Ramblings of a Tampa engineer
photo of person reach out above the water
Photo by nikko macaspac / Unsplash

Every so often I see an issue on GitHub that can best be explained as someone attempting to get help with doing next to zero research on their own. This happened quite recently when a Java update rolled out and broke a selection of applications that were built slightly against the zip specification.

This had a pretty specific crash if affected of some form of "invalid CEN header" and something I blogged about a bit ago.

When the first duplicate issue came in - I was pretty calm and broke it down.

This is actually caused by a recent Java update that made things quite strict. It was reported here and slated for a hotfix this weekend - #3170

Fixed in this commit: #3178

If you want to workaround it. Set these system properties for your java:

    jdk.util.zip.disableZip64ExtraFieldValidation - true

The 2nd time I saw a duplicate question, I went with:

You'll want to update your helper scripts to opt-out of this increased Java ZIP validation.

#3198

It may be fixed anyway once this patch is released to various versions - openjdk/jdk#15273

The 3rd time within days I went with:

Quite a few duplicates of this issue :/

#3290 (comment)

Closing.

It seemed like folks didn't search the issue tracker or even look at the last few merges. At some point this starts to eat at you as it shows the individual looking for support hasn't spent any effort in attempting to resolve the issue on their own.

I'm part of this one community and this individual I feel acts a lot like myself - they are straight to the point and don't stand for anyone not meeting them halfway.

This maintainer responded in under 10 minutes from this report and made it clear that if you don't use the issue template - you aren't getting help.

At times I feel it must be a language barrier as this user came back wondering why the issue was closed. In this specific case - this was this user's first ever post on GitHub and the account was less than a day old when it was posted. Personally, I don't see how you get to the point of installing a plugin for React Native without having a GitHub account yet - just seems a bit odd.

As shown above - when I tried to submit a issue on this repository I was met with a template to fill out. It wasn't as fancy as some of the newer GitHub templates that are more form driven, but it still looked like you had to do quite a lot of deleting to ignore the template.

So yeah I feel for that maintainer.

Take React Native itself which gets 10-20 issues a day on top of all the pull requests. They've got plenty of automation in play that automatically does the following:

  • Checks if the template was followed.
  • Checks if a newer version is available (compared to what you put).
  • Checks if a reproducer was added.
  • Checks if version is out of support.

If you make a mistake - a bot adds a label and in 90-180 days that label will cause the issue to be closed unless you fix it. This has helped them take a backlog of thousands of thousands of issues down to roughly 1,500.

At the scale they are working in - they'd either have to hire a few full-time folks to manage the bug tracker or have automation and machines keep it in check.

I just think back to the era of less structure with forums and Halo modding almost 2 decades ago. Even as a 13 or 14 year old kid - I was beyond nervous of people on the forum knowing I didn't know things. What that meant is I searched, read and tried to find answers to my own problems before I resulted to posting a question.

Chances are there back then - even when I did exhaust my research and post a question - I was still insulted and told something terse. Something has just changed over the last decade or so that folks are more likely to ask a question for someone to direct them / answer them instead of digging into it themselves.

This is a tough middle point to find though, especially in the workplace. You don't want someone struggling researching for hours for something they can simply ask another co-worker. On the flip - you don't want constant questions when a bit of self-research helps build the ability to resolve problems on your own.

This applies directly to GitHub - if you've shown you are willing to help the maintainer(s) help you then you'll have greater luck in getting a resolution. Once you do that a few times - take the next step and attempt to fix the problem itself with a pull request.

You’ve successfully subscribed to Connor Tumbleson
Welcome back! You’ve successfully signed in.
Great! You’ve successfully signed up.
Success! Your email is updated.
Your link has expired
Success! Check your email for magic link to sign-in.