← Documentation index

Stable order for outputs

Data structures such as Perl hashes, Python dictionaries and sets, Rust std::collections::HashMap and std::collections::Hashset, or Ruby Hash objects will list their keys in a different order on every run to limit algorithmic complexity attacks.

Perl

The following Perl code will output the list in a different order on every run:

foreach my $package (keys %deps) {
    print MANIFEST, "$package: $deps[$packages]";
}

To get a deterministic output, the easiest way is to explicitly sort the keys:

foreach my $package (sort keys %deps) {
    print MANIFEST, "$package: $deps[$packages]";
}

For Perl, it is possible to set PERL_HASH_SEED=0 in the environment. This will result in hash keys always being in the same order. See perlrun(1) for more information.

Python

Python users can similarly set the environment variable PYTHONHASHSEED. When set to a given integer value, orders in dictionaries and sets will be the same on every run.

Rust

When iterating over the keys or entries of a HashMap, the order is explicitly undefined and depends on a random seed:

By default, HashMap uses a hashing algorithm selected to provide resistance against HashDoS attacks. The algorithm is randomly seeded, and a reasonable best-effort is made to generate this seed from a high quality, secure source of randomness provided by the host without blocking the program.

Iterating over a HashMap can cause reproducible builds issue when:

  • done inside a build.rs file
  • done in a function that’s directly or indirectly called from a build.rs file

This can often be fixed by replacing HashMap with BTreeMap. There’s a real-world example of how this was fixed.

General

Beware that the locale settings might affect the output of some sorting functions or the sort command.

← Documentation index

Documentation home

Introduction
Achieve deterministic builds
Managing variance
Define a build environment
Distribute the environment
Verification
Specifications