Lirc import: remotes with and without timing information; lirccode vs mode2

[bengtmartensson]

Import of Lirc remotes

Background, the two types of Lirc remotes

There are two types of Lirc remotes: those with timing information and those without, sometimes called mode2 and lirccode.

The Lirc files (in general) contain the keywords one and zero. It may look like

one           642   470
zero          642  1600

(from https://sourceforge.net/p/lirc-remotes/code/ci/master/tree/remotes/yamaha/RX-V995.lircd.conf.)

These state how ones and zeros in the payload are encoded: first the (modulated) IR is turned on for a certain time (here 642 micro seconds), then off (here 470 or 1600 micro seconds). Instead, timingless remotes may say

  one             0     0
  zero            0     0

Of course, these "timings" make no sense, and such a remote is strictly speaking meaningless, Instead, they are dependent on the used driver, in general listed as the driver keyword, see for example https://sourceforge.net/p/lirc-remotes/code/ci/master/tree/remotes/atiusb/atilibusb.lircd.conf . Such a remote can in general be used only with that particular driver, which knows the actual timings used in the signals.

There are also raw remotes, where each command has its own sequence of durations in micro seconds, see https://sourceforge.net/p/lirc-remotes/code/ci/master/tree/remotes/zenith/ZN110.lircd.conf for an example. These, of course, contain timing information.

IrScrutinizer

IrScrutinizer has very advanced capabilities for importing Lirc remotes with timing information. But, by necessity, it fails on Lirc remotes without timing information.

Option "Reject Lirc imports without timing"

IrScrutinizer 2.4.0 contains an option "Reject Lirc imports without timing" (under Options/Import options), which defaults to true. This is a very badly chosen name (sounds like "reject entries without valid signature"). It suggests that there is a silly test that rejects otherwise sane inputs, and clever people can, or should, turn that test off. This is not how it works. Instead, commands were generated with the driver name as dummy protocol name, and the Lirc command code as a parameter, called lirc, like

(from #520). This is pretty useless as-is, and requires quite non-trivial fixes to be useful.

For this reason, I consider the option as more confusing than useful, and the option will be removed in IrScrutinizer 2.4.1. Feedback welcome.