Hello Julien,
On Sun, 21 Nov 2010, Julien BLACHE wrote:
Tomas Pospisek <[email protected]> wrote:
Hi,
When looking at the log, the difference seems to be that both
"Preview" and the ``immediate'' "Scan" have an event flow like this:
[...]
Notice the "Setting of the non-blocking mode" changing it's place in
the event sequence order.
The log is not synchronized between the threads/processes, so that
doesn't actually mean much.
There were 2 "fixes" made to the pixma backend for pthread builds before
1.0.21, so maybe that should be looked into:
13641ac5b3e50f275da6bf39ede59da55edfae93
I undid that patch, still the same problem.
eb66f734baa42619ece201d94852247788215a35
I undid that patch as well, still the same problem.
Alternatively, you could try a build of sane-backends without pthread
and see if that build works better.
I did that too, same problem.
Also, I tried to compile with eb66f734baa42619ece201d94852247788215a35
reversed and with "--disable-pthread" (i.e. without threads) at the same
time and still had the same problem.
So no, sorry, that path led to no success here.
I tried to install Ubuntu Lucid's libsane (V 1.0.20-13ubuntu2)[1] (I had
to use --force-all because it conflicted with newer udev) ... and scanning
worked again properly!!!
Then I tried Ubuntu's newest 1.0.21-2ubuntu2 and scanning under xsane
stopped working with the same problem as before.
So I'd say it seems to be a problem with libsane and that the problem was
introduced between 1.0.20-13ubuntu2 and 1.0.21-2ubuntu2 aka
1.0.21-7(debian).
Should I take this to the sane devel list? Or try to compile out of git?
If I copy debian/* over to the git checkout, then should a
"dpkg-buildpackage" work from there?
Do you have any ideas what would be the most efficient and helpful way to
proceed?
?
*t
[1] http://packages.ubuntu.com/lucid/libsane
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
