LIRC produces multiple events from one keypress

[eladner]

lsb_release:
Description: Ubuntu 14.04.3 LTS
Release: 14.04

# apt-cache policy lirc
lirc:
  Installed: 0.9.0-0ubuntu1
  Candidate: 0.9.0-0ubuntu5
  Version table:
     0.9.0-0ubuntu5 0
        500 http://us.archive.ubuntu.com/ubuntu/ trusty/universe amd64 Packages
 *** 0.9.0-0ubuntu1 0
        100 /var/lib/dpkg/status

What happened:

After upgrading from 12.04 to 14.04.3, lirc 0.9.0-0ubuntu5 was generating multiple non-unique key press events. "irw" showed that the multiple key presses didn't have event numbers (second column of irw output). No hardware changed. Only an upgrade to 14.04.3.

No combination of min_repeat or suppress_repeats in lirc.conf would suppress the multiple events. No combination of repeat or delay options in the .lircrc file worked either.

This same configuration was working fine under 0.9.0-0ubuntu1 on a 12.04 box.

Sample IRW output:

# Note the zero event number in the second column.
0000000000001781 00 1 Hauppauge_350
0000000000001781 00 1 Hauppauge_350
0000000000001782 00 2 Hauppauge_350
0000000000001782 00 2 Hauppauge_350
0000000000001783 00 3 Hauppauge_350
0000000000001783 00 3 Hauppauge_350

I reverted the package back to 0.9.0-0ubuntu1 and the problem did not completely go away (there were occasional multiple events, but they were unique with an event number in the second field). Adding suppress_repeats to the lirc.conf fixed this issue easily.

One caveat is that I'm using an IguanaIR USB receiver wit the supplied igdaemon and the lirc iguanaIR driver blacklisted so the igdaemon can use libUSB to grab and talk to the receiver (iguana's recommended solution).

Apparently there is some regression between 0.9.0-0ubuntu1 and 0.9.0-0ubuntu5 that's not identifying unique events correctly.

What I expected:

LIRC would work with the config files from the previous machine (no other hardware changes other than an upgrade to 14.04.3).

[Alec Leamas (leamas-alec)]

Other users in similar situations has been able to solve the problem by using irrecord to make a new lircd.conf.

[Alec Leamas (leamas-alec)]

In order to advance this bug more testing is needed. As a first step, what happens if you use irrecord(1) to create a new lircd.conf and test that? For a test, just a 3-4 buttons needs to be recorded (but not less). If you do this, and use the generated lircd.conf, what are the results?

[Launchpad Janitor (janitor)]

[Expired for lirc (Ubuntu) because there has been no activity for 60 days.]