Re: [PATCH] fetch-pack: unify ref in and out param

On Wed, Aug 01, 2018 at 01:13:20PM -0700, Jonathan Tan wrote:

When a user fetches:
 - at least one up-to-date ref and at least one non-up-to-date ref,
 - using HTTP with protocol v0 (or something else that uses the fetch
   command of a remote helper)
some refs might not be updated after the fetch.

This bug was introduced in commit 989b8c4452 ("fetch-pack: put shallow
info in output parameter", 2018-06-28) which allowed transports to
report the refs that they have fetched in a new out-parameter
"fetched_refs". If they do so, transport_fetch_refs() makes this
information available to its caller.

Users of "fetched_refs" rely on the following 3 properties:
 (1) it is the complete list of refs that was passed to
     transport_fetch_refs(),
 (2) it has shallow information (REF_STATUS_REJECT_SHALLOW set if
     relevant), and
 (3) it has updated OIDs if ref-in-want was used (introduced after
     989b8c4452).
[...]

Thanks, this is a very clear and well-organized commit message. It
answers my questions, and I agree with the general notion of "we can
figure out the right API for ref patterns later" approach.

 builtin/clone.c             |  4 ++--
 builtin/fetch.c             | 28 ++++------------------------
 fetch-object.c              |  2 +-
 fetch-pack.c                | 30 +++++++++++++++---------------
 t/t5551-http-fetch-smart.sh | 18 ++++++++++++++++++
 transport-helper.c          |  6 ++----
 transport-internal.h        |  9 +--------
 transport.c                 | 34 ++++++----------------------------
 transport.h                 |  3 +--

The patch itself looks sane to me, and obviously fixes the problem. I
cannot offhand think of any reason that munging the existing list would
be a problem (though it has been a while since I have dealt with this
code, so take that with the appropriate grain of salt).

-Peff