From pierre-yves.david at ens-lyon.org  Thu Oct  2 15:06:48 2025
From: pierre-yves.david at ens-lyon.org (Pierre-Yves David)
Date: Thu, 2 Oct 2025 17:06:48 +0200
Subject: Mini Sprint at Pycon.fr (October 30th-31th in Lyon.fr)
Message-ID: <5cb63e84-9960-43b3-ab23-cca28e1f099c@ens-lyon.org>

Hello everyone

We have been given some space for sprinting at the PyconFr Conference in 
Lyon at the end of October

The sprint are held on Thursday October 30th? and Friday October 31th.

The sprints are followed by two days of talks for those interested. The 
whole conference is free to attend (but registration is mandatory).

PyconFr is primarily a French speaking conference, but has a good share 
of talks in English too.


Georges and I will be attending the mini sprint, so we expect to cover 
both Mercurial and Heptapod related topics. (sadly, Rapha?l will be out 
of his very home town that week).


If you are planning to attend, please with the Conference and add 
yourself to the sprint etherpad.

https://semestriel.framapad.org/p/ <DELETE THIS PART> ????? 57hy9r6qdf-ah2c


More details on the conference can be found there 
https://www.pycon.fr/2025/en/index.html


Cheers

-- 
Pierre-Yves David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251002/b647205b/attachment.htm>

From oub at mat.ucm.es  Mon Oct  6 15:27:28 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 06 Oct 2025 17:27:28 +0200
Subject: how to share a file between 2 repositories?  Hardlinks?
Message-ID: <87sefwqcdb.fsf@mat.ucm.es>


Hi 


Here is the scenario. I have a single file that I would like to use in two repositories:

```
repo1
repo2
```

Both repositories will, however, be pushed to Helix.

Using symbolic links between the files is of little use, since the content of the symbolic link will not be transmitted.

I therefore considered using hard links. Is this a sensible approach?

I would prefer to avoid employing a subrepository, as I have encountered problems with them in the past.


Thanks

Uwe Brauer 

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251006/1359ec54/attachment.p7s>

From pierre-yves.david at ens-lyon.org  Tue Oct  7 08:26:41 2025
From: pierre-yves.david at ens-lyon.org (Pierre-Yves David)
Date: Tue, 7 Oct 2025 10:26:41 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87sefwqcdb.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
Message-ID: <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>

Using a hardlink will work, as the hardlink works a the filesystem level 
and Mercurial will be oblivious of that "implementation" details.

However, it will be up to you to preserve that hardlink in sync. Other 
program like your test editor, or generation of the file might break 
that hardlink, Even `hg update` could end up doing that.

So it depends of what kind of file it is, how and how often it gets 
updated. And depending on that you might want to setup hooks to make 
sure the two "file" (i.e. file system path) still point to the same 
"content" (i.e. file system inode) before committing, or after update.

Can you tell use more about this file?


On 10/6/25 17:27, Uwe Brauer via Mercurial wrote:
> Hi
>
>
> Here is the scenario. I have a single file that I would like to use in two repositories:
>
> ```
> repo1
> repo2
> ```
>
> Both repositories will, however, be pushed to Helix.
>
> Using symbolic links between the files is of little use, since the content of the symbolic link will not be transmitted.
>
> I therefore considered using hard links. Is this a sensible approach?
>
> I would prefer to avoid employing a subrepository, as I have encountered problems with them in the past.
>
>
> Thanks
>
> Uwe Brauer
>
>
> _______________________________________________
> Mercurial mailing list
> Mercurial at lists.mercurial-scm.org
> https://lists.mercurial-scm.org/mailman/listinfo/mercurial

-- 
Pierre-Yves David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251007/4e4b8aed/attachment.htm>

From oub at mat.ucm.es  Tue Oct  7 08:39:29 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Tue, 07 Oct 2025 10:39:29 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
Message-ID: <87v7krcdha.fsf@mat.ucm.es>

>>> "PD" == Pierre-Yves David <pierre-yves.david at ens-lyon.org> writes:

> Using a hardlink will work, as the hardlink works a the filesystem
> level and Mercurial will be oblivious of that "implementation"
> details.

> However, it will be up to you to preserve that hardlink in sync. Other
> program like your test editor, or generation of the file might break
> that hardlink, Even `hg update` could end up doing that.

> So it depends of what kind of file it is, how and how often it gets
> updated. And depending on that you might want to setup hooks to make
> sure the two "file" (i.e. file system path) still point to the same
> "content" (i.e. file system inode) before committing, or after update.

> Can you tell use more about this file?

Thanks for the detailed response. I?m probably something of an outlier
here, as I mainly use Mercurial for scientific collaboration involving
LaTeX?along with some Emacs Lisp and MATLAB coding.

The file in question is a BibTeX file containing entries such as:


--8<---------------cut here---------------start------------->8---
@InCollection{christodoulou90:_minkow,
  title        = {The global nonlinear stability of the {M}inkowski
                  space},
  author       = {Christodoulou, D. and Klainerman, S.},
  booktitle    = {S?minaire sur les ?quations aux {D}?riv?es
                  {P}artielles, 1989--1990},
  publisher    = {?cole Polytech., Palaiseau},
  year         = 1990,
  pages        = {Exp.\ No.\ XIII, 31},
  Mrclass      = {83C05 (58G16 58G17)},
  Mrnumber     = 1073188
}
--8<---------------cut here---------------end--------------->8---


The editor I use is GNU Emacs. When I committed the file in one
repository using Emacs VC and then opened the hard-linked file in the
other repository, Emacs refused to commit, claiming there were no
changes?even though I could commit from the command line.

I assume this is an Emacs issue.

Regards

Uwe Brauer 
-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 


From arne_bab at web.de  Tue Oct  7 12:12:02 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Tue, 07 Oct 2025 14:12:02 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87v7krcdha.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es>
Message-ID: <87y0pm3o8d.fsf@web.de>

Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:

> The editor I use is GNU Emacs. When I committed the file in one
> repository using Emacs VC and then opened the hard-linked file in the
> other repository, Emacs refused to commit, claiming there were no
> changes?even though I could commit from the command line.
>
> I assume this is an Emacs issue.

This sounds like vc-mode detecting that it?s a hardlink and using the
"canonical" location -- which wasn?t the file you opened, but the other
repository.

I don?t have a good solution for the problem: I tend to just copy the
bibtex files and merge them manually from time to time using M-x
ediff-merge.

An option could be to use a common bibtex repository, always commit the
changes to the bibtex file individually in the work repository and
export+import those changes to the other repository. Then you can always
import+merge the bibtex repo in all other repos.

To automate it, you could create a commit hook that does the
export+import; in .hg/hgrc:

[hooks]
commit = hg log -r $HG_NODE --template '{files}' | grep -q 'bibtex.bib' && hg export $HG_NODE > /tmp/bibtex.patch && hg -R ../bib import /tmp/bibtex.patch && rm /tmp/bibtex.patch || true

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251007/e3e294c0/attachment.sig>

From arne_bab at web.de  Tue Oct  7 12:12:02 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Tue, 07 Oct 2025 14:12:02 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87v7krcdha.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es>
Message-ID: <87y0pm3o8d.fsf@web.de>

Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:

> The editor I use is GNU Emacs. When I committed the file in one
> repository using Emacs VC and then opened the hard-linked file in the
> other repository, Emacs refused to commit, claiming there were no
> changes?even though I could commit from the command line.
>
> I assume this is an Emacs issue.

