#4012: Compilation results are not deterministic -------------------------------------+------------------------------------- Reporter: kili | Owner: niteria Type: bug | Status: patch Priority: high | Milestone: 8.0.1 Component: Compiler | Version: 6.12.2 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: Other | Test Case: Blocked By: | Blocking: Related Tickets: #10424 | Differential Rev(s): Phab:D910, | Phab:D1073, Phab:D1133, Phab:D1192, Wiki Page: | Phab:D1268 -------------------------------------+------------------------------------- Comment (by Austin Seipp ): In [changeset:"ffcdd84f86727c1621be0a0e522eb6a3a64e479f/ghc" ffcdd84/ghc]: {{{ #!CommitTicketReference repository="ghc" revision="ffcdd84f86727c1621be0a0e522eb6a3a64e479f" Sort field labels before fingerprint hashing `fsEnvElts :: FastStringEnv a -> [a]` returns a list of `[a]` in the order of `Unique`s which is arbitrary. In this case it gives a list of record fields in arbitrary order, from which we then extract the field labels to contribute to the record fingerprint. The arbitrary ordering of field labels introduces unnecessary nondeterminism in interface files as demonstrated by the test case. We sort `FastString` here. It's safe, because the only way that the `Unique` associated with the `FastString` is used in comparison is for equality. If the `Unique`s are different it fallbacks to comparing the actual `ByteString`. Reviewed By: ezyang, thomie, bgamari, austin Differential Revision: https://phabricator.haskell.org/D1373 GHC Trac Issues: #4012 }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/4012#comment:128 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler