`master` is now `main`

`master` is now `main`

[Stefan Monnier]

The default branch of both the `elpa.git` and `nongnu.git` repositories
has been changed to `main`.  The old `master` branches have been
replaced by placeholders pointing to the new branch.

People who clone the repository anew should not be noticeably affected,
but for those who already have a clone, you'll have to switch to the
new branch.

For `nongnu.git` the new branch is a "fast-forward" of the old branch,
so all it takes is:

    git branch -set-upstream-to=origin/main
    git merge

For `elpa.git` OTOH the new branch doesn't share any history with the
new one, so you'll need something like:

    git checkout -b main origin/main

If you have local changes, you'll want to `git stash` before doing that
and `git stash pop` afterwards and it will likely give you non-trivial
conflicts to resolve.  Don't hesitate to ask if you need help resolving
such conflicts.

Note also that the new branch uses new code, so you'll need to do one

    make

as a first step (to fetch the scripts and setup the real makefile) and
after that it should work more or less as before, but you'll likely see
various differences.  I'm still working on improving those new scripts,
so if you encounter any particular problem/regression please holler:
there's a good chance I'm already aware of it, but it'll help me focus
on the cases that matter.


        Stefan

Re: `master` is now `main`

[Pankaj Jangid]

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

> For `nongnu.git` the new branch is a "fast-forward" of the old branch,
> so all it takes is:
>
>     git branch -set-upstream-to=origin/main
>     git merge

Worked. My ‘master’ is now tracking origin/main.

> For `elpa.git` OTOH the new branch doesn't share any history with the
> new one, so you'll need something like:
>
>     git checkout -b main origin/main

$ git checkout -b main origin/main
$ git pull
$ git checkout master
$ git pull
fatal: refusing to merge unrelated histories

> If you have local changes, you'll want to `git stash` before doing that
> and `git stash pop` afterwards and it will likely give you non-trivial
> conflicts to resolve.  Don't hesitate to ask if you need help resolving
> such conflicts.
>

Re: `master` is now `main`

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 964 bytes --]

* 2020-12-19 12:47:27+0530, Pankaj Jangid wrote:

> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> For `elpa.git` OTOH the new branch doesn't share any history with the
>> new one, so you'll need something like:
>>
>>     git checkout -b main origin/main
>
> $ git checkout -b main origin/main
> $ git pull
> $ git checkout master
> $ git pull
> fatal: refusing to merge unrelated histories

Your local "master" is now different from the upstream (origin)
"master". They don't have related history anymore as the error message
says. If your local "master" does not have any of your important commits
you can reset it and its working tree with "git reset --hard":

    git checkout master
    git reset --hard origin/master      # Provided that your upstream is
                                          called "origin".

-- 
/// Teemu Likonen - .-.. https://www.iki.fi/tlikonen/
// OpenPGP: 4E1055DC84E9DFF613D78557719D69D324539450

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 251 bytes --]

Re: `master` is now `main`

[Eli Zaretskii]

> From: Pankaj Jangid <pankaj@codeisgreat.org>
> Date: Sat, 19 Dec 2020 12:47:27 +0530
> Cc: emacs-devel@gnu.org
> 
> $ git checkout -b main origin/main
> $ git pull
> $ git checkout master
> $ git pull
> fatal: refusing to merge unrelated histories

Why do you need the master branch after Stefan's changes?

Re: `master` is now `main`

[Pankaj Jangid]

Teemu Likonen <tlikonen@iki.fi> writes:

>>> For `elpa.git` OTOH the new branch doesn't share any history with the
>>> new one, so you'll need something like:
>>>
>>>     git checkout -b main origin/main
>>
>> $ git checkout -b main origin/main
>> $ git pull
>> $ git checkout master
>> $ git pull
>> fatal: refusing to merge unrelated histories
>
> Your local "master" is now different from the upstream (origin)
> "master". They don't have related history anymore as the error message
> says. If your local "master" does not have any of your important commits
> you can reset it and its working tree with "git reset --hard":
>
>     git checkout master
>     git reset --hard origin/master      # Provided that your upstream is
>                                           called "origin".

$ git checkout master
$ git reset --hard origin/master
HEAD is now at f1727b0c61 New dummy branch to announce deletion
$ git pull
Already up to date.
$ ls
’BRANCH DELETED’ ’USE BRANCH ‘main‘ INSTEAD’

I don’t have any content in this repository. But just curious, where can
people get the old history; which branch?

