Differences between revisions 12 and 13
Revision 12 as of 2013-08-16 08:07:12
Size: 4617
Editor: Lunar
Comment: formating fix
Revision 13 as of 2013-08-16 08:24:26
Size: 5762
Editor: Lunar
Comment: document more possible issues
Deletions are marked like this. Additions are marked like this.
Line 89: Line 89:
Build systems tend to capture information about the environment that
makes them produce different results accross different systems, despite
having the same architecture and software installed.

Ideally, such variations should be fixed in the build system itself, but
it might sometimes not be possible.

=== Build paths ===

These should really be patched out in one way or another. This is not useful
information and can actually hide real bugs.

For debug files, use DebianPts:debugedit.

=== Locale ===

Builds should be made with `LC_ALL=C.UTF-8`.

It's quite unpractical to force such value in `debian/rules` and there is
actually no reason this should not be the default.

'''Actions:'''

 * We could make dpkg-buildpackage exports this variable; but we would
   need to change the policy to make dpkg-buildpackage be the canonical
   solution to build package.

=== hostname, uname output, username ===

'''Actions''':

 * We could write a `LD_PRELOAD` library that could answers consistent results
for several system calls on the same model as `libfaketime`. Bdale suggested
we call it `liblietome`.

The goal

It should be possible to reproduce, byte for byte, every build of every package in Debian.

For now, we will start with a few maintainers who want to opt in to this goal as we flesh out the details of what will make it possible. This page tracks our progress.

Drivers

Status

  • Proof of concept success: hello package: Contents of data.tar.gz and control.tar.gz can be made reproducible when 'gzip' replaced by 'gzip -n' in debian/rules. (#xyz)

  • Buy-in within Debian: 5 packages from 5 maintainers are interested, of which 0 so far have reproducible contents of {data,control}.tar.gz

  • Waiting on a few dpkg bugs for avoiding timestamps and file order inconsistency in {data,control}.tar.gz (or .xz)
  • You can use a script to rebuild a package, with the same build-depends that were used by the build daemons. See "How to reproduce a build" below.
  • Things that need further investigation (by e.g. you!)
    • Document how to use Lunar's script to reproduce a build.
    • Find out if {control,data}.tar.gz files created by dpkg 1.17.1+ have a timestamp embedded.

Use cases

  • If the Debian build daemons are compromised, end users can assure themselves that their binaries are OK if they can regenerate them (and their build dependencies). (You could use a more complicated equivalence test than "do the hashes match?" but if the hashes do match, this is simple.)

Detailed package status list

  • alpine (Asheesh Laroia)
    • Status: Untested
  • haveged (Lunar)
    • Status: Unknown
  • iotop (pabs)
    • Status: Unknown
  • debhelper (joeyh)
    • Status: Unknown
  • magit (lindi)
    • Status: Unknown

Reproducing builds

There are two sides to the problem: first we need to record the initial build environment, and then we need a way to set up the same environment.

Recording the environment

The right place to record the build environment is the .changes file. Rationale: it lists the checksums of the build products and is signed by either the maintainer or the buildd operator.

To add a field to the .changes file, it is possible to add the following in debian/control:

XC-Build-Environment: ${misc:Build-Environment}

The substvars can be filled by using something like:

COLUMNS=999 | dpkg -l | awk '
            BEGIN { printf "misc:Build-Environment=" }
            /^ii/ { ORS=", "; print $2 " (= " $3 ")" }' |
        sed -e 's/, $//' >> debian/substvars

Ideally, this should go in debhelper. But in can be added manually to debian/rules in the meantime.

This does not work currently as dpkg-genchanges does not substitute the variable before adding the field in .changes! It can be fixed by a trivial patch against dpkg, see 719854.

Reproduce the build environment

Known bugs we are waiting on

  • dpkg: (719844) about gzip timestamps

  • dpkg: (719845) tar directory order

Different problems, and their solutions

Build systems tend to capture information about the environment that makes them produce different results accross different systems, despite having the same architecture and software installed.

Ideally, such variations should be fixed in the build system itself, but it might sometimes not be possible.

Build paths

These should really be patched out in one way or another. This is not useful information and can actually hide real bugs.

For debug files, use debugedit.

Locale

Builds should be made with LC_ALL=C.UTF-8.

It's quite unpractical to force such value in debian/rules and there is actually no reason this should not be the default.

Actions:

  • We could make dpkg-buildpackage exports this variable; but we would
    • need to change the policy to make dpkg-buildpackage be the canonical solution to build package.

hostname, uname output, username

Actions:

  • We could write a LD_PRELOAD library that could answers consistent results

for several system calls on the same model as libfaketime. Bdale suggested we call it liblietome.

Non-problems

  • You might think ELF binaries (e.g. /usr/bin/hello in the hello package) have embedded timestamps. Luckily, they don't!

Data files in data.tar.gz have timestamps

  • Recommended solution:
    • Use the timestamp of the of the last debian/changelog entry as reference.
    • touch all files to the reference timestamp before building the binary packages.
    • gzip -n when gzipping anything
    • get rid of non-determinisim (yup...)
    • Alternate solutions:
      • (or) libfaketime (probably breaks some things) (sudo apt-get install faketime)

{data,control}.tar.{gz,xz,bz2} may have timestamps

  • dpkg 1.17.1 might or might not store a timestamp for the .gz versions of these files.
  • *.xz and *.bz2 seem to provide no ability to store a timestamp.

{data,control}.tar.{gz,xz,bz2} will store files in readdir order

This is dependent on an accident of filesystem layout at build time, so it would sometimes not be reproducible.

We should probably fix this in dpkg by sorting the contents of the tar files.

References