1 2013-09-13 00:00:18 <sipa> cfields: why is that testgmp.cpp necessary in the sample patch you showed me yesterday?
2 2013-09-13 00:18:20 <cfields> sipa: it's not, i was just showing you how to properly guard in case the lib wasn't found
3 2013-09-13 00:20:01 <warren> sipa: we got secp256k1 on 0.8 to build on 32bit mac yesterday, we have some hacks to add to your old configure script, but it sounds like cfields is redoing it with autotools now?
4 2013-09-13 00:20:40 <sipa> not soon, i suppose
5 2013-09-13 00:20:51 <sipa> cfields: ok, got it working
6 2013-09-13 00:21:30 <cfields> sipa: you can use those guards to make it portable, so that it works either in autotools or standalone
7 2013-09-13 00:21:53 <sipa> i don't see how it makes a difference
8 2013-09-13 00:22:06 <cfields> HAVE_CONFIG_H specifically
9 2013-09-13 00:22:39 <sipa> i'm not following
10 2013-09-13 00:23:04 <sipa> what does that check file add? it'll fail building in case gmp isn't installed, but the build would fail in that case anyway?
11 2013-09-13 00:23:58 <cfields> sipa: ah, i was thinking you had some optional deps among gmp/openssl/etc
12 2013-09-13 00:24:06 <sipa> yes, sure
13 2013-09-13 00:24:43 <sipa> but i don't see how this helps?
14 2013-09-13 00:25:03 <sipa> (not that it matters to me, i just need it to work - configurability can come later)
15 2013-09-13 00:25:54 <cfields> sipa: don't worry about it, i have a feeling i made some bad assumptions
16 2013-09-13 00:26:15 <sipa> ok
17 2013-09-13 00:47:54 <gavinandresen> Uploading 0.8.5 release binaries to sourceforge....
18 2013-09-13 00:48:18 <gavinandresen> Can y'all proofread release notes for me: https://bitcoincore.org/~gavin/0.8.5Notes.txt
19 2013-09-13 00:48:41 <an3k> sipa: i successfully reindexed all blocks and wallet is synched now :)
20 2013-09-13 00:50:37 <sipa> gavinandresen: ACK
21 2013-09-13 00:52:16 <sipa> an3k: ok, hope it doesn't happen again
22 2013-09-13 00:54:34 <gmaxwell> gavinandresen: passes my read
23 2013-09-13 00:54:53 <gavinandresen> gmaxwell sipa : awesome, thanks
24 2013-09-13 00:54:59 <gavinandresen> 82% uploadedâ¦.
25 2013-09-13 00:55:31 <an3k> yeah. maybe it happened because last shutdown before may be a little bit hard for bitcoin-qt. ubuntu shutdown the notebook because of battery. It doesn't wait that long for an application to shut down ;)
26 2013-09-13 00:56:52 <an3k> i just deleted the last blk12345.dat file and reindexed
27 2013-09-13 00:57:25 <sipa> an3k: the .dat file was almost certainly ok; you'd have gotten a different error otherwise
28 2013-09-13 00:57:54 <an3k> i hope there's just a bug in the code, not a defect in my hardware :)
29 2013-09-13 01:00:06 <an3k> is it a bitcoin-qt bug or ubuntu bug when at first launch bitcoin-qt doesn't place it's icon into ubuntu unity menu bar?
30 2013-09-13 01:00:34 <sipa> no clue
31 2013-09-13 01:02:45 <Luke-Jr> gavinandresen: looks good to me
32 2013-09-13 01:08:01 <gavinandresen> Looks good to me, too; announcement time
33 2013-09-13 01:13:46 <jgarzik> gavinandresen, ACK
34 2013-09-13 01:20:15 <gavinandresen> saivann: awake?
35 2013-09-13 01:22:11 <an3k> developers: want to make an announcement at 3:30 am :)
36 2013-09-13 01:27:33 <gmaxwell> some of us are not in .. italy?
37 2013-09-13 01:29:21 <saivann> gavinandresen : yep
38 2013-09-13 01:29:54 <gavinandresen> saivann: cool! I'm editing _releases/2013-09-13-v0.8.5.md ....
39 2013-09-13 01:30:13 <gavinandresen> saivann: I gather I just need to push that to get the download page on bitcoin.org to magically change
40 2013-09-13 01:30:58 <saivann> gavinandresen: Yes, the _releases/2013-09-13-v0.8.5.md is all you need to add, the RSS / latest version will auto update
41 2013-09-13 01:32:08 <gavinandresen> saivann: I hope I did that right....
42 2013-09-13 01:32:36 <saivann> I will fix it in the following minute if anything is wrong
43 2013-09-13 01:33:15 <TheLordOfTime> what determines how quickly bitcoin-qt indexes the blocks?
