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

Davis, Matthew <Matthew.Davis.2@...>

Hi all,


For those who didn’t see my previous emails, I’d like to tell my story. (I think it’s part of the motivation for Vicky’s question.)

I have about 16 patches I have been trying to submit since February.

I have already submitted some patches to the docs on GitHub, and they have already been merged.

That was simple and straightforward, like it should be.

That was just like submitting to any other non-LF project.

But the rest of the code is on Gerrit, which is not straightforward.


Trying to understand Gerrit is time consuming, because it’s a steep learning curve.

That can be ok if there are useful features of Gerrit which outweigh that high barrier to entry.

So, what are those advantages?


Worst of all is the CCA issue, which is tied into Gerrit. After 4 months of trying, I still haven’t been approved in the system.


The old approval system is not working any more. The new gerrit doesn’t yet have any code. Everyone talks about how things are about to change, but it’s been like this for at least 4 months.

So I am unable to submit patches or raise issue tickets for the main body of code.


I am about to move on from my current role to a new one, and my team at work is being disbanded.

So after 4 months of trying to get patches submitted, I will have to abandon my patches.

Over the last 4 months I’ve had to deploy Tungsten using a custom Ansible playbook which applies custom patches that I cannot contribute upstream.


If Tungsten was on Github, those patches would have been submitted and merged months ago.

But because the code is on Gerrit, (and because of the overbearing CCA process) the community will not get those patches, ever.

That is the impact of using Gerrit.


In a prior discussion someone mentioned self-hosting as a benefit of Gerrit.

GitHub and GitLab both have self-hosted options if that is a concern.


Other project such as Ansible (on GitHub) have an automatic CCA. (“By submitting this PR you agree to … “) So the project still gets the same legal protections, but without the bureaucratic hurdles.


I think that is the motivation for the question.





From: dev@... [mailto:dev@...] On Behalf Of Sukhdev Kapur via Lists.Tungsten.Io
Sent: Tuesday, 11 June 2019 9:34 AM
To: dev@...; discuss@...; dev@...
Subject: Re: [TF Dev] [discuss] So what does Gerrit actually get us, anyway?


This is a very open ended question. What is the motivation of asking this question?





From: dev@... on behalf of VM (Vicky) Brasseur via Lists.Tungsten.Io <>
Sent: Monday, June 10, 2019 9:37 AM
To: discuss@...; dev@...
Subject: Re: [TF Dev] [discuss] So what does Gerrit actually get us, anyway?


Adding dev@, since the devs have skin in this game.


From: <discuss@...> on behalf of "VM (Vicky) Brasseur via Lists.Tungsten.Io" <>
Reply-To: "discuss@..." <discuss@...>
Date: Monday, June 10, 2019 at 12:32 PM
To: "discuss@..." <discuss@...>
Subject: [discuss] So what does Gerrit actually get us, anyway?


Will and I are sitting here in the lovely CloudOps offices in beautiful Montreal, having a little 2-person TF hackathon, and we got to talking…


Real Soon Now™,  we'll be moving the repos from Juniper to Gerrit and then mirroring them from Gerrit to GitHub.


But what exactly does Gerrit do for us, anyway?


Wouldn't just going directly from Juniper to GitHub and not do Gerrit at all entice more people to contribute and dramatically improve the contributor experience?


Neither of us have yet seen an explanation for why Gerrit *needs* to be involved at all.


Could someone please explain why it's needed and, if it turns out it's not really needed, could we please just not subject our community to the misery of having to use Gerrit?



Join to automatically receive all group messages.