Hi Peter, > He means GHC's ABI hashes. GHC itself is known to be non-deterministic > in certain cases. However, most of these issues have been fixed over > the years, and in the context of Nix, I've never suffered from this > issue in practice. So it would be interesting to verify if that's > really the culprit here.
Yes, I meant the ABI hashes (visible in the GHC package database). It could happen that with two builds the Nix hashes would be identical (because they arise from inputs only) while the GHC ABI hashes would be different (due to nondeterminism in the GHC compile). _______________________________________________ nix-dev mailing list [email protected] http://lists.science.uu.nl/mailman/listinfo/nix-dev