This sounds like vc-mode detecting that it?s a hardlink and using the
"canonical" location -- which wasn?t the file you opened, but the other
repository.

I don?t have a good solution for the problem: I tend to just copy the
bibtex files and merge them manually from time to time using M-x
ediff-merge.

An option could be to use a common bibtex repository, always commit the
changes to the bibtex file individually in the work repository and
export+import those changes to the other repository. Then you can always
import+merge the bibtex repo in all other repos.

To automate it, you could create a commit hook that does the
export+import; in .hg/hgrc:

[hooks]
commit = hg log -r $HG_NODE --template '{files}' | grep -q 'bibtex.bib' && hg export $HG_NODE > /tmp/bibtex.patch && hg -R ../bib import /tmp/bibtex.patch && rm /tmp/bibtex.patch || true

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251007/e3e294c0/attachment-0001.sig>

From arne_bab at web.de  Wed Oct  8 22:11:32 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 00:11:32 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87y0pm3o8d.fsf@web.de>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
Message-ID: <87seftcacr.fsf@web.de>

"Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:

> An option could be to use a common bibtex repository, always commit the
> changes to the bibtex file individually in the work repository and
> export+import those changes to the other repository.

That?s actually far too complex. Feels like my brain got poisoned by the
dozens of git clutches at work.

The clean way would be a bibtex repo with a named branch "bibtex" which
holds the bibtex file.

To get bibtex initially, you simply hg pull --force from the bibtex repo
and then hg merge bibtex into default.

To get changes back into the bibtex repo, just hg pull from the bibtex
repo, hg graft the commit onto the bibtex branch and hg push -b bibtex
(only push the bibtex branch).

(? adapt the commit hook to do that)

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/8a6fec69/attachment.sig>

From arne_bab at web.de  Wed Oct  8 22:11:32 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 00:11:32 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87y0pm3o8d.fsf@web.de>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
Message-ID: <87seftcacr.fsf@web.de>

"Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:

> An option could be to use a common bibtex repository, always commit the
> changes to the bibtex file individually in the work repository and
> export+import those changes to the other repository.

That?s actually far too complex. Feels like my brain got poisoned by the
dozens of git clutches at work.

The clean way would be a bibtex repo with a named branch "bibtex" which
holds the bibtex file.

To get bibtex initially, you simply hg pull --force from the bibtex repo
and then hg merge bibtex into default.

To get changes back into the bibtex repo, just hg pull from the bibtex
repo, hg graft the commit onto the bibtex branch and hg push -b bibtex
(only push the bibtex branch).

(? adapt the commit hook to do that)

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/8a6fec69/attachment-0001.sig>

From oub at mat.ucm.es  Thu Oct  9 18:25:15 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 09 Oct 2025 20:25:15 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de>
Message-ID: <87bjmgq6es.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>> An option could be to use a common bibtex repository, always commit the
>> changes to the bibtex file individually in the work repository and
>> export+import those changes to the other repository.

> That?s actually far too complex. Feels like my brain got poisoned by the
> dozens of git clutches at work.

I am deeply sorry ?


> The clean way would be a bibtex repo with a named branch "bibtex" which
> holds the bibtex file.


The point is the bibtex repo had only the default branch, I could use
evolve to rename the branch. Right now I only added the bibtex branch
and one  new change set 
> To get bibtex initially, you simply hg pull --force from the bibtex repo
> and then hg merge bibtex into default.

I run a test, cloned my latex repository and then in this cloned one.


hg pull /home/user/HGs/tex/bib/ --force -b bibtex


--8<---------------cut here---------------start------------->8---
pulling from /home/oub/ALLES/HGs/tex/bib/
searching for changes
warning: repository is unrelated
adding changesets
adding manifests
adding file changes
abort: empty revision on one side of range
warning: outgoing hook exited with status 255
added 399 changesets with 529 changes to 22 files (+1 heads)
32 new obsolescence markers
new changesets e3d1de58c858:9b7a5fbc6b01
(run 'hg heads' to see heads, 'hg merge' to merge)
abort: empty revision on one side of range
warning: changegroup.diffstat hook exited with status 255
--8<---------------cut here---------------end--------------->8---

So that would add quite a bit of changesets to my latex repository.


Regards


Uwe Brauer 
-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/91313f82/attachment.p7s>

From oub at mat.ucm.es  Thu Oct  9 18:25:15 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 09 Oct 2025 20:25:15 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de>
Message-ID: <87bjmgq6es.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>> An option could be to use a common bibtex repository, always commit the
>> changes to the bibtex file individually in the work repository and
>> export+import those changes to the other repository.

> That?s actually far too complex. Feels like my brain got poisoned by the
> dozens of git clutches at work.

I am deeply sorry ?


> The clean way would be a bibtex repo with a named branch "bibtex" which
> holds the bibtex file.


The point is the bibtex repo had only the default branch, I could use
evolve to rename the branch. Right now I only added the bibtex branch
and one  new change set 
> To get bibtex initially, you simply hg pull --force from the bibtex repo
> and then hg merge bibtex into default.

I run a test, cloned my latex repository and then in this cloned one.


hg pull /home/user/HGs/tex/bib/ --force -b bibtex


--8<---------------cut here---------------start------------->8---
pulling from /home/oub/ALLES/HGs/tex/bib/
searching for changes
warning: repository is unrelated
adding changesets
adding manifests
adding file changes
abort: empty revision on one side of range
warning: outgoing hook exited with status 255
added 399 changesets with 529 changes to 22 files (+1 heads)
32 new obsolescence markers
new changesets e3d1de58c858:9b7a5fbc6b01
(run 'hg heads' to see heads, 'hg merge' to merge)
abort: empty revision on one side of range
warning: changegroup.diffstat hook exited with status 255
--8<---------------cut here---------------end--------------->8---

So that would add quite a bit of changesets to my latex repository.


Regards


Uwe Brauer 
-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/91313f82/attachment-0001.p7s>

From arne_bab at web.de  Thu Oct  9 19:22:09 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 21:22:09 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87bjmgq6es.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
Message-ID: <878qhjdgny.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:

>> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>>> An option could be to use a common bibtex repository, always commit the
>>> changes to the bibtex file individually in the work repository and
>>> export+import those changes to the other repository.
>
>> That?s actually far too complex. Feels like my brain got poisoned by the
>> dozens of git clutches at work.
>
> I am deeply sorry ?

?

>> To get bibtex initially, you simply hg pull --force from the bibtex repo
>> and then hg merge bibtex into default.
> hg pull /home/user/HGs/tex/bib/ --force -b bibtex
?
> So that would add quite a bit of changesets to my latex repository.

I think you pulled the bibtex branch and all its dependencies.

To be safe, best init a new repository and create the bibtex branch
there.

If you need the bibtex history, you can use convert with an include rule
to create a repo that *only* has the bibtex file.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/34cbf5f9/attachment.sig>

From arne_bab at web.de  Thu Oct  9 19:22:09 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 21:22:09 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87bjmgq6es.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
Message-ID: <878qhjdgny.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:

>> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>>> An option could be to use a common bibtex repository, always commit the
>>> changes to the bibtex file individually in the work repository and
>>> export+import those changes to the other repository.
>
>> That?s actually far too complex. Feels like my brain got poisoned by the
>> dozens of git clutches at work.
>
> I am deeply sorry ?

?

