depend on cryptography >= 1.0 and cffi >= 1.0 #2475

New Issue

tahoe-lafs · 2015-07-21T23:58:06Z

daira commented

2015-07-21 23:58:06 +00:00

According to dstufft, this combination will fix various cffi build problems, in particular the attempts to compile C files at paths for which that can't possibly work.

However, cryptography 1.0 has not been released yet; the release is blocked on https://github.com/pyca/cryptography/issues/2061.

According to dstufft, this combination will fix various cffi build problems, in particular the attempts to compile C files at paths for which that can't possibly work. However, cryptography 1.0 has not been released yet; the release is blocked on <https://github.com/pyca/cryptography/issues/2061>.

tahoe-lafs added the

labels 2015-07-21 23:58:06 +00:00

tahoe-lafs added this to the undecided milestone 2015-07-21 23:58:06 +00:00

warner commented

2016-03-26 23:19:15 +00:00

Tahoe doesn't depend on cryptography directly, but the path by which we do need it (foolscap -> twistedtls -> pyopenssl -> cryptography) requires cryptography >= 1.3, and the current cryptography-1.3.1 requires cffi >= 1.4.1. So all of these constraints are now in place.

Closing this out.

Tahoe doesn't depend on cryptography directly, but the path by which we *do* need it (`foolscap -> twistedtls -> pyopenssl -> cryptography`) requires `cryptography >= 1.3`, and the current cryptography-1.3.1 requires `cffi >= 1.4.1`. So all of these constraints are now in place. Closing this out.