Building Debian in a GitLab Pipeline

After thinking about multi-stage Debian rebuilds I wanted to implement the idea. Recall my illustration:

Earlier I rebuilt all packages that make up the difference between Ubuntu and Trisquel. It turned out to be a 42% bit-by-bit identical similarity. To check the generality of my approach, I rebuilt the difference between Debian and Devuan too. That was the debdistreproduce project. It “only” had to orchestrate building up to around 500 packages for each distribution and per architecture.

Differential reproducible rebuilds doesn’t give you the full picture: it ignore the shared package between the distribution, which make up over 90% of the packages. So I felt a desire to do full archive rebuilds. The motivation is that in order to trust Trisquel binary packages, I need to trust Ubuntu binary packages (because that make up 90% of the Trisquel packages), and many of those Ubuntu binaries are derived from Debian source packages. How to approach all of this? Last year I created the debdistrebuild project, and did top-50 popcon package rebuilds of Debian bullseye, bookworm, trixie, and Ubuntu noble and jammy, on a mix of amd64 and arm64. The amount of reproducibility was lower. Primarily the differences were caused by using different build inputs.

Last year I spent (too much) time creating a mirror of snapshot.debian.org, to be able to have older packages available for use as build inputs. I have two copies hosted at different datacentres for reliability and archival safety. At the time, snapshot.d.o had serious rate-limiting making it pretty unusable for massive rebuild usage or even basic downloads. Watching the multi-month download complete last year had a meditating effect. The completion of my snapshot download co-incided with me realizing something about the nature of rebuilding packages. Let me below give a recap of the idempotent rebuilds idea, because it motivate my work to build all of Debian from a GitLab pipeline.

One purpose for my effort is to be able to trust the binaries that I use on my laptop. I believe that without building binaries from source code, there is no practically feasible way to trust binaries. To trust any binary you receive, you can de-assemble the bits and audit the assembler instructions for the CPU you will execute it on. Doing that on a OS-wide level this is unpractical. A more practical approach is to audit the source code, and then confirm that the binary is 100% bit-by-bit identical to one that you can build yourself (from the same source) on your own trusted toolchain. This is similar to a reproducible build.

My initial goal with debdistrebuild was to get to 100% bit-by-bit identical rebuilds, and then I would have trustworthy binaries. Or so I thought. This also appears to be the goal of reproduce.debian.net. They want to reproduce the official Debian binaries. That is a worthy and important goal. They achieve this by building packages using the build inputs that were used to build the binaries. The build inputs are earlier versions of Debian packages (not necessarily from any public Debian release), archived at snapshot.debian.org.

I realized that these rebuilds would be not be sufficient for me: it doesn’t solve the problem of how to trust the toolchain. Let’s assume the reproduce.debian.net effort succeeds and is able to 100% bit-by-bit identically reproduce the official Debian binaries. Which appears to be within reach. To have trusted binaries we would “only” have to audit the source code for the latest version of the packages AND audit the tool chain used. There is no escaping from auditing all the source code — that’s what I think we all would prefer to focus on, to be able to improve upstream source code.

The trouble is about auditing the tool chain. With the Reproduce.debian.net approach, that is a recursive problem back to really ancient Debian packages, some of them which may no longer build or work, or even be legally distributable. Auditing all those old packages is a LARGER effort than auditing all current packages! Doing auditing of old packages is of less use to making contributions: those releases are old, and chances are any improvements have already been implemented and released. Or that improvements are no longer applicable because the projects evolved since the earlier version.

See where this is going now? I reached the conclusion that reproducing official binaries using the same build inputs is not what I’m interested in. I want to be able to build the binaries that I use from source using a toolchain that I can also build from source. And preferably that all of this is using latest version of all packages, so that I can contribute and send patches for them, to improve matters.

The toolchain that Reproduce.Debian.Net is using is not trustworthy unless all those ancient packages are audited or rebuilt bit-by-bit identically, and I don’t see any practical way forward to achieve that goal. Nor have I seen anyone working on that problem. It is possible to do, though, but I think there are simpler ways to achieve the same goal.

My approach to reach trusted binaries on my laptop appears to be a three-step effort:

  • Encourage an idempotently rebuildable Debian archive, i.e., a Debian archive that can be 100% bit-by-bit identically rebuilt using Debian itself.
  • Construct a smaller number of binary *.deb packages based on Guix binaries that when used as build inputs (potentially iteratively) leads to 100% bit-by-bit identical packages as in step 1.
  • Encourage a freedom respecting distribution, similar to Trisquel, from this idempotently rebuildable Debian.

How to go about achieving this? Today’s Debian build architecture is something that lack transparency and end-user control. The build environment and signing keys are managed by, or influenced by, unidentified people following undocumented (or at least not public) security procedures, under unknown legal jurisdictions. I always wondered why none of the Debian-derivates have adopted a modern GitDevOps-style approach as a method to improve binary build transparency, maybe I missed some project?

If you want to contribute to some GitHub or GitLab project, you click the ‘Fork’ button and get a CI/CD pipeline running which rebuild artifacts for the project. This makes it easy for people to contribute, and you get good QA control because the entire chain up until its artifact release are produced and tested. At least in theory. Many projects are behind on this, but it seems like this is a useful goal for all projects. This is also liberating: all users are able to reproduce artifacts. There is no longer any magic involved in preparing release artifacts. As we’ve seen with many software supply-chain security incidents for the past years, where the “magic” is involved is a good place to introduce malicious code.

To allow me to continue with my experiment, I thought the simplest way forward was to setup a GitDevOps-centric and user-controllable way to build the entire Debian archive. Let me introduce the debdistbuild project.

Debdistbuild is a re-usable GitLab CI/CD pipeline, similar to the Salsa CI pipeline. It provide one “build” job definition and one “deploy” job definition. The pipeline can run on GitLab.org Shared Runners or you can set up your own runners, like my GitLab riscv64 runner setup. I have concerns about relying on GitLab (both as software and as a service), but my ideas are easy to transfer to some other GitDevSecOps setup such as Codeberg.org. Self-hosting GitLab, including self-hosted runners, is common today, and Debian rely increasingly on Salsa for this. All of the build infrastructure could be hosted on Salsa eventually.

The build job is simple. From within an official Debian container image build packages using dpkg-buildpackage essentially by invoking the following commands.

sed -i 's/ deb$/ deb deb-src/' /etc/apt/sources.list.d/*.sources
apt-get -o Acquire::Check-Valid-Until=false update
apt-get dist-upgrade -q -y
apt-get install -q -y --no-install-recommends build-essential fakeroot
env DEBIAN_FRONTEND=noninteractive \
    apt-get build-dep -y --only-source $PACKAGE=$VERSION
useradd -m build
DDB_BUILDDIR=/build/reproducible-path
chgrp build $DDB_BUILDDIR
chmod g+w $DDB_BUILDDIR
su build -c "apt-get source --only-source $PACKAGE=$VERSION" > ../$PACKAGE_$VERSION.build
cd $DDB_BUILDDIR
su build -c "dpkg-buildpackage"
cd ..
mkdir out
mv -v $(find $DDB_BUILDDIR -maxdepth 1 -type f) out/

The deploy job is also simple. It commit artifacts to a Git project using Git-LFS to handle large objects, essentially something like this:

if ! grep -q '^pool/**' .gitattributes; then
    git lfs track 'pool/**'
    git add .gitattributes
    git commit -m"Track pool/* with Git-LFS." .gitattributes
fi
POOLDIR=$(if test "$(echo "$PACKAGE" | cut -c1-3)" = "lib"; then C=4; else C=1; fi; echo "$DDB_PACKAGE" | cut -c1-$C)
mkdir -pv pool/main/$POOLDIR/
rm -rfv pool/main/$POOLDIR/$PACKAGE
mv -v out pool/main/$POOLDIR/$PACKAGE
git add pool
git commit -m"Add $PACKAGE." -m "$CI_JOB_URL" -m "$VERSION" -a
if test "${DDB_GIT_TOKEN:-}" = ""; then
    echo "SKIP: Skipping git push due to missing DDB_GIT_TOKEN (see README)."
else
    git push -o ci.skip
fi

That’s it! The actual implementation is a bit longer, but the major difference is for log and error handling.

