On Wed, Nov 16, 2022 at 1:12 AM Kent Gibson <warthog618@xxxxxxxxx> wrote: > > On Tue, Nov 15, 2022 at 03:32:57PM +0100, Bartosz Golaszewski wrote: > > On Tue, Nov 15, 2022 at 4:44 AM Kent Gibson <warthog618@xxxxxxxxx> wrote: > > > > > > On Mon, Nov 14, 2022 at 11:12:25PM +0800, Kent Gibson wrote: > > > > On Mon, Nov 14, 2022 at 03:26:38PM +0100, Bartosz Golaszewski wrote: > > > > > On Mon, Nov 14, 2022 at 5:01 AM Kent Gibson <warthog618@xxxxxxxxx> wrote: > > > > > > > > > > > > > > > > I played with the tools a bit and really like the way they look now. I > > > > > think they're ready to hop into master, I'll do some more testing and > > > > > they should be in this week. Just one last request from my side: would > > > > > you mind updating the TOOLS section of the README? I'm aware it's not > > > > > yet updated for v2 and I plan to do it soon but we could already start > > > > > with the tools examples. You can send an incremental patch on top of > > > > > this series. > > > > > > > > > > > > > Good point - I totally forgot about the README. > > > > I'll take a look at it. > > > > > > > > > > And while collecting examples for the README I find the interactive mode > > > tab completion is now broken. It seems I broke it when fixing the prompt > > > behaviour when called from a script. > > > And when fixing that I notice some other odd tab completion behaviour - > > > seems there are at least two separate bugs in in_line_buffer(). > > > All things that gpio-tools-tests doesn't currently test for, so I > > > should extend the test coverage as well. > > > > > > And gpiomon --unquoted aborts as I forgot to cut-and-paste the parser > > > handler - and forgot to add a test for it too. > > > So another fix and more tests to add. > > > > > > I probably wont get that done for a day or two, so you might want to hold > > > off on merging - or I can always provide patches later. > > > > > > I also note that interactive mode does not support quoted line names on > > > the interactive command line yet. Adding that is sure to be entertaining. > > > Do you want to wait for me to patch that, or add it to the TODO list? > > > > > > > Yeah, I can wait. > > > > It seems that, as per the prompt, libedit doesn't provide tab completion > when stdin/stdout isn't a tty, so automated testing of it will require a > different approach such as a pty. Not sure how to do that from bash. > Unless I have an epiphany, or you have a better suggestion, for the moment > I'll just test it manually as it isn't critical functionality - it is just > a nice to have. > Sounds good, just add it to the TODO file in the repo. Bart > > While at it: can you move the following test case: > > > > @test "gpioget: with consumer" { > > gpiosim_chip sim0 num_lines=4 line_name=1:foo line_name=2:bar > > gpiosim_chip sim1 num_lines=8 line_name=3:baz line_name=4:xyz > > > > dut_run gpionotify --banner -F "%l %E %C" foo baz > > > > run_tool gpioget --consumer gpio-tools-tests foo baz > > status_is 0 > > > > dut_read > > output_regex_match "foo requested gpio-tools-tests" > > output_regex_match "baz requested gpio-tools-tests" > > } > > > > to after you actually add gpionotify? Otherwise the test suite fails > > and bisectability is lost. > > > > Well that is iritating. I could rework the test to not rely on > gpionotify - it could use the settling period to make gpioget hold the > line while gpioinfo checks - but it would be slower and uglier. > So I'll add that test with the gpionotify tests, as you suggest. > > Cheers, > Kent.