Re: [edk2-announce] Research Request

[Laszlo Ersek <lersek@redhat.com>]

On 12/06/18 15:05, Laszlo Ersek wrote:
> On 12/05/18 18:26, Rebecca Cran wrote:
>> On Wednesday, 5 December 2018 05:55:41 MST Laszlo Ersek wrote:
>>
>>> (1) Pls. explain to me how I can create an edk2 clone at
>>> <code.bluestop.org>. :)
>>
>> You don't. In a production system it may be possible to clone from either 
>> GitHub or code.bluestop.org (which mirrors github), but the clone URL given 
>> when you click "Clone" on https://code.bluestop.org/diffusion/EDK/ doesn't 
>> work (since I've not configured it).
> 
> Well, I don't specifically desire creating an edk2 clone on
> <code.bluestop.org> *using the WebUI*. However, in order to share my
> work (i.e. to submit a pull request that refers to a topic branch of
> mine), I need to have a publicly available / fetchable git repository.
> (This is no different from mailing list based pull requests BTW.)
> 
> So, let me reformulate: can I *have* (by any means) a personal edk2
> clone on <code.bluestop.org>, and can I push my topic branches there?
> (Obviously it's not specifically about me nor specifically about
> <code.bluestop.org>, but about any contributor with a topic branch to
> submit, and about any site that would possibly run the central edk2
> Phabricator instance.)
> 
> Right now I'm confused whether Phabricator (in general) offers
> repository storage for contributors, or if that would have to come from
> another service. (That wouldn't be too convenient.)
> 
> Anyway, the goal of a personal edk2 repo (clone) for me on
> <code.bluestop.org> would be that I should be able to receive pull
> requests against it.
> 
> Am I misunderstanding something?
> 
>>  
>>> (2) Please create a throw-away account for yourself.
>>
>> Done (though not throw-away).
> 
> I suggested "throw-away" because one of the later steps involves
> deleting it.
> 
>>  
>>> (3) Submit a pullreq against (1), with a topic branch that has two
>>> commits, and simple text file modifications.
>>
>> https://code.bluestop.org/D1
>> Since Phabricator doesn't care about topic branches (just patches),
> 
> Please wait a second, I don't understand. What do you mean by "doesn't
> care about topic branches, just patches"?
> 
> Compare three scenarios:
> 
> (a) Someone implements a new feature in 10 patches, and sends each patch
> individually to the mailing list, without a common cover letter, and
> without numbering in the subject lines. That's what I'd call "doesn't
> care about branches just patches", and it's unusable for development.
> 
> (b) Someone sends a normal patch series, they just don't state what
> upstream commit the series applies to. I can sort-of see this as "no
> topic branch, just patches". Is this what you mean? Does Phabricator
> maintain the series of patches as such (without a base commit), i.e. the
> set of patches in the series, and their relative order?
> 
> (c) An actual pull request that refers to a specific commit hash (which
> may or may not, although it almost always is, identified by a branch
> head or tag).

Sigh, I failed to finish (c). I meant to ask, assuming I upload my work
(a topic branch) via git-push to the phabricator instance that hosts my
personal clone -- when I submit the pull request, does the pull req
preserve the specific commit hash (and hence the full git history
leading up to it), or does the request decay to (b), similarly to the
"normal" mailing list posting?

Thanks,
Laszlo