1 2012-03-21 00:00:16 <BlueMatt> gmaxwell: yea, i think we need to barade jgarzik on that one, his caches are broken...
  2 2012-03-21 00:00:30 <BlueMatt> or are those above the level of sata drivers?
  3 2012-03-21 00:00:40 <sipa> DBordello: meh; i'd have to add a license "not permitted to use for checking green addresses" ;)
  4 2012-03-21 00:01:01 <DBordello> sipa, too bad :)
  5 2012-03-21 00:01:12 <DBordello> if it makes you feel better, I learned a lot about bitcoin transactions :)
  6 2012-03-21 00:02:56 <sipa> BlueMatt: i've made a script that you just give a commit id, and it builds, packages and gpg-signs the output, using git describe to select file names
  7 2012-03-21 00:03:37 <BlueMatt> gmaxwell: not a bad idea...
  8 2012-03-21 00:03:39 <BlueMatt> sipa: mmm
  9 2012-03-21 00:04:05 <DBordello> hah
 10 2012-03-21 00:04:06 <BlueMatt> anyone else watching the "2012 Isaac Asimov Memorial Debate"?
 11 2012-03-21 00:04:16 <BlueMatt> its good, though they are on closing questions now...
 12 2012-03-21 00:05:19 <gmaxwell> okay, 40 minutes to get to height=164946 with 25MB  (vs ~an hour with 4MB)
 13 2012-03-21 00:07:19 <sipa> anyone know a nice separation symbol that does not typically occur in URLs?
 14 2012-03-21 00:08:18 <luke-jr> space
 15 2012-03-21 00:08:49 <sipa> luke-jr: also an option, indeed
 16 2012-03-21 00:08:56 <andytoshi> :
 17 2012-03-21 00:09:08 <BlueMatt> is the % encoding crap url-standard or http-url-standard?
 18 2012-03-21 00:09:30 <BlueMatt> andytoshi: port?
 19 2012-03-21 00:09:36 <luke-jr> andytoshi: & : is in EVERY URL
 20 2012-03-21 00:09:42 <luke-jr> heck, every URI even
 21 2012-03-21 00:10:15 <andytoshi> my bad, i read "typically occur" as "occurs often per URL"
 22 2012-03-21 00:10:25 <andytoshi> also forgot about ports :P
 23 2012-03-21 00:10:39 <sipa> ok, , and space seem to fail; try ???
 24 2012-03-21 00:11:08 <luke-jr> sipa: for what?
 25 2012-03-21 00:11:30 <sipa> luke-jr: -c bitcoin=COMMIT???URL
 26 2012-03-21 00:11:35 <sipa> in gbuild
 27 2012-03-21 00:11:45 <luke-jr> pfft
 28 2012-03-21 00:11:59 <luke-jr> can use comma
 29 2012-03-21 00:12:00 <sipa> wait, that's not a pipe?
 30 2012-03-21 00:12:02 <luke-jr> no
 31 2012-03-21 00:12:07 <luke-jr> and pipe has to be escaped&
 32 2012-03-21 00:12:07 <sipa> wtf?
 33 2012-03-21 00:13:52 <gmaxwell> =
 34 2012-03-21 00:14:13 <luke-jr> ??
 35 2012-03-21 00:14:30 <sipa> ????
 36 2012-03-21 00:15:01 <luke-jr> ??????$??????B??
 37 2012-03-21 00:15:18 <Cory> Any reports of rc4 freezing on startup on Windows 7 x64?
 38 2012-03-21 00:15:28 <BlueMatt> yes
 39 2012-03-21 00:15:30 <luke-jr> yes
 40 2012-03-21 00:15:41 <Cory> Oh.
 41 2012-03-21 00:15:48 <Cory> Heh, seems to have just started today.
 42 2012-03-21 00:15:53 <luke-jr> Cory: you should fix it
 43 2012-03-21 00:15:54 <BlueMatt> https://github.com/bitcoin/bitcoin/issues/956
 44 2012-03-21 00:16:09 <BlueMatt> post there if the posted workaround fixes your issue as well, please
 45 2012-03-21 00:16:23 <luke-jr> save your .bitcoin first tho
 46 2012-03-21 00:17:54 <DBordello> Cory, yes, on mine :)
 47 2012-03-21 00:18:54 <Cory> It was after a crash, yup. Will this be fixed for 0.6?
 48 2012-03-21 00:19:22 <BlueMatt> yea
 49 2012-03-21 00:19:32 <BlueMatt> wait crash on startup or freeze?
 50 2012-03-21 00:20:31 <gavinandresen> BlueMatt Cory : sequence of events is:  crash, then freeze at next startup?
 51 2012-03-21 00:20:54 <Cory> Yes.
 52 2012-03-21 00:21:09 <BlueMatt> mmm, well the crash is disturbing
 53 2012-03-21 00:21:12 <Cory> After Windows crashes, I meant.
 54 2012-03-21 00:21:18 <BlueMatt> freeze probably fixable, crash...
 55 2012-03-21 00:21:20 <DBordello> I also was experiencing the crash, then freeze on all subsequent startups
 56 2012-03-21 00:21:27 <gavinandresen> BlueMatt: that's the boost IPC doo-hickey for opening URI's not being cleaned up.... (I think?)
