'refs/for/*' versus 'refs/publish/*'

[Derek]

In the current version of git-review, it's defaulting to the 'refs/publish/*' namespace instead of 'refs/for/*'. I've got a bug report and in progress review, https://review.openstack.org/#/c/584607/ , and I'm trying to make sure this is the right method forward. Would someone confirm and possibly even add some context as to the initial talk of deprecation of the 'refs/for/*' namespace in favor of 'refs/publish/*', then the subsequent de-deprecation? I found some older dicussion on this topic here, https://groups.google.com/d/topic/repo-discuss/yf3dV8R6opE/discussion but wanted to make sure something hasn't changed in the years that followed.

I was also curious as to what the plan was for 'refs/publish/*' based on moving back to 'refs/for/*'?

-Derek

[Han-Wen Nienhuys]

I have a proposal out to change this magic URL to "refs/review"

https://gerrit-review.googlesource.com/c/gerrit/+/187782

so your mail piqued my curiosity, and I asked around.

In https://gerrit-review.googlesource.com/c/gerrit/+/24720/ , a new
DRAFT status was introduced, and the idea was that people would push
first as

git push HEAD:refs/draft/master

and then later publish through

git push HEAD:refs/publish/master

this was before Git or Gerrit (not sure which) supported push options,
so each feature required a new magic branch.

It turns out draft weren't that widely used, and we removed support
for drafts patchsets completely in recent gerrit releases.

> --
> --
> To unsubscribe, email repo-discuss...@googlegroups.com
> More info at http://groups.google.com/group/repo-discuss?hl=en
>
> ---
> You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--

Google Germany GmbH, Erika-Mann-Strasse 33, 80636 Munich

Registergericht und -nummer: Hamburg, HRB 86891

Sitz der Gesellschaft: Hamburg

Geschäftsführer: Paul Manicle, Halimah DeLaine Prado

[Derek]

I'm still trying to form a cohesive timeline of these events. In 2013, it was said that trying to deprecate refs/for/* was a mistake, https://groups.google.com/d/msg/repo-discuss/yf3dV8R6opE/Fb-2hYnhEjkJ. So my current understanding is that refs/for/* was intended to be deprecated in favor of refs/draft/* and refs/publish/*, it was then undeprecated https://groups.google.com/d/topic/repo-discuss/yf3dV8R6opE/discussion and the namespaces refs/publish/* and refs/draft/* were abandoned in favor of refs/for/* again. Is there any reason this isn't the case?

[Jonathan Nieder]

Hi,

I think there are two parts to the answer to your question:

 1. Yes, people tend to use refs/for/ more than refs/publish/ these days, and "git review" probably should follow suit.

 2. As a heads up, we're probably going to introduce refs/review/ soon, but you don't have any obligation to migrate to it quickly in the internals of tools like "git review". When running sufficiently new versions of Git, we're likely to make something like "git push -o for-review origin master" work, too --- see https://public-inbox.org/git/20180717210915....@google.com/.

Thanks and hope that clarifies,

Jonathan

вт, 24 июл. 2018 г. в 11:21, Derek <derek.wa...@gmail.com>:

[Derek]

Yes, that greatly clarifies. Thanks

On Tuesday, July 24, 2018 at 1:25:50 PM UTC-5, Jonathan Nieder wrote:

Hi,

I think there are two parts to the answer to your question:

 1. Yes, people tend to use refs/for/ more than refs/publish/ these days, and "git review" probably should follow suit.

 2. As a heads up, we're probably going to introduce refs/review/ soon, but you don't have any obligation to migrate to it quickly in the internals of tools like "git review". When running sufficiently new versions of Git, we're likely to make something like "git push -o for-review origin master" work, too --- see https://public-inbox.org/git/20180717210915.139521-1-bmw...@google.com/.