You may review the source code of the base Debdistbuild pipeline definition, the base Debdistbuild script and the rc.d/-style scripts implementing the build.d/ process and the deploy.d/ commands.

There was one complication related to artifact size. GitLab.org job artifacts are limited to 1GB. Several packages in Debian produce artifacts larger than this. What to do? GitLab supports up to 5GB for files stored in its package registry, but this limit is too close for my comfort, having seen some multi-GB artifacts already. I made the build job optionally upload artifacts to a S3 bucket using SHA256 hashed file hierarchy. I’m using Hetzner Object Storage but there are many S3 providers around, including self-hosting options. This hierarchy is compatible with the Git-LFS .git/lfs/object/ hierarchy, and it is easy to setup a separate Git-LFS object URL to allow Git-LFS object downloads from the S3 bucket. In this mode, only Git-LFS stubs are pushed to the git repository. It should have no trouble handling the large number of files, since I have earlier experience with Apt mirrors in Git-LFS.

To speed up job execution, and to guarantee a stable build environment, instead of installing build-essential packages on every build job execution, I prepare some build container images. The project responsible for this is tentatively called stage-N-containers. Right now it create containers suitable for rolling builds of trixie on amd64, arm64, and riscv64, and a container intended for as use the stage-0 based on the 20250407 docker images of bookworm on amd64 and arm64 using the snapshot.d.o 20250407 archive. Or actually, I’m using snapshot-cloudflare.d.o because of download speed and reliability. I would have prefered to use my own snapshot mirror with Hetzner bandwidth, alas the Debian snapshot team have concerns about me publishing the list of (SHA1 hash) filenames publicly and I haven’t been bothered to set up non-public access.

Debdistbuild has built around 2.500 packages for bookworm on amd64 and bookworm on arm64. To confirm the generality of my approach, it also build trixie on amd64, trixie on arm64 and trixie on riscv64. The riscv64 builds are all on my own hosted runners. For amd64 and arm64 my own runners are only used for large packages where the GitLab.com shared runners run into the 3 hour time limit.

What’s next in this venture? Some ideas include:

  • Optimize the stage-N build process by identifying the transitive closure of build dependencies from some initial set of packages.
  • Create a build orchestrator that launches pipelines based on the previous list of packages, as necessary to fill the archive with necessary packages. Currently I’m using a basic /bin/sh for loop around curl to trigger GitLab CI/CD pipelines with names derived from https://popcon.debian.org/.
  • Create and publish a dists/ sub-directory, so that it is possible to use the newly built packages in the stage-1 build phase.
  • Produce diffoscope-style differences of built packages, both stage0 against official binaries and between stage0 and stage1.
  • Create the stage-1 build containers and stage-1 archive.
  • Review build failures. On amd64 and arm64 the list is small (below 10 out of ~5000 builds), but on riscv64 there is some icache-related problem that affects Java JVM that triggers build failures.
  • Provide GitLab pipeline based builds of the Debian docker container images, cloud-images, debian-live CD and debian-installer ISO’s.
  • Provide integration with Sigstore and Sigsum for signing of Debian binaries with transparency-safe properties.
  • Implement a simple replacement for dpkg and apt using /bin/sh for use during bootstrapping when neither packaging tools are available.

What do you think?

Verified Reproducible Tarballs

Remember the XZ Utils backdoor? One factor that enabled the attack was poor auditing of the release tarballs for differences compared to the Git version controlled source code. This proved to be a useful place to distribute malicious data.

The differences between release tarballs and upstream Git sources is typically vendored and generated files. Lots of them. Auditing all source tarballs in a distribution for similar issues is hard and boring work for humans. Wouldn’t it be better if that human auditing time could be spent auditing the actual source code stored in upstream version control instead? That’s where auditing time would help the most.

Are there better ways to address the concern about differences between version control sources and tarball artifacts? Let’s consider some approaches:

  • Stop publishing (or at least stop building from) source tarballs that differ from version control sources.
  • Create recipes for how to derive the published source tarballs from version control sources. Verify that independently from upstream.

While I like the properties of the first solution, and have made effort to support that approach, I don’t think normal source tarballs are going away any time soon. I am concerned that it may not even be a desirable complete solution to this problem. We may need tarballs with pre-generated content in them for various reasons that aren’t entirely clear to us today.

So let’s consider the second approach. It could help while waiting for more experience with the first approach, to see if there are any fundamental problems with it.

How do you know that the XZ release tarballs was actually derived from its version control sources? The same for Gzip? Coreutils? Tar? Sed? Bash? GCC? We don’t know this! I am not aware of any automated or collaborative effort to perform this independent confirmation. Nor am I aware of anyone attempting to do this on a regular basis. We would want to be able to do this in the year 2042 too. I think the best way to reach that is to do the verification continuously in a pipeline, fixing bugs as time passes. The current state of the art seems to be that people audit the differences manually and hope to find something. I suspect many package maintainers ignore the problem and take the release source tarballs and trust upstream about this.

We can do better.

I have launched a project to setup a GitLab pipeline that invokes per-release scripts to rebuild that release artifact from git sources. Currently it only contain recipes for projects that I released myself. Releases which where done in a controlled way with considerable care to make reproducing the tarballs possible. The project homepage is here:

https://gitlab.com/debdistutils/verify-reproducible-releases

The project is able to reproduce the release tarballs for Libtasn1 v4.20.0, InetUtils v2.6, Libidn2 v2.3.8, Libidn v1.43, and GNU SASL v2.2.2. You can see this in a recent successful pipeline. All of those releases were prepared using Guix, and I’m hoping the Guix time-machine will make it possible to keep re-generating these tarballs for many years to come.

I spent some time trying to reproduce the current XZ release tarball for version 5.8.1. That would have been a nice example, wouldn’t it? First I had to somehow mimic upstream’s build environment. The XZ release tarball contains GNU Libtool files that are identified with version 2.5.4.1-baa1-dirty. I initially assumed this was due to the maintainer having installed libtool from git locally (after making some modifications) and made the XZ release using it. Later I learned that it may actually be coming from ArchLinux which ship with this particular libtool version. It seems weird for a distribution to use libtool built from a non-release tag, and furthermore applying patches to it, but things are what they are. I made some effort to setup an ArchLinux build environment, however the now-current Gettext version in ArchLinux seems to be more recent than the one that were used to prepare the XZ release. I don’t know enough ArchLinux to setup an environment corresponding to an earlier version of ArchLinux, which would be required to finish this. I gave up, maybe the XZ release wasn’t prepared on ArchLinux after all. Actually XZ became a good example for this writeup anyway: while you would think this should be trivial, the fact is that it isn’t! (There is another aspect here: fingerprinting the versions used to prepare release tarballs allows you to infer what kind of OS maintainers are using to make releases on, which is interesting on its own.)

I made some small attempts to reproduce the tarball for GNU Shepherd version 1.0.4 too, but I still haven’t managed to complete it.

