The Few, the Tired, the Open Source Coders

Re:The Few, the Tired, the Open Source Coders by _greim_ on Thursday November 19, 2020 @10:46PM (#25157137)

Merging a PR is work too, assuming the maintainer cares enough to review the code. Why not sponsor the project by placing a $300 bounty on the PR merge? As an OSS maintainer who's been on the other side of this in the past, that would have been quite a motivation for me.

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by rectang on Thursday November 19, 2020 @11:59PM (#25157439)

Most pull requests are crap. They are drive-by fixes that do not take maintainability into account, because they only solve the problem that the PR submitter needs solved right now.

Transforming a sloppy drive-by patch into something which doesn't pile on technical debt is hard. For all but the most brilliantly architected projects, it requires someone who can keep the entire project in their head — a core maintainer.

The expectation that unlimited PRs will be reviewed for free is corrosive and contributes to the core-maintainer burnout described in the article.

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by throwawaywrench on Friday November 20, 2020 @12:10AM (#25157478)

>because they only solve the problem that the PR submitter needs solved right now.

This is something that can lead to angry interactions between maintainers and pull request raisers. Never mind adding technical debt, pull requests can outright break uses cases they don't care about in order to implement the single use case they do care about.

raiser: "Merge my PR. It fixes this issue."

maintainer: "It fixes this single issue but you've broken this for everyone else."

raiser: table flip

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by sergeykish on Friday November 20, 2020 @01:06AM (#25157726)

Why? Just use own fork for a while. PR is sharing fix with others, it is visible even if not merged.

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by rusticpenn on Friday November 20, 2020 @08:03AM (#25160045)

A good test strategy would fix the problem. "Your PR fails all these tests, so can you make them all work before submitting the request?"

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by throwawaywrench on Monday November 23, 2020 @08:21PM (#25194138)

Definitely. You just need to keep an eye out for people removing failing tests because they think they are irrelevant :|

[ Parent | on HN ]

Re:The Few, the Tired, the Open Source Coders by _greim_ on Saturday November 21, 2020 @02:17PM (#25171998)

Some of this burden can be offloaded to robots though, via static types, unit tests, linting, and automatic code formatting. I even go so far as to envision my unit tests as a squad of soldiers laying in ambush around my use cases, waiting to pounce on transgressors.

[ Parent | on HN ]