1 2012-01-29 00:15:11 <CIA-97> bitcoin: Con Kolivas * rae044ec31ea6 cgminer/NEWS: Update NEWS. http://tinyurl.com/7xabb5w
  2 2012-01-29 00:15:12 <CIA-97> bitcoin: Con Kolivas * r09ea6bd21a53 cgminer/ (NEWS configure.ac): Bump version number up 2.2.0 http://tinyurl.com/6wplpjc
  3 2012-01-29 00:40:45 <splatster> etotheipi_: trying to run on ubuntu 11.10 getting "Cpp is not defined nameerror"
  4 2012-01-29 00:44:03 <etotheipi_> splatster, was there an earlier error about failing to import CppBlockUtils?
  5 2012-01-29 00:44:15 <splatster> ya something like that
  6 2012-01-29 00:45:13 <etotheipi_> sooo, there's a cryptopp/crypto++ linking issue
  7 2012-01-29 00:45:31 <etotheipi_> it always gets the wrong library... I'm hoping I can fix this with a cmake build system
  8 2012-01-29 00:45:37 <luke-jr> except that we haven't used crypto++ in months
  9 2012-01-29 00:45:59 <luke-jr> or you mean another client?
 10 2012-01-29 00:46:01 <luke-jr> <.<
 11 2012-01-29 00:46:03 <etotheipi_> I don't why the two libraries exist and always get mixed up differently on different systems
 12 2012-01-29 00:46:32 <splatster> ok so what do i do?
 13 2012-01-29 00:46:35 <luke-jr> IIRC, libcrypto++ has a questionable license :P
 14 2012-01-29 00:46:57 <etotheipi_> instead of libcryptopp?
 15 2012-01-29 00:47:18 <luke-jr> yeah, crypto++ isn't licensed for even distribution
 16 2012-01-29 00:47:41 <splatster> etotheipi_: What is the remedy for this problem?
 17 2012-01-29 00:47:44 <etotheipi_> most of crypto<plus><plus> is public domain
 18 2012-01-29 00:47:53 <etotheipi_> splatster, I actually have 11.0 open right now in VM, too
 19 2012-01-29 00:48:03 <etotheipi_> I think there's a simple export command... hold on
 20 2012-01-29 00:48:37 <luke-jr> etotheipi_: not all
 21 2012-01-29 00:53:18 <etotheipi_> splatster, I'm not sure why there's a problem, my default 11.10 install has both libcryptopp and libcrypto++
 22 2012-01-29 00:53:24 <etotheipi_> but mine fails to load, as well
 23 2012-01-29 00:55:24 <splatster> "TheBDM = Cpp.BlockDataManager().getBDM()  NameError: name 'Cpp' is not defined"
 24 2012-01-29 00:56:25 <etotheipi_> splatster... open a python shell in the BitcoinArmory directory and type "import CppBlockUtils"
 25 2012-01-29 00:58:21 <splatster> well my vm just spazzed out
 26 2012-01-29 00:59:22 <splatster> ImportError
 27 2012-01-29 01:01:00 <etotheipi_> splatster, I'm looking at it, now
 28 2012-01-29 01:01:50 <etotheipi_> btw, did you do the "make swig" step?
 29 2012-01-29 01:01:56 <splatster> ya
 30 2012-01-29 01:02:03 <etotheipi_> okay, just making sure :)
 31 2012-01-29 01:05:06 <etotheipi_> damnit... I can't get it to work with any library...
 32 2012-01-29 01:05:18 <etotheipi_> is sipa here?