44 2013-09-13 01:33:43 <TheLordOfTime> or rather, what's the primary governing factor of how quickly the blocks get indexed?
45 2013-09-13 01:34:20 <saivann> gavinandresen: Perfect, the site will take ~15 minutes to update. You can consider it done.
46 2013-09-13 01:34:41 <gavinandresen> ok, wiki updated, mail sent to bitcoin-dev, topic posted to bitcointalk, bitcoin.org updatedâ¦
47 2013-09-13 01:34:55 <gavinandresen> ⦠can somebody update the Latest version in the topic here?
48 2013-09-13 01:43:53 <TheLordOfTime> gavinandresen, if nobody's helped in a few minutes i can go poke the topic for you, just tell me what you need fixed.
49 2013-09-13 01:44:07 <TheLordOfTime> ACTION is currently poking around in a linux kernel and is only partly focused on IRC.
50 2013-09-13 01:44:29 <gavinandresen> s/Latest version: 0.8.4/Latest version: 0.8.5/
51 2013-09-13 01:47:00 <TheLordOfTime> there you go
52 2013-09-13 01:49:57 <ageis> cool. any reason why there isn't a 64-bit windows build?
53 2013-09-13 01:51:59 <phantomcircuit> gavinandresen, what was the solution for the text relocation issue?
54 2013-09-13 01:52:01 <phantomcircuit> compile with -O2?
55 2013-09-13 01:52:21 <phantomcircuit> Luke-Jr, same question
56 2013-09-13 01:53:16 <gavinandresen> ageis: because nobody has shown any advantages (performance or compatibility) to a 64-bit build.
57 2013-09-13 01:53:40 <gavinandresen> phantomcircuit: text relocation issue⦠isn't ringing any bells with me
58 2013-09-13 01:53:54 <gmaxwell> if we ever switch to libsecp256k1 there is a big performance advantage for 64 bit... but not today.
59 2013-09-13 01:53:54 <phantomcircuit> /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/bin/ld: leveldb/libleveldb.a(db_impl.o): relocation R_X86_64_32S against `vtable for leveldb::DB' can not be used when making a shared object; recompile with -fPIC
60 2013-09-13 01:54:20 <phantomcircuit> i'll just disable hardening for now
61 2013-09-13 01:54:27 <gmaxwell> phantomcircuit: go delete leveldb/libleveldb.a
62 2013-09-13 01:54:38 <phantomcircuit> gmaxwell, oh is that all?
63 2013-09-13 01:54:42 <phantomcircuit> can do
64 2013-09-13 01:54:46 <gmaxwell> phantomcircuit: this isn't a hardening issue IIRC, its just a switching from pre-autofools issue.
65 2013-09-13 01:54:49 <gavinandresen> That is a "you should make clean before switching" issue?
66 2013-09-13 01:54:54 <sipa> yes
67 2013-09-13 01:54:56 <phantomcircuit> yeah
68 2013-09-13 01:54:57 <gmaxwell> I think so.
69 2013-09-13 01:55:00 <phantomcircuit> i see it now..
70 2013-09-13 01:55:01 <gavinandresen> cool
71 2013-09-13 01:55:19 <sipa> gmaxwell: 64-bit openssl is already ~2 times faster than 32-bit
72 2013-09-13 01:55:34 <gavinandresen> maybe autogen.sh should "if there is a Makefile, make clean, then do my thing"
73 2013-09-13 01:55:41 <saivann> Is there a OS specific "latest version" link on sourceforce (like http://sourceforge.net/projects/bitcoin/files/latest/download?source=files&default=windows)
74 2013-09-13 01:55:42 <gmaxwell> sipa: hm!
75 2013-09-13 01:55:52 <saivann> weusecoins is always outdated and this would probably fix this problem
76 2013-09-13 01:56:05 <phantomcircuit> gavinandresen, shrug short term issue not worth added complexity
77 2013-09-13 01:56:11 <phantomcircuit> autotools is already complicated enough...
78 2013-09-13 01:56:19 <sipa> who owns/manages weusecoins?
79 2013-09-13 01:56:36 <gavinandresen> phantomcircuit: agreed. I think. Depends on how much time we tell people how to fix it...
80 2013-09-13 01:56:36 <Luke-Jr> gavinandresen: 'make clean' won't work anymore at that point
81 2013-09-13 01:56:38 <sipa> still justmoon?
82 2013-09-13 01:56:41 <gmaxwell> Can we put a link to the signatures file: http://bitcoin.org/en/download it's really hard to find, and we do actually get people looking for them (actually, calling us nasty things for not having one)
83 2013-09-13 01:56:45 <saivann> sipa : It changed ownership, but I don't know the real owner. I can email him though
84 2013-09-13 01:56:58 <gmaxwell> Luke-Jr: make clean before updating not after..
