[PATCH 0/3] Handle signed tags with 'export' remote helpers

From: John Keeping <john@keeping.me.uk>
Date: Apr 14, 2013, 10:57 UTC
Message-ID: <cover.1365936811.git.john@keeping.me.uk>
In-Reply-To: <CAGdFq_g+kk-Fy1fcV6D5x4kroRXX63T8wjKNUqqfu39wUkSO6A@mail.gmail.com>

On Sun, Apr 07, 2013 at 05:02:48PM -0700, Sverre Rabbelier wrote:

Show 11 quoted lines

> On Sun, Apr 7, 2013 at 2:46 PM, Jonathan Nieder <jrnieder@gmail.com> wrote:
> > The remote helper infrastructure is certainly being unhelpful here.  I
> > wonder if transport-helper should just pass --signed-tag=strip and be
> > done with it (leaving open the possibility of a capability to switch
> > to --signed-tag=verbatim when someone wants to teach the testgit
> > helper to support that).  What do you think?
> 
> I think that's (at least for now) the right thing to do. Passing
> anything but signed-tag=strip should be triggered by a capability from
> the helper, since most helpers won't know how to deal with signed
> tags.

I don't like the idea of silently stripping tags, so how about this?

Patch 1 adds a new 'warn-strip' mode to 'fast-export --signed-tags=...' which strips tags but issues a warning when doing so. Then we make transport-helper use that before finally adding a new capability to allow a remote helper to change '--signed-tags=warn-strip' into '--signed-tags=verbatim'.

John Keeping (3):
  fast-export: add --signed-tags=warn-strip mode
  transport-helper: pass --signed-tags=warn-strip to fast-export
  transport-helper: add 'signed-tags' capability

 Documentation/git-fast-export.txt   | 10 ++++++----
 Documentation/gitremote-helpers.txt |  4 ++++
 builtin/fast-export.c               |  8 +++++++-
 git-remote-testgit                  |  1 +
 t/t5801-remote-helpers.sh           | 20 ++++++++++++++++++++
 t/t9350-fast-export.sh              |  6 ++++++
 transport-helper.c                  |  7 ++++++-
 7 files changed, 50 insertions(+), 6 deletions(-)

-- 
1.8.2.694.ga76e9c3.dirty