just updated to the v0.0.35 f-droid build and jerboa started to consistently crash at seemingly random times. likewise, I would get web UI error notifications while using the web frontend.

did a little testing and noticed that it may be related to the “show avatars” setting being enabled.

turning “show avatars” off and saving the settings seems to have eliminated the issues on both web and app.

will continue to test and update the post if needed. any others able to confirm or refute this particular issue might be helpful.

edit: there are still crashes related to editing a post, but crashes while scrolling through posts and comments seem to be much reduced. any feedback is welcome.

edit2: further poking suggests the underlying cause may be jerboa not handling server response timeouts well. turning off avatars may reduce the number of API calls and therefore the number of potentially mishandled timeouts. crash frequency is likely variable as a result of variable server load. if others also experience reduced crash frequency with avatars turned off then I can either create a github issue or add info to an existing one.

  • godless@latte.isnot.coffee
    link
    fedilink
    English
    arrow-up
    9
    ·
    2 years ago

    Make sure you report the issues on github as well, so the devs can actively work on them.

    https://github.com/dessalines/jerboa/issues/ (just check if a duplicate already exists). Is the instance where your account is registered already migrated to 0.18.0? Apparently using Jerboa 0.0.35 with 0.17.4 (and older) causes problems due to a change in API handling. At least that was the reply I got to some of my reported issues, since my home instance is not yet upgraded.

    • qprimedOP
      link
      fedilink
      English
      arrow-up
      6
      ·
      2 years ago

      I am on lemmy.ml which is currently at v0.18.0 so it should be cool. wanted to crowd source some feedback before I open a github issue - dont want to waste dev time on non-reproducible ghosts.

      • godless@latte.isnot.coffee
        link
        fedilink
        English
        arrow-up
        5
        ·
        2 years ago

        Well, if it crashes your phone repeatedly, it’s already worth reporting even if it’s not affecting anyone else for now. Better to fix it early than to leave a critical bug in the code that might affect other functions further down the line, you never know how and where they eventually interlink.

        • qprimedOP
          link
          fedilink
          English
          arrow-up
          1
          ·
          2 years ago

          agreed. the issue also seems to rear its head on the web ui as well (but with no crash, just a json error), so its already possibly larger than just the jerboa. I suspect its due to the switch from websockets to http. so any info I get here will go into a potential github issue.

    • qprimedOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      2 years ago

      do you happen to have a specific commit or series of commits? I took a very quick look through and didnt see any specifc commit.

  • ElTacoEsMiPastor
    link
    fedilink
    English
    arrow-up
    6
    ·
    2 years ago

    yes, it’s been crashing consistently and randomly since the last update. I will try your solution of the avatars and report back.

    thanks in advance!

    • qprimedOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      2 years ago

      its possible that the crash issues are related to server load and jerboa’s handling of invalid server responses. I have noticed that crashes are much reduced over the last 24 hours, so perhaps something has changed on my instance (lemmy.ml).

      any changes for you?

      • ElTacoEsMiPastor
        link
        fedilink
        English
        arrow-up
        2
        ·
        edit-2
        1 year ago

        I share the same instance and the crashes have been way less often. Using your recommendation, they stopped while I was scrolling the feed, but were consistent when checking communities on other instances. Thanks!

        At the moment, it happens even less. Now it’s mostly when a post is loading, but it seems independent from the instance.

        • qprimedOP
          link
          fedilink
          English
          arrow-up
          2
          ·
          2 years ago

          thanks so much for the update. greatly appreciated.

  • FrankTheHealer
    link
    fedilink
    English
    arrow-up
    4
    ·
    2 years ago

    Can confirm this. Upgraded to .35 today via F Droid and have been getting consistent crashes.

    • qprimedOP
      link
      fedilink
      English
      arrow-up
      2
      ·
      2 years ago

      its possible that the crash issues are related to server load and jerboa’s handling of invalid server responses. I have noticed that crashes are much reduced over the last 24 hours, so perhaps something has changed on my instance (lemmy.ml).

      any changes for you?

      • FrankTheHealer
        link
        fedilink
        English
        arrow-up
        1
        ·
        2 years ago

        Yeah, earlier today, the crashes seemed to have stopped, but soon came back and would keep happening every few minutes

  • qprimedOP
    link
    fedilink
    English
    arrow-up
    3
    ·
    edit-2
    1 year ago

    quick update on this. have been playing with the jerboa v0.0.36 f-droid build and it seems to handle network errors and junk data in a much more sane way - still not ideal, but it now throws visual error notifications instead of an outright crash.

    looks like the jerboa robustness issue is slowly being addressed. lets see what we get with v0.0.37.

    @toototabon@lemmy.ml, @FrankTheHealer@lemmy.ml, @NettoHikari@social.fossware.space, @godless@latte.isnot.coffee

    • qprimedOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      2 years ago

      so its possible that the crash issues are related to server load and jerboa’s handling of invalid server responses. I have noticed that crashes are much reduced over the last 24 hours, so perhaps something has changed on my instance (lemmy.ml). if you friend still has issues, you might have them try a different (possibly less congested) server and see if that help?

      its quite likely that jerboa devs and committers are already on this issue, but its really useful to get broad userbase reports of symptoms and quirks to help really squash bugs.

  • qprimedOP
    link
    fedilink
    English
    arrow-up
    1
    ·
    2 years ago

    so, I am pretty sure (well, tentatively pretty sure - yeah, thats a thing) that the crashing issues are a combination invalid server responses due to server overload and jerboa’s mishandling of those responses. crashes are transient (but rhythmic) and come in clusters, a pattern you might expect when dealing with stochastic server load.

    I am not an android dev, but I have a coding background so I going to take a look at recent jerboa git issues and commits and see what the activity looks like. my focus is on jerboa right now, as its handling of unepected data from any source should be robust and sane - for many, many reasons.

    of course, regardless of what I do, I am sure these issues will be resolved anyway… just gonna try and do whatever I can to help out, even if its just monitoring git commits. everyone with a little bit of free time and an interest in a fedi alternative to reddit should do the same.