 33 2012-01-29 01:05:18 <luke-jr> crypto++ and cryptopp are the same thing
 34 2012-01-29 01:05:42 <splatster> I have to go for a bit
 35 2012-01-29 01:05:43 <etotheipi_> luke-jr, there are two separate libararies, (don't ask me why), some systems have one some have the other
 36 2012-01-29 01:05:49 <splatster> tell me what happens
 37 2012-01-29 01:05:58 <etotheipi_> in the past I've resolved the problem by switching to the other
 38 2012-01-29 01:06:36 <etotheipi_> and I actually see both on my 11.10 /usr/lib dir... but my compiler-fu is not strong enough to resolve this
 39 2012-01-29 01:15:52 <sipa> etotheipi_: now i am
 40 2012-01-29 01:31:09 <etotheipi_> sipa, you had run Armory, and said there was a library issue (libcryptopp/++)
 41 2012-01-29 01:31:28 <etotheipi_> I'm seeing that now, but I'm not very good with the linking/library stuff to re-figure out how to fix it
 42 2012-01-29 01:31:35 <sipa> what's the error?
 43 2012-01-29 01:31:58 <etotheipi_> loading the module, it claims "undefined symbol: ..."
 44 2012-01-29 01:32:14 <etotheipi_> which would normally come from, I presume, the libcryptopp.so or libcrypto++.so
 45 2012-01-29 01:32:27 <etotheipi_> but I have both in /usr/lib
 46 2012-01-29 01:32:37 <sipa> oh yes the python code simply didn't load that library
 47 2012-01-29 01:32:43 <sipa> so linking of you Cpp lib failed
 48 2012-01-29 01:32:59 <sipa> no idea how python is supposed to deal with these problems
 49 2012-01-29 01:33:14 <etotheipi_> we're talking about runtime/dynamic linking, right?
 50 2012-01-29 01:33:25 <sipa> yes
 51 2012-01-29 01:33:32 <etotheipi_> because I keep wondering if I should just statically link the library into the app
 52 2012-01-29 01:34:25 <sipa> that's a possibility
 53 2012-01-29 01:34:27 <etotheipi_> but regardless of static-vs-dynamic... you had done something to make sure it loads the right library
 54 2012-01-29 01:34:33 <etotheipi_> do you remember what it was?
 55 2012-01-29 01:35:25 <sipa> i preloaded it
 56 2012-01-29 01:35:33 <sipa> LD_PRELOAD=...
 57 2012-01-29 01:35:42 <sipa> but that definitely not the correct solution
 58 2012-01-29 01:37:02 <shazooun> http://www.google.com/trends/?q=bitcoin,napster&ctab=0&geo=all&date=ytd&sort=0
 59 2012-01-29 01:38:01 <etotheipi_> unfortunately, all this linking stuff is not my forte... that's why static linking is enticing... I know I can figure that out
 60 2012-01-29 01:41:21 <luke-jr> done something = avoid crypto++
 61 2012-01-29 01:41:44 <sipa> crypto++ is really just a bunch of public-domain source files bundles
 62 2012-01-29 01:41:56 <sipa> just take the ones you need, and put them in your source code :)
 63 2012-01-29 01:42:10 <luke-jr> so wait, is Armory Python or C? <.<
 64 2012-01-29 01:42:13 <etotheipi_> have you looked through the source code?  it's pretty... intense
 65 2012-01-29 01:42:16 <sipa> Python and C++
 66 2012-01-29 01:42:26 <luke-jr> sipa: not quite. some are straight copyright
 67 2012-01-29 01:42:35 <etotheipi_> I actually used to use a pure-python ECDSA
 68 2012-01-29 01:42:41 <etotheipi_> it worked, but WAAAY too slow
 69 2012-01-29 01:42:42 <sipa> ok, you need to check the individial files of course
 70 2012-01-29 01:42:46 <etotheipi_> luke-jr, I'm looking this up now
 71 2012-01-29 01:43:04 <etotheipi_> it's fairly explicit that the source files are public domain
 72 2012-01-29 01:43:07 <luke-jr> etotheipi_: maybe GnuTLS or libgcrypt?
 73 2012-01-29 01:43:53 <luke-jr> etotheipi_: 3way, cast, seal, des, diamond, md2, md4, md5, md5mac, rc6, rijndael, skipjack, square, safer, and twofish are PD
 74 2012-01-29 01:44:02 <luke-jr> mars is unlicensed
 75 2012-01-29 01:44:21 <luke-jr> nfc where ECDSA fits into those
 76 2012-01-29 01:44:36 <etotheipi_> luke-jr, do you have a reference?
 77 2012-01-29 01:44:40 <luke-jr> and for more fun, PD isn't legal in some countries
 78 2012-01-29 01:44:47 <luke-jr> etotheipi_: Gentoo makes licenses easy.
 79 2012-01-29 01:45:09 <CIA-97> bitcoin: Con Kolivas debug * r622788563777 cgminer/ (9 files in 2 dirs): Merge branch 'master' into debug http://tinyurl.com/6mlsvkc
 80 2012-01-29 01:45:56 <luke-jr> http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/licenses/cryptopp?view=markup
 81 2012-01-29 01:45:57 <etotheipi_> crap... I completely forgot about crypto-algorithm export
 82 2012-01-29 01:46:21 <sipa> ?
 83 2012-01-29 01:46:32 <etotheipi_> anyone know anything about the export laws on crypto-algorithms?
 84 2012-01-29 01:47:10 <etotheipi_> http://en.wikipedia.org/wiki/Cryptography#Export_controls
 85 2012-01-29 01:47:17 <etotheipi_> (I don't know why I ask when I can just google it)
 86 2012-01-29 01:48:05 <sipa> http://en.wikipedia.org/wiki/Export_of_cryptography_in_the_United_States#Current_status
 87 2012-01-29 01:49:26 <etotheipi_> For instance, the BIS must be notified before open-source cryptographic software is made publicly available on the Internet, though no review is required
 88 2012-01-29 01:49:51 <roconnor> luke-jr: under US law, open source licenses can be revoked after 35 years.
 89 2012-01-29 01:50:07 <roconnor> well, not just open source liecenses
 90 2012-01-29 01:50:11 <roconnor> any licenses
 91 2012-01-29 01:53:16 <luke-jr> roconnor: O.o
 92 2012-01-29 01:53:31 <luke-jr> as in, full copyright status again?
 93 2012-01-29 01:53:39 <roconnor> yep
 94 2012-01-29 01:53:40 <luke-jr> or just the copyright (and therefore copyleft) expires?
 95 2012-01-29 01:53:40 <roconnor> http://www.copyright.gov/title17/92chap2.html#203
 96 2012-01-29 01:53:42 <luke-jr> wtf
 97 2012-01-29 01:54:01 <roconnor> after the 35 years you have 5 years to register to reclaim your copyright
 98 2012-01-29 01:54:13 <roconnor> or something like that
 99 2012-01-29 01:54:25 <gmaxwell> etotheipi_: authentication software is broadly excluded by EAR and doesn't need the ECCN dance.
100 2012-01-29 01:54:25 <roconnor> oh, and the work has to be made after 1978
101 2012-01-29 01:54:39 <luke-jr> roconnor: looks difficult
102 2012-01-29 01:54:43 <roconnor> etotheipi_: Ya, I was about to say, there is no encryption in bitcoin anyways
103 2012-01-29 01:55:01 <gmaxwell> etotheipi_: so, e.g. so long as the cryptographic stuff in your commercial software was only used for authentication (as is the care for bitcoin) you wouldn't need anything even if you weren't OSS.
104 2012-01-29 01:55:02 <luke-jr> roconnor: they have to notify each person in writing&
105 2012-01-29 01:55:09 <CIA-97> bitcoin: Con Kolivas * r83a836765432 cgminer/adl.c: Linux's ADL uses a busnumber in descending order for devices so enumerate them in the opposite order to windows. http://tinyurl.com/6mk63jw
106 2012-01-29 01:55:25 <roconnor> luke-jr: would notifying Sun be adequate?