85 2013-09-13 01:57:09 <Luke-Jr> gmaxwell: autogen.sh doesn't exist before updating ;)
86 2013-09-13 01:57:19 <gmaxwell> Luke-Jr: make -f makefile.unix clean
87 2013-09-13 01:57:23 <gavinandresen> ok, never mind, bad idea. Maybe autogen.sh should check for the obj/ directories and nuke them⦠no, bad idea, too....
88 2013-09-13 01:57:43 <Luke-Jr> gmaxwell: that won't exist if autogen.sh does
89 2013-09-13 01:58:01 <gmaxwell> gavinandresen: my initial reaction was "meh, one time cost, not worth doing anything about" but considering how many people its burningâ¦
90 2013-09-13 01:58:12 <gavinandresen> mmm
91 2013-09-13 01:58:28 <gavinandresen> could update the README temporarily, but nobody reads READMEs...
92 2013-09-13 01:58:43 <phantomcircuit> --without-qt will disable building bitcoin-qt right
93 2013-09-13 01:58:49 <sipa> phantomcircuit: indeed
94 2013-09-13 01:59:09 <sipa> warren: i updated my secp256k1 branch
95 2013-09-13 01:59:23 <sipa> warren: can you tell me which commits are still necessary on top of it?
96 2013-09-13 02:00:19 <sipa> gitignore sounds useful, but the rest needs rebasing, i think
97 2013-09-13 02:00:37 <warren> sipa: updated to master?
98 2013-09-13 02:01:02 <sipa> yes
99 2013-09-13 02:01:25 <gmaxwell> could just rename the leveldb directory in git entirely...
100 2013-09-13 02:02:51 <sipa> gavinandresen: interestingly, your DoS_tests timing test fail completely with libsecp256k1
101 2013-09-13 02:03:17 <sipa> i just deleted the offending code for now (it's not intended to be merged soon anyway), but i don't really follow what it was doing
102 2013-09-13 02:14:48 <saivann> gmaxwell : Is a link to the appropriate SHA256SUMS.asc file on bitcoin.org/en/download enough?
103 2013-09-13 02:15:10 <Luke-Jr> IMO, linking to gitian.sigs repo would be better
104 2013-09-13 02:15:25 <Luke-Jr> .asc is just one signature right now
105 2013-09-13 02:16:02 <saivann> Luke-Jr: Where is that file located?
106 2013-09-13 02:16:22 <Luke-Jr> https://github.com/bitcoin/gitian.sigs
107 2013-09-13 02:18:01 <cfields> i can whip up a quick timestamp target that would fix the leveldb autotools-switch problem
108 2013-09-13 02:18:09 <cfields> completely uninvasive, up to you guys if it's necessary
109 2013-09-13 02:19:09 <gmaxwell> Luke-Jr: that doesn't tell people how to verify the files they just downloaded.
110 2013-09-13 02:19:26 <Luke-Jr> gmaxwell: unfortunately not
111 2013-09-13 02:19:35 <Luke-Jr> maybe it deserves its own page on bitcoin.org explaining that
112 2013-09-13 02:19:56 <gmaxwell> for now the SHA256SUMS.asc is good.
113 2013-09-13 02:20:01 <Luke-Jr> ACTION shrugs
114 2013-09-13 02:20:06 <gmaxwell> Luke-Jr: perhaps we should write those instructions into SHA256SUMS.asc
115 2013-09-13 02:27:57 <cfields> http://pastebin.com/raw.php?i=zSLzX259 to fix the migration, if desired. forces a one-time clean
116 2013-09-13 02:35:36 <jgarzik> I continue to be tempted to break bitcoind JSON-RPC, to convert all decimal values to satoshis
117 2013-09-13 02:35:48 <jgarzik> a pre-1.0 change
118 2013-09-13 02:39:12 <gmaxwell> we could make every place it writes value also write intvalue but that doesn't solve input provided by the user.
119 2013-09-13 02:47:19 <gavinandresen> sipa: DoS_bantime fails? That's weird-- it tests the "ban is lifted after 24 hours by default" rule
120 2013-09-13 02:55:21 <Luke-Jr> jgarzik: +1
121 2013-09-13 02:59:05 <saivann> gmaxwell, Luke-Jr: https://github.com/bitcoin/bitcoin.org/pull/241
122 2013-09-13 02:59:32 <Luke-Jr> saivann: looks fine
123 2013-09-13 05:07:28 <Micha> iPhone|This is probably a bug, but I just thought I'd mention that sourceforge is claiming 0.8.1 is the latest version (though ofc there *is* no latest version for iOS): http://i.imgur.com/RJe57JY.png