git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff King <peff@peff.net>
To: Ken Moffat <zarniwhoop@ntlworld.com>
Cc: "Junio C Hamano" <gitster@pobox.com>,
	"Vicent Martí" <tanoku@gmail.com>,
	"Stefan Näwe" <stefan.naewe@atlas-elektronik.com>,
	"Javier Domingo Cansino" <javierdo1@gmail.com>,
	"git@vger.kernel.org" <git@vger.kernel.org>,
	"Linux Kernel" <linux-kernel@vger.kernel.org>
Subject: Re: [ANNOUNCE] Git v1.9-rc0
Date: Wed, 22 Jan 2014 21:09:13 -0500	[thread overview]
Message-ID: <20140123020913.GF17254@sigill.intra.peff.net> (raw)
In-Reply-To: <20140122203030.GB14211@milliways>

On Wed, Jan 22, 2014 at 08:30:30PM +0000, Ken Moffat wrote:

>  Two questions: Does regenerating (e.g. if the tarball has dropped
> out of the cache) change its sums (md5sum or similar) ?  In (beyond)
> linuxfromscratch we use md5sums to verify that a tarball has not
> changed.

The tarballs we auto-generate from tags are cached, but they can
change if the cached version expires _and_ the archive-generation code
changes.

We use "git archive" to generate the tarballs themselves, and then gzip
the with "gzip -n". So it should be consistent from run to run. However,
very occasionally there are bugfixes in "git archive" which can affect
the output. E.g., commit 22f0dcd (archive-tar: split long paths more
carefully, 2013-01-05) changes the representation of certain long paths,
and generating a tarball with and without it will result in different
checksums (for some repos).

So if you are planning on baking md5sums into a package-build system, it
is much better to point at "official" releases which are rolled once by
the project maintainer, rather than the automatic tag page.

Junio, since you prepare such tarballs[1] anyway for kernel.org, it
might be worth uploading them to the "Releases" page of git/git.  I
imagine there is a programmatic way to do so via GitHub's API, but I
don't know offhand. I can look into it if you are interested.

-Peff

  parent reply	other threads:[~2014-01-23  2:09 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-16 21:03 What's cooking in git.git (Jan 2014, #03; Thu, 16) Junio C Hamano
2014-01-21 22:14 ` [ANNOUNCE] Git v1.9-rc0 Junio C Hamano
2014-01-22 12:53   ` Javier Domingo Cansino
2014-01-22 15:42     ` Junio C Hamano
2014-01-22 15:49     ` Stefan Näwe
2014-01-22 16:11       ` Junio C Hamano
2014-01-22 17:26         ` Vicent Martí
2014-01-22 18:10           ` Junio C Hamano
2014-01-22 20:30             ` Ken Moffat
2014-01-22 21:04               ` Junio C Hamano
2014-01-22 23:11                 ` Ken Moffat
2014-01-23  2:09               ` Jeff King [this message]
2014-01-23 18:15                 ` Junio C Hamano
2014-01-24 23:36                   ` Jeff King
2014-01-27 22:56                     ` Junio C Hamano
2014-01-28  0:04                       ` Jeff King
2014-01-27 17:59   ` Kacper Kornet
2014-01-27 18:58     ` Jonathan Nieder
2014-01-28 12:28       ` Kacper Kornet
2014-01-30 22:20         ` Sitaram Chamarty

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=20140123020913.GF17254@sigill.intra.peff.net \
    --to=peff@peff.net \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=javierdo1@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=stefan.naewe@atlas-elektronik.com \
    --cc=tanoku@gmail.com \
    --cc=zarniwhoop@ntlworld.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).