--- Day changed Sat Aug 25 2018
20:18 < andytoshi> BlueMatt: in the rust-secp fuzz target, the output of sign will not pass verify .. is that intentional?
20:19 < andytoshi> sign does msg||sk, verify looks for msg||pk
21:04 < BlueMatt> andytoshi: sk == pk
21:04 < BlueMatt> so, it should
21:16 < andytoshi> ah ok
21:59 < andytoshi> BlueMatt: can i have an ack on https://github.com/rust-bitcoin/rust-bitcoin/pull/152 ? then i'll release 0.14.1
22:03 < BlueMatt> what abut https://github.com/rust-bitcoin/rust-bitcoin/issues/149
22:03 < BlueMatt> andytoshi: no tess?
22:03 < BlueMatt> ts
22:04 < andytoshi> i can add a test, sure
22:04 < andytoshi> it's a #derive on a newtyped Vec, i'm very confident in its behavior
22:04 < BlueMatt> also what about fixing 149 for .1?
22:05 < BlueMatt> heh, I hate confidence about rust's std on stuff like that
22:05 < BlueMatt> maybe I'm just too new
22:05 < BlueMatt> mostly cause its not a vec
22:05 < BlueMatt> its a Box<[u8]>
22:09 < andytoshi> added test
22:15 < andytoshi> hmmm... #149 is hard to fix for .1 actually
22:15 < andytoshi> because Decimal::new does not return an error type
22:16 < andytoshi> i guess i could change the Display impl to just not crash..
22:18 < BlueMatt> that, probably
22:18 < BlueMatt> or, at least initially
22:33 < andytoshi> i guess at some point it should refuse to Display ... if you try to make a million-digit Decimal you're probably just causing trouble
22:33 < andytoshi> so i wonder if i should make anything with more than 18 digits display "number too large" or something silly
22:34 < andytoshi> this is not supposed to be a BigNum type, it's supposed to losslessly encode bitcoin amounts between 0 and 21M BTC
22:38 < andytoshi> sigh, freeking everything overflows .. Display, +, <, integer_value
22:38 < andytoshi> i guess that's only four things. i can fix them each
22:47 < andytoshi> i think integer_value() is going to have to panic if you try to get it to return a too-large value
22:47 < andytoshi> i don't see any other sensible behavior
22:53 < BlueMatt> yea, clearly gotta not allow creation of a bogus value
22:58 < andytoshi> i could just make new() panic for now
22:58 < andytoshi> i mean, i have never used it with a non-fixed value
22:59 < andytoshi> in fact i have never used it except with a hardcoded 8..
23:02 < andytoshi> i also don't want to add an error return because then i'll need .unwrap() everywhere
23:15 < andytoshi> so i think that's the best solution. if anyone creates a Decimal with a giant exponent they'll get an immediate panic and know that they have to validate their input or use a real bignum library or something
23:15 < andytoshi> and normal people who are just putting constant 8s everywhere to deal with Core RPC can keep doing so