#45 closed defect (fixed)
upgrade to latest Crypto++ with SHA-256 bugfixes
| Reported by: | zooko | Owned by: | |
|---|---|---|---|
| Priority: | major | Milestone: | |
| Version: | 0.5.19 | Keywords: | |
| Cc: | Launchpad Bug: |
Description
copy in the latest Crypto++ patches for SHA-256 and any other new patches that we ought to bundle
Change History (3)
comment:1 Changed at 2010-07-06T05:24:55Z by zooko
comment:2 Changed at 2010-09-10T19:44:06Z by zooko
The next step on this is to copy in (at least some of) raeburn's unit test from http://sourceforge.net/apps/trac/cryptopp/ticket/2 which demonstrates the bug in our current embedded copy of Crypto++ (which corresponds to upstream Crypto++ v5.6.0).
comment:3 Changed at 2010-09-19T05:23:25Z by zooko
- Resolution set to fixed
- Status changed from new to closed
raeburn's unit tests didn't demonstrate any bugs on my amd64 linux machine in virtualbox on mac os x on amd64 (intel core 2 duo). I added raeburn's tests -- test-sha256.cpp -- to the source tree, but it isn't getting compiled right now because it breaks the build on Windows.
David-Sarah committed the fixes to SHA-256 computation from Crypto++ v5.6.1: b1605a9540ef203e4dbcc44451ac7cc8aa61c880/git, 6136a60bec5744ca5a0fe2d4af4e3b6f0e253d27/git, 018e5ce0a3843717f8020c37bc1d66545c1030bb/git.
http://tahoe-lafs.org/pipermail/tahoe-dev/2010-July/004604.html