Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [servlet-dev] Removing restrictions on direct commits

On Wed, 3 Nov 2021 at 20:28, Mark Thomas <markt@xxxxxxxxxx> wrote:
On 03/11/2021 01:48, Stuart Douglas wrote:
> On Wed, 3 Nov 2021 at 11:02, Greg Wilkins <gregw@xxxxxxxxxxx
> <mailto:gregw@xxxxxxxxxxx>> wrote:
>     I kind of like having it, as we should very rarely be in a situation
>     where a PR must be committed ASAP.
>     If a PR really needs to be committed ASAP, then I think project
>     admins have the power to override on a case by case basis anyway.
> None of us are admins though AFAIK.


> Personally I would prefer that Mark
> be made an admin and the 'Include Administrators' box be unticked. It
> would mean that Mark could merge if he really needs something for the
> release, but it is a big red button with a warning.

That would address my immediate concern regarding being able to address
release issues in a timely manner. The main thing I don't like about
this is that it gives me karma other committers don't have.

>     As a committer to many projects, I have often self determined that
>     my changes were ready to be committed... only to be surprised by
>     some silly omission, accidental inclusion or plain stupid mistake.
>          We are all fallible and even if not, then does the extra few
>     hours required to wait for a review really hurt?
>     cheers
>     P.S. If you do turn it off, I'm still going to pester others for
>     review of any of my PRs, as I for one have broken too many builds to
>     have any delusions of infallibility.

Oh indeed. I am by no means infallible either. However, my experience of
the last few years where EL does not have this restriction is that
always requiring it adds unnecessary friction to the project. I am
confident that all the committers can make good decisions on when to
just commit something and when a PR + discussion is required. Even if
they do get it wrong (very unusual in my experience) we can always
revert things if necessary.

> The process should remain the same either way, and we should definitely
> keep the 'Require a pull request before merging' protection even if you
> are allowed to merge your own PR's, just to prevent accidental pushes.

I'm not convinced (working on multiple projects where PRs are not
required) that accidental pushes are that much of a problem but I could
live with that.

I looks like my initial proposal goes further than the community is
comfortable with but that the community would accept some degree of
relaxation. With that in mind I'm amending my proposal to the following.

Remove the GitHub enforced requirement that all PRs must be reviewed.
All changes will still be via PR. Committers will be expected to use
their judgement as to how long to allow for feedback before merging a PR.

I am ok with this, I trust other committers' judgement here, and a PR also means an email notification so everyone will be aware of it.

If there are questions after the fact the PR also provides a place for this discussion to happen.

I agree that accidental pushes are not really a problem in practice, however it's also that the PR provides a place where additional metadata about the change can live.



> Stuart
>     On Wed, 3 Nov 2021 at 03:30, Mark Thomas <markt@xxxxxxxxxx
>     <mailto:markt@xxxxxxxxxx>> wrote:
>         All,
>         In approx 24 hours time I intend to request that the branch
>         restrictions
>         that prevent committers committing directly to the master branch
>         and
>         those that require every PR to be reviewed before merge are removed.
>         My reasoning is as follows:
>         - I have seen the benefits of these restrictions not being
>         present in EL
>         - I'm expecting a number of non-substantive changes will be
>         required to
>             successfully complete the release process and PR + review
>         for all of
>             them will significantly slow us down
>         - Committers are perfectly capable of determining which changes
>         need a
>             PR and review and which can be made directly - and if they
>         get it
>             wrong changes can easily be reverted
>         I was intending to propose this change after the Jakarta 10
>         release but
>         on reflection, I think the sooner, the better.
>         Thoughts? Comments? Objections?
>         Mark
>         _______________________________________________
>         servlet-dev mailing list
>         servlet-dev@xxxxxxxxxxx <mailto:servlet-dev@xxxxxxxxxxx>
>         To unsubscribe from this list, visit
>         <>
>     --
>     Greg Wilkins <gregw@xxxxxxxxxxx <mailto:gregw@xxxxxxxxxxx>> CTO
> <>
>     _______________________________________________
>     servlet-dev mailing list
>     servlet-dev@xxxxxxxxxxx <mailto:servlet-dev@xxxxxxxxxxx>
>     To unsubscribe from this list, visit
>     <>
> _______________________________________________
> servlet-dev mailing list
> servlet-dev@xxxxxxxxxxx
> To unsubscribe from this list, visit

servlet-dev mailing list
To unsubscribe from this list, visit

Back to the top