Re: `master` is now `main`

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Date: Sat, 19 Dec 2020 00:33:47 -0500
> 
>     git checkout -b main origin/main
> 
> If you have local changes, you'll want to `git stash` before doing that
> and `git stash pop` afterwards and it will likely give you non-trivial
> conflicts to resolve.  Don't hesitate to ask if you need help resolving
> such conflicts.
> 
> Note also that the new branch uses new code, so you'll need to do one
> 
>     make
> 
> as a first step (to fetch the scripts and setup the real makefile) and
> after that it should work more or less as before

Could you please show a list of commands to run manually instead of
that "make" step?  For various boring reasons, I'd prefer not to run
Make in the repository, but instead perform the commands it invokes by
hand.

Thanks.

Re: `master` is now `main`

[Andreas Schwab]

$ make
make: *** No rule to make target 'elcs', needed by 'all-in-place'. Stop.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: `master` is now `main`

[Teemu Likonen]

[-- Attachment #1: Type: text/plain, Size: 922 bytes --]

* 2020-12-19 14:06:16+0530, Pankaj Jangid wrote:

> I don’t have any content in this repository. But just curious, where
> can people get the old history; which branch?

I think the old master is not in the upstream anymore but everybody who
used to have it can still access the objects before they are garbage
collected. The latest commit I can find from my repository's old master
branch is 6749a2982a43923. I used "git log --walk-reflogs" to find where
my branches' head has been in time. You could tag that commit so that it
and all its parents stay in your repository:

    git tag old-master 6749a2982a43923cb4523519f251858a96cf3337

But I can't tell if you have that commit. Use

    git log --walk-reflogs origin/master

to see where your version of the branch pointed before.

-- 
/// Teemu Likonen - .-.. https://www.iki.fi/tlikonen/
// OpenPGP: 4E1055DC84E9DFF613D78557719D69D324539450

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 251 bytes --]

Re: `master` is now `main`

[Stefan Monnier]

>> Note also that the new branch uses new code, so you'll need to do one
>> 
>>     make
>> 
>> as a first step (to fetch the scripts and setup the real makefile) and
>> after that it should work more or less as before
>
> Could you please show a list of commands to run manually instead of
> that "make" step?

Here's the complete `Makefile`:

    # "Dummy" initial Makefile for GNU Emacs Lisp Package Archive.
    setup: GNUmakefile admin
    GNUmakefile: admin
            ln -s admin/GNUmakefile ./
    admin:
            git worktree add -b elpa-admin admin origin/elpa-admin


;-)


        Stefan

Re: `master` is now `main`

[Stefan Monnier]

>> so if you encounter any particular problem/regression please holler:
>> there's a good chance I'm already aware of it, but it'll help me focus
>> on the cases that matter.

> $ make
> make: *** No rule to make target 'elcs', needed by 'all-in-place'. Stop.

So this is indeed a case I'm already aware of.  I'll try and fix it ASAP.


        Stefan

Re: `master` is now `main`

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org
> Date: Sat, 19 Dec 2020 10:44:17 -0500
> 
> > Could you please show a list of commands to run manually instead of
> > that "make" step?
> 
> Here's the complete `Makefile`:
> 
>     # "Dummy" initial Makefile for GNU Emacs Lisp Package Archive.
>     setup: GNUmakefile admin
>     GNUmakefile: admin
>             ln -s admin/GNUmakefile ./
>     admin:
>             git worktree add -b elpa-admin admin origin/elpa-admin

Granted, I looked into that, in the hope that I could figure it out
myself.  But there's no admin subdirectory in my clone of the
repository, so I got confused...

I figured it out now.

Re: `master` is now `main`

[John Wiegley]

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

SM> The default branch of both the `elpa.git` and `nongnu.git` repositories
SM> has been changed to `main`. The old `master` branches have been replaced
SM> by placeholders pointing to the new branch.

Thank you, Stefan. While I agree with the sentiment that this but a drop in a
very large bucket and has no substantial impact for the cause of racial unity,
still it reflect our intention to genuinely care about and consider these
issue, and to voice our support for those who do fight the real fight.

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

Re: `master` is now `main`

[Stefan Monnier]

>> make: *** No rule to make target 'elcs', needed by 'all-in-place'. Stop.
> So this is indeed a case I'm already aware of.  I'll try and fix it ASAP.

Should work better now,


        Stefan

Re: `master` is now `main`

[Arthur Miller]

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

>>> make: *** No rule to make target 'elcs', needed by 'all-in-place'. Stop.
>> So this is indeed a case I'm already aware of.  I'll try and fix it ASAP.
>
> Should work better now,
>
>
>         Stefan

I didn't want to bother with renaming etc, just cloned repo, minut ago,
cd into Emacs and was still in 'master':

 2020-12-20 14:13:48 ⌚  pascal in ~/repos/new
○ → git clone https://git.savannah.gnu.org/git/emacs.git
Klonar till "emacs"...
remote: Counting objects: 914123, done.
remote: Compressing objects: 100% (162363/162363), done.
remote: Total 914123 (delta 751347), reused 913283 (delta 750557)
Tar emot objekt: 100% (914123/914123), 299,77 MiB | 10,38 MiB/s, klart.
Analyserar delta: 100% (751347/751347), klart.