>> To get bibtex initially, you simply hg pull --force from the bibtex repo
>> and then hg merge bibtex into default.
> hg pull /home/user/HGs/tex/bib/ --force -b bibtex
?
> So that would add quite a bit of changesets to my latex repository.

I think you pulled the bibtex branch and all its dependencies.

To be safe, best init a new repository and create the bibtex branch
there.

If you need the bibtex history, you can use convert with an include rule
to create a repo that *only* has the bibtex file.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/34cbf5f9/attachment-0001.sig>

From oub at mat.ucm.es  Thu Oct  9 20:00:27 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 09 Oct 2025 22:00:27 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de>
Message-ID: <875xcnrgkk.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>>> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>>> An option could be to use a common bibtex repository, always commit the
>>> changes to the bibtex file individually in the work repository and
>>> export+import those changes to the other repository.
>> 
>>> That?s actually far too complex. Feels like my brain got poisoned by the
>>> dozens of git clutches at work.
>> 
>> I am deeply sorry ?

> ?

>>> To get bibtex initially, you simply hg pull --force from the bibtex repo
>>> and then hg merge bibtex into default.
>> hg pull /home/user/HGs/tex/bib/ --force -b bibtex
> ?
>> So that would add quite a bit of changesets to my latex repository.

> I think you pulled the bibtex branch and all its dependencies.

> To be safe, best init a new repository and create the bibtex branch
> there.

Well this complicates things a bit and also defies the idea of VC

> If you need the bibtex history, you can use convert with an include rule
> to create a repo that *only* has the bibtex file.


This is also quite large.

I came up with the following ?bright? idea

In the bibtex repo:
 hg up 0
 hg branch bibtex
 hg revert bibtex -r 409 (which is tip)
 hg commit -m "Start the bibtex branch"

Then in the latex repository I run

 hg pull /home/user/bib -b bibtex --force

But then in my latex repository I see 
hg log -G

?  changeset:     1994:6d3d9467e443
?  tag:           tip
?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
?  Branch:        bibtex
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
?  Phase:         public
?  Summary:       Start the bibtex branch
?
?  changeset:     1993:e3d1de58c858
   Branch:        default
   Author:        oub
   Date:          Sun, 19 Dec 2004 10:37:47 +0000
   Phase:         public
   Parent:        -1:000000000000
   Summary:       Initial revision

?  changeset:     1992:a15e729491d4
?  Remote_Branch: default/default
?  Branch:        default
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
?  Phase:         public
?  Summary:       replace \mu'_p by \epsilon_p'
?

That looks also odd.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/afa67c47/attachment.p7s>

From oub at mat.ucm.es  Thu Oct  9 20:00:27 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 09 Oct 2025 22:00:27 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de>
Message-ID: <875xcnrgkk.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>>> "Dr. Arne Babenhauserheide" <arne_bab at web.de> writes:
>>> An option could be to use a common bibtex repository, always commit the
>>> changes to the bibtex file individually in the work repository and
>>> export+import those changes to the other repository.
>> 
>>> That?s actually far too complex. Feels like my brain got poisoned by the
>>> dozens of git clutches at work.
>> 
>> I am deeply sorry ?

> ?

>>> To get bibtex initially, you simply hg pull --force from the bibtex repo
>>> and then hg merge bibtex into default.
>> hg pull /home/user/HGs/tex/bib/ --force -b bibtex
> ?
>> So that would add quite a bit of changesets to my latex repository.

> I think you pulled the bibtex branch and all its dependencies.

> To be safe, best init a new repository and create the bibtex branch
> there.

Well this complicates things a bit and also defies the idea of VC

> If you need the bibtex history, you can use convert with an include rule
> to create a repo that *only* has the bibtex file.


This is also quite large.

I came up with the following ?bright? idea

In the bibtex repo:
 hg up 0
 hg branch bibtex
 hg revert bibtex -r 409 (which is tip)
 hg commit -m "Start the bibtex branch"

Then in the latex repository I run

 hg pull /home/user/bib -b bibtex --force

But then in my latex repository I see 
hg log -G

?  changeset:     1994:6d3d9467e443
?  tag:           tip
?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
?  Branch:        bibtex
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
?  Phase:         public
?  Summary:       Start the bibtex branch
?
?  changeset:     1993:e3d1de58c858
   Branch:        default
   Author:        oub
   Date:          Sun, 19 Dec 2004 10:37:47 +0000
   Phase:         public
   Parent:        -1:000000000000
   Summary:       Initial revision

?  changeset:     1992:a15e729491d4
?  Remote_Branch: default/default
?  Branch:        default
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
?  Phase:         public
?  Summary:       replace \mu'_p by \epsilon_p'
?

That looks also odd.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/afa67c47/attachment-0001.p7s>

From arne_bab at web.de  Thu Oct  9 21:31:52 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 23:31:52 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <875xcnrgkk.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de> <875xcnrgkk.fsf@mat.ucm.es>
Message-ID: <87wm53bw3b.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:
> This is also quite large.

Did you convert with a filemap?

? https://wiki.mercurial-scm.org/ConvertExtension#A--filemap

you need

include "path/to/bibtex.bib"

ANd then hg convert --filemap repo bibtex-repo

Then you can hg branch bibtex in the bibtex-repo.

> I came up with the following ?bright? idea
>
> In the bibtex repo:
>  hg up 0
>  hg branch bibtex
>  hg revert bibtex -r 409 (which is tip)
>  hg commit -m "Start the bibtex branch"
>
> Then in the latex repository I run
>
>  hg pull /home/user/bib -b bibtex --force
>
> But then in my latex repository I see 
> hg log -G
>
> ?  changeset:     1994:6d3d9467e443
> ?  tag:           tip
> ?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
> ?  Branch:        bibtex
> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
> ?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
> ?  Phase:         public
> ?  Summary:       Start the bibtex branch
> ?
> ?  changeset:     1993:e3d1de58c858
>    Branch:        default
>    Author:        oub
>    Date:          Sun, 19 Dec 2004 10:37:47 +0000
>    Phase:         public
>    Parent:        -1:000000000000
>    Summary:       Initial revision
>
> ?  changeset:     1992:a15e729491d4
> ?  Remote_Branch: default/default
> ?  Branch:        default
> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
> ?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
> ?  Phase:         public
> ?  Summary:       replace \mu'_p by \epsilon_p'
> ?
>
> That looks also odd.

That just means that you have multiple roots. Mercurial can deal with that.

But it means that the bibtex branch has all files from default, which
means that it?s far bigger than needed to version the bibtex file.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/e2f6754a/attachment.sig>

From arne_bab at web.de  Thu Oct  9 21:31:52 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Thu, 09 Oct 2025 23:31:52 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <875xcnrgkk.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de> <875xcnrgkk.fsf@mat.ucm.es>
Message-ID: <87wm53bw3b.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:
> This is also quite large.

Did you convert with a filemap?

? https://wiki.mercurial-scm.org/ConvertExtension#A--filemap

you need

include "path/to/bibtex.bib"

ANd then hg convert --filemap repo bibtex-repo

Then you can hg branch bibtex in the bibtex-repo.