Do you want a supply-chain challenge for the Easter weekend? Pick some well-known software and try to re-create the official release tarballs from the corresponding Git checkout. Is anyone able to reproduce anything these days? Bonus points for wrapping it up as a merge request to my project.

Happy Supply-Chain Security Hacking!

Reproducible Software Releases

Around a year ago I discussed two concerns with software release archives (tarball artifacts) that could be improved to increase confidence in the supply-chain security of software releases. Repeating the goals for simplicity:

  • Release artifacts should be built in a way that can be reproduced by others
  • It should be possible to build a project from source tarball that doesn’t contain any generated or vendor files (e.g., in the style of git-archive).

While implementing these ideas for a small project was accomplished within weeks – see my announcement of Libntlm version 1.8 – adressing this in complex projects uncovered concerns with tools that had to be addressed, and things stalled for many months pending that work.

I had the notion that these two goals were easy and shouldn’t be hard to accomplish. I still believe that, but have had to realize that improving tooling to support these goals takes time. It seems clear that these concepts are not universally agreed on and implemented generally.

I’m now happy to recap some of the work that led to releases of libtasn1 v4.20.0, inetutils v2.6, libidn2 v2.3.8, libidn v1.43. These releases all achieve these goals. I am working on a bunch of more projects to support these ideas too.

What have the obstacles so far been to make this happen? It may help others who are in the same process of addressing these concerns to have a high-level introduction to the issues I encountered. Source code for projects above are available and anyone can look at the solutions to learn how the problems are addressed.

First let’s look at the problems we need to solve to make “git-archive” style tarballs usable:

Version Handling

To build usable binaries from a minimal tarballs, it need to know which version number it is. Traditionally this information was stored inside configure.ac in git. However I use gnulib’s git-version-gen to infer the version number from the git tag or git commit instead. The git tag information is not available in a git-archive tarball. My solution to this was to make use of the export-subst feature of the .gitattributes file. I store the file .tarball-version-git in git containing the magic cookie like this:

$Format:%(describe)$

With this, git-archive will replace with a useful version identifier on export, see the libtasn1 patch to achieve this. To make use of this information, the git-version-gen script was enhanced to read this information, see the gnulib patch. This is invoked by ./configure to figure out which version number the package is for.

Translations

We want translations to be included in the minimal source tarball for it to be buildable. Traditionally these files are retrieved by the maintainer from the Translation project when running ./bootstrap, however there are two problems with this. The first one is that there is no strong authentication or versioning information on this data, the tools just download and place whatever wget downloaded into your source tree (printf-style injection attack anyone?). We could improve this (e.g., publish GnuPG signed translations messages with clear versioning), however I did not work on that further. The reason is that I want to support offline builds of packages. Downloading random things from the Internet during builds does not work when building a Debian package, for example. The translation project could solve this by making a monthly tarball with their translations available, for distributors to pick up and provide as a separate package that could be used as a build dependency. However that is not how these tools and projects are designed. Instead I reverted back to storing translations in git, something that I did for most projects back when I was using CVS 20 years ago. Hooking this into ./bootstrap and gettext workflow can be tricky (ideas for improvement most welcome!), but I used a simple approach to store all directly downloaded po/*.po files directly as po/*.po.in and make the ./bootstrap tool move them in place, see the libidn2 commit followed by the actual ‘make update-po’ commit with all the translations where one essential step is:

# Prime po/*.po from fall-back copy stored in git.
for poin in po/*.po.in; do
    po=$(echo $poin | sed 's/.in//')
    test -f $po || cp -v $poin $po
done
ls po/*.po | sed 's|.*/||; s|\.po$||' > po/LINGUAS

Fetching vendor files like gnulib

Most build dependencies are in the shape of “You need a C compiler”. However some come in the shape of “source-code files intended to be vendored”, and gnulib is a huge repository of such files. The latter is a problem when building from a minimal git archive. It is possible to consider translation files as a class of vendor files, since they need to be copied verbatim into the project build directory for things to work. The same goes for *.m4 macros from the GNU Autoconf Archive. However I’m not confident that the solution for all vendor files must be the same. For translation files and for Autoconf Archive macros, I have decided to put these files into git and merge them manually occasionally. For gnulib files, in some projects like OATH Toolkit I also store all gnulib files in git which effectively resolve this concern. (Incidentally, the reason for doing so was originally that running ./bootstrap took forever since there is five gnulib instances used, which is no longer the case since gnulib-tool was rewritten in Python.) For most projects, however, I rely on ./bootstrap to fetch a gnulib git clone when building. I like this model, however it doesn’t work offline. One way to resolve this is to make the gnulib git repository available for offline use, and I’ve made some effort to make this happen via a Gnulib Git Bundle and have explained how to implement this approach for Debian packaging. I don’t think that is sufficient as a generic solution though, it is mostly applicable to building old releases that uses old gnulib files. It won’t work when building from CI/CD pipelines, for example, where I have settled to use a crude way of fetching and unpacking a particular gnulib snapshot, see this Libntlm patch. This is much faster than working with git submodules and cloning gnulib during ./bootstrap. Essentially this is doing:

GNULIB_REVISION=$(. bootstrap.conf >&2; echo $GNULIB_REVISION)
wget -nv https://gitlab.com/libidn/gnulib-mirror/-/archive/$GNULIB_REVISION/gnulib-mirror-$GNULIB_REVISION.tar.gz
gzip -cd gnulib-mirror-$GNULIB_REVISION.tar.gz | tar xf -
rm -fv gnulib-mirror-$GNULIB_REVISION.tar.gz
export GNULIB_SRCDIR=$PWD/gnulib-mirror-$GNULIB_REVISION
./bootstrap --no-git
./configure
make

Test the git-archive tarball

This goes without saying, but if you don’t test that building from a git-archive style tarball works, you are likely to regress at some point. Use CI/CD techniques to continuously test that a minimal git-archive tarball leads to a usable build.

Mission Accomplished

So that wasn’t hard, was it? You should now be able to publish a minimal git-archive tarball and users should be able to build your project from it.

I recommend naming these archives as PROJECT-vX.Y.Z-src.tar.gz replacing PROJECT with your project name and X.Y.Z with your version number. The archive should have only one sub-directory named PROJECT-vX.Y.Z/ containing all the source-code files. This differentiate it against traditional PROJECT-X.Y.Z.tar.gz tarballs in that it embeds the git tag (which typically starts with v) and contains a wildcard-friendly -src substring. Alas there is no consistency around this naming pattern, and GitLab, GitHub, Codeberg etc all seem to use their own slightly incompatible variant.

Let’s go on to see what is needed to achieve reproducible “make dist” source tarballs. This is the release artifact that most users use, and they often contain lots of generated files and vendor files. These files are included to make it easy to build for the user. What are the challenges to make these reproducible?

Build dependencies causing different generated content

The first part is to realize that if you use tool X with version A to generate a file that goes into the tarball, version B of that tool may produce different outputs. This is a generic concern and it cannot be solved. We want our build tools to evolve and produce better outputs over time. What can be addressed is to avoid needless differences. For example, many tools store timestamps and versioning information in the generated files. This causes needless differences, which makes audits harder. I have worked on some of these, like Autoconf Archive timestamps but solving all of these examples will take a long time, and some upstream are reluctant to incorporate these changes. My approach meanwhile is to build things using similar environments, and compare the outputs for differences. I’ve found that the various closely related forks of GNU/Linux distributions are useful for this. Trisquel 11 is based on Ubuntu 22.04, and building my projects using both and comparing the differences only give me the relevant differences to improve. This can be extended to compare AlmaLinux with RockyLinux (for both versions 8 and 9), Devuan 5 against Debian 12, PureOS 10 with Debian 11, and so on.

