non-fast-forward pushes to Emacs Git repo "scratch" branches?

non-fast-forward pushes to Emacs Git repo "scratch" branches?

[Ted Zlatanov]

Can the Emacs Git repo allow non-fast-forward pushes to "scratch/*"
branches? Those are never released so it wouldn't compromise the
integrity of the public source code.

Thanks
Ted

Re: non-fast-forward pushes to Emacs Git repo "scratch" branches?

[Stefan Monnier]

> Can the Emacs Git repo allow non-fast-forward pushes to "scratch/*"
> branches?

No, because the non-fast-forward constraint is applied globally.
But you can delete a branch and then push a new branch under the same
name, with the same end result.

I don't know if John would object to it, and I don't think this has
already been used in the past, but when I suggested the use of
the "scratch/" prefix, I did assume that such "rebasing" would be used
there occasionally.


        Stefan

Re: non-fast-forward pushes to Emacs Git repo "scratch" branches?

[John Wiegley]

[-- Attachment #1: Type: text/plain, Size: 540 bytes --]

>>>>> Stefan Monnier <monnier@iro.umontreal.ca> writes:

> I don't know if John would object to it, and I don't think this has already
> been used in the past, but when I suggested the use of the "scratch/"
> prefix, I did assume that such "rebasing" would be used there occasionally.

I don't object to anything done in scratch. It's scratch; track its branches
with caution.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 629 bytes --]

Re: non-fast-forward pushes to Emacs Git repo "scratch" branches?

[Ted Zlatanov]

On Fri, 17 Jun 2016 18:25:13 -0400 Stefan Monnier <monnier@iro.umontreal.ca> wrote: 

>> Can the Emacs Git repo allow non-fast-forward pushes to "scratch/*"
>> branches?

SM> No, because the non-fast-forward constraint is applied globally.
SM> But you can delete a branch and then push a new branch under the same
SM> name, with the same end result.

SM> I don't know if John would object to it, and I don't think this has
SM> already been used in the past, but when I suggested the use of
SM> the "scratch/" prefix, I did assume that such "rebasing" would be used
SM> there occasionally.

Perfect, thanks.

Ted

Re: non-fast-forward pushes to Emacs Git repo "scratch" branches?

[Phillip Lord]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> Can the Emacs Git repo allow non-fast-forward pushes to "scratch/*"
>> branches?
>
> No, because the non-fast-forward constraint is applied globally.
> But you can delete a branch and then push a new branch under the same
> name, with the same end result.

It is possible to enforce this on a per branch basis. Just applying it
on master and anything starting with "emacs" (for release branches)
would do the trick.

It would also be possible to prevent, for example, accidental deletions
of emacs-25, or master as well as enforce the "don't check into emacs-25
without asking" rule.

All good things for newer developers.

Phil

Re: non-fast-forward pushes to Emacs Git repo "scratch" branches?

[Ted Zlatanov]

On Mon, 20 Jun 2016 16:21:25 +0100 phillip.lord@russet.org.uk (Phillip Lord) wrote: 

PL> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>>> Can the Emacs Git repo allow non-fast-forward pushes to "scratch/*"
>>> branches?
>> 
>> No, because the non-fast-forward constraint is applied globally.
>> But you can delete a branch and then push a new branch under the same
>> name, with the same end result.

PL> It is possible to enforce this on a per branch basis. Just applying it
PL> on master and anything starting with "emacs" (for release branches)
PL> would do the trick.

PL> It would also be possible to prevent, for example, accidental deletions
PL> of emacs-25, or master as well as enforce the "don't check into emacs-25
PL> without asking" rule.

I'd be excited to see these improvements. Another is to enforce the
commit message formatting.

I don't have access to the relevant infrastructure, unfortunately. I
imagine it's a matter of modifying the receive hook scripts. Can I help
make it happen?

Thanks
Ted