107 2012-01-29 01:55:40 <luke-jr> roconnor: hmm, I guess it would work against big companies
108 2012-01-29 01:56:03 <luke-jr> whoever wrote the BSD networking stack should revoke the license from MS
109 2012-01-29 01:56:06 <luke-jr> <.<
110 2012-01-29 01:56:54 <gmaxwell> Microsoft includes copyright notices for the BSD stuff they include.
111 2012-01-29 01:57:22 <roconnor> gmaxwell: that won't help with the licence is revoked
112 2012-01-29 01:57:26 <roconnor> *when
113 2012-01-29 02:00:38 <gmaxwell> roconnor: Nonsense, the tranfer termination doesn't apply to derivative works.
114 2012-01-29 02:01:39 <roconnor> good point
115 2012-01-29 02:01:55 <luke-jr> ooooooo
116 2012-01-29 02:02:24 <gmaxwell> roconnor: you made me think for a second though!
117 2012-01-29 02:03:33 <roconnor> gmaxwell: but at that the code can no longer be modified
118 2012-01-29 02:03:44 <roconnor> BSD stack becomes frozen
119 2012-01-29 02:03:56 <roconnor> well, the stack formally known as BSD
120 2012-01-29 02:04:31 <gmaxwell> There is apparently an paper on this subject on SSRN (where most law review articles show up), that I don't remember reading: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1461859
121 2012-01-29 02:06:57 <roconnor> gmaxwell: actually it is a bit unclear to me
122 2012-01-29 02:07:10 <roconnor> ``A derivative work prepared under authority of the grant before its termination may continue to be utilized under the terms of the grant after its termination
123 2012-01-29 02:07:26 <roconnor> which I guess literally means that the derived code continues to be under the BSD
124 2012-01-29 02:08:12 <luke-jr> quick! be sure no work retains a single copyright holder for 35 years!
125 2012-01-29 02:08:37 <roconnor> I'd like to see a court ruling to be clear
126 2012-01-29 02:08:39 <roconnor> :P
127 2012-01-29 02:09:09 <gmaxwell> roconnor: the paper on SSRN seemes to advocate that congress should pass a law fixing this.. though it also suggests some judicial approaches that make it a non-issue.
128 2012-01-29 02:11:39 <etotheipi_> luke-jr, that link you sent me shows that everything in the Crypto++ library is public domain, except for maybe mars.cpp
129 2012-01-29 02:12:33 <gmaxwell> MARS.. been a while since I heard of that.. poor AES losers.
130 2012-01-29 02:12:44 <luke-jr> roconnor: &beginning at the end of thirty-five years __from the date of execution of the grant__&
131 2012-01-29 02:12:59 <luke-jr> execution for free software licenses is the first time YOU copy it, right?
132 2012-01-29 02:13:01 <gmaxwell> I liked the mixing rounds in mars a lot.
133 2012-01-29 02:13:58 <roconnor> I don't know; is the execution the time you make the licence?
134 2012-01-29 02:14:19 <roconnor> what is 1978 + 35 years?
135 2012-01-29 02:14:29 <sipa> 2013
136 2012-01-29 02:14:55 <roconnor> I need to tell my academic friends to start reverting their licences on thier academic papes
137 2012-01-29 02:14:57 <roconnor> *papers
138 2012-01-29 02:15:06 <roconnor> next year
139 2012-01-29 02:15:16 <roconnor> thought I guess notices can be sent out today
140 2012-01-29 02:16:26 <gmaxwell> It would be fun to write boiler plate notices to sending to elsevier.
141 2012-01-29 02:16:35 <sipa> roconnor: not sure whether academic publications are  license or a transfer of copyright itself
142 2012-01-29 02:16:51 <roconnor> sipa: it's a transfer, but the same law applies IIRC
143 2012-01-29 02:16:55 <gmaxwell> sipa: both can be terminated, in fact. They're not works for hire, for sure.
