Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Stefan Monnier]

> +    particular version control system.  (At time of writing Emacs is
> +    about to move to its third VCS and another move in the future is
                            ^^^^^
                            fourth
RCS -> CVS -> Bzr

GNU Arch was semi-officially used for a while between CVS and Bzr, tho
not for the main branch.


        Stefan

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Eric S. Raymond]

Stefan Monnier <monnier@iro.umontreal.ca>:
> > +    particular version control system.  (At time of writing Emacs is
> > +    about to move to its third VCS and another move in the future is
>                             ^^^^^
>                             fourth
> RCS -> CVS -> Bzr
> 
> GNU Arch was semi-officially used for a while between CVS and Bzr, tho
> not for the main branch.

Wow.  That's a record; I've never actually seen more than three transitions
before.  I will correct.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[David Kastrup]

"Eric S. Raymond" <esr@thyrsus.com> writes:

> Stefan Monnier <monnier@iro.umontreal.ca>:
>> > +    particular version control system.  (At time of writing Emacs is
>> > +    about to move to its third VCS and another move in the future is
>>                             ^^^^^
>>                             fourth
>> RCS -> CVS -> Bzr
>> 
>> GNU Arch was semi-officially used for a while between CVS and Bzr, tho
>> not for the main branch.

Ah right.  I already forgot about the ubiquitous arch tags in every
file.  Looks like they are gone?

> Wow.  That's a record; I've never actually seen more than three
> transitions before.  I will correct.

Emacs is at the forefront of everything.

-- 
David Kastrup

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Eric S. Raymond]

David Kastrup <dak@gnu.org>:
> > Wow.  That's a record; I've never actually seen more than three
> > transitions before.  I will correct.
> 
> Emacs is at the forefront of everything.

Or hindfront, or something.  RCS.  That's...impressive.

Most of the venerably ancient project histories I've converted only
have the following geological strata

1. Subversion
2. Random crap generated by cvs2svn
3. CVS

Layer 2 is the nasty one.  Those conversion artifacts tend to trip up
Subversion-to-anything converters, especially git-svn. I originally
wrote reposurgeon to have a way to clean those up.

What I've never seen before is a plain RCS layer visible below the CVS
(now that I've been clued in some things in older ChangeLogs make sense.) I
guess that's because there just aren't that many projects old enough to
predate CVS.  Only three out of over forty of mine are.

/me wouldn't be surprised to stumble over a fossil trilobite next.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Andreas Schwab]

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

>> +    particular version control system.  (At time of writing Emacs is
>> +    about to move to its third VCS and another move in the future is
>                             ^^^^^
>                             fourth
> RCS -> CVS -> Bzr
>
> GNU Arch was semi-officially used for a while between CVS and Bzr, tho
> not for the main branch.

Eg the multi-tty branch originates from an arch branch.

(I didn't actually convert directly from CVS to bzr, but used git as an
intermediary.)

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Eric S. Raymond]

Andreas Schwab <schwab@linux-m68k.org>:
> (I didn't actually convert directly from CVS to bzr, but used git as an
> intermediary.)

Interesting. What CVS lifter did you use?
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Andreas Schwab]

"Eric S. Raymond" <esr@thyrsus.com> writes:

> Andreas Schwab <schwab@linux-m68k.org>:
>> (I didn't actually convert directly from CVS to bzr, but used git as an
>> intermediary.)
>
> Interesting. What CVS lifter did you use?

I think I started with parsecvs, but I cannot remember the exact
details.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Paul Eggert]

On 01/16/2014 08:23 AM, Eric S. Raymond wrote:

> there just aren't that many projects old enough to predate CVS.

Hah!  The last public software project I converted was from
SCCS to Git.  And this was genuine circa-1980 Programmer's
Workbench SCCS, not some pale imitation.  None of the
repository converters I found on the web worked, so I had to
write my own.  I wish I'd known about your repository
toolset -- maybe some of its Swiss Army knife blades would
have saved me some work.

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Eric S. Raymond]

Andreas Schwab <schwab@linux-m68k.org>:
> > Interesting. What CVS lifter did you use?
> 
> I think I started with parsecvs, but I cannot remember the exact
> details.

Good.  That is one of only two I know of that don't suck.  If you had
given me almost any other answer I would have begun to fear that I would
have to re-lift the CVS myself.

In fact, I now maintain what used to be parsecvs.  It's called
cvs-fast-export now and has several improvements, including being able
to emit a fast-import stream directly.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: [Emacs-diffs] trunk r116038: On VCS-independent ways of identifying commits, and why they are desirable.

[Eric S. Raymond]

Paul Eggert <eggert@cs.ucla.edu>:
> On 01/16/2014 08:23 AM, Eric S. Raymond wrote:
> 
> > there just aren't that many projects old enough to predate CVS.
> 
> Hah!  The last public software project I converted was from
> SCCS to Git.  And this was genuine circa-1980 Programmer's
> Workbench SCCS, not some pale imitation.  None of the
> repository converters I found on the web worked, so I had to
> write my own.  I wish I'd known about your repository
> toolset -- maybe some of its Swiss Army knife blades would
> have saved me some work.

Quite likely.  But if this was before 2010 reposurgeon didn't exist yet.

I maintain a Python version of sccs2rcs (yes, it's descended from the 
C-shell version you know about).  If I had to do SCCS -> git I would
run it and then lift the resulting RCS with cvs-fast-export (which
is a significantly improved version of the parsecvs code Keith
Packard wrote for lifting X). 
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>