> I came up with the following ?bright? idea
>
> In the bibtex repo:
>  hg up 0
>  hg branch bibtex
>  hg revert bibtex -r 409 (which is tip)
>  hg commit -m "Start the bibtex branch"
>
> Then in the latex repository I run
>
>  hg pull /home/user/bib -b bibtex --force
>
> But then in my latex repository I see 
> hg log -G
>
> ?  changeset:     1994:6d3d9467e443
> ?  tag:           tip
> ?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
> ?  Branch:        bibtex
> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
> ?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
> ?  Phase:         public
> ?  Summary:       Start the bibtex branch
> ?
> ?  changeset:     1993:e3d1de58c858
>    Branch:        default
>    Author:        oub
>    Date:          Sun, 19 Dec 2004 10:37:47 +0000
>    Phase:         public
>    Parent:        -1:000000000000
>    Summary:       Initial revision
>
> ?  changeset:     1992:a15e729491d4
> ?  Remote_Branch: default/default
> ?  Branch:        default
> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
> ?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
> ?  Phase:         public
> ?  Summary:       replace \mu'_p by \epsilon_p'
> ?
>
> That looks also odd.

That just means that you have multiple roots. Mercurial can deal with that.

But it means that the bibtex branch has all files from default, which
means that it?s far bigger than needed to version the bibtex file.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251009/e2f6754a/attachment-0001.sig>

From oub at mat.ucm.es  Fri Oct 10 07:50:01 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Fri, 10 Oct 2025 09:50:01 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de> <875xcnrgkk.fsf@mat.ucm.es>
 <87wm53bw3b.fsf@web.de>
Message-ID: <87sefrp55i.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>> This is also quite large.

> Did you convert with a filemap?

Well no (but I tried now, see below). The reason why I did not:

    + the repo is already pushed and shared since a couple of years

    + the vast mayority of changes (400 changesets) were only done in bibgraf.bib. So converting might not really reduce the number of changesets

> ? https://wiki.mercurial-scm.org/ConvertExtension#A--filemap

> you need

> include "path/to/bibtex.bib"

> ANd then hg convert --filemap repo bibtex-repo

Well I did what I always did (in a sandbox)

    + hg clone /home/user/HGs/tex/bib/ bib-clone

    + filemap contains
      include bibgraf.bib

    + cp filemap bib-clone/

    + g convert --datesort /home/oub/tmp4/bib-clone/ /home/oub/tmp4/bib-repo --filemap /home/oub/tmp4/bib-clone/filemap  

Then bib-repo has 363 instead of 406 changesets.
And in addition I have to delete my helix repository and create a new
one. The problem is that this will confuse my collaborators who have no
idea about mercurial at all, save push, fetch (not pull+merge), commit
and clone.




> Then you can hg branch bibtex in the bibtex-repo.

>> I came up with the following ?bright? idea
>> 
>> In the bibtex repo:
>> hg up 0
>> hg branch bibtex
>> hg revert bibtex -r 409 (which is tip)
>> hg commit -m "Start the bibtex branch"
>> 
>> Then in the latex repository I run
>> 
>> hg pull /home/user/bib -b bibtex --force
>> 
>> But then in my latex repository I see 
>> hg log -G
>> 
>> ?  changeset:     1994:6d3d9467e443
>> ?  tag:           tip
>> ?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
>> ?  Branch:        bibtex
>> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
>> ?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
>> ?  Phase:         public
>> ?  Summary:       Start the bibtex branch
>> ?
>> ?  changeset:     1993:e3d1de58c858
>> Branch:        default
>> Author:        oub
>> Date:          Sun, 19 Dec 2004 10:37:47 +0000
>> Phase:         public
>> Parent:        -1:000000000000
>> Summary:       Initial revision
>> 
>> ?  changeset:     1992:a15e729491d4
>> ?  Remote_Branch: default/default
>> ?  Branch:        default
>> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
>> ?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
>> ?  Phase:         public
>> ?  Summary:       replace \mu'_p by \epsilon_p'
>> ?
>> 
>> That looks also odd.

> That just means that you have multiple roots. Mercurial can deal with that.

But if I push that now my collaborators pull hg log -G might confuse them

> But it means that the bibtex branch has all files from default, which
> means that it?s far bigger than needed to version the bibtex file.

hg pull /home/user/bib-repo -b bibtex --force


Now results also in 

  changeset:     1996:262b91e38ff2
?  tag:           tip
?  Remote_Branch: /home/oub/tmp4/bib-repo/bibtex
?  Branch:        bibtex
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Fri, 10 Oct 2025 09:47:27 +0200
?  Phase:         public
?  Summary:       Add first commit to bibtex branch
?
?  changeset:     1995:e3d1de58c858
   Branch:        default
   Author:        oub
   Date:          Sun, 19 Dec 2004 10:37:47 +0000
   Phase:         public
   Parent:        -1:000000000000
   Summary:       Initial revision

?  changeset:     1994:4886ee5a8900
?  Remote_Branch: default/default
?  Branch:        default
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Fri, 10 Oct 2025 09:15:18 +0200
?  Phase:         public
?  Summary:       Rewrite the abstract
?


As nifty as the idea sounds, I am not sure I can use it, thanks

regards

Uwe 

> Best wishes,
> Arne

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251010/92a48c7a/attachment.p7s>

From oub at mat.ucm.es  Fri Oct 10 07:50:01 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Fri, 10 Oct 2025 09:50:01 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87seftcacr.fsf@web.de> <87bjmgq6es.fsf@mat.ucm.es>
 <878qhjdgny.fsf@web.de> <875xcnrgkk.fsf@mat.ucm.es>
 <87wm53bw3b.fsf@web.de>
Message-ID: <87sefrp55i.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>> This is also quite large.

> Did you convert with a filemap?

Well no (but I tried now, see below). The reason why I did not:

    + the repo is already pushed and shared since a couple of years

    + the vast mayority of changes (400 changesets) were only done in bibgraf.bib. So converting might not really reduce the number of changesets

> ? https://wiki.mercurial-scm.org/ConvertExtension#A--filemap

> you need

> include "path/to/bibtex.bib"

> ANd then hg convert --filemap repo bibtex-repo

Well I did what I always did (in a sandbox)

    + hg clone /home/user/HGs/tex/bib/ bib-clone

    + filemap contains
      include bibgraf.bib

    + cp filemap bib-clone/

    + g convert --datesort /home/oub/tmp4/bib-clone/ /home/oub/tmp4/bib-repo --filemap /home/oub/tmp4/bib-clone/filemap  

Then bib-repo has 363 instead of 406 changesets.
And in addition I have to delete my helix repository and create a new
one. The problem is that this will confuse my collaborators who have no
idea about mercurial at all, save push, fetch (not pull+merge), commit
and clone.




> Then you can hg branch bibtex in the bibtex-repo.

>> I came up with the following ?bright? idea
>> 
>> In the bibtex repo:
>> hg up 0
>> hg branch bibtex
>> hg revert bibtex -r 409 (which is tip)
>> hg commit -m "Start the bibtex branch"
>> 
>> Then in the latex repository I run
>> 
>> hg pull /home/user/bib -b bibtex --force
>> 
>> But then in my latex repository I see 
>> hg log -G
>> 
>> ?  changeset:     1994:6d3d9467e443
>> ?  tag:           tip
>> ?  Remote_Branch: /home/oub/tmp4/bib2/bibtex
>> ?  Branch:        bibtex
>> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
>> ?  Date:          Thu, 09 Oct 2025 21:52:06 +0200
>> ?  Phase:         public
>> ?  Summary:       Start the bibtex branch
>> ?
>> ?  changeset:     1993:e3d1de58c858
>> Branch:        default
>> Author:        oub
>> Date:          Sun, 19 Dec 2004 10:37:47 +0000
>> Phase:         public
>> Parent:        -1:000000000000
>> Summary:       Initial revision
>> 
>> ?  changeset:     1992:a15e729491d4
>> ?  Remote_Branch: default/default
>> ?  Branch:        default
>> ?  Author:        Uwe Brauer <oub at mat.ucm.es>
>> ?  Date:          Thu, 09 Oct 2025 10:07:22 +0300
>> ?  Phase:         public
>> ?  Summary:       replace \mu'_p by \epsilon_p'
>> ?
>> 
>> That looks also odd.

