--- Day changed Thu Sep 10 2015
03:38 -!- gmaxwell [greg@wikimedia/KatWalsh/x-0001] has quit [Ping timeout: 260 seconds]
03:38 -!- gmaxwell [greg@mf4-xiph.osuosl.org] has joined #secp256k1
03:39 -!- gmaxwell is now known as Guest7769
03:49 -!- Guest7769 [greg@mf4-xiph.osuosl.org] has quit [Changing host]
03:49 -!- Guest7769 [greg@wikimedia/KatWalsh/x-0001] has joined #secp256k1
03:49 -!- Guest7769 is now known as gmaxwell
06:47 < fkhan> just curious, has anyone attempted to verify #302 yet?
07:22 < sipa> 'verify' ?
07:47 < fkhan> well, i suppose i wasn't sure but i was just thinking it could be nice to ack-check the calculations agianst the standard
07:48 < fkhan> i personally havent but if i had the time it would be something i would do for understanding
07:48 < fkhan> would like to* do
07:48 < sipa> you mean review the code?
07:49 < sipa> the script in 302 works, it verifies that the formulas in the group implementation are correct, and it correctly detects a bug that an older implementation had
07:49 < fkhan> yes, that is... what i was trying to convey, heh
07:49 < sipa> of course, the script may be buggy
07:50 < fkhan> right, and i realize you are the author of that pull, so i was mostly querying for other reviewers
07:50 < fkhan> mostly to follow along, if anything
07:51 < sipa> i'm in the process of changing it significantly
07:51 < fkhan> i would like the time to review it but i have not had much success is reviewing low level crypto implementations
07:51 < sipa> this is mostly math
07:51 < fkhan> so the time investment for me is kinda large
07:51 < sipa> verifying that polynomials are identical, assuming certain identities
07:52 < fkhan> yes
07:53 < sipa> i want to get to a point where the sage script (which is really python) is able to interpret the actual C code
07:53 < sipa> instead needing manual translation between them
07:54 < fkhan> haha yeah... i was looking at sage and had that thought too (this is very similar to python..)
07:54 < sipa> it's preprocessed python
07:54 < fkhan> hmm, interesting
07:54 < fkhan> yeah i can see how that would make verifying more accurate
07:54 < sipa> but right now, not all branches are checked
07:54 < fkhan> or, checking the code* really
07:54 < fkhan> i see
07:54 < sipa> as the script cannot verify additions with infinity etc
07:55 < sipa> i've refactored it now and rewritten pretty much every line already
07:55 < sipa> but it doesn't work anymore :)
07:55 < fkhan> hehe :)
07:56 < fkhan> yeah, i really love following along, i think this library from a security/verification perspective is really educational, at least for me
07:56 < sipa> introduced classes to represent points, insyead of passing around ax ay az etc all the time
07:56 < sipa> oh, for me too :)
08:05 < fkhan> hehe yes, no i would really like to review the internals, i might kick that idea around this weekend
08:06 < fkhan> this conversation has been compelling
11:48 -!- nullbyte [~NSA@cpe-66-68-54-206.austin.res.rr.com] has joined #secp256k1
11:53 -!- nullbyte [~NSA@cpe-66-68-54-206.austin.res.rr.com] has quit [Ping timeout: 260 seconds]
11:54 -!- nullbyte [~NSA@193.138.219.233] has joined #secp256k1
17:34 < gmaxwell> sipa: I recieved some more positive review comments. Someone who'd asked some third party to evaluate it and apparently they were very complementary.
17:38 < sipa> nice
17:45 < midnightmagic> I wonder who it was.