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
x/review/git-codereview: automatically assign reviewers #17555
Comments
FTR: Some contributors, me included, do not normally assign a reviewer because they are told not to do so:
|
I agree with cznic: this feels like a documentation issue. I don't assign a reviewer because i am told to do so. If we just change the guidelines I feel even I as a newbie contributor could figure out a possible reviewer. |
I don't think changing the documentation will be enough to solve this problem. Suppose we write something like
Then what? I've contributed for like 2 years I don't really know who should I add as a reviewer on half of the changes I upload... I doubt a novice contributor will know better. |
One possible solution could be to write a wiki page with a map
etc... but if we can compile such a list easily, then it could be a good idea to also include it in the |
Related issue: #12364 |
Related to #12364, we should probably can create a list of reviewers by looking at who has approved reviews for the package in the past X months and blame the previous author of the lines. Then, pick 2 or 3 people and assign the CL to them. Btw, none of this has to be done in the codereview program. We should probably use Gerrit hooks if we want to automate the assignment. |
I like the idea of having a auto-assignment bot.
gerrit is used by a lot of other projects too, so people
coming from other gerrit-using projects might not be
using our own tool.
|
I should mention that we now have https://golang.org/s/owners which is a rough list of owners per package. |
oh sweet!
…On Sat, Jul 15, 2017 at 2:03 PM, Joe Tsai ***@***.***> wrote:
I should mention that we now have https://golang.org/s/owners which is a
rough list of owners per package.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
<#17555 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/ABYNbHRwFVamDXI14wtw8GaTCDoNSROiks5sOP7sgaJpZM4KeC8O>
.
--
Jessie Frazelle
4096R / D4C4 DD60 0D66 F65A 8EFC 511E 18F3 685C 0022 BFF3
pgp.mit.edu <http://pgp.mit.edu/pks/lookup?op=get&search=0x18F3685C0022BFF3>
|
Is this being worked on actively? My thinking is that having owners in spreadsheet is a maintenance burden. That doc has not beed updated in 1 year. What if there could be some kind of package which can return suggested reviews based on the change, older changes along with reviewes? It could be a package using which we can build additions to Gerrit, git-codereview, and other tools. I am a new contributor to the project and have really hard times trying to figure out the reviewers. |
cc @andybons |
I created a CL a while ago to introduce an owners package, but I am too busy to keep working on it at this point. See https://go-review.googlesource.com/c/build/+/39390. |
There is a plan to create an endpoint on dev.golang.org that can be queried both by humans and machines to find owners. @rakyll’s prior work on this will be helpful. Thanks! |
Many contributors don't assign a reviewer to their CLs, either because they don't think they need to or they don't know who would be a good reviewer. In theory, those CLs are supposed to get promptly triaged and assigned reviewers. In practice, they aren't, and CLs fall through the cracks.
I suggest that when 'git mail' is invoked without specifying a reviewer with the
-r
flag, it automatically selects one.The selection algorithm can be tuned over time, but as an initial pass, I'd suggest a recency-weighted sum of who has committed to and/or reviewed the files touched by the CL, falling back to directories/parent directories as needed for new files/directories. The initial reviewer can always redirect to another reviewer if the algorithm guesses wrong.
If the CL contains DO NOT REVIEW or DO NOT SUBMIT, it will not automatically select a reviewer.
This won't help with people who don't use git-codereview. An alternative that would cover them is to set up a bot that monitors CLs and does the automatic reviewer assignment. Doing it in git-codereview seems simpler and easier and probably sufficient, though.
Feedback requested.
cc @bradfitz @adg @quentinmit @spf13
The text was updated successfully, but these errors were encountered: