lock issues

[dougwilson]

locking issues is like the best feature in github. we should always lock an issue/pr when we close it and think the user is satisfied and doesn't need to respond. this prevents people from randomly posting in some old issue about an unrelated problem.

[jonathanong]

i wouldn't say "always", but 99% of the time.

[dougwilson]

right. basically use discretion and try not to do it when it would cut someone off the convo/make them mad :)

[rlidwka]

how about never locking them? from user's point of view, it's very annoying when someone decides whether you need to respond or not for you

people respond in old issues because they think it is related, I very rarely see completely unrelated problems

besides, it's kinda puts contributors above everybody else, when you can bring back some discussion, but other people don't

[dougwilson]

I very rarely see completely unrelated problems

they happen constantly in all the repos i am a part of (posting on an issue from 2012--4 major versions old, come on!). people can easily make a new issue and paste a link to the old one if they feel it's related (we are on the web, after all).

that thread was open for comments for 15 days, and was only locked after it actually landed. the discussion for that example then moved to the release 4.7 pr (or any new issue someone opened if they wanted).

and i'm suggesting specifically not locking something not giving a person a chance to respond--especially when it's a debate.

[rlidwka]

posting on an issue from 2012--4 major versions old, come on!

well okay, maybe it'll make sense to close issues last active more than a year ago

people can easily make a new issue and paste a link to the old one if they feel it's related (we are on the web, after all)

this would cause a lot of duplicated threads about the same thing, so people usually don't post new issue at all even if they can add something useful to the old one

[dougwilson]

posting on an issue from 2012--4 major versions old, come on!

well okay, maybe it'll make sense to close issues last active more than a year ago

fwiw, this is what i'm trying to prevent here; not to quell down discussions. but without actively locking the threads, they get forgotten about until someone posts in it just like we didn't want in the first place, lol. also, it sucks that you cannot use the checkbox stuff in github to lock a bunch of threads at once

[rlidwka]

I but without actively locking the threads, they get forgotten about until someone posts in it just like we didn't want in the first place

I still don't get why new responses to old issues cause any trouble. If it's useless, just ignore that.

[jonathanong]

i like locking because i want new convos in new issues. adding comments to closed issues kill visibility. and many times, they are actually new issues.

[dougwilson]

I but without actively locking the threads, they get forgotten about until someone posts in it just like we didn't want in the first place

I still don't get why new responses to old issues cause any trouble. If it's useless, just ignore that.

sometimes the users just don't realize what they are doing--ignoring it would make us look like we don't care about them :)

adding comments to closed issues kill visibility. and many times, they are actually new issues.

this

[dougwilson]

@rlidwka if you needed an example: https://github.com/strongloop/express/issues/1181#issuecomment-52599087

[Fishrock123]

Should I make a thing that locks all closed issues that are over 6 months old?

[dougwilson]

sure

[rlidwka]

@dougwilson , well user got his answer, issue locked, no harm done, except TJ got one unwanted notification (out of hundreds he is receiving every day). What is so bad about that question, that you have to lock down each and every one closed issue to avoid it?

If you want to talk about examples, here is the opposite one: https://github.com/jashkenas/coffeescript/issues/1752, which was reopened on user report after 1.5 years (I didn't specifically searched for anything, it was updated 2 days ago, so I noted that down)

[dougwilson]

@rlidwka what about @Fishrock123 's "thing" suggestion?

[rlidwka]

@dougwilson , locking something after N months of inactivity looks like the best solution if github won't generate any notifications about it.

[dougwilson]

right. the only reason i wanted to be aggressive about it was otherwise they would just never get locked. if we had a script that would go through all the non-locked issue and lock them if it's close and been x days since the last comment, i think that's good.