Too many open issues (and PRs)

[Krzmbrzl]

From my point of view, there are currently way too many open issues and confusingly many open PRs.

PullRequests
One of the first questions that came to my mind when looking at the PRs was "why are there so many open PRs?". To a new potential contributor this could seem intimidating in the sense that it could be conceived as PRs not getting processed properly (which I now know is not the case). I think it would have a motivating effect for potential contributors to see only a handful of open PRs in this Repo, as this clearly shows that PRs are processed in time symbolizing that they are much appreciated.

Not don't get me wrong. I'm not complaining about how PRs are actually handled. I'm just pointing out how it might appear at first glance.

Given that the oldest open PR is from 2011, I think that a lot of those PRs can be closed as (temporarily) dead/discontinued. If the author of that PR steps forward again at some point later, it would be a piece of cake to re-open it, but as long as nobody is working on it, I don't see a need for keeping it open.
There could be a label (something like incomplete-PR) assigned to them so that if in the future someone wants to see which PRs where closed because of this reason, they can easily find those PRs.

I also have the feeling that some of these PRs are changes that didn't make it into 1.3 so they haven't been merged yet, as the development focused on the 1.3 release. As the release is done now though, I think now is a good time to merge these PRs in the master branch.

Issues
I scanned over the issues briefly in the last couple of days and I think that there a few duplicates in there, some that are probably no longer relevant and some missing information because the author of the issue never bothered to give the details they have been asked for.

If the issues are to be used to structure the development of Mumble, I think they absolutely have to be reduced and properly categorized. With well over 700 open issues, I doubt that there is anyone having even a rough overview of what is an open issue already and what is not.

My proposal for tackling this problem is the following:
Those issues should be revisited to make sure that only the relevant ones staying open. If there is an issue with too few information or if it is unclear, whether this still applies, ask again for more information and add a special label to them.

Then a bot like https://github.com/probot/no-response could be used to give the authors of those issues a certain amount of time (e.g. 30 days) to respond to the request. If there is no response in that time, the issue will be automatically closed by this bot.
Note that the bot will only process issues that have been manually assigned to that specific label. Thus it won't accidentally close off important issues.

After having gone through all issues, I'd hope there will be less open issues there and the ones remaining will provide the necessary information to be processed. And most importantly: The remaining issues will be categorized in a more or less uniform way.

---

Processed issue-pages:

• 1
• 2
• 3
• 4
• 5
• 6
• 7
• 8
• 9
• 10
• 11
• 12
• 13
• 14
• 15
• 16
• 17
• 18
• 19
• 20
• 21
• 22
• 23
• 24
• 25
• 26
• 27
• 28
• 29
• 30

[bendem]

Kind of related to #3279

[Krzmbrzl]

Kind of, but not quite. With the suggested stale bot, even issues that are valid but are simply old and already have all info, would be marked/closed. With the one I suggested here, only issues marked manually would be processed so one doesn't have to worry about a bot being overly eagerly and closing a bunch of valid issues

[Krzmbrzl]

@davidebeatrici do you have objections against such a robot? I'd suggest using the needs-more-input label to be used for this purpose.

[Mikaela]

I feel you on the amount of open issues, I have been requesting another project to add stalebot as I have seen good experiences of it around GitHub. I don't know if the bot you mentioned would be that useful unless you often have issues where authors don't comment the requested information.

• https://github.com/probot/stale

[Krzmbrzl]

As stated above: With such a bot, the validity and completeness of an issue is not taken into account. If there is an issue with all information, then that won't change just because nobody had the time to work on it yet.
Issues should either be closed as too unimportant or they should stay open until fixed.

I have the feeling though that we have a lot of issues that are incomplete and in which the author hasn't responded to requests. These are the kinds of issues I want to see closed.

[davidebeatrici]

@davidebeatrici do you have objections against such a robot? I'd suggest using the needs-more-input label to be used for this purpose.

Sounds good.

[Krzmbrzl]

As I don't think that I have the respective permissions, I think it has to be you, who adds the bot :)

[Krzmbrzl]

Note for self: I have processed until including page 29 of the issues so far

[Krzmbrzl]

Quick status update: I haven't even processed a third of our issues and already I was able to close about 30 immediately and marked a lot of issues as incomplete. These should be closed automatically within the next month unless someone steps up and provides the necessary information.
Therefore I'd like to be optimistic and say that we might get rid of quite a few issues by this housekeeping 🤞

[Kissaki]

Seeing that the bot is already active we should document somewhere our desired/defined state of how we configured/set it up/designed it to work. Be it public or internal.

[Krzmbrzl]

Agreed. Promoting #3942 I think this would be a perfect candiate for being put into a GitHub-wiki (if we decided to create one) ☝️

As a quick-reference one can look at https://github.com/mumble-voip/mumble/blob/master/.github/no-response.yml for how the bot is set-up. In short it processes issues with the needs-more-input label and if the original author of that issue doesn't comment within 30 days after the issue as been labelled, it gets closed.

[Krzmbrzl]

I don't think this issue is needed anymore. Guess the topic is obvious ^^