All of lore.kernel.org
 help / color / mirror / Atom feed
From: Junio C Hamano <gitster@pobox.com>
To: Russell Steicke <russellsteicke@gmail.com>
Cc: Dongas <dongas86@gmail.com>, git@vger.kernel.org
Subject: Re: A question about changing remote repo name
Date: Tue, 05 Jan 2010 17:02:58 -0800	[thread overview]
Message-ID: <7vd41oqc3h.fsf@alter.siamese.dyndns.org> (raw)
In-Reply-To: <c1b8b6671001051607j6b6398d8gf08f7615a38de64e@mail.gmail.com> (Russell Steicke's message of "Wed\, 6 Jan 2010 08\:07\:48 +0800")

Russell Steicke <russellsteicke@gmail.com> writes:

> Ah yes, I forgot about those files.  I _think_ (but I'm not certain
> without digging into the code) that it's always safe to rm
> packed-refs, as that file is a cache for quicker access to the refs.

No what you think is incorrect, no it is not safe at all, and no it is not
a cache.

refs are created loose as files under $GIT_DIR/refs/.  When pack-refs is
run (either from the toplevel or as part of gc), these loose ref files are
removed and instead written to $GIT_DIR/packed-refs file.  When a new ref
is created or an existing ref is updated, only $GIT_DIR/refs/ is modified.

When reading or enumerating, a loose ref, if found, is used.  Otherwise
a record in packed-refs is consulted.  The consequences are:

 - If you have $GIT_DIR/refs/heads/master and at the same time, an entry
   for refs/heads/master exists in $GIT_DIR/packed-refs, removing the
   entry in packed-refs happens to be safe, because it was already masked.

 - If you do not have $GIT_DIR/refs/heads/old, but an entry for
   refs/heads/old exists in $GIT_DIR/packed-refs, removing the entry in
   packed-refs means you lose the "old" branch.  You don't want to do
   that.

 - When deleting a ref, we remove it from $GIT_DIR/refs/, but at the same
   time, if packed-refs contains an entry for it, we remove it from
   there.  Otherwise, the removal of the loose one will unmask the stale
   entry in packed-refs file.

  parent reply	other threads:[~2010-01-06  1:03 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-04  6:45 A question about changing remote repo name Dongas
2010-01-04 20:09 ` Miklos Vajna
2010-01-05  1:53   ` Dongas
2010-01-05  1:57     ` Erik Faye-Lund
2010-01-05  2:25       ` Dongas
2010-01-05  2:52 ` Russell Steicke
2010-01-05  4:30   ` Dongas
2010-01-05 21:15     ` David Aguilar
2010-01-06  3:55       ` Dongas
2010-01-06  0:07     ` Russell Steicke
2010-01-06  0:27       ` Johannes Schindelin
2010-01-07  1:53         ` Russell Steicke
2010-01-06  1:02       ` Junio C Hamano [this message]
2010-01-06  3:58       ` Dongas

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=7vd41oqc3h.fsf@alter.siamese.dyndns.org \
    --to=gitster@pobox.com \
    --cc=dongas86@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=russellsteicke@gmail.com \
    /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.