Re: Continuing the Gitlab experiment: Single-patch PRs for gitlab

[Bertrand Marquis <Bertrand.Marquis@xxxxxxx>]

> On 4 Sep 2020, at 13:42, Jan Beulich <jbeulich@xxxxxxxx> wrote:
> 
> On 04.09.2020 12:43, Bertrand Marquis wrote:
>>> On 4 Sep 2020, at 11:20, Jan Beulich <jbeulich@xxxxxxxx> wrote:
>>> 
>>> On 04.09.2020 11:54, George Dunlap wrote:
>>>> At the community call last month as well as this, we discussed whether to 
>>>> continue the “Gitlab experiment”.  It was generally agreed that reviewing 
>>>> Juergen’s long series was fairly sub-optimal, and that email was more 
>>>> suited to that sort of series.
>>>> 
>>>> That said, there was general agreement that requiring all patches to go 
>>>> through email was going to limit contribution, particularly of one-off 
>>>> “drive-by” contributions.  As such, it was proposed that we consider 
>>>> allowing both Gitlab PRs, and email: that for one-off or short series, 
>>>> Gitlab PRs would be accepted, but that for longer series and/or longer 
>>>> term contributors, we would encourage people to switch to patchbombing the 
>>>> mailing list.
>>>> 
>>>> We decided to continue the “Gitlab Experiment”, but with short PRs.  As 
>>>> such, Andy Cooper has posted two PRs:
>>>> 
>>>> https://gitlab.com/xen-project/xen/-/merge_requests/2
>>> 
>>> This looks to be confusing, to me at least. Following this link I
>>> can't see the actual change directly. Following either of the links
>>> after "Request to merge" gives a 404 error (after gitlab not being
>>> able to sign me in via Google, but then being able to sign me in
>>> via github) on both
>>> 
>>> https://gitlab.com/xen-project/people/andyhhp/xen
>>> https://gitlab.com/xen-project/people/andyhhp/xen/-/tree/xen-pv-segbase
>> 
>> I think you missed on top the “Commits” and “Changes"
> 
> I did indeed. Nevertheless I think the other links ought to be
> working, as they're (I assume) going to give me a view of the
> tree as a whole with the patch/series applied.
> 
> That said, from an engineer's point of view I think the main
> thing of interest should be shown in the "primary" view, and
> that's (imo) clearly the entire patch.

If you look at the serie sent by Trammel this is a bit more clear:
- overview is for global comments
- commits list the different patches of the serie
- when you click on a commit you see the diff for this commit and you can 
comment

Having everything on one page would fit for a small patch but for a serie that 
would be to much.

I agree though the interface could be better but maybe we just have to get used 
to it.

Cheers

Bertrand

> 
> Jan