144 2012-01-29 02:17:10 <roconnor> In the case of any work other than a work made for hire, the exclusive or nonexclusive grant of a transfer or license of copyright or of any right under a copyright, executed by the author on or after January 1, 1978, otherwise than by will, is subject to termination under the following conditions
145 2012-01-29 02:17:40 <gmaxwell> cool. Then we can count on elsevier to pay for the lobbying required to correct the law before it's an issue for free software.
146 2012-01-29 02:17:45 <sipa> interesting
147 2012-01-29 02:18:26 <roconnor> gmaxwell: the music labels will do it.  There are still a few songs that get quite a bit of play from 1978
148 2012-01-29 02:19:25 <sipa> gmaxwell: will you and Elsevier agree on what the necessary "correction" is?
149 2012-01-29 02:19:57 <gmaxwell> sipa: I think we can reach a compromise, yes.
150 2012-01-29 02:24:06 <shazooun> elsevier kickbacks to uni administrators
151 2012-01-29 02:25:03 <gmaxwell> shazooun: I think if you could document that it would create controversy, considering how much universities pay for elsevier subscriptions.
152 2012-01-29 02:25:17 <gmaxwell> (fun grass roots awareness raising: http://mako.cc/fun/overpricetags/ )
153 2012-01-29 02:28:08 <shazooun> cool
154 2012-01-29 02:31:15 <sipa> story about head of my former research group: http://www.earlham.edu/~peters/fos/newsletter/05-11-01.htm
155 2012-01-29 02:39:40 <shazooun> Historical analysis of higher ed.: The growth in the ranks of administrators (85 percent) and associated professional staff (240 percent) has far outstripped the increase in faculty (51 percent) between 1975 and 2005.
156 2012-01-29 02:40:18 <shazooun> administrators with budgetary authority, will naturally higher more administrators to consolidate/bolster their own power
157 2012-01-29 02:49:15 <gribble> The average time to generate a block at 1100 Khps, given the supplied difficulty of 8.8, is 9 hours, 32 minutes, and 39 seconds
158 2012-01-29 02:49:15 <roconnor> ;;bc,calcd 1100 8.8
159 2012-01-29 02:57:30 <jondoe> I'm writing an application in Python. I'm storing values as Decimal and I want to send it with RPC to the bitcoin client
160 2012-01-29 02:57:36 <jondoe> the problem is, I get an exception:
161 2012-01-29 02:57:39 <jondoe> <type 'exceptions.TypeError'>, TypeError("Decimal('5') is not JSON serializable"
162 2012-01-29 02:58:01 <jondoe> apparantly it wants the value as a real; how do I convert?
163 2012-01-29 02:58:45 <etotheipi_> jondoe, probably something like decObj.to_integral()
164 2012-01-29 02:59:13 <etotheipi_> wait, that rounds to an integer
165 2012-01-29 02:59:50 <etotheipi_> jondoe,  you can just wrap it with float(decObj)
166 2012-01-29 02:59:56 <etotheipi_> or int(decObj)
167 2012-01-29 03:00:33 <shazooun> chillagua
168 2012-01-29 03:00:54 <jondoe> wont I get rounding problem if I convert it to float?
169 2012-01-29 03:03:30 <jondoe> oh, the rounding is too small?
170 2012-01-29 03:03:37 <jondoe> for it to be a problem?
171 2012-01-29 03:03:47 <sipa> you need to round it to a IEEE 754 double
172 2012-01-29 03:03:56 <sipa> then you won't have precision problems
173 2012-01-29 03:04:48 <jondoe> how do I convert it to a double?
174 2012-01-29 03:04:58 <jondoe> i'm kind of new to python >.<
175 2012-01-29 03:06:23 <roconnor> etotheipi_: armory doesn't appear to survive serious block reorganization
176 2012-01-29 03:06:37 <roconnor> well it doesn't seem to behave properly
177 2012-01-29 03:07:00 <roconnor> oh wait
178 2012-01-29 03:07:02 <roconnor> nevermind
179 2012-01-29 03:07:03 <etotheipi_> how so?  I have a re-rog unittest
180 2012-01-29 03:07:08 <roconnor> my bad
181 2012-01-29 03:07:12 <etotheipi_> it's been pretty heavily tested
182 2012-01-29 03:07:31 <roconnor> sorry
183 2012-01-29 03:07:44 <etotheipi_> ...though I haven't tested zero-conf transaction handling after a reorg
184 2012-01-29 03:08:11 <etotheipi_> fwhew!
