On 09/15/2013 10:44 PM, Gergely Risko wrote:
According to nixpkgs git log Vladimir fixed this already:
Well, it's rather a work-around -- not using the bogus "expect" package.
But the queue is full of old builds without this fix. There are 5069 *-freebsd builds containing the package name python and 2431 containing pypy. Since the commit message is about libffi, it's reasonable to think that pypy is affected. Since one timeout failure is exactly 1 hour, this means that we will have to wait 1200 hours (i686 and x86_64 builds in parallel, so not 2400 hours) at least. Thats 50 days. Can someone please fix this? Or alternatively tell me why I'm full of sh*t and why my calculation is broken? :)
The main problem is that Hydra does *not* do negative caching of this failure for some reason. As "expect" was a transitive dependency of e.g. mesa, it's contained in almost every package using X. There will surely be thousands of runs if someone doesn't stop it!
Vlada
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ nix-dev mailing list nix-dev@lists.science.uu.nl http://lists.science.uu.nl/mailman/listinfo/nix-dev
