13:22 <re_irc> < (@barafael:matrix.org)> I'm using serial_port wrapped by linux-embedded-hal. First, it didn't work because I didn't do the particular "echo 12 > /sys/class/gpio/export". When I did, it worked nicely :) However, these exports are not persisted over reboots etc. Is this by design of linux, should I do this programmatically, or just set it by shell script and leave it to the "integrator"?

13:41 <re_irc> < (@barafael:matrix.org)> I figured it out... Need to put the settings in/etc/sysfs.d/local-gpio.conf, then they keep being reapplied. Nice. Though I'll totally forget this. Even if I put it in the readme.

14:10 <re_irc> < (@sirhcel:matrix.org)> : At first, i'm sorry that i'm still late with your pr. Sorry! To my understanding it works like that: the kernel does not care about persisting changes made at runtime and you have to apply your delta to its defaults on every startup. Either through software (as you did) or by startup scripts.

14:11 <re_irc> < (@sirhcel:matrix.org)> A lot of applications ship with fragments for udev or startup scripts and integrating the setup into the system is up to the user.

14:17 <re_irc> < (@barafael:matrix.org)> : not sure which PR you mean! Maybe a misunderstanding? Thanks for the clarification! I am including an example config file for sysfsutils.

14:24 <re_irc> < (@sirhcel:matrix.org)> You ar right, i confused your work with embedded-hal serial traits with this pr for serialport-rs (https://github.com/serialport/serialport-rs/pull/59) from . I'm still under some heavy work load and did not manage to complete and integrate this pr. My bad feelings for that seem to start hauning me ... 😅

14:58 <re_irc> < (@dbrgn:matrix.coredump.ch)> is there a way to implement both embedded_hal::serial::Error and embedded_hal_nb::serial::Error for the same type:

14:58 <re_irc> error[E0119]: conflicting implementations of trait `embedded_hal_nb::serial::Error` for type `SerialError`

14:58 <re_irc> < (@dbrgn:matrix.coredump.ch)> * type?

14:59 <re_irc> < (@sirhcel:matrix.org)> * haunting

15:09 <re_irc> < (@dbrgn:matrix.coredump.ch)> hmm, I tried using separate error types for blocking and nb implementations, but then it clashes at the implementation of "ErrorType"... I guess separate wrapper types would be required for blocking and non-blocking serialport implementations? that seems a bit unfortunate :/

15:31 <re_irc> < (@adamgreig:matrix.org)> embedded_hal_nb::serial::Error is a pub re-export of embedded_hal::serial::Error

15:32 <re_irc> < (@adamgreig:matrix.org)> so if you just implement embedded_hal::serial::Error hopefully that works?

16:44 <re_irc> < (@dbrgn:matrix.coredump.ch)> : aaaah, that explains it! thanks.

19:59 <re_irc> < (@9names:matrix.org)> : Hmm, why do you need sysfs gpio for serial_port?

22:07 <re_irc> < (@barafael:matrix.org)> : good point, I was confused at that moment. I am using the serial port, but my question is solely about sysfs gpios :D

22:12 <re_irc> < (@9names:matrix.org)> Okay :)

22:12 <re_irc> If you haven't got a good reason to use sysfs I recommend you try ti use the character device gpio interface instead - this is cdev in the Linux embedded hal.

22:12 <re_irc> The sysfs interface is slow and is hard to handle errors with

22:20 <re_irc> < (@barafael:matrix.org)> : my main goal was just to get rid of rppal in order to support any Linux board. This is the first time in using gpios on something that isn't a microcontroller, so I really appreciate your suggestion!

22:21 <re_irc> < (@barafael:matrix.org)> * 9names: my main goal was just to get rid of rppal in order to support any Linux board(not just RPI).

22:21 <re_irc> < (@barafael:matrix.org)> * I'm

23:40 <re_irc> <bortolotti> hi everyone

23:41 <re_irc> <bortolotti> this is my 1st time using this platform (element)

23:41 <re_irc> <bortolotti> what would be some other interesting communities for those who are interested in programming (not only embedded)?

23:42 <re_irc> <bortolotti> plus, apparently, there's a linux terminal version of this platform - does anyone here use it?