Timestamps

Sometimes tools store timestamps in files in a way that is harder to fix. Two notable examples of this are *.po translation files and Texinfo manuals. For translation files, I have resolved this by making sure the files use a predictable POT-Creation-Date timestamp, and I set it to the modification timestamps of the NEWS file in the repository (which I set to the git commit of the latest commit elsewhere) like this:

dist-hook: po-CreationDate-to-mtime-NEWS
.PHONY: po-CreationDate-to-mtime-NEWS
po-CreationDate-to-mtime-NEWS: mtime-NEWS-to-git-HEAD
  $(AM_V_GEN)for p in $(distdir)/po/*.po $(distdir)/po/$(PACKAGE).pot; do \
    if test -f "$$p"; then \
      $(SED) -e 's,POT-Creation-Date: .*\\n",POT-Creation-Date: '"$$(env LC_ALL=C TZ=UTC0 stat --format=%y $(srcdir)/NEWS | cut -c1-16,31-)"'\\n",' < $$p > $$p.tmp && \
      if cmp $$p $$p.tmp > /dev/null; then \
        rm -f $$p.tmp; \
      else \
        mv $$p.tmp $$p; \
      fi \
    fi \
  done

Similarily, I set a predictable modification time of the texinfo source file like this:

dist-hook: mtime-NEWS-to-git-HEAD
.PHONY: mtime-NEWS-to-git-HEAD
mtime-NEWS-to-git-HEAD:
  $(AM_V_GEN)if test -e $(srcdir)/.git \
                && command -v git > /dev/null; then \
    touch -m -t "$$(git log -1 --format=%cd \
      --date=format-local:%Y%m%d%H%M.%S)" $(srcdir)/NEWS; \
  fi

However I’ve realized that this needs to happen earlier and probably has to be run during ./configure time, because the doc/version.texi file is generated on first build before running ‘make dist‘ and for some reason the file is not rebuilt at release time. The Automake texinfo integration is a bit inflexible about providing hooks to extend the dependency tracking.

The method to address these differences isn’t really important, and they change over time depending on preferences. What is important is that the differences are eliminated.

ChangeLog

Traditionally ChangeLog files were manually prepared, and still is for some projects. I maintain git2cl but recently I’ve settled with gnulib’s gitlog-to-changelog because doing so avoids another build dependency (although the output formatting is different and arguable worse for my git commit style). So the ChangeLog files are generated from git history. This means a shallow clone will not produce the same ChangeLog file depending on how deep it was cloned. For Libntlm I simply disabled use of generated ChangeLog because I wanted to support an even more extreme form of reproducibility: I wanted to be able to reproduce the full “make dist” source archives from a minimal “git-archive” source archive. However for other projects I’ve settled with a middle ground. I realized that for ‘git describe‘ to produce reproducible outputs, the shallow clone needs to include the last release tag. So it felt acceptable to assume that the clone is not minimal, but instead has some but not all of the history. I settled with the following recipe to produce ChangeLog's covering all changes since the last release.

dist-hook: gen-ChangeLog
.PHONY: gen-ChangeLog
gen-ChangeLog:
  $(AM_V_GEN)if test -e $(srcdir)/.git; then			\
    LC_ALL=en_US.UTF-8 TZ=UTC0					\
    $(top_srcdir)/build-aux/gitlog-to-changelog			\
       --srcdir=$(srcdir) --					\
       v$(PREV_VERSION)~.. > $(distdir)/cl-t &&			\
       { printf '\n\nSee the source repo for older entries\n'	\
         >> $(distdir)/cl-t &&					\
         rm -f $(distdir)/ChangeLog &&				\
         mv $(distdir)/cl-t $(distdir)/ChangeLog; }		\
  fi

I’m undecided about the usefulness of generated ChangeLog files within ‘make dist‘ archives. Before we have stable and secure archival of git repositories widely implemented, I can see some utility of this in case we lose all copies of the upstream git repositories. I can sympathize with the concept of ChangeLog files died when we started to generate them from git logs: the files no longer serve any purpose, and we can ask people to go look at the git log instead of reading these generated non-source files.

Long-term reproducible trusted build environment

Distributions comes and goes, and old releases of them goes out of support and often stops working. Which build environment should I chose to build the official release archives? To my knowledge only Guix offers a reliable way to re-create an older build environment (guix gime-machine) that have bootstrappable properties for additional confidence. However I had two difficult problems here. The first one was that I needed Guix container images that were usable in GitLab CI/CD Pipelines, and this side-tracked me for a while. The second one delayed my effort for many months, and I was inclined to give up. Libidn distribute a C# implementation. Some of the C# source code files included in the release tarball are generated. By what? You guess it, by a C# program, with the source code included in the distribution. This means nobody could reproduce the source tarball of Libidn without trusting someone elses C# compiler binaries, which were built from binaries of earlier releases, chaining back into something that nobody ever attempts to build any more and likely fail to build due to bit-rot. I had two basic choices, either remove the C# implementation from Libidn (which may be a good idea for other reasons, since the C and C# are unrelated implementations) or build the source tarball on some binary-only distribution like Trisquel. Neither felt appealing to me, but a late christmas gift of a reproducible Mono came to Guix that resolve this.

Embedded images in Texinfo manual

For Libidn one section of the manual has an image illustrating some concepts. The PNG, PDF and EPS outputs were generated via fig2dev from a *.fig file (hello 1985!) that I had stored in git. Over time, I had also started to store the generated outputs because of build issues. At some point, it was possible to post-process the PDF outputs with grep to remove some timestamps, however with compression this is no longer possible and actually the grep command I used resulted in a 0-byte output file. So my embedded binaries in git was no longer reproducible. I first set out to fix this by post-processing things properly, however I then realized that the *.fig file is not really easy to work with in a modern world. I wanted to create an image from some text-file description of the image. Eventually, via the Guix manual on guix graph, I came to re-discover the graphviz language and tool called dot (hello 1993!). All well then? Oh no, the PDF output embeds timestamps. Binary editing of PDF’s no longer work through simple grep, remember? I was back where I started, and after some (soul- and web-) searching I discovered that Ghostscript (hello 1988!) pdfmarks could be used to modify things here. Cooperating with automake’s texinfo rules related to make dist proved once again a worthy challenge, and eventually I ended up with a Makefile.am snippet to build images that could be condensed into:

info_TEXINFOS = libidn.texi
libidn_TEXINFOS += libidn-components.png
imagesdir = $(infodir)
images_DATA = libidn-components.png
EXTRA_DIST += components.dot
DISTCLEANFILES = \
  libidn-components.eps libidn-components.png libidn-components.pdf
libidn-components.eps: $(srcdir)/components.dot
  $(AM_V_GEN)$(DOT) -Nfontsize=9 -Teps < $< > $@.tmp
  $(AM_V_at)! grep %%CreationDate $@.tmp
  $(AM_V_at)mv $@.tmp $@
libidn-components.pdf: $(srcdir)/components.dot
  $(AM_V_GEN)$(DOT) -Nfontsize=9 -Tpdf < $< > $@.tmp
# A simple sed on CreationDate is no longer possible due to compression.
# 'exiftool -CreateDate' is alternative to 'gs', but adds ~4kb to file.
# Ghostscript add <1kb.  Why can't 'dot' avoid setting CreationDate?
  $(AM_V_at)printf '[ /ModDate ()\n  /CreationDate ()\n  /DOCINFO pdfmark\n' > pdfmarks
  $(AM_V_at)$(GS) -q -dBATCH -dNOPAUSE -sDEVICE=pdfwrite -sOutputFile=$@.tmp2 $@.tmp pdfmarks
  $(AM_V_at)rm -f $@.tmp pdfmarks
  $(AM_V_at)mv $@.tmp2 $@
libidn-components.png: $(srcdir)/components.dot
  $(AM_V_GEN)$(DOT) -Nfontsize=9 -Tpng < $< > $@.tmp
  $(AM_V_at)mv $@.tmp $@
pdf-recursive: libidn-components.pdf
dvi-recursive: libidn-components.eps
ps-recursive: libidn-components.eps
info-recursive: $(top_srcdir)/.version libidn-components.png

Surely this can be improved, but I’m not yet certain in what way is the best one forward. I like having a text representation as the source of the image. I’m sad that the new image size is ~48kb compared to the old image size of ~1kb. I tried using exiftool -CreateDate as an alternative to GhostScript, but using it to remove the timestamp added ~4kb to the file size and naturally I was appalled by this ignorance of impending doom.

