all messages for Guix-related lists mirrored at yhetil.org
 help / color / mirror / code / Atom feed
From: "Clément Lassieur" <clement@lassieur.org>
To: 26734@debbugs.gnu.org
Subject: bug#26734: Snippets (even empty ones) of tar sources reset the timestamps of all files
Date: Mon, 01 May 2017 15:57:00 +0200	[thread overview]
Message-ID: <874lx4d6j7.fsf@lassieur.org> (raw)

Please consider the following package:

--8<---------------cut here---------------start------------->8---
(define-public hello
  (package
    (name "hello")
    (version "2.10")
    (source (origin
              (method url-fetch)
              (uri (string-append "mirror://gnu/hello/hello-" version
                                  ".tar.gz"))
              (sha256
               (base32
                "0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i"))
              (modules '((guix build utils)))
              (snippet '(begin #t))))
    (build-system gnu-build-system)
    (synopsis "Hello, GNU world: An example GNU package")
    (description
     "GNU Hello prints the message \"Hello, world!\" and then exits.  It
serves as an example of standard GNU coding practices.  As such, it supports
command-line arguments, multiple languages, and so on.")
    (home-page "https://www.gnu.org/software/hello/")
    (license gpl3+)))
--8<---------------cut here---------------end--------------->8---

It is the 'hello' package with two extra lines that add an empty
snippet.

--8<---------------cut here---------------start------------->8---
(modules '((guix build utils)))
(snippet '(begin #t))
--8<---------------cut here---------------end--------------->8---

Now here is the output of 'ls -l' within the original (extracted)
'hello':

--8<---------------cut here---------------start------------->8---
-rw-r--r-- 1 clement users  93787 Nov 16  2014 ABOUT-NLS
-rw-r--r-- 1 clement users  43830 Nov 16  2014 aclocal.m4
-rw-r--r-- 1 clement users    593 Jul 19  2014 AUTHORS
drwxr-xr-x 3 clement users   4096 Nov 16  2014 build-aux/
-rw-r--r-- 1 clement users  12988 Nov 16  2014 ChangeLog
-rw-r--r-- 1 clement users  30632 Jul 19  2014 ChangeLog.O
-rw-r--r-- 1 clement users  32871 Nov 16  2014 config.in
-rwxr-xr-x 1 clement users 449922 Nov 16  2014 configure
-rw-r--r-- 1 clement users   2408 Oct  5  2014 configure.ac
drwxr-xr-x 2 clement users   4096 Nov 16  2014 contrib/
-rw-r--r-- 1 clement users  35147 Dec 12  2013 COPYING
drwxr-xr-x 2 clement users   4096 Nov 16  2014 doc/
-rw-r--r-- 1 clement users   4573 Nov 16  2014 GNUmakefile
-r--r--r-- 1 clement users   1400 Nov 16  2014 hello.1
-rw-r--r-- 1 clement users  15752 Dec 29  2013 INSTALL
drwxr-xr-x 2 clement users   4096 Nov 16  2014 lib/
drwxr-xr-x 2 clement users   4096 Nov 16  2014 m4/
-rw-r--r-- 1 clement users  63223 Nov 16  2014 maint.mk
-rw-r--r-- 1 clement users   4367 Aug  5  2014 Makefile.am
-rw-r--r-- 1 clement users 143282 Nov 16  2014 Makefile.in
drwxr-xr-x 2 clement users   4096 Nov 16  2014 man/
-rw-r--r-- 1 clement users   4023 Nov 16  2014 NEWS
drwxr-xr-x 2 clement users   4096 Nov 16  2014 po/
-rw-r--r-- 1 clement users   3582 Jul 19  2014 README
-rw-r--r-- 1 clement users   1571 Jul 19  2014 README-dev
-rw-r--r-- 1 clement users   3014 Nov 16  2014 README-release
drwxr-xr-x 2 clement users   4096 Nov 16  2014 src/
drwxr-xr-x 2 clement users   4096 Nov 16  2014 tests/
-rw-r--r-- 1 clement users    898 Jul 19  2014 THANKS
-rw-r--r-- 1 clement users     75 Jul 19  2014 TODO
--8<---------------cut here---------------end--------------->8---

And here is the same output within the modified 'hello' (with the empty
snippet):

