GitLab

So that we don't regress

Kunal Gangakhedkar authored Sep 08, 2010 and Bastien Nocera committed Sep 08, 2010

The libfprint library is generally built with C calling conventions.
Which makes it difficult to link to C++ programs as it is.

This patch adds the support for linking with C++ code - by telling the C++
compiler to use C calling conventions for libfprint functions.
Reported-by: Guus Ellenkamp <guus@activediscovery.net>
Signed-off-by: Kunal Gangakhedkar <kunal.gangakhedkar@gmail.com>

Hugo Grostabussiat authored Aug 31, 2010 and Bastien Nocera committed Aug 31, 2010

Inserted "_2016" in structures, enums and functions names
that are related with device 147e:2016.

Hugo Grostabussiat authored Aug 31, 2010 and Bastien Nocera committed Aug 31, 2010

https://bugs.freedesktop.org/show_bug.cgi?id=29719

HAL is deprecated, and the rules are broken when multiple
drivers use the same USB IDs.

So that you can compile against ImageMagick even if the
distter doesn't.

At the end of the configure run.

As OpenSSL cannot really be used in an LGPLv2+ piece of software
without an exception.

Adapted from the example code at:
http://www.mozilla.org/projects/security/pki/nss/sample-code/sample2.html
and:
http://github.com/bagder/curl/commit/f3b77e5

Alexia Death authored Dec 26, 2009 and Bastien Nocera committed Aug 19, 2010

Don't consider the scan complete unless theres at least
MIN_ROWS recorded or very long blank read occurred.

Typical problem spot: one brief touch before starting the
actual scan. Happens most commonly if scan is started from
before the first joint resulting in a gap after the inital touch.

http://lists.reactivated.net/pipermail/fprint/2009-December/001406.html

Alexia Death authored Dec 25, 2009 and Bastien Nocera committed Aug 19, 2010

Rather than when reading the rows. Doing it there seems a lot
more sensible and does not get in the way of minimizing the
impact of lost USB packets.

http://lists.reactivated.net/pipermail/fprint/2009-December/001404.html

Alexia Death authored Dec 24, 2009 and Bastien Nocera committed Aug 19, 2010

If a packet goes missing, use the data from previous row
to minimize distortion, or use some neutral dummy data.

http://lists.reactivated.net/pipermail/fprint/2009-December/001404.html

Alexia Death authored Dec 26, 2009 and Bastien Nocera committed Aug 19, 2010

This segfault was caused by proceeding to free the conversion result
without checking for an error condition in image to print
data conversion.

http://lists.reactivated.net/pipermail/fprint/2009-December/001405.html

upeksonly handles revision 1 devices, upeke2 revision 2 devices.

https://bugzilla.redhat.com/show_bug.cgi?id=504399

Check all the drivers for one that'll drive our device in
question, and prefer ones that have a discover() method that runs
successfully to a driver without such a method.

This allow drivers to both reject devices (if 2 drivers handle
the same USB IDs), or a single driver to handle variants of the
same device (through the devtype out value of the discover method).

Jorge Suarez de Lis authored Aug 17, 2010 and Bastien Nocera committed Aug 17, 2010

As reported working in:
http://www.mail-archive.com/fprint@reactivated.net/msg01426.html

Jorge Suarez de Lis authored Aug 17, 2010 and Bastien Nocera committed Aug 17, 2010

The device driver is always disabled as it uses the same Device ID
as another working reader:
https://bugzilla.redhat.com/show_bug.cgi?id=504399#c38

Guido Grazioli authored Aug 17, 2010 and Bastien Nocera committed Aug 17, 2010

Ported to the new internal libfprint API.

https://bugzilla.redhat.com/show_bug.cgi?id=499732

Matthew Garrett authored Oct 01, 2009 and Bastien Nocera committed Aug 17, 2010

Most fingerprint readers can be suspended safely, so use
the udev rules to autosuspend them when they appear.

The script itself contains whitelists and blacklists.

When imaging support is required, we prefer to use gdk-pixbuf,
as it's already on things like Live CDs.
Also fix the examples building against the system libfprint.

Bastien Nocera authored Nov 25, 2008 and Daniel Drake committed Nov 25, 2008

It was increasing i too early, and writing past the allocated
structures.

Bastien Nocera authored Nov 25, 2008 and Daniel Drake committed Nov 25, 2008

One less warning.

Bastien Nocera authored Nov 24, 2008 and Daniel Drake committed Nov 24, 2008

Also export the scan type through HAL.

Alon Bar-Lev's device behaves differently, but seems to work fine with
upekts even if we ignore the difference.

Don't log anything by default, and let the applications change the log
level.

Bastien Nocera authored Nov 20, 2008 and Daniel Drake committed Nov 20, 2008

For each driver, export the scan type supported by the devices it
handles. This allows front-ends to direct the user better.

because Bastien asked nicely. No guarantees about API or stability.

Reported to work with Motion Computing M1400 tablet by Daniel Hazelbaker
and Alex Van Deusen.

Bastien Nocera authored Oct 30, 2008 and Daniel Drake committed Nov 02, 2008

Tag all the fprint supported devices in HAL. Patch from
myself and Timo Hoenig <thoenig@suse.de>.

Pointed out by Bastien Nocera

Cleans up the conditional compilation system

While we're a noisy library we might as well ask libusb to be as well

The long awaited 147e:2016 reverse engineering efforts.
Image quality is OK, not brilliant - a little fuzzy, but it works
quite well.