Restriction to become OWNER in Chromium project

109 views
Skip to first unread message

Hirokazu Honda

unread,
Sep 26, 2018, 6:03:38 AM9/26/18
to Chromium-dev
Hi Chromies,

Last time I considered to become OWNER in some directory, I noticed there is the restriction for OWNERS.
One of them is,
Be a Chromium project member with full commit access of at least 6 months tenure.

If a person has been working a lot in a directory and is familiar with the code base, I think the person should deserve to be OWNERS in the directory.
Although I don't know how this term was decided, I feel six months after becoming a committer is too long.
I even wonder if the term restriction is reasonable.

I think a shortage of OWNERS is a general issue in Chromium project.
Could we elaborate these restrictions?

Sincerely,
-Hiro

Peter Beverloo

unread,
Sep 26, 2018, 6:12:13 AM9/26/18
to hi...@chromium.org, Chromium-dev
+1, the six month requirement always felt a bit arbitrary to me. Obviously an OWNER needs to be a committer, but the rate at which someone ramps up with the design, requirements, constraints and dependencies of a particular area of code is highly variable - a decision best made by existing owners.

Thanks,
Peter


--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAO5uPHOiGr19e5xL9_EiLcr%2Bx5F9H1NHDF53CL_LKpk5ot2wUA%40mail.gmail.com.

Alexandre Courbot

unread,
Sep 26, 2018, 10:34:21 AM9/26/18
to pe...@chromium.org, Hirokazu Honda, chromi...@chromium.org
I'd like to second this as well. There are situations where the 6
months restriction could become a problem, for example:

* The unique owner of a part of the code with contributors but no
knowledgeable committer decides to step down,
* The amount of contributions to the same specific part of the code
suddenly ramps up, leaving the unique owner swamped to do the reviews
for at least 6 months (assuming a knowledgeable contributor can
immediately be promoted to committer).

Is this 6 months rule inflexible? If it is, or if we decide to lax
this, who is in position to make the call?

Cheers,
Alex.
> To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAE8XsdjgGqW-GjFyG4F3h_xhm8QGMpJM7ZAX2g30aZxXKWdk1Q%40mail.gmail.com.

Nico Weber

unread,
Sep 26, 2018, 10:36:53 AM9/26/18
to acou...@chromium.org, Peter Beverloo, hi...@chromium.org, Chromium-dev
On Wed, Sep 26, 2018 at 10:33 AM Alexandre Courbot <acou...@chromium.org> wrote:
I'd like to second this as well. There are situations where the 6
months restriction could become a problem, for example:

* The unique owner of a part of the code with contributors but no
knowledgeable committer decides to step down,

Mostly unrelated to this discussion, but: No code should have a single owner. Every OWNERS files should contain at least two entries. If you see OWNERS files with a single entry being added, add for more co-owners.
 

Jochen Eisinger

unread,
Sep 26, 2018, 10:37:53 AM9/26/18
to acou...@chromium.org, pe...@chromium.org, Hirokazu Honda, chromi...@chromium.org
Is that in practice a problem, or is the real problem that somebody contributes a lot of code, but we somehow forgot to make them a committer?

Hirokazu Honda

unread,
Sep 26, 2018, 11:18:05 AM9/26/18
to sma...@chromium.org, Chromium-dev, pe...@chromium.org
Thanks folks for replying.

Among the restrictions, six months rule is well-defined and also
cannot be resolved by someone's action/effort.
("last 90 days" rule makes sense and can be cleared by the person's effort.)
That also means the rule is not reflected a person's ability.

Peter's opinion sounds good to me.
> Obviously an OWNER needs to be a committer, but the rate at which someone ramps up with the design, requirements, constraints and dependencies of a particular area of code is highly variable - a decision best made by existing owners.

There are many directories and OWNER files in Chromium directories.
There are, thus, many situations on these directories.
The existing OWNERs have a responsibility in their directories,
including to make a flexible decision some committer can become OWNER.

> Is that in practice a problem, or is the real problem that somebody contributes a lot of code, but we somehow forgot to make them a committer?
I would not say either of them. There are many situations.

> From my experience, the 6 month requirement is pretty flexible. I still haven't been on Chrome for 5 months, and I've been added to 2 OWNERS files. It's up to the current OWNERS to make the call and approve the CL.

Hmm, this rule is flexible, while the doc states "must'? Indeed, I
don't intend at all to point out this rule is not kept sometimes since
some people may miss the rule.
If we really respect this rule, we should hook them in presubmit script.

