https://www.dendron.so/ logo
Join Discord
Powered by
# dev
  • l

    lukecarrier

    12/30/2020, 5:35 AM
    I moved jobs without taking my holiday at my former employer so I'm pretty grateful for this week πŸ™ˆ
  • k

    kevins8

    12/30/2020, 5:45 AM
    no cold start at the new gig. glad you're getting time now πŸŽ„
  • l

    lukecarrier

    12/30/2020, 12:48 PM
    Had to use
    jest --forceExit
    because of a handle leak somewhere, I suspect down to tmp.DirSync somewhere in the
    dendron-cli
    package's tests but I can't confirm because Jest hangs when I specify
    --runInBand --detectOpenHandles
    . I'll come back to this later
  • l

    lukecarrier

    12/30/2020, 12:49 PM
    I don't recall seeing this until I rebased on top of alpha 1, so I might be able to catch it with a git bisect
  • k

    kevins8

    12/30/2020, 4:30 PM
    the green and red are very holiday. I like the greens though :) is windows were the leak happens?
  • l

    lukecarrier

    12/30/2020, 4:30 PM
    I could repro this with Linux and macOS too, if you look toward the end of the dendron-cli log you can see Jest complain and suggest the detect open handles switch
  • l

    lukecarrier

    12/30/2020, 4:31 PM
    I did manage to track it down earlier so I can retrace my steps a bit, I just wanted to get Windows up to speed before I got into real yak shaving
  • k

    kevins8

    12/30/2020, 4:36 PM
    got it. yeah, i can totally imagine
    tmpDir
    causing issues. if i have time today will also take a look
  • l

    lukecarrier

    12/30/2020, 4:37 PM
    I'm aiming to get Windows builds mostly passing today, at least for the CLI
  • l

    lukecarrier

    12/30/2020, 4:37 PM
    I've moved the scripts into npm scripts for now and written a node equivalent of sync_vault, which removes the bash dependency on windows
  • l

    lukecarrier

    12/30/2020, 4:48 PM
    I think Parallels breaks AirPods... so much for the Apple ecosystem "just working"
  • k

    kevins8

    12/30/2020, 5:01 PM
    my airpods have been getting a cracking sound for the last few months. it's a known issue (http://feedproxy.google.com/~r/Techcrunch/~3/bYet25H3Zco/), i just rather talk about it on discord then actually get them replaced πŸ˜…
  • l

    lukecarrier

    12/30/2020, 5:11 PM
    I just find Apple's whole "seamless" handoff experience really poor. I have a pair of PXC 550s that can connect to two devices at once and they do a far better job across ecosystems than Apple can manage within their walled garden
  • l

    lukecarrier

    12/30/2020, 5:11 PM
    Apple support keep offering replacement pairs, but they never seem interested in troubleshooting the issue, which I think is with the cloud pairing service in macOS
  • l

    lukecarrier

    12/30/2020, 5:20 PM
    Does your MBP suffer from the audio popping issue? πŸ™ˆ
  • k

    kevins8

    12/30/2020, 5:34 PM
    yeam, hand off is a hit or miss, more often a miss
  • k

    kevins8

    12/30/2020, 5:34 PM
    and yes, it does. it seems app specific though. spotify is good. vlc usually pops
  • l

    lukecarrier

    12/30/2020, 5:35 PM
    A suggested workaround that seems to help in some apps is to change the output format for the internal speaker to 48KHz in MIDI setup, I have no idea why
  • l

    lukecarrier

    12/30/2020, 5:36 PM
    Catalina and Big Sur are reminiscent of late 2010s desktop Linux... I think people need to be more critical of Apple's quality control
  • k

    kevins8

    12/30/2020, 5:39 PM
    yeah, there's a lot of issues that you don't usually see coming from apple. that being said, apple sillicon seems to be amazing and i'm looking forward to arm laptops being a mainstream thing
  • l

    lukecarrier

    12/30/2020, 5:40 PM
    for sure, I'm surprised how badly Microsoft dropped the ball on this. Rosetta outperforming native Intel hardware in some cases is seriously impressive
  • k

    kevins8

    12/30/2020, 6:28 PM
    yeah, every review of apple sillicon has been raving about it
  • l

    lukecarrier

    12/30/2020, 10:43 PM
    Still some snapshot issues to fix with the Windows CLI build, but I have fixed all of the issues related to hardcoded /tmp by using the
    TEMP
    environment variable. I'll pick this up again tomorrow and fix any breakage I caused in the Linux build πŸ˜…
  • l

    lukecarrier

    12/30/2020, 10:44 PM
    It's running natively under PowerShell, so we might want to add a separate cygwin pipeline to pick up any weird edge cases too
  • k

    kevins8

    12/30/2020, 10:57 PM
    nice! when you say natively under power shell, do you mean the ci runs it under power shell?
  • l

    lukecarrier

    12/30/2020, 11:28 PM
    Yup! That’s the default under GitHub actions. The one issue I have seen a few times is that we do string comparisons on which don’t account for case insensitivity of drive letters, but that should be an easy fix 😁
  • k

    kevins8

    12/31/2020, 1:10 AM
    oh yeah. normalize all the tests!
  • l

    lukecarrier

    12/31/2020, 11:20 AM
    Okay, this is more complex than I had hoped... strap in. On Windows, there's no
    /tmp
    , but the user's temp dir is exposed with the
    TEMP
    env var. On macOS, there is
    /tmp
    , and it's exposed in the
    TMPDIR
    env var. On Ubuntu, there is
    /tmp
    and it's not exposed via any env var. Would you be happy with us relocating logs to e.g.
    <project root>/logs
    ?
  • l

    lukecarrier

    12/31/2020, 11:20 AM
    I had mistakenly assumed Linux distros set TMPDIR, but it seems they do not...
  • k

    kevins8

    12/31/2020, 4:40 PM
    standards are a mythin πŸ˜… yeah, that would be great. i think that's how most projects do it. the
    /tmp
    was just so I could get logs when running locally. not sure why it didn't crash on travis back in the day :/
1...414243...108Latest