Test reproducibility of tarball

Again, you need to continuously test the properties you desire. This means building your project twice using different environments and comparing the results. I’ve settled with a small GitLab CI/CD pipeline job that perform bit-by-bit comparison of generated ‘make dist’ archives. It also perform bit-by-bit comparison of generated ‘git-archive’ artifacts. See the Libidn2 .gitlab-ci.yml 0-compare job which essentially is:

0-compare:
  image: alpine:latest
  stage: repro
  needs: [ B-AlmaLinux8, B-AlmaLinux9, B-RockyLinux8, B-RockyLinux9, B-Trisquel11, B-Ubuntu2204, B-PureOS10, B-Debian11, B-Devuan5, B-Debian12, B-gcc, B-clang, B-Guix, R-Guix, R-Debian12, R-Ubuntu2404, S-Trisquel10, S-Ubuntu2004 ]
  script:
  - cd out
  - sha256sum */*.tar.* */*/*.tar.* | sort | grep    -- -src.tar.
  - sha256sum */*.tar.* */*/*.tar.* | sort | grep -v -- -src.tar.
  - sha256sum */*.tar.* */*/*.tar.* | sort | uniq -c -w64 | sort -rn
  - sha256sum */*.tar.* */*/*.tar.* | grep    -- -src.tar. | sort | uniq -c -w64 | grep -v '^      1 '
  - sha256sum */*.tar.* */*/*.tar.* | grep -v -- -src.tar. | sort | uniq -c -w64 | grep -v '^      1 '
# Confirm modern git-archive tarball reproducibility
  - cmp b-almalinux8/src/*.tar.gz b-almalinux9/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz b-rockylinux8/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz b-rockylinux9/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz b-debian12/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz b-devuan5/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz r-guix/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz r-debian12/src/*.tar.gz
  - cmp b-almalinux8/src/*.tar.gz r-ubuntu2404/src/*v2.*.tar.gz
# Confirm old git-archive (export-subst but long git describe) tarball reproducibility
  - cmp b-trisquel11/src/*.tar.gz b-ubuntu2204/src/*.tar.gz
# Confirm really old git-archive (no export-subst) tarball reproducibility
  - cmp b-debian11/src/*.tar.gz b-pureos10/src/*.tar.gz
# Confirm 'make dist' generated tarball reproducibility
  - cmp b-almalinux8/*.tar.gz b-rockylinux8/*.tar.gz
  - cmp b-almalinux9/*.tar.gz b-rockylinux9/*.tar.gz
  - cmp b-pureos10/*.tar.gz b-debian11/*.tar.gz
  - cmp b-devuan5/*.tar.gz b-debian12/*.tar.gz
  - cmp b-trisquel11/*.tar.gz b-ubuntu2204/*.tar.gz
  - cmp b-guix/*.tar.gz r-guix/*.tar.gz
# Confirm 'make dist' from git-archive tarball reproducibility
  - cmp s-trisquel10/*.tar.gz s-ubuntu2004/*.tar.gz

Notice that I discovered that ‘git archive’ outputs differ over time too, which is natural but a bit of a nuisance. The output of the job is illuminating in the way that all SHA256 checksums of generated tarballs are included, for example the libidn2 v2.3.8 job log:

$ sha256sum */*.tar.* */*/*.tar.* | sort | grep -v -- -src.tar.
368488b6cc8697a0a937b9eb307a014396dd17d3feba3881e6911d549732a293  b-trisquel11/libidn2-2.3.8.tar.gz
368488b6cc8697a0a937b9eb307a014396dd17d3feba3881e6911d549732a293  b-ubuntu2204/libidn2-2.3.8.tar.gz
59db2d045fdc5639c98592d236403daa24d33d7c8db0986686b2a3056dfe0ded  b-debian11/libidn2-2.3.8.tar.gz
59db2d045fdc5639c98592d236403daa24d33d7c8db0986686b2a3056dfe0ded  b-pureos10/libidn2-2.3.8.tar.gz
5bd521d5ecd75f4b0ab0fc6d95d444944ef44a84cad859c9fb01363d3ce48bb8  s-trisquel10/libidn2-2.3.8.tar.gz
5bd521d5ecd75f4b0ab0fc6d95d444944ef44a84cad859c9fb01363d3ce48bb8  s-ubuntu2004/libidn2-2.3.8.tar.gz
7f1dcdea3772a34b7a9f22d6ae6361cdcbe5513e3b6485d40100b8565c9b961a  b-almalinux8/libidn2-2.3.8.tar.gz
7f1dcdea3772a34b7a9f22d6ae6361cdcbe5513e3b6485d40100b8565c9b961a  b-rockylinux8/libidn2-2.3.8.tar.gz
8031278157ce43b5813f36cf8dd6baf0d9a7f88324ced796765dcd5cd96ccc06  b-clang/libidn2-2.3.8.tar.gz
8031278157ce43b5813f36cf8dd6baf0d9a7f88324ced796765dcd5cd96ccc06  b-debian12/libidn2-2.3.8.tar.gz
8031278157ce43b5813f36cf8dd6baf0d9a7f88324ced796765dcd5cd96ccc06  b-devuan5/libidn2-2.3.8.tar.gz
8031278157ce43b5813f36cf8dd6baf0d9a7f88324ced796765dcd5cd96ccc06  b-gcc/libidn2-2.3.8.tar.gz
8031278157ce43b5813f36cf8dd6baf0d9a7f88324ced796765dcd5cd96ccc06  r-debian12/libidn2-2.3.8.tar.gz
acf5cbb295e0693e4394a56c71600421059f9c9bf45ccf8a7e305c995630b32b  r-ubuntu2404/libidn2-2.3.8.tar.gz
cbdb75c38100e9267670b916f41878b6dbc35f9c6cbe60d50f458b40df64fcf1  b-almalinux9/libidn2-2.3.8.tar.gz
cbdb75c38100e9267670b916f41878b6dbc35f9c6cbe60d50f458b40df64fcf1  b-rockylinux9/libidn2-2.3.8.tar.gz
f557911bf6171621e1f72ff35f5b1825bb35b52ed45325dcdee931e5d3c0787a  b-guix/libidn2-2.3.8.tar.gz
f557911bf6171621e1f72ff35f5b1825bb35b52ed45325dcdee931e5d3c0787a  r-guix/libidn2-2.3.8.tar.gz

I’m sure I have forgotten or suppressed some challenges (sprinkling LANG=C TZ=UTC0 helps) related to these goals, but my hope is that this discussion of solutions will inspire you to implement these concepts for your software project too. Please share your thoughts and additional insights in a comment below. Enjoy Happy Hacking in the course of practicing this!