> That just means that you have multiple roots. Mercurial can deal with that.

But if I push that now my collaborators pull hg log -G might confuse them

> But it means that the bibtex branch has all files from default, which
> means that it?s far bigger than needed to version the bibtex file.

hg pull /home/user/bib-repo -b bibtex --force


Now results also in 

  changeset:     1996:262b91e38ff2
?  tag:           tip
?  Remote_Branch: /home/oub/tmp4/bib-repo/bibtex
?  Branch:        bibtex
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Fri, 10 Oct 2025 09:47:27 +0200
?  Phase:         public
?  Summary:       Add first commit to bibtex branch
?
?  changeset:     1995:e3d1de58c858
   Branch:        default
   Author:        oub
   Date:          Sun, 19 Dec 2004 10:37:47 +0000
   Phase:         public
   Parent:        -1:000000000000
   Summary:       Initial revision

?  changeset:     1994:4886ee5a8900
?  Remote_Branch: default/default
?  Branch:        default
?  Author:        Uwe Brauer <oub at mat.ucm.es>
?  Date:          Fri, 10 Oct 2025 09:15:18 +0200
?  Phase:         public
?  Summary:       Rewrite the abstract
?


As nifty as the idea sounds, I am not sure I can use it, thanks

regards

Uwe 

> Best wishes,
> Arne

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251010/92a48c7a/attachment-0001.p7s>

From rainer.woitok at gmail.com  Sat Oct 11 08:56:38 2025
From: rainer.woitok at gmail.com (Dr Rainer Woitok)
Date: Sat, 11 Oct 2025 10:56:38 +0200
Subject: Deprecated command "tip"
Message-ID: <26858.7110.608245.596384@woitok.gmail.com>

Greetings,

running "hg help tip" yields

   show the tip revision (DEPRECATED)

   ...

   This command is deprecated, please use 'hg heads' instead.

But if there are more than one heads in the current branch,  command "hg
heads ." will list them all.   So wouldn't be  "hg log -r tip"  a better
replacement?

Sincerely,
  Rainer

From oub at mat.ucm.es  Sat Oct 11 12:35:59 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Sat, 11 Oct 2025 14:35:59 +0200
Subject: Deprecated command "tip"
References: <26858.7110.608245.596384@woitok.gmail.com>
Message-ID: <87ldlhobtc.fsf@mat.ucm.es>

>>> "RW" == Rainer Woitok <rainer.woitok at gmail.com> writes:

> Greetings,
> running "hg help tip" yields

>    show the tip revision (DEPRECATED)

>    ...

>    This command is deprecated, please use 'hg heads' instead.

> But if there are more than one heads in the current branch,  command "hg
> heads ." will list them all.   So wouldn't be  "hg log -r tip"  a better
> replacement?

What is about 

 hg up tip 

Then?

That is a command I use very very often

Uwe Brauer 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251011/336451c5/attachment.p7s>

From pierre-yves.david at ens-lyon.org  Sat Oct 11 12:51:27 2025
From: pierre-yves.david at ens-lyon.org (Pierre-Yves David)
Date: Sat, 11 Oct 2025 14:51:27 +0200
Subject: Deprecated command "tip"
In-Reply-To: <26858.7110.608245.596384@woitok.gmail.com>
References: <26858.7110.608245.596384@woitok.gmail.com>
Message-ID: <ed0b8c2e-daeb-4abb-8428-ec32d4075483@ens-lyon.org>

`hg tip` is deprecated because the concept of tip is deprecated.

On 10/11/25 10:56, Dr Rainer Woitok wrote:
> Greetings,
>
> running "hg help tip" yields
>
>     show the tip revision (DEPRECATED)
>
>     ...
>
>     This command is deprecated, please use 'hg heads' instead.
>
> But if there are more than one heads in the current branch,  command "hg
> heads ." will list them all.   So wouldn't be  "hg log -r tip"  a better
> replacement?
>
> Sincerely,
>    Rainer
> _______________________________________________
> Mercurial mailing list
> Mercurial at lists.mercurial-scm.org
> https://lists.mercurial-scm.org/mailman/listinfo/mercurial

-- 
Pierre-Yves David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251011/5bd3c5a6/attachment.htm>

From jwdevel at gmail.com  Sun Oct 12 06:07:16 2025
From: jwdevel at gmail.com (John W)
Date: Sat, 11 Oct 2025 23:07:16 -0700
Subject: Deprecated command "tip"
In-Reply-To: <ed0b8c2e-daeb-4abb-8428-ec32d4075483@ens-lyon.org>
References: <26858.7110.608245.596384@woitok.gmail.com>
 <ed0b8c2e-daeb-4abb-8428-ec32d4075483@ens-lyon.org>
Message-ID: <CADO30so3qFg6KFchVuNU=jB3RM5BNpBET4rtHGtF1oDn15uc9A@mail.gmail.com>

On Sat, Oct 11, 2025 at 6:09?AM Pierre-Yves David
<pierre-yves.david at ens-lyon.org> wrote:
> ... the concept of tip is deprecated.

If that is the case, its usage should be removed from various current
docs as well, no?

eg: https://www.mercurial-scm.org/help/topics/revisions
or: https://www.mercurial-scm.org/help/commands/log
or: $ hg help revsets

See various mentions of 'tip' in specifying revisions, as well as in
plain English ? "the tipmost ...", "prior to the tip...", "The
reserved name 'tip' ...", etc.

Just my 2?

But in truth, I'm a little curious why it (the concept) is deprecated,
since I've generally found it to be a useful notion.
But I'm not looking to fight anyone about it (:

The 'tip' command itself was deprecated in 2013, but the 'tip'
revision, in the revset language, and in its use in the docs, has been
around a long time since.

Do you have more information about this deprecation of the *concept*?
Is it a recent change?

Thanks
-John

From oub at mat.ucm.es  Sun Oct 12 07:23:14 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Sun, 12 Oct 2025 09:23:14 +0200
Subject: Deprecated command "tip"
References: <26858.7110.608245.596384@woitok.gmail.com>
 <ed0b8c2e-daeb-4abb-8428-ec32d4075483@ens-lyon.org>
Message-ID: <87frbomvml.fsf@mat.ucm.es>

>>> "PD" == Pierre-Yves David <pierre-yves.david at ens-lyon.org> writes:

> `hg tip` is deprecated because the concept of tip is deprecated.

So what is the replacement of 

 hg up tip

then? And what is the substitution for tip?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251012/74e4088f/attachment.p7s>

From oub at mat.ucm.es  Sun Oct 12 07:25:44 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Sun, 12 Oct 2025 09:25:44 +0200
Subject: when hardlinks between files of two repos break
Message-ID: <87a51wmvif.fsf@mat.ucm.es>


Hi

I did more experiments on the hardlink of 2 files I wanted to share
between to repositories. It turned out that

    + hg commit -m "Whatever" is save

    + hg push helix is save

    + but hg up otherbranch, breaks the hardlink even if I run hg up default afterwards

Is this a design decision or and unwanted consequence?


