#189 FTDI Regression since duty_cycle support

[William Manley]

The recent commit c9afda9 added support for duty_cycle to the ftdi driver.
This worked for lircd config files that actually specified duty_cycle
but broke it for config files that didn't as it would default to 0 rather
than 50.

Before this commit it was the responsibility of each driver that implements
duty_cycle to know about this default. This fix centralises the defaults
in config_file.c, simplifying the drivers. As a result, even though this
patch is to fix a bug in ftdi it doesn't actually touch ftdi.c.

This issue didn't come up in testing c9afda9 because I'd only tested it
with config files that actually explicitly set duty_cycle. I've tested this
fix with both, but only with the ftdi driver.

[Alec Leamas]

hm... First of all: good catch!

That said, I'm not enterely happy with this patch. For better or worse, the overall convention in lircd.conf and the remote struct is that the value 0 reflects the default, unconfigured value. Furthermore, the lircd.conf manpage explicitly states the the configured value is between 1 and 100 i. e. , not 0.

I think it would be better if you kept the current handling in config_file.c and friends and instead parsed the value 0 in ftdix.c as 'unconfigured' i, e., 50 in the same way as the default driver. Some more text in lirc.conf.5 on this issue would of course also be great.

[William Manley]

I think it would be better if you kept the current handling in config_file.c and friends and instead parsed the value 0 in ftdix.c as 'unconfigured' i, e., 50 in the same way as the default driver.

Ok. I figured that duty_cycle was like freq in that it is a property of the carrier, thus should always be set, but I'm happy to change it. I still like the idea of centralising the default however so I'll make an accessor function in ir_remote.h if that would be acceptable?

[Alec Leamas]

An accessor providing the default value if it's set to 0? Looks reasonable. But if you do, please also update the default.c driver so it's used in a consistent manner.

[William Manley]

Here's a patch implementing this with an accessor. Please don't apply it yet, I won't have a chance to test it until tomorrow.

[David Röthlisberger]

I have tested Will's patch second patch ("0001-plugins-ftdi-Specify-duty_cycle-to-fix-ftdi-driver.patch" in the previous comment) successfully will all the following variations:

• ftdi driver with duty_cycle specified in the config file
• ftdi driver with duty_cycle not specified
• ftdix driver with duty_cycle specified
• ftdix driver with duty_cycle not specified
• default driver with duty_cycle specified
• default driver with duty_cycle not specified

[William Manley]

Thanks Dave

[Alec Leamas]

Thanks to both of you for a nice piece of work! Committed as [f164c8]