GitHub pull requests

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

GitHub pull requests

Julian Foad-5
Here's a thread for discussing the github (GH) pull-request (PR) method of contributing.

My views, briefly, include:

* I don't want open source projects to support GitHub; there is at least one good open source alternative, GitLab. So I don't want us to keep the existing GH semi-integration as-is, for that reason at least. I don't see a real need to have any level of presence on GitHub beyond probably a place-holder that points to our preferred options.

* When the facility to contribute in this way (PRs) is available, it is because We, the ASF community, have made it so, and we can turn it off if we don't like it. It would be very rude of us to criticize contributors for using a route that we have made available. (The same applies to writing to the mailing list from google groups, for example.)

* I think drive-by contributions can be a valuable route for new contributors to get started. I do acknowledge that they sometimes require a level of hand-holding that can be tedious compared with contributions from more "seasoned" contributors. I think that's OK for our community. Nobody has to do any more than ensure a brief, polite response is given.

* I think the PR style of contribution is useful.

* I would love to help figure out how we could make a PR style of contribution work well, perhaps using GitLab, certainly with better integration than the existing GH semi-integration.

--
- Julian
Reply | Threaded
Open this post in threaded view
|

Re: GitHub pull requests

Branko Čibej
On 24.01.2019 12:29, Julian Foad wrote:
> Here's a thread for discussing the github (GH) pull-request (PR) method of contributing.
>
> My views, briefly, include:
>
> * I don't want open source projects to support GitHub; there is at least one good open source alternative, GitLab. So I don't want us to keep the existing GH semi-integration as-is, for that reason at least. I don't see a real need to have any level of presence on GitHub beyond probably a place-holder that points to our preferred options.


Hmph. I don't want open-source projects to "support" git, and I suppose
my wish is just as likely to happen as yours.


> * When the facility to contribute in this way (PRs) is available, it is because We, the ASF community, have made it so, and we can turn it off if we don't like it.


This point is far too general for this list. members@ might be the right
forum, but there are many, many people at the ASF that believe GitHub is
the whole world (seems to go in hand with the "git == version control"
and "GitHub == git" viewpoints).


On the other hand, we can request that Infra removes *our* GitHub
mirror, just as we requested in the past that it creates one. I'd rather
see the ability to submit pull requests disabled for that mirror, though.


>  It would be very rude of us to criticize contributors for using a route that we have made available. (The same applies to writing to the mailing list from google groups, for example.)
>
> * I think drive-by contributions can be a valuable route for new contributors to get started. I do acknowledge that they sometimes require a level of hand-holding that can be tedious compared with contributions from more "seasoned" contributors. I think that's OK for our community. Nobody has to do any more than ensure a brief, polite response is given.
>
> * I think the PR style of contribution is useful.


I think it is not. This style seems to be designed specifically so that
contributors have to interact with the community as *little* as
possible. It's completely at odds with how we do things here. The
example that prompted the creation of  this thread illustrates this
quite well.


> * I would love to help figure out how we could make a PR style of contribution work well, perhaps using GitLab, certainly with better integration than the existing GH semi-integration.

Surely you mean non-integration. :)

-- Brane
Reply | Threaded
Open this post in threaded view
|

Re: GitHub pull requests

Julian Foad-5
Branko Čibej wrote:
> Julian Foad wrote:
> > * When the facility to contribute in this way (PRs) is available, it is because We, the ASF community, have made it so, and we can turn it off if we don't like it.
>
> [...] we can request that Infra removes *our* GitHub
> mirror, just as we requested in the past that it creates one.

Exactly: our project gets to say what goes for our project.

> I'd rather
> see the ability to submit pull requests disabled for that mirror, though.

Me too. GH, as a closed system, intentionally limits what we can do with it. I'm pretty sure I asked in the past to turn off PRs and was told that's not possible. I'm pretty sure with GitLab we could turn off PRs.

> > * I think the PR style of contribution is useful.
>
> I think it is not. This style seems to be designed specifically so that
> contributors have to interact with the community as *little* as
> possible. [...]

That is certainly the effect we get from GitHub's PRs and the ASF non-integration, but it's not necessarily true of the PR concept in general.

--
- Julian