Regards


Uwe Brauer

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251012/ef633115/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 13 07:13:27 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 13 Oct 2025 09:13:27 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
Message-ID: <87bjmbl1ew.fsf@mat.ucm.es>

>>> "ABvM" == Arne Babenhauserheide via Mercurial <mercurial at lists.mercurial-scm.org> writes:


> An option could be to use a common bibtex repository, always commit the
> changes to the bibtex file individually in the work repository and
> export+import those changes to the other repository. Then you can always
> import+merge the bibtex repo in all other repos.

> To automate it, you could create a commit hook that does the
> export+import; in .hg/hgrc:

> [hooks]
> commit = hg log -r $HG_NODE --template '{files}' | grep -q
> 'bibtex.bib' && hg export $HG_NODE > /tmp/bibtex.patch && hg -R ../bib
> import /tmp/bibtex.patch && rm /tmp/bibtex.patch || true

I tried that step by step, but 
hg  -R /home/user/bib import bib.diff


Results in 

applying bib.diff
patching file bibgraf.bib
Hunk #1 FAILED at 0
1 out of 1 hunks FAILED -- saving rejects to file bibgraf.bib.rej
abort: patch failed to apply


So the export/import option seems not to work, either

    + I have to check the hardlinks very often

    + or I create bibtex branch in my bib repository that contains very
      few changesets and pull that with the force option.



-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251013/3cb18bc5/attachment.p7s>

From arne_bab at web.de  Mon Oct 13 08:07:38 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Mon, 13 Oct 2025 10:07:38 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
In-Reply-To: <87bjmbl1ew.fsf@mat.ucm.es>
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87bjmbl1ew.fsf@mat.ucm.es>
Message-ID: <87frbnkywl.fsf@web.de>

Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:

>>>> "ABvM" == Arne Babenhauserheide via Mercurial <mercurial at lists.mercurial-scm.org> writes:
>> [hooks]
>> commit = hg log -r $HG_NODE --template '{files}' | grep -q
>> 'bibtex.bib' && hg export $HG_NODE > /tmp/bibtex.patch && hg -R ../bib
>> hg import /tmp/bibtex.patch && rm /tmp/bibtex.patch || true
>
> I tried that step by step, but 
> hg  -R /home/user/bib import bib.diff
>
>
> Results in 
>
> applying bib.diff
> patching file bibgraf.bib
> Hunk #1 FAILED at 0

This means that there were already other changes in bibgraf.bib

Did the two files already go out of sync earlier?

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251013/51864a1d/attachment.sig>

From oub at mat.ucm.es  Tue Oct 14 06:05:34 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Tue, 14 Oct 2025 08:05:34 +0200
Subject: how to share a file between 2 repositories? Hardlinks?
References: <87sefwqcdb.fsf@mat.ucm.es>
 <084b0e0e-1e8c-49a9-b284-673c7128c81c@ens-lyon.org>
 <87v7krcdha.fsf@mat.ucm.es> <87y0pm3o8d.fsf@web.de>
 <87bjmbl1ew.fsf@mat.ucm.es> <87frbnkywl.fsf@web.de>
Message-ID: <87ldlehvbl.fsf@mat.ucm.es>


> Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:
>> 
>> I tried that step by step, but 
>> hg  -R /home/user/bib import bib.diff
>> 
>> 
>> Results in 
>> 
>> applying bib.diff
>> patching file bibgraf.bib
>> Hunk #1 FAILED at 0

> This means that there were already other changes in bibgraf.bib

> Did the two files already go out of sync earlier?

Ah you are right, I should have exported 437:438 not 483 alone.
BTW what is the purpose of "-R /home/user/bib" shouldn't it work without?

So right now I have to think which approach is better

    + the ?graft from another repository? via export/import (and write a alias or script)

    + or pull --force a branch, having two roots and then graft but *not* merge.

Thanks

Uwe 



-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251014/12d66755/attachment.p7s>

From mkember at janestreet.com  Thu Oct 16 14:58:38 2025
From: mkember at janestreet.com (Mitchell Kember)
Date: Thu, 16 Oct 2025 10:58:38 -0400
Subject: Proposal: move from issueNNNN to #NNNN
Message-ID: <CAGxpoFML3MK32jdv1d-aMtCUF7RsazUZLBGGdO2bbiH+7G8RJA@mail.gmail.com>

Hi,

Now that we track Mercurial issues on Heptapod
<https://foss.heptapod.net/mercurial/mercurial-devel/-/issues>, I propose
we should:

1. Reference issues as "#NNNN" instead of "issueNNNN" in code and commit
messages. These have the benefit of being hyperlinked when reading from
Heptapod.

2. Use "(fixes #NNNN)" instead of "(issueNNNN)" at the end of a commit
message subject line that fixes an issue. This auto-closes the issue when
the MR is merged. I assume Heptapod uses the same Gitlab defaults
<https://docs.gitlab.com/user/project/issues/managing_issues/#default-closing-pattern>
for
closing patterns.

3. Leave the existing 1080 occurrences of /issue[0-9]+/ in the codebase
alone, since these are the sorts of lines you often want to hg annotate.
Though I could also see an argument for changing them.

A possible downside is that #N comes up naturally in printf syntax, HTML
entities, CSS colors, or talking about a list of items, making it harder to
search for. An alternative would be to always write out "issue #NNNN" for
(1).


Mitchell
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251016/e8deef05/attachment.htm>

From oub at mat.ucm.es  Thu Oct 16 19:23:54 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 16 Oct 2025 21:23:54 +0200
Subject: how to clone/convert/pull only 1 branch push to github takes ages
Message-ID: <874iryzm45.fsf@mat.ucm.es>



Hi all

I need to push my mercurial repository to github using hg-git. The
repository contains 533 changesets and various branches. 

The push takes hours and never finished successfully. I also tried hg
push -b some-branch github

But it did not work neither (I am using the change set of hg-git that
supports named branches)

So I thought to extract either by pull or clone or convert just the last
branch.

I tried 

hg clone -r 533 -b 2024-25 /home/oub/ALLES/HGs/tex/vorlesungen/HGCal-Num-Transp-old /home/oub/tmp5/HGCal-new 

And similar commands but I always end up with the complete repository 

Any ideas?

Thanks

Uwe Brauer 

From pierre-yves.david at ens-lyon.org  Thu Oct 16 19:52:06 2025
From: pierre-yves.david at ens-lyon.org (Pierre-Yves David)
Date: Thu, 16 Oct 2025 21:52:06 +0200
Subject: how to clone/convert/pull only 1 branch push to github takes ages
In-Reply-To: <874iryzm45.fsf@mat.ucm.es>
References: <874iryzm45.fsf@mat.ucm.es>
Message-ID: <8796f197-a594-4402-b4c5-20227d62b313@ens-lyon.org>

(taking hours for 553 changesets? seems very fishy.)

the command use show (using --rev) should only select that revision and 
its ancestors. How many ancestors does 533 has?