Thanks,
-Hiro
On Wed, Sep 26, 2018 at 11:42 PM <sma...@chromium.org> wrote:
>
> From my experience, the 6 month requirement is pretty flexible. I still haven't been on Chrome for 5 months, and I've been added to 2 OWNERS files. It's up to the current OWNERS to make the call and approve the CL.

Dirk Pranke

unread,
Sep 26, 2018, 12:56:42 PM9/26/18
to hi...@chromium.org, sma...@chromium.org, Chromium-dev, Peter Beverloo
There's no automatic tooling that enforces the six-month rule, so in practice there's flexibility.

However, we should make the wording match what we actually want to do, of course.

I think it would be reasonable to change the wording to say that six months is recommended as a guideline, but that exceptions can be made under the consensus of the existing owners.

Does anyone strongly disagree with that proposal?

As to who can make that call, it's probably some combination of the people listed in //ENG_REVIEW_OWNERS and myself, since I've owned most of the actual owner-enforcement code and processes.

-- Dirk

>> >> To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev+unsubscribe@chromium.org.

>> >> To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAO5uPHOiGr19e5xL9_EiLcr%2Bx5F9H1NHDF53CL_LKpk5ot2wUA%40mail.gmail.com.
>> >
>> > --
>> > --
>> > Chromium Developers mailing list: chromi...@chromium.org
>> > View archives, change email options, or unsubscribe:
>> > http://groups.google.com/a/chromium.org/group/chromium-dev
>> > ---
>> > You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
>> > To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAE8XsdjgGqW-GjFyG4F3h_xhm8QGMpJM7ZAX2g30aZxXKWdk1Q%40mail.gmail.com.

--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
    http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.

sma...@chromium.org

unread,
Sep 26, 2018, 1:10:13 PM9/26/18
to Chromium-dev, pe...@chromium.org, hi...@chromium.org
From my experience, the 6 month requirement is pretty flexible. I still haven't been on Chrome for 5 months, and I've been added to 2 OWNERS files. It's up to the current OWNERS to make the call and approve the CL.


On Wednesday, September 26, 2018 at 10:34:21 AM UTC-4, Alexandre Courbot wrote:
>> To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev+unsubscribe@chromium.org.

Alexandre Courbot

unread,
Sep 26, 2018, 1:23:34 PM9/26/18
to dpr...@chromium.org, Hirokazu Honda, sma...@chromium.org, chromi...@chromium.org, pe...@chromium.org
Thanks for the clarification. Your pragmatic view on that issue sounds
very reasonable indeed, and I see that in practice things seem to have
been more flexible than stated. Would it make sense to adapt the
wording of https://chromium.googlesource.com/chromium/src/+/master/docs/code_reviews.md#expectations-of-owners
to reflect this? Maybe "6 months of consistent and non-trivial
activity in the area ones applies to" or something close would be
appropriate here.

Cheers,
Alex.
>> >> >> To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.
>> >> >> To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAO5uPHOiGr19e5xL9_EiLcr%2Bx5F9H1NHDF53CL_LKpk5ot2wUA%40mail.gmail.com.
>> >> >
>> >> > --
>> >> > --
>> >> > Chromium Developers mailing list: chromi...@chromium.org
>> >> > View archives, change email options, or unsubscribe:
>> >> > http://groups.google.com/a/chromium.org/group/chromium-dev
>> >> > ---
>> >> > You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
>> >> > To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAE8XsdjgGqW-GjFyG4F3h_xhm8QGMpJM7ZAX2g30aZxXKWdk1Q%40mail.gmail.com.
>>
>> --
>> --
>> Chromium Developers mailing list: chromi...@chromium.org
>> View archives, change email options, or unsubscribe:
>> http://groups.google.com/a/chromium.org/group/chromium-dev
>> ---
>> You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
>> To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAO5uPHPtaVMVHGhO3kdn1BGsZxY_d6JfRiUtJcnQFG%2B-uO5QhQ%40mail.gmail.com.
>>
>
> --
> --
> Chromium Developers mailing list: chromi...@chromium.org
> View archives, change email options, or unsubscribe:
> http://groups.google.com/a/chromium.org/group/chromium-dev
> ---
> You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
> To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAEoffTDvQKv2qi2AXf2t%2Bah25e98YogGR%2Bt6stB1Uk8ZzuC2-Q%40mail.gmail.com.