185 2012-01-29 03:09:55 <roconnor> ;;bc,calcd 1100 0.125
186 2012-01-29 03:09:56 <gribble> The average time to generate a block at 1100 Khps, given the supplied difficulty of 0.125, is 8 minutes and 8 seconds
187 2012-01-29 03:10:44 <jondoe> How large is the testnet blockchain?
188 2012-01-29 03:10:50 <etotheipi_> about 45k blocks
189 2012-01-29 03:11:04 <etotheipi_> ~ 30 MB
190 2012-01-29 03:11:15 <sipa> etotheipi_: is python float a double?
191 2012-01-29 03:11:16 <gmaxwell> takes a long time to validate though, I guess because of stupid transactions.
192 2012-01-29 03:11:36 <etotheipi_> sipa, there is no distinction in python... though I don't know which one they use
193 2012-01-29 03:11:45 <etotheipi_> I would have to assume, if they have one, they would pick double
194 2012-01-29 03:12:45 <etotheipi_> I just checked, it's double
195 2012-01-29 03:13:34 <etotheipi_> that's pretty cool:  sys.float_info
196 2012-01-29 03:13:52 <etotheipi_> >>> sys.float_info
197 2012-01-29 03:13:53 <etotheipi_> sys.floatinfo(max=1.7976931348623157e+308, max_exp=1024, max_10_exp=308, min=2.2250738585072014e-308, min_exp=-1021, min_10_exp=-307, dig=15, mant_dig=53, epsilon=2.2204460492503131e-16, radix=2, rounds=1)
198 2012-01-29 03:14:09 <jondoe> neato
199 2012-01-29 03:14:10 <jondoe> thanks
200 2012-01-29 03:14:41 <jondoe> with that detailed information I might as well add an assertion as well
201 2012-01-29 03:15:10 <shazooun> cools
202 2012-01-29 03:15:24 <sipa> you need 52 bits of precision :)
203 2012-01-29 03:15:47 <etotheipi_> >>> sys.byteorder
204 2012-01-29 03:15:54 <etotheipi_> I love python :)
205 2012-01-29 03:15:59 <splatster> etotheipi_: So what happened with it?
206 2012-01-29 03:16:00 <sipa> i wonder if there is a VAX python
207 2012-01-29 03:16:11 <sipa> and whether sys.byteorder says "braid-dead" there
208 2012-01-29 03:16:13 <jondoe> i've never programmed python before this project, but I already love it
209 2012-01-29 03:16:14 <shazooun> simplejson 2.1 and higher has native support for Decimal type
210 2012-01-29 03:16:24 <sipa> *brain-dead
211 2012-01-29 03:16:32 <jondoe> shazooun, ah nice
212 2012-01-29 03:16:39 <etotheipi_> splatster, I might have a temp solution... but I really need to work out a better one
213 2012-01-29 03:16:48 <etotheipi_> let me try the temp solution, so you can at least test it
214 2012-01-29 03:17:10 <splatster> Ya, I'll test out your solution
215 2012-01-29 03:18:03 <roconnor> sipa: how come block explorer http://blockexplorer.com/testnet/b/1 claims this difficuly is 1/8th the difficulty of mainnet?
216 2012-01-29 03:18:33 <etotheipi_> splatster
217 2012-01-29 03:18:36 <sipa> because it is?
218 2012-01-29 03:18:49 <splatster> etotheipi_: ya?
219 2012-01-29 03:18:59 <etotheipi_> ack, can't copy from VM