Towards Idempotent Rebuilds?

After rebuilding all added/modified packages in Trisquel, I have been circling around the elephant in the room: 99% of the binary packages in Trisquel comes from Ubuntu, which to a large extent are built from Debian source packages. Is it possible to rebuild the official binary packages identically? Does anyone make an effort to do so? Does anyone care about going through the differences between the official package and a rebuilt version? Reproducible-build.org‘s effort to track reproducibility bugs in Debian (and other systems) is amazing. However as far as I know, they do not confirm or deny that their rebuilds match the official packages. In fact, typically their rebuilds do not match the official packages, even when they say the package is reproducible, which had me surprised at first. To understand why that happens, compare the buildinfo file for the official coreutils 9.1-1 from Debian bookworm with the buildinfo file for reproducible-build.org’s build and you will see that the SHA256 checksum does not match, but still they declare it as a reproducible package. As far as I can tell of the situation, the purpose of their rebuilds are not to say anything about the official binary build, instead the purpose is to offer a QA service to maintainers by performing two builds of a package and declaring success if both builds match.

I have felt that something is lacking, and months have passed and I haven’t found any project that address the problem I am interested in. During my earlier work I created a project called debdistreproduce which performs rebuilds of the difference between two distributions in a GitLab pipeline, and display diffoscope output for further analysis. A couple of days ago I had the idea of rewriting it to perform rebuilds of a single distribution. A new project debdistrebuild was born and today I’m happy to bless it as version 1.0 and to announces the project! Debdistrebuild has rebuilt the top-50 popcon packages from Debian bullseye, bookworm and trixie, on amd64 and arm64, as well as Ubuntu jammy and noble on amd64, see the summary status page for links. This is intended as a proof of concept, to allow people experiment with the concept of doing GitLab-based package rebuilds and analysis. Compare how Guix has the guix challenge command.

Or I should say debdistrebuild has attempted to rebuild those distributions. The number of identically built packages are fairly low, so I didn’t want to waste resources building the rest of the archive until I understand if the differences are due to consequences of my build environment (plain apt-get build-dep followed by dpkg-buildpackage in a fresh container), or due to some real difference. Summarizing the results, debdistrebuild is able to rebuild 34% of Debian bullseye on amd64, 36% of bookworm on amd64, 32% of bookworm on arm64. The results for trixie and Ubuntu are disappointing, below 10%.

So what causes my rebuilds to be different from the official rebuilds? Some are trivial like the classical problem of varying build paths, resulting in a different NT_GNU_BUILD_ID causing a mismatch. Some are a bit strange, like a subtle difference in one of perl’s headers file. Some are due to embedded version numbers from a build dependency. Several of the build logs and diffoscope outputs doesn’t make sense, likely due to bugs in my build scripts, especially for Ubuntu which appears to strip translations and do other build variations that I don’t do. In general, the classes of reproducibility problems are the expected. Some are assembler differences for GnuPG’s gpgv-static, likely triggered by upload of a new version of gcc after the original package was built. There are at least two ways to resolve that problem: either use the same version of build dependencies that were used to produce the original build, or demand that all packages that are affected by a change in another package are rebuilt centrally until there are no more differences.

The current design of debdistrebuild uses the latest version of a build dependency that is available in the distribution. We call this a “idempotent rebuild“. This is usually not how the binary packages were built originally, they are often built against earlier versions of their build dependency. That is the situation for most binary distributions.

Instead of using the latest build dependency version, higher reproducability may be achieved by rebuilding using the same version of the build dependencies that were used during the original build. This requires parsing buildinfo files to find the right version of the build dependency to install. We believe doing so will lead to a higher number of reproducibly built packages. However it begs the question: can we rebuild that earlier version of the build dependency? This circles back to really old versions and bootstrappable builds eventually.

While rebuilding old versions would be interesting on its own, we believe that is less helpful for trusting the latest version and improving a binary distribution: it is challenging to publish a new version of some old package that would fix a reproducibility bug in another package when used as a build dependency, and then rebuild the later packages with the modified earlier version. Those earlier packages were already published, and are part of history. It may be that ultimately it will no longer be possible to rebuild some package, because proper source code is missing (for packages using build dependencies that were never part of a release); hardware to build a package could be missing; or that the source code is no longer publicly distributable.

I argue that getting to 100% idempotent rebuilds is an interesting goal on its own, and to reach it we need to start measure idempotent rebuild status.

One could conceivable imagine a way to rebuild modified versions of earlier packages, and then rebuild later packages using the modified earlier packages as build dependencies, for the purpose of achieving higher level of reproducible rebuilds of the last version, and to reach for bootstrappability. However, it may be still be that this is insufficient to achieve idempotent rebuilds of the last versions. Idempotent rebuilds are different from a reproducible build (where we try to reproduce the build using the same inputs), and also to bootstrappable builds (in which all binaries are ultimately built from source code). Consider a cycle where package X influence the content of package Y, which in turn influence the content of package X. These cycles may involve several packages, and it is conceivable that a cycle could be circular and infinite. It may be difficult to identify these chains, and even more difficult to break them up, but this effort help identify where to start looking for them. Rebuilding packages using the same build dependency versions as were used during the original build, or rebuilding packages using a bootsrappable build process, both seem orthogonal to the idempotent rebuild problem.

Our notion of rebuildability appears thus to be complementary to reproducible-builds.org’s definition and bootstrappable.org’s definition. Each to their own devices, and Happy Hacking!

Addendum about terminology: With “idempotent rebuild” I am talking about a rebuild of the entire operating system, applied to itself. Compare how you build the latest version of the GNU C Compiler: it first builds itself using whatever system compiler is available (often an earlier version of gcc) which we call step 1. Then step 2 is to build a copy of itself using the compiler built in step 1. The final step 3 is to build another copy of itself using the compiler from step 2. Debian, Ubuntu etc are at step 1 in this process right now. The output of step 2 and step 3 ought to be bit-by-bit identical, or something is wrong. The comparison between step 2 and 3 is what I refer to with an idempotent rebuild. Of course, most packages aren’t a compiler that can compile itself. However entire operating systems such as Trisquel, PureOS, Ubuntu or Debian are (hopefully) a self-contained system that ought to be able to rebuild itself to an identical copy. Or something is amiss. The reproducible build and bootstrappable build projects are about improve the quality of step 1. The property I am interested is the identical rebuild and comparison in step 2 and 3. I feel the word “idempotent” describes the property I’m interested in well, but I realize there may be better ways to describe this. Ideas welcome!

Reproducible and minimal source-only tarballs

With the release of Libntlm version 1.8 the release tarball can be reproduced on several distributions. We also publish a signed minimal source-only tarball, produced by git-archive which is the same format used by Savannah, Codeberg, GitLab, GitHub and others. Reproducibility of both tarballs are tested continuously for regressions on GitLab through a CI/CD pipeline. If that wasn’t enough to excite you, the Debian packages of Libntlm are now built from the reproducible minimal source-only tarball. The resulting binaries are reproducible on several architectures.

What does that even mean? Why should you care? How you can do the same for your project? What are the open issues? Read on, dear reader…

This article describes my practical experiments with reproducible release artifacts, following up on my earlier thoughts that lead to discussion on Fosstodon and a patch by Janneke Nieuwenhuizen to make Guix tarballs reproducible that inspired me to some practical work.