 2020-12-20 14:17:54 ⌚  pascal in ~/repos/new
○ → cd ./emacs/

 2020-12-20 14:59:20 ⌚  pascal in ~/repos/new/emacs
± |master ✓| → git checkout feature/native-comp
Grenen feature/native-comp ställdes in att spåra fjärrgrenen "feature/native-comp" från "origin".
Växlade till en ny gren "feature/native-comp"

Re: `master` is now `main`

[Filipp Gunbin]

In elpa, GNUMakefile contains:

EMACS=emacs --batch

I don't have Emacs installed (use the source tree build), it'd nice to
be able to account for that.

Thanks,
Filipp

Re: `master` is now `main`

[Stefan Monnier]

> EMACS=emacs --batch
> I don't have Emacs installed (use the source tree build), it'd nice to
> be able to account for that.

How do you suggest we do that?


        Stefan

Re: `master` is now `main`

[Filipp Gunbin]

On 21/12/2020 17:15 -0500, Stefan Monnier wrote:

>> EMACS=emacs --batch
>> I don't have Emacs installed (use the source tree build), it'd nice to
>> be able to account for that.
>
> How do you suggest we do that?

Maybe

EMACS=${EMACS:-emacs}
EMACS_OPTS=${EMACS_OPTS:---batch}

So you could override them in the shell command.

Filipp

Re: `master` is now `main`

[Thien-Thi Nguyen]

[-- Attachment #1: Type: text/plain, Size: 1015 bytes --]


Anyone else seeing a failed fresh clone?  :-(

 $ git clone -o sv git+ssh://git.sv.gnu.org/srv/git/emacs/elpa
 Clone in 'elpa' in corso...
 remote: Counting objects: 335978, done.        
 remote: Compressing objects: 100% (94283/94283), done.        
 remote: Total 335978 (delta 238158), reused 335412 (delta 237677)        
 Ricezione degli oggetti: 100% (335978/335978), 227.92 MiB | 4.61 MiB/s, fatto.
 Risoluzione dei delta: 100% (238158/238158), fatto.
 fatal: non si è ricevuto l'oggetto atteso 1b04f0fcd58ea0c82e522bd6e49dc08a3c7912ec
 fatal: index-pack non riuscito

(Grumble, old computers, bad RAM, grumble.)

-- 
Thien-Thi Nguyen -----------------------------------------------
 (defun responsep (query)               ; (2021) Software Libero
   (pcase (context query)               ;       = Dissenso Etico
     (`(technical ,ml) (correctp ml))
     ...))                              748E A0E8 1CB8 A748 9BFA
--------------------------------------- 6CE4 6703 2224 4C80 7502


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 219 bytes --]

Re: `master` is now `main`

[Stefan Kangas]

Thien-Thi Nguyen <ttn@gnuvola.org> writes:

> Anyone else seeing a failed fresh clone?  :-(
>
>  $ git clone -o sv git+ssh://git.sv.gnu.org/srv/git/emacs/elpa
>  Clone in 'elpa' in corso...
>  remote: Counting objects: 335978, done.
>  remote: Compressing objects: 100% (94283/94283), done.
>  remote: Total 335978 (delta 238158), reused 335412 (delta 237677)
>  Ricezione degli oggetti: 100% (335978/335978), 227.92 MiB | 4.61 MiB/s, fatto.
>  Risoluzione dei delta: 100% (238158/238158), fatto.
>  fatal: non si è ricevuto l'oggetto atteso 1b04f0fcd58ea0c82e522bd6e49dc08a3c7912ec
>  fatal: index-pack non riuscito
>
> (Grumble, old computers, bad RAM, grumble.)

I believe you need to use this:

    git -c fetch.fsckObjects=false clone .../elpa

The fetch.fsckObjects setting works around a glitch in the ELPA
repository <https://debbugs.gnu.org/22690>.

(IMO, we should just bite the bullet and recreate the repository with
the correct settings, but that's me.)

Re: `master` is now `main`

[Thien-Thi Nguyen]

[-- Attachment #1: Type: text/plain, Size: 1768 bytes --]


() Stefan Kangas <stefankangas@gmail.com>
() Sun, 3 Jan 2021 10:29:22 +0100

   I believe you need to use this:

       git -c fetch.fsckObjects=false clone .../elpa

   The fetch.fsckObjects setting works around a glitch in the
   ELPA repository <https://debbugs.gnu.org/22690>.

   (IMO, we should just bite the bullet and recreate the
   repository with the correct settings, but that's me.)

