Jacob Nevins
2017-04-01 18:23:26 UTC
With Gna shutting down, we intend to migrate the official source
repository from Subversion to Git. (Currently we expect to host the
result on Github.)
I'm just getting started on this conversion (offline), but one thing
that will be needed is full committer details of the form
Jacob Nevins <***@chiark.greenend.org.uk>
for every commit, where currently we just have local usernames like
'jtn'. Once we have published an official git history, these details
will be essentially unchangeable for existing commits.
If you have a preference as to how your commits will appear, please let
me know, including your svn/CVS user ID. (If you send me a preference
you'll probably save me time.)
By default I will make some effort to track down real names and current
public email addresses, starting with the most prolific contributors and
probably stopping when I run out of time.
I will use Gna real names where available, but I will not use your
registered Gna email address if it's not public on Gna, unless I'm
reasonably convinced it is otherwise a public email address for you.
I'll also be using <http://www.freeciv.org/wiki/People> as a resource.
For anyone I can't track down (or if I run out of time) I expect I'll
use <svn-username>@freeciv.org (email to which will in most cases
bounce; I don't intend to set up forwarding in the first instance).
I don't currently intend to go through the history and populate separate
author/committer fields correctly (except perhaps for the very early CVS
history which was all committed by a user called 'freeciv'), so above
I'm talking only about people who've actually committed to CVS or SVN.
I've attached a list of distinct usernames found in the repository as of
r35175, sorted alphabetically.
=-=-=-=-=-
I intend to use the reposurgeon tool, and broadly follow the method of
<http://www.catb.org/~esr/reposurgeon/dvcs-migration-guide.html>,
although not slavishly (for instance I'm not going to rewrite commit
messages to have git-ish summaries, and I'd like to preserve svn
revision information).
I hope to put candidate histories up somewhere for review as they become
available. I'm aiming to be done by the end of April, because that is
the best guess we have about when Gna will shut down.
At some point I'll have to declare svn.gna.org closed to further commits
to enable the final migration; probably there will be a period where
committing new code will not be possible.
=-=-=-=-=-
Help wanted:
- Do you know of pre-2006 mailing list archives (complete or partial)?
(From when the mailing lists were @freeciv.org, hosted by
complete.org.)
gmane.org had them, but is not currently reliable.
- Do you know anything about the CVS->Subversion migration?
Various evidence points to it happening 2005-10-11 and being done
with 'cvs2svn'.
If the pre-migration CVS repository (pile of ,v files) is preserved
somewhere (even partially) that would be useful.
(So that I can take the opportunity to investigate and rectify any
history mangling during this conversion; everyone I know of who's
done a svn->git migration says they found history badness from a
previous CVS->svn migration. But this will not block the migration if
it doesn't turn up.)
repository from Subversion to Git. (Currently we expect to host the
result on Github.)
I'm just getting started on this conversion (offline), but one thing
that will be needed is full committer details of the form
Jacob Nevins <***@chiark.greenend.org.uk>
for every commit, where currently we just have local usernames like
'jtn'. Once we have published an official git history, these details
will be essentially unchangeable for existing commits.
If you have a preference as to how your commits will appear, please let
me know, including your svn/CVS user ID. (If you send me a preference
you'll probably save me time.)
By default I will make some effort to track down real names and current
public email addresses, starting with the most prolific contributors and
probably stopping when I run out of time.
I will use Gna real names where available, but I will not use your
registered Gna email address if it's not public on Gna, unless I'm
reasonably convinced it is otherwise a public email address for you.
I'll also be using <http://www.freeciv.org/wiki/People> as a resource.
For anyone I can't track down (or if I run out of time) I expect I'll
use <svn-username>@freeciv.org (email to which will in most cases
bounce; I don't intend to set up forwarding in the first instance).
I don't currently intend to go through the history and populate separate
author/committer fields correctly (except perhaps for the very early CVS
history which was all committed by a user called 'freeciv'), so above
I'm talking only about people who've actually committed to CVS or SVN.
I've attached a list of distinct usernames found in the repository as of
r35175, sorted alphabetically.
=-=-=-=-=-
I intend to use the reposurgeon tool, and broadly follow the method of
<http://www.catb.org/~esr/reposurgeon/dvcs-migration-guide.html>,
although not slavishly (for instance I'm not going to rewrite commit
messages to have git-ish summaries, and I'd like to preserve svn
revision information).
I hope to put candidate histories up somewhere for review as they become
available. I'm aiming to be done by the end of April, because that is
the best guess we have about when Gna will shut down.
At some point I'll have to declare svn.gna.org closed to further commits
to enable the final migration; probably there will be a period where
committing new code will not be possible.
=-=-=-=-=-
Help wanted:
- Do you know of pre-2006 mailing list archives (complete or partial)?
(From when the mailing lists were @freeciv.org, hosted by
complete.org.)
gmane.org had them, but is not currently reliable.
- Do you know anything about the CVS->Subversion migration?
Various evidence points to it happening 2005-10-11 and being done
with 'cvs2svn'.
If the pre-migration CVS repository (pile of ,v files) is preserved
somewhere (even partially) that would be useful.
(So that I can take the opportunity to investigate and rectify any
history mangling during this conversion; everyone I know of who's
done a svn->git migration says they found history badness from a
previous CVS->svn migration. But this will not block the migration if
it doesn't turn up.)