Let’s look at how a maintainer release some software, and how a user can reproduce the released artifacts from the source code. Libntlm provides a shared library written in C and uses GNU Make, GNU Autoconf, GNU Automake, GNU Libtool and gnulib for build management, but these ideas should apply to most project and build system. The following illustrate the steps a maintainer would take to prepare a release:

git clone https://gitlab.com/gsasl/libntlm.git
cd libntlm
git checkout v1.8
./bootstrap
./configure
make distcheck
gpg -b libntlm-1.8.tar.gz

The generated files libntlm-1.8.tar.gz and libntlm-1.8.tar.gz.sig are published, and users download and use them. This is how the GNU project have been doing releases since the late 1980’s. That is a testament to how successful this pattern has been! These tarballs contain source code and some generated files, typically shell scripts generated by autoconf, makefile templates generated by automake, documentation in formats like Info, HTML, or PDF. Rarely do they contain binary object code, but historically that happened.

The XZUtils incident illustrate that tarballs with files that are not included in the git archive offer an opportunity to disguise malicious backdoors. I blogged earlier how to mitigate this risk by using signed minimal source-only tarballs.

The risk of hiding malware is not the only motivation to publish signed minimal source-only tarballs. With pre-generated content in tarballs, there is a risk that GNU/Linux distributions such as Trisquel, Guix, Debian/Ubuntu or Fedora ship generated files coming from the tarball into the binary *.deb or *.rpm package file. Typically the person packaging the upstream project never realized that some installed artifacts was not re-built through a typical autoconf -fi && ./configure && make install sequence, and never wrote the code to rebuild everything. This can also happen if the build rules are written but are buggy, shipping the old artifact. When a security problem is found, this can lead to time-consuming situations, as it may be that patching the relevant source code and rebuilding the package is not sufficient: the vulnerable generated object from the tarball would be shipped into the binary package instead of a rebuilt artifact. For architecture-specific binaries this rarely happens, since object code is usually not included in tarballs — although for 10+ years I shipped the binary Java JAR file in the GNU Libidn release tarball, until I stopped shipping it. For interpreted languages and especially for generated content such as HTML, PDF, shell scripts this happens more than you would like.

Publishing minimal source-only tarballs enable easier auditing of a project’s code, to avoid the need to read through all generated files looking for malicious content. I have taken care to generate the source-only minimal tarball using git-archive. This is the same format that GitLab, GitHub etc offer for the automated download links on git tags. The minimal source-only tarballs can thus serve as a way to audit GitLab and GitHub download material! Consider if/when hosting sites like GitLab or GitHub has a security incident that cause generated tarballs to include a backdoor that is not present in the git repository. If people rely on the tag download artifact without verifying the maintainer PGP signature using GnuPG, this can lead to similar backdoor scenarios that we had for XZUtils but originated with the hosting provider instead of the release manager. This is even more concerning, since this attack can be mounted for some selected IP address that you want to target and not on everyone, thereby making it harder to discover.

With all that discussion and rationale out of the way, let’s return to the release process. I have added another step here:

make srcdist
gpg -b libntlm-1.8-src.tar.gz

Now the release is ready. I publish these four files in the Libntlm’s Savannah Download area, but they can be uploaded to a GitLab/GitHub release area as well. These are the SHA256 checksums I got after building the tarballs on my Trisquel 11 aramo laptop:

91de864224913b9493c7a6cec2890e6eded3610d34c3d983132823de348ec2ca  libntlm-1.8-src.tar.gz
ce6569a47a21173ba69c990965f73eb82d9a093eb871f935ab64ee13df47fda1  libntlm-1.8.tar.gz

So how can you reproduce my artifacts? Here is how to reproduce them in a Ubuntu 22.04 container:

podman run -it --rm ubuntu:22.04
apt-get update
apt-get install -y --no-install-recommends autoconf automake libtool make git ca-certificates
git clone https://gitlab.com/gsasl/libntlm.git
cd libntlm
git checkout v1.8
./bootstrap
./configure
make dist srcdist
sha256sum libntlm-*.tar.gz

You should see the exact same SHA256 checksum values. Hooray!

This works because Trisquel 11 and Ubuntu 22.04 uses the same version of git, autoconf, automake, and libtool. These tools do not guarantee the same output content for all versions, similar to how GNU GCC does not generate the same binary output for all versions. So there is still some delicate version pairing needed.

Ideally, the artifacts should be possible to reproduce from the release artifacts themselves, and not only directly from git. It is possible to reproduce the full tarball in a AlmaLinux 8 container – replace almalinux:8 with rockylinux:8 if you prefer RockyLinux:

podman run -it --rm almalinux:8
dnf update -y
dnf install -y make wget gcc
wget https://download.savannah.nongnu.org/releases/libntlm/libntlm-1.8.tar.gz
tar xfa libntlm-1.8.tar.gz
cd libntlm-1.8
./configure
make dist
sha256sum libntlm-1.8.tar.gz

The source-only minimal tarball can be regenerated on Debian 11:

podman run -it --rm debian:11
apt-get update
apt-get install -y --no-install-recommends make git ca-certificates
git clone https://gitlab.com/gsasl/libntlm.git
cd libntlm
git checkout v1.8
make -f cfg.mk srcdist
sha256sum libntlm-1.8-src.tar.gz 

As the Magnus Opus or chef-d’œuvre, let’s recreate the full tarball directly from the minimal source-only tarball on Trisquel 11 – replace docker.io/kpengboy/trisquel:11.0 with ubuntu:22.04 if you prefer.

podman run -it --rm docker.io/kpengboy/trisquel:11.0
apt-get update
apt-get install -y --no-install-recommends autoconf automake libtool make wget git ca-certificates
wget https://download.savannah.nongnu.org/releases/libntlm/libntlm-1.8-src.tar.gz
tar xfa libntlm-1.8-src.tar.gz
cd libntlm-v1.8
./bootstrap
./configure
make dist
sha256sum libntlm-1.8.tar.gz

Yay! You should now have great confidence in that the release artifacts correspond to what’s in version control and also to what the maintainer intended to release. Your remaining job is to audit the source code for vulnerabilities, including the source code of the dependencies used in the build. You no longer have to worry about auditing the release artifacts.

I find it somewhat amusing that the build infrastructure for Libntlm is now in a significantly better place than the code itself. Libntlm is written in old C style with plenty of string manipulation and uses broken cryptographic algorithms such as MD4 and single-DES. Remember folks: solving supply chain security issues has no bearing on what kind of code you eventually run. A clean gun can still shoot you in the foot.

Side note on naming: GitLab exports tarballs with pathnames libntlm-v1.8/ (i.e.., PROJECT-TAG/) and I’ve adopted the same pathnames, which means my libntlm-1.8-src.tar.gz tarballs are bit-by-bit identical to GitLab’s exports and you can verify this with tools like diffoscope. GitLab name the tarball libntlm-v1.8.tar.gz (i.e., PROJECT-TAG.ARCHIVE) which I find too similar to the libntlm-1.8.tar.gz that we also publish. GitHub uses the same git archive style, but unfortunately they have logic that removes the ‘v’ in the pathname so you will get a tarball with pathname libntlm-1.8/ instead of libntlm-v1.8/ that GitLab and I use. The content of the tarball is bit-by-bit identical, but the pathname and archive differs. Codeberg (running Forgejo) uses another approach: the tarball is called libntlm-v1.8.tar.gz (after the tag) just like GitLab, but the pathname inside the archive is libntlm/, otherwise the produced archive is bit-by-bit identical including timestamps. Savannah’s CGIT interface uses archive name libntlm-1.8.tar.gz with pathname libntlm-1.8/, but otherwise file content is identical. Savannah’s GitWeb interface provides snapshot links that are named after the git commit (e.g., libntlm-a812c2ca.tar.gz with libntlm-a812c2ca/) and I cannot find any tag-based download links at all. Overall, we are so close to get SHA256 checksum to match, but fail on pathname within the archive. I’ve chosen to be compatible with GitLab regarding the content of tarballs but not on archive naming. From a simplicity point of view, it would be nice if everyone used PROJECT-TAG.ARCHIVE for the archive filename and PROJECT-TAG/ for the pathname within the archive. This aspect will probably need more discussion.

