All of lore.kernel.org
 help / color / mirror / Atom feed
From: Eric Wong <e@80x24.org>
To: Craig McQueen <craig.mcqueen@innerrange.com>
Cc: git@vger.kernel.org
Subject: Re: git svn and SVN property svn:original-date
Date: Thu, 23 Mar 2017 17:31:56 +0000	[thread overview]
Message-ID: <20170323173156.GA29732@starla> (raw)
In-Reply-To: <5500469A22567C4BAF673A6E86AFA3A40295E65B4365@IR-CENTRAL.corp.innerrange.com>

Craig McQueen <craig.mcqueen@innerrange.com> wrote:
> Eric Wong wrote:
> > Craig McQueen <craig.mcqueen@innerrange.com> wrote:
> > > Apparently SVN revisions can have an "svn:original-date" property, which
> > would be good to set on dcommit, to preserve the timestamp from the git
> > repository.
> > >
> > >
> > https://subversion.apache.org/docs/api/1.7/group__svn__props__revision
> > > __props.html#ga8f17351dd056149da9cb490f1daf4018
> > 
> > Any idea if which versions of SVN it's supported in and how recent the
> > feature is?
> 
> I see discussion about it in 2003, so I guess it's been there right from 1.0.0.
> 
> > Perhaps we can enable it everywhere, and maybe only old clients won't
> > understand it, but won't fail; and we could start using it as the author date
> > with "git svn fetch".
> 
> Using it for author date sounds sensible.

OK, I'd be inclined to accept a patch + tests for that;
my plate might be full with non-git stuff until April.

> > OTOH, that would break the (perhaps unofficial) independently-created-git-
> > svn-mirrors-should-have-same-oids-by-default
> > rule when people run different versions of git, so maybe it could be an
> > option...
> 
> Hmm, good question. Maybe it should be an option, though I
> hope it would be enabled by default (since the feature would
> be more metadata-preserving, which is a good thing), with an
> option to disable it to allow backwards compatibility with
> people running an older version of git. That's my opinion
> anyway, and I realise my opinion is not necessarily
> well-informed regarding all considerations.

Given git-svn behavior hasn't changed much in over a decade, I'd
rather be conservative with regards to changing any existing
behavior.  We could offer one warning about it to inform users
about it if original-date is seen, however.

> Note, I'm unclear as to whether Subversion is willing to store
> timezone information in svn:original-date. But I guess having
> git author date correspond to svn:original-date is an
> improvement for preserving more metadata, even if the timezone
> is lost in the process.

Can you find any examples of it in the wild?  I prefer having
real-world examples to look at.  Thanks.

      reply	other threads:[~2017-03-23 17:32 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-22 23:26 git svn and SVN property svn:original-date Craig McQueen
2017-03-23  2:56 ` Eric Wong
2017-03-23  4:18   ` Craig McQueen
2017-03-23 17:31     ` Eric Wong [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170323173156.GA29732@starla \
    --to=e@80x24.org \
    --cc=craig.mcqueen@innerrange.com \
    --cc=git@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.