 57 2012-03-21 00:21:36 <BlueMatt> gavinandresen: apparently, yea
 58 2012-03-21 00:21:36 <Cory> DBordello: Exactly. Did you try the fix?
 59 2012-03-21 00:21:42 <DBordello> Cory, trying it now
 60 2012-03-21 00:22:08 <BlueMatt> gavinandresen: damn, and when I wrote that I thought it worked well...
 61 2012-03-21 00:22:12 <BlueMatt> it did on linux...
 62 2012-03-21 00:22:52 <BlueMatt> maybe someone knows of something better than boost::interprocess
 63 2012-03-21 00:23:08 <gavinandresen> Qt has equivalents, I believe
 64 2012-03-21 00:23:10 <luke-jr> probably a way to tell boost not to block?
 65 2012-03-21 00:23:48 <DBordello> it is only the gui that freezes.  The client continues to process blocks
 66 2012-03-21 00:24:29 <BlueMatt> luke-jr: in theory, the code (as written) should never block the main thread?
 67 2012-03-21 00:24:36 <luke-jr> boost::interprocess::named_mutex::remove( "MutexName" );
 68 2012-03-21 00:24:43 <luke-jr> can we do that without having the mutex yet?
 69 2012-03-21 00:24:54 <luke-jr> BlueMatt: doesn't sound like it does& just the GUI
 70 2012-03-21 00:25:07 <BlueMatt> luke-jr: well, according to the docs written, it doesnt block
 71 2012-03-21 00:25:14 <BlueMatt> in reality, boost sucks (apparently)
 72 2012-03-21 00:25:44 <luke-jr> maybe, but if we can fix it with ::remove for 0.6, sounds the best solution&
 73 2012-03-21 00:26:06 <BlueMatt> you cant blindly remove, you can remove the message_queue out from under one which is already listening...
 74 2012-03-21 00:26:21 <BlueMatt> actually, we already do that...
 75 2012-03-21 00:26:29 <Cory> The workaround fixed it.
 76 2012-03-21 00:26:41 <luke-jr> what happens if you blindly remove?
 77 2012-03-21 00:26:46 <BlueMatt> it gets removed
 78 2012-03-21 00:26:50 <BlueMatt> we do, actually
 79 2012-03-21 00:26:53 <luke-jr> ok, that's what we want
 80 2012-03-21 00:26:55 <BlueMatt> maybe try it earlier in init
 81 2012-03-21 00:27:15 <sipa> you can't remove it before checking a process is still listening using it
 82 2012-03-21 00:27:17 <BlueMatt> can anyone with build capabilities reproduce the issue?
 83 2012-03-21 00:27:23 <BlueMatt> sipa: we already do...
 84 2012-03-21 00:27:35 <sipa> do remove or do check?
 85 2012-03-21 00:27:40 <BlueMatt> remove w/o check
 86 2012-03-21 00:27:44 <BlueMatt> remove + replace
 87 2012-03-21 00:27:54 <sipa> :o
 88 2012-03-21 00:28:06 <BlueMatt> ie newest running copy of bitcoin always gets the message_queue
 89 2012-03-21 00:28:30 <sipa> well, that's not a problem
 90 2012-03-21 00:28:35 <DBordello> that workaround fixed it for me.  Thank  you guys
 91 2012-03-21 00:29:04 <luke-jr> why is it "BitcoinURL" when it's not a URL, btw?
 92 2012-03-21 00:29:21 <DBordello> I do not have build capability, but i can test
 93 2012-03-21 00:29:53 <sipa> DBordello: building 0.6.0rc4-15-g89516bd (the fast block download version)
 94 2012-03-21 00:30:13 <luke-jr> BlueMatt: it looks like we do create + delete + create&
 95 2012-03-21 00:30:20 <DBordello> sipa, fantastic
 96 2012-03-21 00:32:05 <luke-jr> BlueMatt: i presume the first create is the problem
 97 2012-03-21 00:32:23 <luke-jr> moreover, I don't see the first one ever deleted&
 98 2012-03-21 00:41:27 <BlueMatt> luke-jr: create shouldnt block either
 99 2012-03-21 00:41:38 <BlueMatt> luke-jr: could be try_send being the problem, no real way to know
100 2012-03-21 01:00:28 <sipa> the gitian vm does not have some strange url filtering firewall, does it?
101 2012-03-21 01:00:43 <luke-jr> yes
102 2012-03-21 01:00:46 <sipa> it seems unable to fetch from https://github.com/sipa/bitcoin.git
103 2012-03-21 01:00:51 <luke-jr> it uses NAT
104 2012-03-21 01:01:02 <BlueMatt> the vm itself has no connection to the outside
105 2012-03-21 01:01:12 <sipa> while https://github.com/bitcoin/bitcoin.git works fine
106 2012-03-21 01:01:25 <sipa> error: RPC failed; result=7, HTTP code = 0
107 2012-03-21 01:01:36 <sipa> in var/build.log
108 2012-03-21 01:01:59 <luke-jr> BlueMatt: it has NAT