Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

This repository is flooding with pull request #1929

Closed
amaank404 opened this issue Oct 24, 2020 · 22 comments
Closed

This repository is flooding with pull request #1929

amaank404 opened this issue Oct 24, 2020 · 22 comments

Comments

@amaank404
Copy link

Hello Dear Maintainers, this repository is now flooding with 600+ pull request. Possibly please pay attention to it. The same is happening with The Python Repository

@prashantdoshi28
Copy link

Yes. It would be good for repo if people start reviewing PRs instead of creating for some time, I'm not saying stop completely, but along with creating PRs, focus on reviewing too, just until number of PRs are in 2 digits.

Like if you create 1 PR, then try to review 2-3 PR also.

@amaank404
Copy link
Author

Great, let's review PRs and ask them if they could help in reviewing PRs

@amaank404
Copy link
Author

I am not a Java Guy, so I would not be able to do anything. I opened this to aware maintainers and other incoming people somewhat like in this.

@prashantdoshi28
Copy link

Yeah saw that few days ago and provided a suggestion over there for maintainers. Hope that everyone participates in this.

Hi @ayaankhan98 , can you please ask others to help review PRs over Gitter or tag relevant people over here to increase awareness. Also, can you mark this issue as Help Needed or anything relevant.

Lets hope that we will soon be able to back to max double digit PRs.

@ayaankhan98
Copy link
Member

Yeah saw that few days ago and provided a suggestion over there for maintainers. Hope that everyone participates in this.

Hi @ayaankhan98 , can you please ask others to help review PRs over Gitter or tag relevant people over here to increase awareness. Also, can you mark this issue as Help Needed or anything relevant.

Lets hope that we will soon be able to back to max double digit PRs.

yes hope so soon we will be able to control this flood.

@prashantdoshi28
Copy link

Yeah saw that few days ago and provided a suggestion over there for maintainers. Hope that everyone participates in this.
Hi @ayaankhan98 , can you please ask others to help review PRs over Gitter or tag relevant people over here to increase awareness. Also, can you mark this issue as Help Needed or anything relevant.
Lets hope that we will soon be able to back to max double digit PRs.

yes hope so soon we will be able to control this flood.

If there are no concerns, then add new issue or mark this issue in some way so that everyone who is working with this repo can know of this conversations and start helping out. Just for broadcasting or notifying purpose.

@ayaankhan98 ayaankhan98 pinned this issue Nov 15, 2020
@prashantdoshi28
Copy link

Hey guys, what if we add a comment to every PR, asking PR author to create a fresh one along with an issue, and we close/decline that PR. This way, we can be sure of that the PRs which we are reviewing, will be fruitful i.e. if we suggest any changes, author will go through those and that PR won't wait there forever.

To sum up what I mean is :

  1. Go to a PR.
  2. Add a comment asking PR author to create a new one along with an issue as we will be declining current one.
  3. If author will be available, then they will follow it. If not, then we can close PRs knowing that even if we review those PRs, author is not going to update it to our required standards and PR will be in open state forever.

Pros :

  1. Number of open PRs will reduce.
  2. We can make sure that no PR will be in open state forever, as author will keep updating as required.
  3. We can remove all spam PRs.

Cons :

  1. It is not an exact solution but a workaround.
  2. Active PR authors will have to do some repeat work.

Also, what if we apply similar strategy to issues (not all issues need to closed as some are meant to be open).

I know it might sound a bit chaotic but we should come up with some sort of technique to solve this issue.

Need your inputs here @ayaankhan98 and @xcodz-dot . We need to tackle this problem as I myself have reviewed some PRs and provided some suggestions but authors don't seem to be active. And reviewing PRs without any proper communication response from authors is just not worth the time.

I hope you consider this and think this through :)

@amaank404
Copy link
Author

I think we should simply comment the changes and note the PR somewhere, if the author doesn't respond, we close it.

@prashantdoshi28
Copy link

I think we should simply comment the changes and note the PR somewhere, if the author doesn't respond, we close it.

Agree, but that's the solution when number of PRs are less and authors are active. I'm not sure that it will work effectively in our case, as there are many spams present. Also the PRs go back till 2018-19, so we can't expect those PR author to update it.

Now as you said, if author does not respond, but if we close PRs irrespective of author's response, he/she will open new anyway as we will ask them to create a new PR in comments.

This will also make sure that redundant PRs are not opened i.e. 2 authors create PR for same concept or bug. Currently due to large number of PRs, no one knows if duplicate PRs are present or not but if number would be smaller, that can be done.

Also let's say, 2 authors create PR for same concept, author A & B, A creates first and then B. Now when we go to PR page, we are shown in last created/updated PR. Now reviewers ask some changes for B's PR and A's PR is skipped although A is active and B is not. So ultimately, for sometime or even forever, both PR will be open, which should not have happened.

That is why I suggested for this drastic measures, as normal measures won't be that effective (they will work but it will take quite a lot of time).

This is just a suggestion for what I think should be done counter this problem. Rest is in the hands of maintainers. I hope they take some initiative for this as I see only a few people participating in review process.

@amaank404
Copy link
Author

For that method, I think we should just type a comment saying them to open a new one, and refer them to this issue. In this method we will have to close all of them, and slowly as the authors come online, new PRs will start coming.

On Python repository, according to my method it is going to take 4 months to get the numbers down to 10.
(estimated number of months = number of PR / the rate of PRs being closed a day.)

