1 2014-11-20 00:01:31 <Luke-Jr> cfields: are you seeing a different outcome there? it's working consistently for me, native and cross
 2 2014-11-20 00:02:50 <cfields> Luke-Jr: i'm looking into the order of things. moving AC_CANONICAL_HOST isn't doing what you think. it runs regardless (likely from automake/libtool init)
 3 2014-11-20 00:03:05 <Luke-Jr> cfields: yes, I know. the outcome is fine though
 4 2014-11-20 00:03:14 <Luke-Jr> I can move it back
 5 2014-11-20 00:04:35 <cfields> yea, you should be able to work it all out without messing with the host. as written, it should fail for cross builds because the cross-cc wouldn't be set before running dumpmachine, but it works because it's injected earlier anyway
 6 2014-11-20 00:07:59 <Luke-Jr> https://github.com/sipa/secp256k1/pull/3
 7 2014-11-20 00:08:04 <Luke-Jr> should I PR against bitcoin/secp256k1?
 8 2014-11-20 00:12:35 <Luke-Jr> this builds for i686 (native) and win64 ♱
 9 2014-11-20 00:12:41 <Luke-Jr> err, (cross)
10 2014-11-20 00:18:44 <cfields> Luke-Jr: you're bypassing all the canonicalization logic by overriding the host in the switch/case. i really wish you wouldn't.
11 2014-11-20 00:19:09 <Luke-Jr> cfields: what logic is relevant there? when is $CC -dumpmachine wrong?
12 2014-11-20 00:19:32 <Luke-Jr> best I can tell, config.guess is a relic from compilers without -dumpmachine (and still used in that case)
13 2014-11-20 00:20:19 <cfields> Luke-Jr: feel free to PR, but please give me a bit to read up on how the other distros handle this. i don't want to bypass a good chunk of autotools compatibility layer (what else it it good for then ;) without seeing how others deal with this case
14 2014-11-20 00:20:57 <Luke-Jr> cfields: IMO, even if there is some edge case this breaks, it's better to break that case than to break the more common i686-on-x8664-hardware case
15 2014-11-20 00:21:24 <Luke-Jr> no rush though
16 2014-11-20 00:49:25 <maaku> sipa: is the gmp dependency in the consensus code?
17 2014-11-20 00:52:10 <gmaxwell> maaku: none of this is consensus code.
18 2014-11-20 00:52:41 <gmaxwell> It's signing only.
19 2014-11-20 01:01:42 <maaku> ok
20 2014-11-20 01:46:57 <Luke-Jr> fair warning: mining on testnet with ASICs
21 2014-11-20 01:51:26 <phantomcircuit> Luke-Jr, why
22 2014-11-20 02:09:59 <Luke-Jr> phantomcircuit: testing solo mining in BFGMiner first, then working on GBT improvements
23 2014-11-20 02:10:48 <Luke-Jr> from time to time, that means I need to load up testnet with hashing :P
24 2014-11-20 02:11:49 <Luke-Jr> (I could do TNIAB, but I'd prefer to restock on all the TNBTC I've been giving away like candy)
25 2014-11-20 02:12:21 <Luke-Jr> only like 10 kTNBTC left :P
26 2014-11-20 02:18:59 <Luke-Jr> lol, the irony >_<
27 2014-11-20 02:20:52 <Luke-Jr> don't merge #1816 just yet, that one check the unit tests can't cover actually fails :x
28 2014-11-20 02:24:06 <Luke-Jr> ok, pushed a fix to #1816
29 2014-11-20 02:24:09 <Luke-Jr> all good again :P
30 2014-11-20 02:25:42 <Luke-Jr> meh, testnet's diff is pretty high already
31 2014-11-20 02:55:40 <gmaxwell> Is there no way to download the windows binaries travis builds?
32 2014-11-20 02:58:34 <Luke-Jr> gmaxwell: not unless we scp them somewhere in the travis build
33 2014-11-20 03:00:12 <Luke-Jr> (which for PRs could expose the SSH key)
34 2014-11-20 03:00:28 <gmaxwell> ugh.
35 2014-11-20 03:00:31 <gmaxwell> that sucks.