oldest commit?

oldest commit?

[Thomas Koch]

Just for the curious, the oldest commit I found (from 1985):

http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=ce5584125c44a1a2fbb46e810459c50b227a95e2

and the (currenty) last page of commits:
http://git.savannah.gnu.org/cgit/emacs.git/log/?ofs=118050

Paginating through such old commits might stress the machine though!

Looking forward to dive into emacs development now that the source code is 
available. :-)

Regards, Thomas Koch

Re: oldest commit?

[Karl Fogel]

Thomas Koch <thomas@koch.ro> writes:
>Just for the curious, the oldest commit I found (from 1985):
>
>http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=ce5584125c44a1a2fbb46e810459c50b227a95e2

Funny that in that commit record, Jim's affiliation (by email address)
is with a company that won't exist for another 8 years :-).  I guess
it's not too surprising that somewhere along Emacs's long chain of
conversions there was at least one metadata chronology issue.

-K

>and the (currenty) last page of commits:
>http://git.savannah.gnu.org/cgit/emacs.git/log/?ofs=118050
>
>Paginating through such old commits might stress the machine though!
>
>Looking forward to dive into emacs development now that the source code is 
>available. :-)

Re: oldest commit?

[Andreas Schwab]

Karl Fogel <kfogel@red-bean.com> writes:

> Funny that in that commit record, Jim's affiliation (by email address)
> is with a company that won't exist for another 8 years :-). 

All mail addresses in the commits that originate from CVS are fake, and
roughly represent the state from the time I did the first conversion.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Re: oldest commit?

[Eric S. Raymond]

Karl Fogel <kfogel@red-bean.com>:
> Funny that in that commit record, Jim's affiliation (by email address)
> is with a company that won't exist for another 8 years :-).  I guess
> it's not too surprising that somewhere along Emacs's long chain of
> conversions there was at least one metadata chronology issue.

The transition comment warns about this. Email addresses are based on
conversion day, not on the time of the commit.  

Here's what I think happened to jimb's address.  At the time of the 
bzr conversion in 2009, there were no email addresses in the CVS,
so whover did that lift supplied jimb's email in 2009, at which point
RH certainly *did* exist. :-) 

I felt it was more important to have consistent identifiers than to
try to dig up historical addresses that might not be recoverable and, if
they were, might no longer be valid.

This is one reason why I asked on the list for preferred form of name
and email address, and why (in cases where it applied) I reconciled
variant forms of peoples' names and used the UTF-8 for whatever
non-ASCII characters they required.  In some cases this requred
multiple patches.  Here are some examples - the =C stands for "all
commits":

=C filter --replace /Adam Sjogren/Adam Sjøgren/g
=C filter --replace /K. Handa/Kenichi Handa/g
=C filter --regexp /[Aa]gustin [Mm]artin/Agustín Martín/g

In the third one I'm reconciling four variant forms and adding an
acute-accented i from the fifth (correct) one, which also occurred.

In some cases human names were entirely missing malformed or had been
swapped with email addresses:

=C filter --regexp /^ *<dancol/Daniel Colascione <dancol/
=C filter --regexp /dancol@dancol.org <>/Daniel Colascione <dancol@dancol.org>/
=C filter --replace /<fabian@anue.biz>/<fgallina@gnu.org>/
=C filter --replace /<fabian@cuca>/<fgallina@gnu.org>/

Just to show that this can happen to the best of us:

=C filter --replace /Richard M. Stallman <>/Richard M. Stallman <rms@gnu.org>/

Here's a fun little triplet - you can infer a fourth address that also occurred:

=C filter --replace /joakim <joakim@exodia>/Joakim Verona <joakim@verona.se>/
=C filter --replace /root <root@exodia>/Joakim Verona <joakim@verona.se>/
=C filter --replace /joakim verona <joakim@verona.se>/Joakim Verona <joakim@vero

And here's the award for most diacriticals:

=C filter --replace /Stepán Nemec/Štěpán Němec/g

When I talk about the difference between a slapdash repo conversion
and a really high-quality, polished one, this is the sort of thing I mean.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>

Re: oldest commit?

[Karl Fogel]

"Eric S. Raymond" <esr@thyrsus.com> writes:
>Karl Fogel <kfogel@red-bean.com>:
>> Funny that in that commit record, Jim's affiliation (by email address)
>> is with a company that won't exist for another 8 years :-).  I guess
>> it's not too surprising that somewhere along Emacs's long chain of
>> conversions there was at least one metadata chronology issue.
>
>The transition comment warns about this. Email addresses are based on
>conversion day, not on the time of the commit.  

Oh, just to be clear, my point was not a criticism -- merely an amusing
side note.  I have no complaints about this transition (or the past
ones, for that matter).

Best,
­K

>Here's what I think happened to jimb's address.  At the time of the 
>bzr conversion in 2009, there were no email addresses in the CVS,
>so whover did that lift supplied jimb's email in 2009, at which point
>RH certainly *did* exist. :-) 
>
>I felt it was more important to have consistent identifiers than to
>try to dig up historical addresses that might not be recoverable and, if
>they were, might no longer be valid.
>
>This is one reason why I asked on the list for preferred form of name
>and email address, and why (in cases where it applied) I reconciled
>variant forms of peoples' names and used the UTF-8 for whatever
>non-ASCII characters they required.  In some cases this requred
>multiple patches.  Here are some examples - the =C stands for "all
>commits":
>
>=C filter --replace /Adam Sjogren/Adam Sjøgren/g
>=C filter --replace /K. Handa/Kenichi Handa/g
>=C filter --regexp /[Aa]gustin [Mm]artin/Agustín Martín/g
>
>In the third one I'm reconciling four variant forms and adding an
>acute-accented i from the fifth (correct) one, which also occurred.
>
>In some cases human names were entirely missing malformed or had been
>swapped with email addresses:
>
>=C filter --regexp /^ *<dancol/Daniel Colascione <dancol/
>=C filter --regexp /dancol@dancol.org <>/Daniel Colascione <dancol@dancol.org>/
>=C filter --replace /<fabian@anue.biz>/<fgallina@gnu.org>/
>=C filter --replace /<fabian@cuca>/<fgallina@gnu.org>/
>
>Just to show that this can happen to the best of us:
>
>=C filter --replace /Richard M. Stallman <>/Richard M. Stallman <rms@gnu.org>/
>
>Here's a fun little triplet - you can infer a fourth address that also occurred:
>
>=C filter --replace /joakim <joakim@exodia>/Joakim Verona <joakim@verona.se>/
>=C filter --replace /root <root@exodia>/Joakim Verona <joakim@verona.se>/
>=C filter --replace /joakim verona <joakim@verona.se>/Joakim Verona <joakim@vero
>
>And here's the award for most diacriticals:
>
>=C filter --replace /Stepán Nemec/Štěpán Němec/g
>
>When I talk about the difference between a slapdash repo conversion
>and a really high-quality, polished one, this is the sort of thing I mean.