Thanks for the hint and the link.  Unfortunately, adding that
option doesn't work; same outcome.  (I notice, however, that the
number of objects has grown since i last tried -- this trend
will surely continue.)

 $ git -c fetch.fsckObjects=false clone -o sv git+ssh://git.sv.gnu.org/srv/git/emacs/elpa
 Clone in 'elpa' in corso...
 remote: Counting objects: 336037, done.        
 remote: Compressing objects: 100% (94335/94335), done.        
 remote: Total 336037 (delta 238188), reused 335407 (delta 237677)        
 Ricezione degli oggetti: 100% (336037/336037), 227.93 MiB | 4.45 MiB/s, fatto.
 Risoluzione dei delta: 100% (238188/238188), fatto.
 fatal: non si è ricevuto l'oggetto atteso 0600204de07aa6c258cfe2700d273106fb2ba3ac
 fatal: index-pack non riuscito

I'm unsure if this is just my computer or a general dropping of
support in newer Git versions for the old repo format -- will
investigate.  Maybe time to get a new disk drive or RAM or ...?

-- 
Thien-Thi Nguyen -----------------------------------------------
 (defun responsep (query)               ; (2021) Software Libero
   (pcase (context query)               ;       = Dissenso Etico
     (`(technical ,ml) (correctp ml))
     ...))                              748E A0E8 1CB8 A748 9BFA
--------------------------------------- 6CE4 6703 2224 4C80 7502


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 219 bytes --]

Re: `master` is now `main`

[Stefan Monnier]

>> Anyone else seeing a failed fresh clone?  :-(
>>
>>  $ git clone -o sv git+ssh://git.sv.gnu.org/srv/git/emacs/elpa
>>  Clone in 'elpa' in corso...
>>  remote: Counting objects: 335978, done.
>>  remote: Compressing objects: 100% (94283/94283), done.
>>  remote: Total 335978 (delta 238158), reused 335412 (delta 237677)
>>  Ricezione degli oggetti: 100% (335978/335978), 227.92 MiB | 4.61 MiB/s, fatto.
>>  Risoluzione dei delta: 100% (238158/238158), fatto.
>>  fatal: non si è ricevuto l'oggetto atteso 1b04f0fcd58ea0c82e522bd6e49dc08a3c7912ec
>>  fatal: index-pack non riuscito
>>
>> (Grumble, old computers, bad RAM, grumble.)
>
> I believe you need to use this:
>
>     git -c fetch.fsckObjects=false clone .../elpa

I think the fsck failure leads to a different error message, so
I suspect that's not his problem.

I don't know what could be the cause of his problem, but I'll note that
you can do the following:

    git --single-branch -b main git://git.sv.gnu.org/emacs/elpa.git

which will download very little info (only the `main` branch) so is more
likely to avoid the problem you're seeing (I think).

Of course, that won't get you all the GNU ELPA packages.  But subsequent

    make packages/[PKGNAME]

should then fetch the corresponding package's info.  So if you're not
interested in downloading all the packages but only a handful of them,
it can be good option.

> The fetch.fsckObjects setting works around a glitch in the ELPA
> repository <https://debbugs.gnu.org/22690>.
> (IMO, we should just bite the bullet and recreate the repository with
> the correct settings, but that's me.)

Actually, the fsck failure is for an object that was in the old `master`
branch (now named `old/master-2020-dec-14`), which is not used any more.
We should purge it from the repository, but I think it's a bit too early
for that.  Alternatively, we could setup another repository where could
keep this branch (maybe along with other old/deprecated branches).

Also if someone is interested in recreating the branch without the fsck
problem, please go ahead.  I don't have the motivation to look into how
to do that (it's probably a small matter of rebasing or something like
that).


        Stefan

Re: `master` is now `main`

[Andreas Schwab]

On Jan 02 2021, Thien-Thi Nguyen wrote:

> (Grumble, old computers, bad RAM, grumble.)

Nothing we can do about.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: `master` is now `main`

[Matt Armstrong]

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

> The default branch of both the `elpa.git` and `nongnu.git` repositories
> has been changed to `main`.  The old `master` branches have been
> replaced by placeholders pointing to the new branch.

Stefan, is there any intention to do the same master->main change for
emacs.git?

Re: `master` is now `main`

[Eli Zaretskii]

> From: Matt Armstrong <matt@rfc20.org>
> Date: Sun, 14 Feb 2021 11:03:56 -0800
> 
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
> 
> > The default branch of both the `elpa.git` and `nongnu.git` repositories
> > has been changed to `main`.  The old `master` branches have been
> > replaced by placeholders pointing to the new branch.
> 
> Stefan, is there any intention to do the same master->main change for
> emacs.git?

No.