Let's wait for maintainers opinion.

@prashantdoshi28
Copy link

For that method, I think we should just type a comment saying them to open a new one, and refer them to this issue. In this method we will have to close all of them, and slowly as the authors come online, new PRs will start coming.

Yep that is what I'm thinking.

On Python repository, according to my method it is going to take 4 months to get the numbers down to 10.
(estimated number of months = number of PR / the rate of PRs being closed a day.)

Nice, its good to know.

Let's wait for maintainers opinion.

Sounds like a plan.

@ayaankhan98
Copy link
Member

For that method, I think we should just type a comment saying them to open a new one, and refer them to this issue. In this method we will have to close all of them, and slowly as the authors come online, new PRs will start coming.

On Python repository, according to my method it is going to take 4 months to get the numbers down to 10.
(estimated number of months = number of PR / the rate of PRs being closed a day.)

Let's wait for maintainers opinion.

Most of the PR are just a joke, their is no valueable code to review and also a lot of them are duplicate.
I think just review the recent PR's which have some valuable implementations and close rest of them direclty (just leave a message there for the author in order to justify their work, if found good we can review that from their onwards).
do the same for old PR's also

@amaank404
Copy link
Author

That is doesn't look nice to me, its easy to check if a code is valuable or not, within 5 minutes we can mention a member to close it if the work is not valuable. simply closing all of them is not a good solution either. Its better if we proceed in the classic way of reviewing PRs

@prashantdoshi28
Copy link

prashantdoshi28 commented Nov 23, 2020

That is doesn't look nice to me, its easy to check if a code is valuable or not, within 5 minutes we can mention a member to close it if the work is not valuable. simply closing all of them is not a good solution either. Its better if we proceed in the classic way of reviewing PRs

@xcodz-dot yes I agree that it does not look nice, but the thing is, I don't think that there are many members reviewing PRs and can just close them as they go. Currently only @ayaankhan98 is the one who is reviewing PRs and has rights to close/mark PRs. Now let's say we do get some people who reviews PRs but what's its use if they can't mark/close them, so ultimately it will come down to maintainers, who I think are already overloaded with their works.

Most of the PR are just a joke, their is no valueable code to review and also a lot of them are duplicate.
I think just review the recent PR's which have some valuable implementations and close rest of them direclty (just leave a message there for the author in order to justify their work, if found good we can review that from their onwards).
do the same for old PR's also

@ayaankhan98 that is what I wanted to convey and I still think that it's the only way to get ahold of this situation.

Also, I have an idea to get out of this dilemma. What if we drop comment to each and every PR tagging authors to reply to it if they are active and willing to update PRs according to standards. We give them a few days like maybe 5. Then if they don't respond back in that time period, we just simply close it with a comment that if author wishes to provide their work to this repo, they will have to create a new PR along with an issue for the same.

So I think if all 3 of us start adding comments to PRs, we can get it done by this weekend and we start closing PRs next weekend.

We would have to do this regardless of what approach we take i.e. if we review and wait for author or we simply close PRs, we will have to add comments first for informing. So why not split it into 2 parts, add comments (not review them) and then after a time period close them.

What do you guys say @xcodz-dot @ayaankhan98 ?

@ayaankhan98
Copy link
Member

So why not split it into 2 parts, add comments (not review them) and then after a time period close them.

why a 2-step process? you can just directly leave a suitable comment that if the author wishes to continue this PR then he/she can reopen the PR and from thereafter we will consider that and will review that in a proper way. Or if he/she does not replies to that then also there is no problem because PR is already closed.
This was for the old PR's

For the recent PR's just review them properly because author will respond to them very likely.

@prashantdoshi28
Copy link

@ayaankhan98 yep that is what I think should be done as I mentioned before. But the decision is up to you guys.

@ayaankhan98
Copy link
Member

That is doesn't look nice to me, its easy to check if a code is valuable or not, within 5 minutes we can mention a member to close it if the work is not valuable. simply closing all of them is not a good solution either. Its better if we proceed in the classic way of reviewing PRs

I don't understand what was not nice in that. I am not closing anyones work i am just closing the PR with a option there if the author wishes to continue then for sure i will consider his/her work and will review that properly.
Because i don't have enough available time to first ping the author and then wait for their response and if they don't response for say 5-7 days then i will go back to that PR and then do that same work which i could have done earlier.

@amaank404
Copy link
Author

Ok, any other person can review, and mention the member for it to be closed

The Plan

Maintainers

Part 1

  • Goes to PR
  • Reviews it
  • done

Part 2

  • Goes where he is mentioned
  • Accepts or closes PR (The other people review it)

The People

  • Goes to PR
  • Reviews it
  • Mentions a Maintainer
  • Done

@amaank404
Copy link
Author

and if talked about pinging the author (Do not ping, simply ask him and continue with other PRs). Those PRs can be closed when all the useful PRs are done.

@amaank404
Copy link
Author

Please try my steps, also The Python repo is now on 293 from 500 which is great

@prashantdoshi28
Copy link

Please try my steps, also The Python repo is now on 293 from 500 which is great

Good to hear that. Keep it up.

@amaank404
Copy link
Author

How about implementing the action named stale (We would not require pinging authors), the stale closes issue and PR over time if there is no activity in it

@ayaankhan98 ayaankhan98 unpinned this issue Feb 26, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants