[jchw]

Reproducible builds really are orthogonal to Nix, even if it seems like they're not. Nix was not built with reproducible builds in mind, and the problems it solves are ultimately independent even though they are related in some ways. The Nix PhD thesis by Eelco Dolstra, the founder of Nix, lays out the motivations for Nix very plainly. I decided I'd probably be better off quoting it directly than summarizing it poorly, so here goes:

From The Purely Functional Software Deployment Model, "1.3. Motivation"[1]:

    From the previous discussion of existing deployment systems it should be clear that they
    lack important features to support safe and efficient deployment. In particular, they have
    some or all of the following problems:

    • Dependency specifications are not validated, leading to incomplete deployment.
    • Dependency specifications are inexact (e.g., nominal).
    • It is not possible to deploy multiple versions or variants of a component side-by-side.
    • Components can interfere with each other.
    • It is not possible to roll back to previous configurations.
    • Upgrade actions are not atomic.
    • Applications must be monolithic, i.e., they must statically contain all their depen-
    dencies.
    • Deployment actions can only be performed by administrators, not by unprivileged
    users.
    • There is no link between binaries and the sources and build processes that built them.
    • The system supports either source deployment or binary deployment, but not both;
    or it supports both but in a non-unified way.
    • It is difficult to adapt components.
    • Component composition is manual.
    • The component framework is narrowly restricted to components written in a specific
    programming language or framework.
    • The system depends on non-portable techniques.

    The objective of the research described in this thesis is to develop a deployment system
    that does not have these problems.

Of course, part of the reason why reproducible builds was not considered for this is probably because it was simply not a hot topic at the time (had the term 'reproducible builds' even been coined yet?) and there were much bigger fish to fry with package management than reproducibility at that time, considering how relatively poorly packages were specified. Since then, "traditional" package managers and package repositories have put substantial work into cleaning up their package manifests and ensuring that they have accurate dependencies and other specifications such that today highly reproducible systems can be and have been built on top of them, limitations from the lack of hermetic guarantees notwithstanding.

Despite this, because Nix does guarantee bit-exact external inputs to a derivation, it does indeed make an excellent starting point for reproducible builds, but Nix itself as a tool is definitely orthogonal to reproducible builds, as it solves an entirely different problem that just happens to be related. You don't need the purity guarantees Nix gives you to get reproducible builds, and having those purity guarantees don't automatically give you reproducible builds (though as seen by Nixpkgs, it isn't uncommon for a build to coincidentally be reproducible just as a result of packaging it into a Nix derivation... just, not really specifically guaranteed by anything.)

[1]: https://edolstra.github.io/pubs/phd-thesis.pdf

[pveierland]

The entire point of purity within Nix is to have reproducibility by its nature: `In this model a binary component is uniquely defined by the declared inputs used to build the component` (section 1.5). It is not clear to me how this is an entirely different problem than that of reproducible builds. Yes, there are more problems to solve regarding achieving bit-for-bit reproducible builds that Nix does not fully solve alone, but reproducible software builds and deployments clearly are goals of Nix as a technology.

[jchw]

> The entire point of purity within Nix is to have reproducibility by its nature: `In this model a binary component is uniquely defined by the declared inputs used to build the component` (section 1.5). It is not clear to me how this is an entirely different problem than that of reproducible builds. Yes, there are more problems to solve regarding achieving bit-for-bit reproducible builds that Nix does not fully solve alone, but reproducible software builds and deployments clearly are goals of Nix as a technology.

Sorry, but you pretty much quoted what I would've quoted to refute your claim: it's exactly right that binary components are uniquely defined by their declared inputs, but note the subtle consequence that has: they are not defined at all by their outputs, only their declared inputs. (Making them be defined by their outputs is entirely possible FWIW, this is pretty much what content-addressed Nix paths are for.) This also applies recursively: the bit-exactness of external inputs are guaranteed by cryptographic hashes, but if you have any inputs that are, themselves, derivations, you can extremely trivially add impurities, because the system is just simply not designed to stop you from doing this. An example:

    stdenv.mkDerivation {
      name = "trivialimpurity";
      unpackPhase = "date>$out";
    }

It would be quite possible to make a system that is specifically designed to be resistant to this; consider that Nix goes great lengths to accomplish what it does, literally ripping apart ELF binaries to force dynamic symbols to be resolved in a static manner. There's no reason why you couldn't go further to force deeper reproducibility in a similar vein, it's just that Nix simply doesn't.

I think it's actually OK that Nix doesn't attempt to resolve this problem, because it is a pretty hairy one. Obviously you can't practically make the build environment 100% completely reproducible as it would be painfully slow and weird, so the best you could really probably do is intercept syscalls and try to make threading behavior deterministic (which would also make builds slow as balls, but less slow than running every build in unaccelerated qemu or something like that.) What Nix does do is solve the living shit out of the problem it was designed to solve, something you can feel very viscerally when you compare the state of Nixpkgs to the state of the AUR (a package repo I consider to be very good, but definitely can give some perspective about the problems Nix is designed to solve IMO, based on some of the problems I've run into with it.)