On 10/16/25 21:23, Uwe Brauer via Mercurial wrote:
>
> Hi all
>
> I need to push my mercurial repository to github using hg-git. The
> repository contains 533 changesets and various branches.
>
> The push takes hours and never finished successfully. I also tried hg
> push -b some-branch github
>
> But it did not work neither (I am using the change set of hg-git that
> supports named branches)
>
> So I thought to extract either by pull or clone or convert just the last
> branch.
>
> I tried
>
> hg clone -r 533 -b 2024-25 /home/oub/ALLES/HGs/tex/vorlesungen/HGCal-Num-Transp-old /home/oub/tmp5/HGCal-new
>
> And similar commands but I always end up with the complete repository
>
> Any ideas?
>
> Thanks
>
> Uwe Brauer
> _______________________________________________
> Mercurial mailing list
> Mercurial at lists.mercurial-scm.org
> https://lists.mercurial-scm.org/mailman/listinfo/mercurial

-- 
Pierre-Yves David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251016/49f58b0f/attachment.htm>

From oub at mat.ucm.es  Thu Oct 16 20:00:52 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 16 Oct 2025 22:00:52 +0200
Subject: how to clone/convert/pull only 1 branch push to github takes ages
References: <874iryzm45.fsf@mat.ucm.es>
 <8796f197-a594-4402-b4c5-20227d62b313@ens-lyon.org>
Message-ID: <87y0pay5u3.fsf@mat.ucm.es>


> (taking hours for 553 changesets? seems very fishy.)
> the command use show (using --rev) should only select that revision
> and its ancestors. How many ancestors does 533 has?

I am not sure what show --rev means, I run 

hg log -r "ancestors(533)" --template "{rev}\n" | wc -l

And it returns
529


Basically when an academic year was finished I started a new branch, so
the history is quite linear.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251016/350e256c/attachment.p7s>

From oub at mat.ucm.es  Thu Oct 16 20:12:09 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Thu, 16 Oct 2025 22:12:09 +0200
Subject: how to clone/convert/pull only 1 branch push to github takes ages
References: <874iryzm45.fsf@mat.ucm.es>
 <8796f197-a594-4402-b4c5-20227d62b313@ens-lyon.org>
 <87y0pay5u3.fsf@mat.ucm.es>
Message-ID: <87sefiy5ba.fsf@mat.ucm.es>

>>> "UB" == Uwe Brauer <oub at mat.ucm.es> writes:

>> (taking hours for 553 changesets? seems very fishy.)
>> the command use show (using --rev) should only select that revision
>> and its ancestors. How many ancestors does 533 has?

> I am not sure what show --rev means, I run 

> hg log -r "ancestors(533)" --template "{rev}\n" | wc -l

> And it returns
> 529


> Basically when an academic year was finished I started a new branch, so
> the history is quite linear.

?One solution? is 

 hg up 0
 hg branch 2025-26
 hg revert --all -r 533
 hg commit "Start a new branch at 0"



-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251016/f758426f/attachment.p7s>

From oub at mat.ucm.es  Sun Oct 19 15:40:34 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Sun, 19 Oct 2025 17:40:34 +0200
Subject: warning when pushing a repository with secret changesets
Message-ID: <87sefesxvx.fsf@mat.ucm.es>

Hi all


This question, strictly speaking, concerns only the *hg-git* converter.
However, since the relevant mailing list is not exactly bustling with
activity?and as at least some users here rely on that plugin?I dare to
ask the following.

Is there any way, perhaps via a hook, for Mercurial to warn me when I am
about to push a repository that contains secret changesets?

That might sound odd in the Mercurial world, but the problem is that Git
lacks the concept of phases. As a result, the *hg-git* converter ends up
pushing secret changesets as well.


Thanks and regards


Uwe Brauer 

From arne_bab at web.de  Mon Oct 20 06:36:12 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Mon, 20 Oct 2025 08:36:12 +0200
Subject: warning when pushing a repository with secret changesets
In-Reply-To: <87sefesxvx.fsf@mat.ucm.es>
References: <87sefesxvx.fsf@mat.ucm.es>
Message-ID: <87bjm23wrn.fsf@web.de>

Hi Uwe,

that sounds like you need a pre-push hook that throws an error when the
bookmark you are on contains any secret changesets.

I hope that already helps you find a solution!
(could you post it here when you fleshed it out?)

Best wishes,
Arne

Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:
> Is there any way, perhaps via a hook, for Mercurial to warn me when I am
> about to push a repository that contains secret changesets?
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/8ab0b679/attachment.sig>

From arne_bab at web.de  Mon Oct 20 06:36:12 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Mon, 20 Oct 2025 08:36:12 +0200
Subject: warning when pushing a repository with secret changesets
In-Reply-To: <87sefesxvx.fsf@mat.ucm.es>
References: <87sefesxvx.fsf@mat.ucm.es>
Message-ID: <87bjm23wrn.fsf@web.de>

Hi Uwe,

that sounds like you need a pre-push hook that throws an error when the
bookmark you are on contains any secret changesets.

I hope that already helps you find a solution!
(could you post it here when you fleshed it out?)

Best wishes,
Arne

Uwe Brauer via Mercurial <mercurial at lists.mercurial-scm.org> writes:
> Is there any way, perhaps via a hook, for Mercurial to warn me when I am
> about to push a repository that contains secret changesets?
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/8ab0b679/attachment-0001.sig>

From oub at mat.ucm.es  Mon Oct 20 06:56:58 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 08:56:58 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <877bwqqcw5.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

It is a named branch, I am using change set 13cf16d4c84f
and right the pre-push hook seems the right one.

I spent the yesterday afternoon quite some time, but nothing really
worked so far.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)


Of course, but... ?

Uwe 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/518d8a92/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 20 06:56:58 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 08:56:58 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <877bwqqcw5.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

It is a named branch, I am using change set 13cf16d4c84f
and right the pre-push hook seems the right one.

I spent the yesterday afternoon quite some time, but nothing really
worked so far.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)


Of course, but... ?

Uwe 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/518d8a92/attachment-0001.p7s>

From oub at mat.ucm.es  Mon Oct 20 07:04:19 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 09:04:19 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <871pmyqcjw.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)

One idea was (Linux) to use 
hg log -r "secret()" --template "{rev} " | wc -w

And if the number is strictly larger that 0, abort. But....
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/358851b8/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 20 07:04:19 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 09:04:19 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <871pmyqcjw.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)

One idea was (Linux) to use 
hg log -r "secret()" --template "{rev} " | wc -w

And if the number is strictly larger that 0, abort. But....
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/358851b8/attachment-0001.p7s>

From oub at mat.ucm.es  Mon Oct 20 08:00:43 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 10:00:43 +0200
Subject: [a partial solution] (was: warning when pushing a repository with
 secret changesets)
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <87sefeovdg.fsf_-_@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)

I tortured my good ?friend? ChatGTP,

This works somehow 

And in the repository in question

[hooks]
pre-push.check-secret = python:/home/oub/scripts/hg_hooks-allbranches.py:check_no_secret


The problem is; if I have a branch with draft changesets and another one with
secret changesets, I cannot push:

 hg push -b draftbranch github

Will abort, although this branch contains only draft changesets. The
generalisation to branches so far does not work so 


?Ethan, your mission, should you choose to accept it, is to generalise
the code to branches? ?

Uwe 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: hg_hooks-allbranches.py
Type: text/x-python
Size: 731 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/210af2da/attachment.py>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/210af2da/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 20 08:00:43 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 10:00:43 +0200
Subject: [a partial solution] (was: warning when pushing a repository with
 secret changesets)
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
Message-ID: <87sefeovdg.fsf_-_@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Hi Uwe,
> that sounds like you need a pre-push hook that throws an error when the
> bookmark you are on contains any secret changesets.

