[sane-devel] segfault in sane_net_get_devicesconnectingto network saned

[m. allan noah]

why not try running saned under xinetd, and see if it plays nicer?

allan

On Tue, May 11, 2010 at 2:42 PM, Brian J. Murrell <brian at interlinx.bc.ca> wrote:
> m. allan noah <kitno455 <at> gmail.com> writes:
>>
>
> Hi,
>
>> Ok- let me first appologize if you took Julien the wrong way. You see,
>> any program that uses sane-backends is effectively pulling the
>> backends into it's running process, giving it all manner of options to
>> screw things up.
>
> Yes, understood, completely.
>
>> In the absence of source code, it can be very
>> difficult to identify the culprit.
>
> Indeed, and I do sympathize.  I really do wish I did not have a binary driver
> scanner, but I guess we all have to deal with our lot.
>
>> As such, it is very frustrating for
>> those of us that provide this free support to you, to hear complaints
>> about cost of such code. You just accidentally hit a nerve, i think :)
>
> OK.  I am happy to consider it all water under the bridge and move on.
>
>> 1. can you enable the test backend on the server, and disable the
>> brother backend in /etc/sane.d/dll.conf?
>
> To be sure it's completely disabled, I even moved the library modules out of
> the way.
>
>> lets see if that at least
>> shows up with scanimage -L on localhost and on the client.
>
> $ scanimage -L
> device `test:0' is a Noname frontend-tester virtual device
> device `test:1' is a Noname frontend-tester virtual device
> device `net:localhost:test:0' is a Noname frontend-tester virtual device
> device `net:localhost:test:1' is a Noname frontend-tester virtual device
>
> xsane scans from the local and network connected "test:0" devices just fine.
>
>> 2. can you run saned with the -d flag, and see what is going on when
>> brother is enabled?
>
> OK.  So this is pretty good.  I run "sudo /usr/sbin/saned -d -a saned" on a tty
> and then "scanimage -L" on another tty and get:
>
> device `brother3:bus1;dev1' is a Brother DCP-7040 USB scanner
> device `test:0' is a Noname frontend-tester virtual device
> device `test:1' is a Noname frontend-tester virtual device
> device `net:localhost:brother3:bus1;dev1' is a Brother DCP-7040 USB scanner
> device `net:localhost:test:0' is a Noname frontend-tester virtual device
> device `net:localhost:test:1' is a Noname frontend-tester virtual device
>
> Which is much better than any result I got with saned running as a daemon.
> xsane also comes up with the list of scanners which it didn't before.  So if we
> step backwards to get to where things break...
>
> Disabling the test scanner still doesn't yield the segfault:
>
> $ scanimage -L
> device `brother3:bus1;dev1' is a Brother DCP-7040 USB scanner
> device `net:localhost:brother3:bus1;dev1' is a Brother DCP-7040 USB scanner
>
> Ah ha!  But restarting saned as a daemon (i.e. service saned start) yields the
> segfault again.
>
> So...
>
> $ sudo /usr/sbin/saned -a saned
>
> segfaults scanimage
>
> $ sudo /usr/sbin/saned -d -a saned
>
> doesn't segfault scanimage
>
> Perhaps the problem will be immediately obvious for you, perhaps not.  Happy to
> gather more info if not.
>
> Cheers,
> b.
>
>
>
> --
> sane-devel mailing list: sane-devel at lists.alioth.debian.org
> http://lists.alioth.debian.org/mailman/listinfo/sane-devel
> Unsubscribe: Send mail with subject "unsubscribe your_password"
>             to sane-devel-request at lists.alioth.debian.org
>



-- 
"The truth is an offense, but not a sin"