[pveierland]

Yes, Nix does not provide you a guarantee that there are no impurities in builds, which the extensional model specifically caters to, and where the intentional model improves upon this by providing a system bases on content hashes instead of input hashes.

However, Nix specially attempts to facilitate a pure software distribution model, which is why it does everything it does, e.g. forcing all input files to be copied to the store and providing mechanisms such as the pure evaluation mode which restricts access to the current system and time information. Yes, there are other ways to introduce impurities, but Nix tries in many ways to systematically remove these sources to increase the purity of the deployment process.

If the entire process of building an artifact is pure, then the artifact would be entirely reproducible, given that you have access to the same inputs. Yes, there are many ways to introduce impurity, however claiming that Nix as a purely functional software distribution model, where the central point is to achieve purity, is fully orthogonal to reproducible builds, seems incorrect.

[xyzzy_plugh]

Reproducibility is a second order effect. It's not a direct goal (at least originally) and as already discussed in this thread, more or less to death, not something that is required. It is very easy to write a derivation that cannot be reproduced. Sometimes it's even useful. Purity isn't really required at all. I wouldn't even say reproducibility is encouraged. It just sort of falls out of the system.

If Nix was worried about impurities that impact reproducible builds, they certainly wouldn't make live timestamps readily available in the pure environment, would they?

[jchw]

> Yes, Nix does not provide you a guarantee that there are no impurities in builds, which the extensional model specifically caters to, and where the intentional model improves upon this by providing a system bases on content hashes instead of input hashes.

Actually, the intensional model doesn't improve matters at all here, I was only pointing it out to demonstrate that the fact that binary components are addressed by their inputs doesn't really have anything to do with reproducibility. Of course the intensional model would mean that if you made the same build twice and it wasn't reproducible, then you'd get a different hash; however, that's not really an improvement over the current approach, which is to just build it twice and compare the output results. If anything, it just makes things more convoluted for reproducibility, due to the fact that you have to factor out self-references to check the hash.

The main advantage of the intensional model, as far as I know, is that it simplifies the trust model a bit. In the extensional model, you have to trust the substitutor, otherwise it could poison your Nix store. In the intensional model, derivations are addressed by their contents, so it's impossible to really poison the Nix store per-se, since you can definitely validate that the store path is correct for a content-addressed path.

Really though, it doesn't have a lot to do with reproducibility, and even in the work done in recent years I've not seen it mentioned at all in relation to reproducible builds, though I fully admit that it's very possible it's somehow useful and I just missed it.

> If the entire process of building an artifact is pure, then the artifact would be entirely reproducible, given that you have access to the same inputs.

That is true. Nix, though, explicitly only makes certain parts of the process pure, and the parts that it makes pure are specifically driven by the motivations outlined above. It is true that if you made the entire process completely pure, the build would be reproducible, and it is also true that Nix very intentionally does not try to do this, because it just simply wasn't in the list of problems Nix was solving at the time.

Likewise, though, you can still make a build reproducible without functional purity, which is exactly what has been done by various other reproducible build projects. They just happen to avoid impurities that would impact the result without any specific guarantees, which happens to be exactly what you have to do to make a reproducible build in Nix.

> Yes, there are many ways to introduce impurity, however claiming that Nix as a purely functional software distribution model, where the central point is to achieve purity, is fully orthogonal to reproducible builds, seems incorrect.

I don't know what "fully orthogonal" means relative to just "orthogonal". I am using "orthogonal" to mean "independent", i.e. what Nix solves is fully independent of reproducible builds. This follows because:

- It is possible to do reproducible builds without "purity" guarantees or sandboxing.

- It is possible to have builds that are not reproducible provided the purity guarantees of Nix.

The closest that Nix's purity comes to being related to reproducible builds is that it does, in fact, prevent some of the causes of unreproducible builds by enforcing the exact inputs, but that's why I consider it to be adjacent but ultimately orthogonal. If it's the word orthogonal that in particular is too strong, then maybe a better word would be "independent".

[pveierland]

Yes, I agree that reproducible builds is something that can be solved fully independently of Nix, and that Nix alone does not solve it, and that as such it is independent of Nix.

However, as the point of Nix is to achieve reproducible software deployments through a pure and functional description of the deployment, where it also provides mechanisms that systematically improves build reproducibility, I feel that orthogonal is misleading (as was my original disagreement) because reproducible builds correlate with Nix, and because achieving reproducible software deployment is a clear original goal of Nix, e.g. by seeking to remove implicit dependencies through mechanisms such as sandboxing of builds.