Side note on git archive output: It seems different versions of git archive produce different results for the same repository. The version of git in Debian 11, Trisquel 11 and Ubuntu 22.04 behave the same. The version of git in Debian 12, AlmaLinux/RockyLinux 8/9, Alpine, ArchLinux, macOS homebrew, and upcoming Ubuntu 24.04 behave in another way. Hopefully this will not change that often, but this would invalidate reproducibility of these tarballs in the future, forcing you to use an old git release to reproduce the source-only tarball. Alas, GitLab and most other sites appears to be using modern git so the download tarballs from them would not match my tarballs – even though the content would.

Side note on ChangeLog: ChangeLog files were traditionally manually curated files with version history for a package. In recent years, several projects moved to dynamically generate them from git history (using tools like git2cl or gitlog-to-changelog). This has consequences for reproducibility of tarballs: you need to have the entire git history available! The gitlog-to-changelog tool also output different outputs depending on the time zone of the person using it, which arguable is a simple bug that can be fixed. However this entire approach is incompatible with rebuilding the full tarball from the minimal source-only tarball. It seems Libntlm’s ChangeLog file died on the surgery table here.

So how would a distribution build these minimal source-only tarballs? I happen to help on the libntlm package in Debian. It has historically used the generated tarballs as the source code to build from. This means that code coming from gnulib is vendored in the tarball. When a security problem is discovered in gnulib code, the security team needs to patch all packages that include that vendored code and rebuild them, instead of merely patching the gnulib package and rebuild all packages that rely on that particular code. To change this, the Debian libntlm package needs to Build-Depends on Debian’s gnulib package. But there was one problem: similar to most projects that use gnulib, Libntlm depend on a particular git commit of gnulib, and Debian only ship one commit. There is no coordination about which commit to use. I have adopted gnulib in Debian, and add a git bundle to the *_all.deb binary package so that projects that rely on gnulib can pick whatever commit they need. This allow an no-network GNULIB_URL and GNULIB_REVISION approach when running Libntlm’s ./bootstrap with the Debian gnulib package installed. Otherwise libntlm would pick up whatever latest version of gnulib that Debian happened to have in the gnulib package, which is not what the Libntlm maintainer intended to be used, and can lead to all sorts of version mismatches (and consequently security problems) over time. Libntlm in Debian is developed and tested on Salsa and there is continuous integration testing of it as well, thanks to the Salsa CI team.

Side note on git bundles: unfortunately there appears to be no reproducible way to export a git repository into one or more files. So one unfortunate consequence of all this work is that the gnulib *.orig.tar.gz tarball in Debian is not reproducible any more. I have tried to get Git bundles to be reproducible but I never got it to work — see my notes in gnulib’s debian/README.source on this aspect. Of course, source tarball reproducibility has nothing to do with binary reproducibility of gnulib in Debian itself, fortunately.

One open question is how to deal with the increased build dependencies that is triggered by this approach. Some people are surprised by this but I don’t see how to get around it: if you depend on source code for tools in another package to build your package, it is a bad idea to hide that dependency. We’ve done it for a long time through vendored code in non-minimal tarballs. Libntlm isn’t the most critical project from a bootstrapping perspective, so adding git and gnulib as Build-Depends to it will probably be fine. However, consider if this pattern was used for other packages that uses gnulib such as coreutils, gzip, tar, bison etc (all are using gnulib) then they would all Build-Depends on git and gnulib. Cross-building those packages for a new architecture will therefor require git on that architecture first, which gets circular quick. The dependency on gnulib is real so I don’t see that going away, and gnulib is a Architecture:all package. However, the dependency on git is merely a consequence of how the Debian gnulib package chose to make all gnulib git commits available to projects: through a git bundle. There are other ways to do this that doesn’t require the git tool to extract the necessary files, but none that I found practical — ideas welcome!

Finally some brief notes on how this was implemented. Enabling bootstrappable source-only minimal tarballs via gnulib’s ./bootstrap is achieved by using the GNULIB_REVISION mechanism, locking down the gnulib commit used. I have always disliked git submodules because they add extra steps and has complicated interaction with CI/CD. The reason why I gave up git submodules now is because the particular commit to use is not recorded in the git archive output when git submodules is used. So the particular gnulib commit has to be mentioned explicitly in some source code that goes into the git archive tarball. Colin Watson added the GNULIB_REVISION approach to ./bootstrap back in 2018, and now it no longer made sense to continue to use a gnulib git submodule. One alternative is to use ./bootstrap with --gnulib-srcdir or --gnulib-refdir if there is some practical problem with the GNULIB_URL towards a git bundle the GNULIB_REVISION in bootstrap.conf.

The srcdist make rule is simple:

git archive --prefix=libntlm-v1.8/ -o libntlm-1.8-src.tar.gz HEAD

Making the make dist generated tarball reproducible can be more complicated, however for Libntlm it was sufficient to make sure the modification times of all files were set deterministically to the timestamp of the last commit in the git repository. Interestingly there seems to be a couple of different ways to accomplish this, Guix doesn’t support minimal source-only tarballs but rely on a .tarball-timestamp file inside the tarball. Paul Eggert explained what TZDB is using some time ago. The approach I’m using now is fairly similar to the one I suggested over a year ago. If there are problems because all files in the tarball now use the same modification time, there is a solution by Bruno Haible that could be implemented.

Side note on git tags: Some people may wonder why not verify a signed git tag instead of verifying a signed tarball of the git archive. Currently most git repositories uses SHA-1 for git commit identities, but SHA-1 is not a secure hash function. While current SHA-1 attacks can be detected and mitigated, there are fundamental doubts that a git SHA-1 commit identity uniquely refers to the same content that was intended. Verifying a git tag will never offer the same assurance, since a git tag can be moved or re-signed at any time. Verifying a git commit is better but then we need to trust SHA-1. Migrating git to SHA-256 would resolve this aspect, but most hosting sites such as GitLab and GitHub does not support this yet. There are other advantages to using signed tarballs instead of signed git commits or git tags as well, e.g., tar.gz can be a deterministically reproducible persistent stable offline storage format but .git sub-directory trees or git bundles do not offer this property.

Doing continous testing of all this is critical to make sure things don’t regress. Libntlm’s pipeline definition now produce the generated libntlm-*.tar.gz tarballs and a checksum as a build artifact. Then I added the 000-reproducability job which compares the checksums and fails on mismatches. You can read its delicate output in the job for the v1.8 release. Right now we insists that builds on Trisquel 11 match Ubuntu 22.04, that PureOS 10 builds match Debian 11 builds, that AlmaLinux 8 builds match RockyLinux 8 builds, and AlmaLinux 9 builds match RockyLinux 9 builds. As you can see in pipeline job output, not all platforms lead to the same tarballs, but hopefully this state can be improved over time. There is also partial reproducibility, where the full tarball is reproducible across two distributions but not the minimal tarball, or vice versa.

If this way of working plays out well, I hope to implement it in other projects too.

What do you think? Happy Hacking!