Hi. On 05/12/2015 12:45 AM, William Kennington wrote: > Maybe it would make more sense to only build the i686 builds if our > tested set of x86_64 binaries build correctly.
This seems the best suggestion I know of now. It shouldn't be difficult to implement, and it also makes important x86_64 parts build before i686, which I find good for fixing problems. BTW, I once opened an issue for X32 ABI discussion https://github.com/NixOS/nixpkgs/issues/1579 It would be nice if we had an option to say "evaluate these platforms less often, e.g. every fifth evaluation". Similarly, some core stuff might be good to evaluate more often, e.g. the tarball (done by Travis ATM). Then we could have some less common platforms without being too expensive (maybe even things like ARM cross-builds of core stuff). But I expect it would bring more complexity to hydra, as not all evaluations would have the same set of jobs... Vladimir
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ nix-dev mailing list [email protected] http://lists.science.uu.nl/mailman/listinfo/nix-dev