> I hope that already helps you find a solution!
> (could you post it here when you fleshed it out?)

I tortured my good ?friend? ChatGTP,

This works somehow 

And in the repository in question

[hooks]
pre-push.check-secret = python:/home/oub/scripts/hg_hooks-allbranches.py:check_no_secret


The problem is; if I have a branch with draft changesets and another one with
secret changesets, I cannot push:

 hg push -b draftbranch github

Will abort, although this branch contains only draft changesets. The
generalisation to branches so far does not work so 


?Ethan, your mission, should you choose to accept it, is to generalise
the code to branches? ?

Uwe 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: hg_hooks-allbranches.py
Type: text/x-python
Size: 731 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/210af2da/attachment-0001.py>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/210af2da/attachment-0001.p7s>

From arne_bab at web.de  Mon Oct 20 08:30:51 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Mon, 20 Oct 2025 10:30:51 +0200
Subject: warning when pushing a repository with secret changesets
In-Reply-To: <871pmyqcjw.fsf@mat.ucm.es>
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es>
Message-ID: <875xca3rgk.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:

>>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:
>> that sounds like you need a pre-push hook that throws an error when the
>> bookmark you are on contains any secret changesets.
> One idea was (Linux) to use 
> hg log -r "secret()" --template "{rev} " | wc -w

maybe

hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

-b . ? current branch

If not using hg-git, this works:

pre-push = if hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && echo "there are secret changesets" >&2; then false; fi

I?m not sure which hook is triggered when pushing to hg-git, though.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/48f04945/attachment.sig>

From arne_bab at web.de  Mon Oct 20 08:30:51 2025
From: arne_bab at web.de (Dr. Arne Babenhauserheide)
Date: Mon, 20 Oct 2025 10:30:51 +0200
Subject: warning when pushing a repository with secret changesets
In-Reply-To: <871pmyqcjw.fsf@mat.ucm.es>
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es>
Message-ID: <875xca3rgk.fsf@web.de>

Uwe Brauer <oub at mat.ucm.es> writes:

>>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:
>> that sounds like you need a pre-push hook that throws an error when the
>> bookmark you are on contains any secret changesets.
> One idea was (Linux) to use 
> hg log -r "secret()" --template "{rev} " | wc -w

maybe

hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

-b . ? current branch

If not using hg-git, this works:

pre-push = if hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && echo "there are secret changesets" >&2; then false; fi

I?m not sure which hook is triggered when pushing to hg-git, though.

Best wishes,
Arne
-- 
Unpolitisch sein
hei?t politisch sein,
ohne es zu merken.
draketo.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 1125 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/48f04945/attachment-0001.sig>

From oub at mat.ucm.es  Mon Oct 20 08:50:37 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 10:50:37 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es> <875xca3rgk.fsf@web.de>
Message-ID: <87frbeot2a.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>>>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:
>>> that sounds like you need a pre-push hook that throws an error when the
>>> bookmark you are on contains any secret changesets.
>> One idea was (Linux) to use 
>> hg log -r "secret()" --template "{rev} " | wc -w

> maybe

> hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

> -b . ? current branch

> If not using hg-git, this works:

> pre-push = if hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && echo "there are secret changesets" >&2; then false; fi

That works if I push to a mercurial repository but it does not work if I
push to git (with hg-git), that is it pushes anyhow.

One the other hand chatgtp's solution works also for git but is unable to
distinguish the branches. 

> I?m not sure which hook is triggered when pushing to hg-git, though.

> Best wishes,
> Arne

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/72af6c89/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 20 08:50:37 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 10:50:37 +0200
Subject: warning when pushing a repository with secret changesets
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es> <875xca3rgk.fsf@web.de>
Message-ID: <87frbeot2a.fsf@mat.ucm.es>

>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:

> Uwe Brauer <oub at mat.ucm.es> writes:
>>>>> "AB" == Arne Babenhauserheide <arne_bab at web.de> writes:
>>> that sounds like you need a pre-push hook that throws an error when the
>>> bookmark you are on contains any secret changesets.
>> One idea was (Linux) to use 
>> hg log -r "secret()" --template "{rev} " | wc -w

> maybe

> hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

> -b . ? current branch

> If not using hg-git, this works:

> pre-push = if hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && echo "there are secret changesets" >&2; then false; fi

That works if I push to a mercurial repository but it does not work if I
push to git (with hg-git), that is it pushes anyhow.

One the other hand chatgtp's solution works also for git but is unable to
distinguish the branches. 

> I?m not sure which hook is triggered when pushing to hg-git, though.

> Best wishes,
> Arne

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/72af6c89/attachment-0001.p7s>

From oub at mat.ucm.es  Mon Oct 20 09:04:17 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 11:04:17 +0200
Subject: [poor man's solution] (was: warning when pushing a repository with
 secret changesets)
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es> <875xca3rgk.fsf@web.de>
Message-ID: <87a51mosfi.fsf_-_@mat.ucm.es>



> maybe

> hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

> -b . ? current branch

> If not using hg-git, this works:

To have a solution that works for a branch, that is circumvent the
pre-push hook is 
[alias]
pushnohook = !sh -c 'hg push --config hooks.pre-push.check-secret= "$@"'

But it is not elegant at all
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/9c81f726/attachment.p7s>

From oub at mat.ucm.es  Mon Oct 20 09:04:17 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 20 Oct 2025 11:04:17 +0200
Subject: [poor man's solution] (was: warning when pushing a repository with
 secret changesets)
References: <87sefesxvx.fsf@mat.ucm.es> <87bjm23wrn.fsf@web.de>
 <871pmyqcjw.fsf@mat.ucm.es> <875xca3rgk.fsf@web.de>
Message-ID: <87a51mosfi.fsf_-_@mat.ucm.es>



> maybe

> hg log -b . -r 'secret()' --template SECRET | grep -q SECRET && fail

> -b . ? current branch

> If not using hg-git, this works:

To have a solution that works for a branch, that is circumvent the
pre-push hook is 
[alias]
pushnohook = !sh -c 'hg push --config hooks.pre-push.check-secret= "$@"'

But it is not elegant at all
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251020/9c81f726/attachment-0001.p7s>

From oub at mat.ucm.es  Mon Oct 27 07:38:05 2025
From: oub at mat.ucm.es (Uwe Brauer)
Date: Mon, 27 Oct 2025 08:38:05 +0100
Subject: alias for changing the author of a changeset
Message-ID: <871pmovlpe.fsf@mat.ucm.es>

Hi,

Occasionally, I find myself needing to change the author of a changeset.

Using the evolve extension, the procedure is straightforward:

hg up 30
hg amend --user "Gandalf <new at example.com>"
hg evolve
hg up tip


I am trying, without much success, to write an alias for this?either in .hgrc or as a shell script.

Does anyone have a suggestion?

Thanks.

-- 
I strongly condemn Hamas heinous despicable pogroms/atrocities on Israel.
Stop the war in Gaza, guarantee humanitarian aid, and bring the hostages back.
I strongly condemn Putin's war of aggression against Ukraine.
I support to deliver weapons to Ukraine's military. 
I support the EU and NATO membership of Ukraine. 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5684 bytes
Desc: not available
URL: <http://lists.mercurial-scm.org/pipermail/mercurial/attachments/20251027/8d4a3a86/attachment.p7s>