax-4.0-8.i386.rpm but I had the same problem
    > when I compiled the source locally.
    
    I am not running into this problem under RedHat 5.0. Perhaps you have a
    previous copy of hfaxd running, a remnant of a previous installation?

No. The strace(1) output shows clearly, that the fd 0 (stdin)
is not a socket fd and the reason for that is explained in the
man page of hfaxd(1M).

	matthias
-- 
firm: matthias.apitz@sisis.de [voc:+49-89-61308-351, fax: +49-89-61308-188]
priv: guru@thias.muc.de
 PGP: Key fingerprint = 0C 01 F2 23 EC 17 A2 D5  46 2D 29 4C 0E 8B 7E 8F
 URL: http://www.sisis.de/~guru/         http://www.muc.de/~thias/

from USENET:
People who run servers understand that flashy interactive interfaces have
nothing to do with the underlying functionality and often get in the way.