Taro Logo
0

Debugging tips - how to stop and take a break

Profile picture
Senior Software Engineer at Taro Community2 days ago

How do you know when to:

1/ Stop and take a break from debugging after spending loads of time on minor errors from a large code base that has been taking too long? It's easy to see hours turn into days. I'm trying to figure out how to save time and prevent wasting too much of it in the future. I sometimes code til 12 midnight, 2-4AM or stop at 10PM and still feel like I was productive, but not as far as I could have gone. Note I'm trying to use this framework across 3 projects, so when one doesn't build on an error, I move onto another, find other errors, but also learn what works in some areas and not others.

I spent too much time on a lot of things that I think the framework team (VC backed startup with open source options) should have fixed, lots of bugs - there's tons of the, mostly syntactical errors, a lot of dependency errors (mismatching versions), and a lack of documentation on some custom components many developers have common issues with). Days turn into almost 1-2 weeks on a lot of the same common errors on a framework (commonly used, large code base)? The framework is amazing, and considered best in market after trying to work with similar ones for years, but it's just...well a lot to work with!

2 / Decide whether or not to use an older project that works with a stable, but not the latest version of a framework?

Context: I've been working with an open source framework a friend said he'd work with me on, which he noted had a lot of bloat. I underestimated how much bloat there was.

  • Every time I was going to deploy, some other component would break and I would find myself debugging for hours on end to fix, and then almost deploy, until I find that there's yet, another bug.
  • The framework's staff (they're a VC backed early stagestartup by a top VC well known) and open source maintainers say not to use the latest version, but a stable version.

It's hard because, then I find another error again, and then fail to deploy (on a build). This is often due also to something less of syntactical error and more like something the open source framework paid staff/team and contributors either has little do no documentation for, or support, which is still being rolled out, but it is a common issue with many developers. (ex. a lot of people use database integrations with GCP, AWS, but not this one even if a lot of people use it, Supabase).

13
2

Discussion

(2 comments)
  • 1
    Profile picture
    Tech Lead @ Robinhood, Meta, Course Hero
    2 days ago

    For the debugging thing, my answer is similar to what I wrote here: "Overworking and code quality"

    2 / Decide whether or not to use an older project that works with a stable, but not the latest version of a framework?

    There's a huge difference between latest and stable, which is why the latest build is often called something like "Canary" or "Beta/Alpha". Unless there's a major new feature you need in the latest build, you should always use stable. It looks like this is very much the case in your scenario where the latest build is driving you crazy.

    Hopefully the stable build is truly stable for you 🙏

    • 0
      Profile picture
      Senior Software Engineer [OP]
      Taro Community
      2 days ago

      Lol it's probably the opposite, the 'stable build' has a lot of deprecated code (lots of deduping) and in the most recent product update from their townhall on Discord, the founder and team said they had to roll out features for deduping and slow and long build times (everyone is experiencing similar problems), so it looks like the recommendation from one of their devs gave me the wrong advice (use the latest instead of stable, which I know is odd, since they said to use stable in an older project, but it didn't work!