Re: [discuss] So what does Gerrit actually get us, anyway?

Will Stevens
 

Agreed. Reviewer time is a big gating factor when dealing with contribution. 

But the community is made up of more than just code contributors. Users have to find the project, be comfortable with it and be willing to adopt and engage. Gerrit is not doing us ANY favors on that front.

I am a professional developer and gerrit is not my happy place. Once you understand the projects and the code (by pulling it down and navigating it locally), Gerrit will start to fade into the background and you will only ever care to look at it when you commit code or review code. It is easy to forget the struggles others face when trying to get familiar with the code you know so well. 

Cheers,

Will

On Tue, Jun 11, 2019, 9:00 AM Jan Gutter <jan.gutter@...> wrote:
On Tue, 11 Jun 2019 at 14:34, Will Stevens <wstevens@...> wrote:
Sukhdev, we are not looking for a philosophical discussion.  We are trying to understand if Gerrit is really the best solution for what the community needs.

Another discussion for over beers, but the fact that the boat is not being rocked may be part of why, a year later, the community is still struggling to work with the project.

Apologies again for chiming in with a very biased view. The discussion currently seems to be blaming Gerrit for the lack of outside contributions. At the risk of sounding flippant, maybe the problem is lack of dedicated reviewers?

From my perspective, when I want to contribute to a new and unfamiliar project, I generally start with git blame on the piece of code I want to work with. That leads me to a git patch, and in OpenStack's case, leads me to a review. Looking through the discussions on the review leads me to who I should contact on email, or chat to, in order to ask for advice and guidance. In OpenStack's case, I received feedback in the same week.

Even with OpenStack, it's recognised that reviewer time is mostly the gating factor for any project. For example, the Nova project recently wrestled with the issue that they might have to compromise on the policy of +2's from two different employers or they run the risk of running into a hard limit.

I cannot claim to have the answers, but I can say it's pretty difficult for a random person to contribute to a project where the bugs on mainline commits reference a private JIRA, little public discussion goes on and development, planning and design isn't held in publicly recorded meetings.

By all means, Gerrit has a learning curve. It's also not a coincidence that two of the top three open source projects use Gerrit. I would strongly advocate against switching to the Linux Kernel development model, however.
 
Will Stevens
Chief Technology Officer
c 514.826.0190



On Mon, Jun 10, 2019 at 11:06 PM Sukhdev Kapur via Lists.Tungsten.Io <sukhdev=juniper.net@...> wrote:
Ha ha. You want to drag me into philosophical discussion:-)
Short answer- this is how the world was when Contrail started.
So, the short answer is -  "that is the way it has been done in the past"

Gerrit is not known to being user friendly, really? Says who? I would entertain that debate over a beer, not here :-):-)

I do not want to engage in to philosophical discussion.  We all have certain religions about tools, programming languages, development methodologies. So, let’s leave it at that.  

Let’s not rock the boat. If someone has extra cycles to spare, I can suggest several  improvements that we can make that will make TF great. 

Cheers 
Sukhdev


From: dev@... on behalf of Will Stevens <wstevens@...>
Sent: Monday, June 10, 2019 4:53 PM
To: dev@...
Cc: discuss@...
Subject: Re: [TF Dev] [discuss] So what does Gerrit actually get us, anyway?
 
The motivation is to understand if there is a real reason we MUST use Gerrit, or if it was chosen simply because "that is the way it has been done in the past". Gerrit is not known for being user friendly or intuitive for contributors, so what are the roots that hold Gerrit in place as our standard?  Are there alternatives which would better suit the communities needs going forward?

There may be many very good reasons for using Gerrit, but if there are not, it seems prudent to understand that now before we adopt it en masse. 

That is my understanding of the motivation behind this question...

Cheers,

Will



--
Jan Gutter
Principal Software Engineer

Netronome | First Floor Suite 1, Block A, Southdowns Ridge Office Park,
Cnr Nellmapius and John Vorster St, Irene, Pretoria, 0157
Phone: +27 (12) 665-4427 | Skype: jangutter |  www.netronome.com

Join dev@lists.tungsten.io to automatically receive all group messages.