This has been forked from the original and I claim no authorship, but I do claim full agreement with these words and the sentiment they describe. If you find the following to be too harsh, opinionated, non-constructive, or are just someone who is easily offended, then I would suggest taking a look at the less firey breakdown at https://fvsch.com/stale-bots.
Stale bot is a GitHub bot account which through a GitHub App integration can be enabled for a repo by its maintainers and then watch issues and PRs and close them if they have seen no activity from the maintainers after a period of time.
I have an issue with this bot, mainly because it is annoying and insulting to spent time trying to improve someone else's project and then find that they decided to solve their inability to stay on top of their issue backlog by not ignoring the outstanding issues like a decent person would, but closing them, punishing the issue reporter / PR author for actions (or rathen an inaction) of the maintainer.
To add insult to the injury, this bot seems to be used mainly by open source projects of big corps where the maintainers do actually get paid to work on the project, yet seem to struggle to find them time to triage issues in a timely manner.
Now for a disclaimer: I don't fault the maintainers here for their lack or time or ability to triage issues in a reasonable amount of time. Being an open source maintainer is hard, whether or not you have corporate backing. No one is entitled to anyone else's time they are not paying for in open source.
What I do not like is the impersonal, rude and extremely unnecessary hostility the Stale bot introduces into the process.
Can't keep up with your backlog? Welcome to the club! Who can?
Sending a bot to tell other people just trying to be helpful their contribution has been promoted from ignored-for-now to abandoned-without-a-word? That sucks!
At the moment, I am experimenting with seeing whether the Stale bot can be
monitored using the GitHub API. A quick look suggests it should be possible
using the https://api.github.com/users/stale[bot]/events
GitHub API endpoint.
The next step could be either creating an anti-bot which would tell the Stale bot to go fuck itself on each issue the Stale bot ruins, or perhaps a script which creates an issue in this repo and merely links the ruined issue from it so that it is less disturbing to the watchers of the original issue (creating no notification), or perhaps there is no next step. As much as the Stale bot sucks and is annoying, creating more hurt in the world is not a way forward. The maintainers of the repositories in question did activate the Stale bot for a reason and as much as I might disagree with that reason, I have no right to vandalize their repos for it.
So, what is this? Not much really. Just a script which watches the Stale bot and reports its acts for me to shake my head at. I am curious how much damage it causes, but I don't think there is a way forward which is less complicated than all the reasons that lead to the existence of the Stale bot in the first place.
Everything. (I have not yet set up CI which updates this readme regularly.)
node .