Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Jc2k
    @Jc2k
    It’s too big for 1 pr though (every time...) so will need some faff to get merged
    Joachim Lusiardi
    @jlusiardi
    i release 0.13.0 (with BLE) today
    Jc2k
    @Jc2k
    That's really exciting news :D
    Jc2k
    @Jc2k
    i am still working on the same big HA patch. I wonder if I will ever be done. Have managed to get 1/4 of it upstream so far.
    really looking getting HA onto 0.13.0 and starting to roll out BLE though :D
    thank you for all your work on this!
    Joachim Lusiardi
    @jlusiardi
    No Problem πŸ˜‰
    Joachim Lusiardi
    @jlusiardi
    @Jc2k i saw your PR for tado. How can we be sure that changing something for device a does not break support for device b?
    Jc2k
    @Jc2k
    its a good question :-/ ideally we'd have a shelf of homekit devices connected to ci and be able to check but i certainly can't afford to not use my HomeKit devices xD
    in practice i think we need to aim for byte for byte compatibility with iOS clients
    Joachim Lusiardi
    @jlusiardi
    hehe same here. On the other hand, i don't expect Apple to build workarounds into there iOS clients...
    Jc2k
    @Jc2k
    if the ios client doesn't sent a request arg, we shouldn't
    Joachim Lusiardi
    @jlusiardi
    correct
    Jc2k
    @Jc2k
    so in the case of the most recent PR, an iPhone doesn't sent the =0 cases in my testing
    Joachim Lusiardi
    @jlusiardi
    ok
    Jc2k
    @Jc2k
    i guess another factor is the tado isn't some unknown brand either - it's Apple certified enough to be on their store:
    Joachim Lusiardi
    @jlusiardi
    agreed, it wasn't a specific issue with that PR more a general "how can we be sure"
    Jc2k
    @Jc2k
    yeah, its definitely a worry
    best i have right now is that going forward part of the PR is capturing logs from a real iOS device and homekit_python and the PR has to move things in the direction of byte-for-byte perfection to be merged.
    Jc2k
    @Jc2k
    we probably need a regression test harness too
    e.g. for previous PR i'd add one test that said pairing headers must equal exactly the ones we currently send otherwise we risk breaking device X
    seperate test for paired comms
    Joachim Lusiardi
    @jlusiardi
    at least some documentation why we did what?
    Jc2k
    @Jc2k
    yes
    ideally links to tickets
    Joachim Lusiardi
    @jlusiardi
    mh yes
    i'll be afk for the evening
    Jc2k
    @Jc2k
    that might even reference HA tickets - so we can try and track down an owner with one if we need them to test it
    ok
    not sure what my plans are yet - want to have a look at my Eve Energy for #125
    Joachim Lusiardi
    @jlusiardi
    What is also not the best way: we only know which devices we had to work on to support them. Not which one worked out of the box πŸ˜•
    Jc2k
    @Jc2k
    @jlusiardi are there any tickets you want to address before we do another release? (it would be good to be able to roll out the tado fixes in HA soon, and to get the pairing flow change added to HA too)
    Joachim Lusiardi
    @jlusiardi
    hi @Jc2k, i will have a look!
    what do you think about the last comment on jlusiardi/homekit_python#129
    Joachim Lusiardi
    @jlusiardi
    I think i can make a 0.14 tomorrow
    enough features are there
    ;)
    Jc2k
    @Jc2k
    \o/
    having a think about #129
    not sure what we can do, short of acquiring a Smartbar
    pcap would help potentially, i guess the failure is happening before the messages are encrypted
    Jc2k
    @Jc2k
    it dropping the connection after our M3 makes me think it could still be a difference between how an iPhone pairing looks vs how ours looks, rather than it being a different pairing system for software authentication vs MFi hardware auth.
    i know there is still at least one deviation from how apples implementation looks on the wire. looks like apple sends the bonjour service name rather than ip in Host header - but would be surprised if that tripped us up at the M3 and not the M1
    Joachim Lusiardi
    @jlusiardi
    It also looks strange to me how this message was communicated. The GitHub account does not seem official or β€žoldβ€œ
    It also would imply there is some difference how devices are paired.
    Joachim Lusiardi
    @jlusiardi
    @Jc2k release done, 0.14.0 is on pypi!
    πŸŽ‰ and stuff
    Jc2k
    @Jc2k
    :fireworks: yay, thank you
    Joachim Lusiardi
    @jlusiardi
    you're welcome ;)
    Joachim Lusiardi
    @jlusiardi
    @/all I started to list supported devices in https://github.com/jlusiardi/homekit_python/tree/add_device_reports/doc/tested_devices can you contribute yours please?