--8<---------------cut here---------------start------------->8---
-rw-r--r-- 1 clement users  93787 Jan  1  1970 ABOUT-NLS
-rw-r--r-- 1 clement users  43830 Jan  1  1970 aclocal.m4
-rw-r--r-- 1 clement users    593 Jan  1  1970 AUTHORS
drwxr-xr-x 3 clement users   4096 Jan  1  1970 build-aux/
-rw-r--r-- 1 clement users  12988 Jan  1  1970 ChangeLog
-rw-r--r-- 1 clement users  30632 Jan  1  1970 ChangeLog.O
-rw-r--r-- 1 clement users  32871 Jan  1  1970 config.in
-rwxr-xr-x 1 clement users 449922 Jan  1  1970 configure
-rw-r--r-- 1 clement users   2408 Jan  1  1970 configure.ac
drwxr-xr-x 2 clement users   4096 Jan  1  1970 contrib/
-rw-r--r-- 1 clement users  35147 Jan  1  1970 COPYING
drwxr-xr-x 2 clement users   4096 Jan  1  1970 doc/
-rw-r--r-- 1 clement users   4573 Jan  1  1970 GNUmakefile
-r--r--r-- 1 clement users   1400 Jan  1  1970 hello.1
-rw-r--r-- 1 clement users  15752 Jan  1  1970 INSTALL
drwxr-xr-x 2 clement users   4096 Jan  1  1970 lib/
drwxr-xr-x 2 clement users   4096 Jan  1  1970 m4/
-rw-r--r-- 1 clement users  63223 Jan  1  1970 maint.mk
-rw-r--r-- 1 clement users   4367 Jan  1  1970 Makefile.am
-rw-r--r-- 1 clement users 143282 Jan  1  1970 Makefile.in
drwxr-xr-x 2 clement users   4096 Jan  1  1970 man/
-rw-r--r-- 1 clement users   4023 Jan  1  1970 NEWS
drwxr-xr-x 2 clement users   4096 Jan  1  1970 po/
-rw-r--r-- 1 clement users   3582 Jan  1  1970 README
-rw-r--r-- 1 clement users   1571 Jan  1  1970 README-dev
-rw-r--r-- 1 clement users   3014 Jan  1  1970 README-release
drwxr-xr-x 2 clement users   4096 Jan  1  1970 src/
drwxr-xr-x 2 clement users   4096 Jan  1  1970 tests/
-rw-r--r-- 1 clement users    898 Jan  1  1970 THANKS
-rw-r--r-- 1 clement users     75 Jan  1  1970 TODO
--8<---------------cut here---------------end--------------->8---

As you can see, the timestamps are not the same, although there is
nothing in the snippet.

I think the code is in guix/packages.scm (patch-and-repack).  Here is my
understanding of what happens: the presence of the snippet leads to 3
things:

1. extraction of the archive,
2. modification of some files (because of patches or snippet),
3. compression of the archive.

Step 3 sets all timestamps to zero so to avoid non-determinism in the
archive.  We obviously don't want archives to depend on the time of
their creation.  But I believe we should only reset the timestamps of
the files we modified in step 2.  Unmodified files should stay
unmodified.

What's the point?  Well, while working on the 0ad package, I realized
that building with 32 cores would always fail with a snippet (even if it
is empty) and always succeed without a snippet.  It took me a few hours
to understand this, and I put a comment in the 0ad package.  Maybe it is
a 0ad bug (some tricky race condition that depends of file timestamps),
but I believe it is also a Guix bug: an empty snippet should not change
in any way the binary output of a package.

I tried to patch 'patch-and-repack', but it triggers a full
rebuild...  WDYT?

             reply	other threads:[~2017-05-01 13:58 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-05-01 13:57 Clément Lassieur [this message]
2017-05-02 13:00 ` bug#26734: Snippets (even empty ones) of tar sources reset the timestamps of all files Ludovic Courtès
2017-05-02 13:17   ` Clément Lassieur
2017-05-03  8:58     ` Ludovic Courtès
2017-05-03 21:45       ` Mark H Weaver
2017-05-03 22:01         ` Ludovic Courtès

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=874lx4d6j7.fsf@lassieur.org \
    --to=clement@lassieur.org \
    --cc=26734@debbugs.gnu.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.
Code repositories for project(s) associated with this external index

	https://git.savannah.gnu.org/cgit/guix.git

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.