Trying to graft the Aquamacs branch

Trying to graft the Aquamacs branch

[Eric S. Raymond]

In order to graft the Aquamacs branch to the git conversion of the
Emacs repository, I need to extract from it an import-stream branch dump
consisting of all the Aquamacs commits and *only* those commits.

Furthermore, in order for merges to Aquamacs to be built correctly,
the parent links in the branch dump cannot be ordinary marks, which
would have no meaning outside the full Aquamacs repository the branch
dump was made from.  Instead, they must be callouts - action stamps
that identify a parent commit by committer email and date and can
refer between repositories.

Identifying the correct set of commits is not trivial either.  To
accomplish this, I have written a new reposurgeon primitive function
#dsc() which takes a selection set and adds to it all descendants of
that set.

In principle, then, it should be possible to extract the correct
branch dump like this:

reposurgeon
% read aquamacs-emacs
% @dsc(<2005-06-06T18:41:57Z!david.reitter@gmail.com>) write --callout >aquamacs-branch.fi

where the @dsc() argument is intended to be the initial Aquamacs
commit from which all subsequent ones are descended.  If this worked,
I should then be able to do this:

% read <aquamacs-branch.fi
% read <emacs.fi
$ graft aquamacs-branch

The latter command would look for commits in the Emacs repository
(which is selected because iot was just read) that match each callout
in the Aquamacs branch dump, and attach them as parents.

Alas, when I try to perform the write --callout operation reposurgeon
runs out of memory.  That's on a machine with 8GB of core.

Smaller callout-dump cases work (they're in my regression-test suite) 
so I think I'm hitting a hardware limit here.

Accordingly, I have two requests:

1) Please verify that I have identified the correct root commit of the
Aquamacs branch. If I have not, I need to know what it actually is.

2) If you have access to a machine with more physical memory, please
try to extract the branch dump and drop it somewhere I can get it.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Every Communist must grasp the truth, 'Political power grows out of
the barrel of a gun.'
        -- Mao Tse-tung, 1938, inadvertently endorsing the Second Amendment.

Re: Trying to graft the Aquamacs branch

[David Reitter]

[-- Attachment #1: Type: text/plain, Size: 707 bytes --]

On Mar 28, 2014, at 2:06 PM, Eric S. Raymond <esr@thyrsus.com> wrote:
> % @dsc(<2005-06-06T18:41:57Z!david.reitter@gmail.com>) write --callout >aquamacs-branch.fi

This, I believe, is the correct initial commit of the two relevant branches.

> 2) If you have access to a machine with more physical memory, please
> try to extract the branch dump and drop it somewhere I can get it.

My memory is only larger by one order of magnitude (=16GB).
I think someone with more memory might have more luck.
If I should try running it overnight, I would need the version of reposurgeon with your special function,
and also the input file (unless it can work with my git repository directly).

- David


[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 4151 bytes --]

Re: Trying to graft the Aquamacs branch

[Eric S. Raymond]

David Reitter <david.reitter@gmail.com>:
> My memory is only larger by one order of magnitude (=16GB).
> I think someone with more memory might have more luck.
> If I should try running it overnight, I would need the version of reposurgeon with your special function,
> and also the input file (unless it can work with my git repository directly).

It can work on your repository directly.  You can download reposurgeon 3.8, 
with the @dsc() function, from http://www.catb.org/esr/reposurgeon/
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>