Peter Kasting

unread,
Sep 26, 2018, 5:44:44 PM9/26/18
to Dirk Pranke, hi...@chromium.org, sma...@chromium.org, Chromium-dev, Peter Beverloo
On Wed, Sep 26, 2018 at 9:55 AM Dirk Pranke <dpr...@chromium.org> wrote:
There's no automatic tooling that enforces the six-month rule, so in practice there's flexibility.

However, we should make the wording match what we actually want to do, of course.

I think it would be reasonable to change the wording to say that six months is recommended as a guideline, but that exceptions can be made under the consensus of the existing owners.

Does anyone strongly disagree with that proposal?

Honestly, I think our recommendation should probably be closer to three months than six.  Otherwise, +1 to this.

PK 

Dirk Pranke

unread,
Sep 26, 2018, 6:03:53 PM9/26/18
to Peter Kasting, hi...@chromium.org, sma...@chromium.org, Chromium-dev, Peter Beverloo
Yeah, actually I think 3 months is probably better as well.

-- Dirk

Hirokazu Honda

unread,
Sep 27, 2018, 10:02:52 PM9/27/18
to dpr...@chromium.org, pkas...@chromium.org, Sam Maier, Chromium-dev, Peter Beverloo
I think our current a proposal is
* Change them as recommendation,
* Change six months to three months.

Right?

Does anyone have any other opinions?
Dirk, could you update the doc in a couple of days if no one will be
against them?

Thanks,
-Hiro

Dirk Pranke

unread,
Sep 27, 2018, 10:44:19 PM9/27/18
to Hirokazu Honda, Peter Kasting, Sam Maier, Chromium-dev, Peter Beverloo
Yes, I will update the docs. 

-- Dirk

Hirokazu Honda

unread,
Oct 4, 2018, 5:09:54 AM10/4/18
to Dirk Pranke, Peter Kasting, Sam Maier, Chromium-dev, Peter Beverloo
Hi, Dirk.

One week have passed. I think nobody is against this change.
Could you update the doc?

Thanks,
-Hiro

Dirk Pranke

unread,
Oct 5, 2018, 9:50:22 PM10/5/18
to Hirokazu Honda, Peter Kasting, Sam Maier, Chromium-dev, Peter Beverloo
Hi,

Yes, I will update the doc; sorry, I was overly busy this week and haven't gotten to this yet, but it's still in my queue.

-- Dirk

Yuri Wiitala

unread,
Oct 5, 2018, 11:01:57 PM10/5/18
to Dirk Pranke, hi...@chromium.org, Peter Kasting, sma...@chromium.org, Chromium-dev, pe...@chromium.org
One other point: While I agree it'd be good to reduce the time restriction in general, I feel that certain areas of the code base would not benefit. For example, the top-level dirs (e.g., <root>, base, content, services, etc.) should probably have a multiple-year requirement. So, maybe the wording should state that the time restriction should be in some "common sense" proportion to the scope of files becoming OWNER'ed?


--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.

Peter Kasting

unread,
Oct 5, 2018, 11:23:25 PM10/5/18
to Yuri Wiitala, Dirk Pranke, hi...@chromium.org, sma...@chromium.org, Chromium-dev, Peter Beverloo
On Fri, Oct 5, 2018 at 8:00 PM Yuri Wiitala <m...@chromium.org> wrote:
One other point: While I agree it'd be good to reduce the time restriction in general, I feel that certain areas of the code base would not benefit. For example, the top-level dirs (e.g., <root>, base, content, services, etc.) should probably have a multiple-year requirement. So, maybe the wording should state that the time restriction should be in some "common sense" proportion to the scope of files becoming OWNER'ed?

I don't think we should write anything in our rules for this.  Being an OWNER requires approval of the existing owners, and I trust the owners in those directories to have good judgment.  It's not likely anyone new to the project would become an owner in these directories anyway, and in the rare case where it happened -- e.g. some C++ committee member works on implementing stuff in base/ full time for months -- I don't think we'd want a procedural rule like this to get in the way.

PK 

Dirk Pranke

unread,
Oct 7, 2018, 1:38:08 PM10/7/18
to Peter Kasting, Yuri Wiitala, Hirokazu Honda, Sam Maier, Chromium-dev, Peter Beverloo
I agree with Peter.

-- Dirk
 

PK 

Reply all
Reply to author
Forward
0 new messages