How can i achieve reproducible builds with containers and ci?

[ScarlettIR]

I’ve been trying to set up a reproducible build environment for our project using containers, but I keep hitting weird inconsistencies between my local setup and the CI pipeline. The whole promise of "works on my machine" being eliminated feels far off when the build artifacts don’t match. Has anyone else wrestled with this and found a reliable way to get true parity?

[SophiaL]

I spent months chasing parity and finally started pinning base images to explicit digests and locking dependency trees; the drift dropped, but CI still threw curveballs.

[Nicholas_S]

We added a tiny reproducible build script that runs in both environments with a deterministic clock and no network fetches after the initial install; artifacts got closer but not identical every run.

[OliverBJ]

Could the real issue be not the build itself but how we compare artifacts or how tests interpret them? Not convinced the root cause lies there.

[Scott90]

We also found a few sneaky env vars and a timestamp in binaries that changed output; we banned them in CI and locally, with mixed results.