[00:04:55] @prole check it out 😄 [00:05:13] https://cdn.discordapp.com/attachments/416000318149754881/574748465809326080/Screenshot_from_2019-05-06_08-04-46.png [00:14:22] @prole windows hashes: ebf9501cbc810bce729f222c632916b23f9f72efa092fc41d5269d03dbc16a5d navcoin-4.6.0-win.tar.gz 5cdc9eef9e394cf015e9da04da17096db871c375f5a9b8be11a230cee7dfc483 navcoin-4.6.0-win32-debug.zip 141d3d50c23f5167e2dd45f7f5957a38c1e9a4729f30082468002820145ed48e navcoin-4.6.0-win32-setup.exe 944f6fa5994a90b7fff90a7392d8f56c8ef5909527d12bfdeb120dc03a1eff8c navcoin-4.6.0-win32.zip [00:14:22] e3690476ea7b495b3acc9d7e36eb94763809970cf7816b9cda2c546ce1731a0a navcoin-4.6.0-win64-debug.zip c3c758985ebf1aa562d4d48ac775a4d9a9b3b9f1d6471e30a5ec9caa42b3e2e4 navcoin-4.6.0-win64-setup.exe e9937c91d7383fa540d350f3008813cd2adaa1795e20f224593cbde4b555627d navcoin-4.6.0-win64.zip e8c56e47e99af1d41eb05ee961258c8ca4c1822b7a5e72c7961cb32650f2c72f src/navcoin-4.6.0.tar.gz 1ab77ec95438457191e70dedd55b59b30a135bb4090ea7650ee1dd327db6a7c0 [00:14:23] navcoin-win-4.0.0-res.yml [00:18:13] @caffeine you have some data in error.log now? 😄 [00:54:47] @prole OSX: 4b85898fbb98e492fc1d1a64152bbddeb73137f3eb6b01103b806110bb243c35 navcoin-4.6.0-osx.dmg 72bceae82fdf36397ebc0787a2529dafd7ae81f259afab371a0b40270d16fcf6 navcoin-4.6.0-osx.tar.gz 5987daea2b9a25f65f8fb2596d290cc7370a664e28c188c18efb69733cae5a4f navcoin-4.6.0-osx64.tar.gz e8c56e47e99af1d41eb05ee961258c8ca4c1822b7a5e72c7961cb32650f2c72f src/navcoin-4.6.0.tar.gz [00:54:47] 571bf86264463e561aefc06acec74e5182b7efd0e96054ab307db6d7103c0a49 navcoin-osx-4.0.0-res.yml [01:33:20] mine have finished 1eaf880ff5fb82e5cb737281b7021a154686f91fa704b27cbb6ec7c2f24d37a4 navcoin-4.6.0-aarch64-linux-gnu-debug.tar.gz 5837fa19f4b5784871f081d06bcef14ff67b9b0bab1037333b7c514fa0e8e652 navcoin-4.6.0-aarch64-linux-gnu.tar.gz c20b706fe9649d07570e2c21c8ff5331057c7136290c01df2ce6c0cc2b111f15 navcoin-4.6.0-arm-linux-gnueabihf-debug.tar.gz 0ec939fa648a08583a40972d652af3d630f198a9eb38557b94bffabe5eb5b32f [01:33:20] navcoin-4.6.0-arm-linux-gnueabihf.tar.gz 84609ed6a0fd23d0b8111d38c7b6f37a59e91d51352d481110e919dae5faf16b navcoin-4.6.0-i686-pc-linux-gnu-debug.tar.gz 5cbba74983c5cdc39d29f726a62b1ad8e3de1c619d5e4117a68bd96883062e1d navcoin-4.6.0-i686-pc-linux-gnu.tar.gz d99e812e574142913607a20ced1a6691018c15b53629a8ca62b2a383dd94b5b0 navcoin-4.6.0-osx.dmg b6fed840638e553f342ea439d33baf78b070b0b767651c156a7743ae6aa65c7d navcoin-4.6.0-osx64.tar.gz [01:33:21] 5cdc9eef9e394cf015e9da04da17096db871c375f5a9b8be11a230cee7dfc483 navcoin-4.6.0-win32-debug.zip a83e61e7a97d2247e3258909236a95fc84451c53ff83e756223f04e18798099b navcoin-4.6.0-win32-setup.exe 944f6fa5994a90b7fff90a7392d8f56c8ef5909527d12bfdeb120dc03a1eff8c navcoin-4.6.0-win32.zip e3690476ea7b495b3acc9d7e36eb94763809970cf7816b9cda2c546ce1731a0a navcoin-4.6.0-win64-debug.zip 35aaa6f4ea30a525a36427cab5a517e407bcf2083d15daa45fd05712f452f55e [01:33:21] navcoin-4.6.0-win64-setup.exe e9937c91d7383fa540d350f3008813cd2adaa1795e20f224593cbde4b555627d navcoin-4.6.0-win64.zip 0d27053136a258bcd989b3265eb47164bda4749fbbac8a4147c1e817d0b8bdd7 navcoin-4.6.0-x86_64-linux-gnu-debug.tar.gz a82ca18c2a3767a1f895080632e43fe65c46465eaad8354ad983d44885f210ba navcoin-4.6.0-x86_64-linux-gnu.tar.gz e8c56e47e99af1d41eb05ee961258c8ca4c1822b7a5e72c7961cb32650f2c72f navcoin-4.6.0.tar.gz [01:39:24] aarch, arm, i686 adn x86 are all matching [01:39:30] windows zip's are matching [01:40:00] windows exe's are not matching but that is a known issue that they are non deterministic [01:40:30] osx tar and dmg are not matching, i would like to figure out why. [01:41:10] which MacOSX SDK are you using @mxaddict ? [01:41:22] I'm running 10.11 [01:41:24] the one used in travis [01:41:25] 10.11 [01:41:29] mm [01:41:49] I wander why [01:42:32] Hahaha, so windows is non deterministic with the build? That means exe can't be verified? Ironic 😄 [01:43:13] yeah, well we don't exactly know what the issue there is. we've had a couple runs at figuring it out but could never get to the bottom of it [01:43:26] always our exe hashes never match [01:43:44] even if you run it again on your gitian you will receive a different hash [01:47:30] That might be an issue with mingw [01:50:09] I wonder why the mac hashes are not matching [01:52:58] Anyways, I've been running 4.6.0 tag on my VPS with cold staking [01:53:04] And it's been working fine [01:53:23] So 4.60 is a keeper @procrastinator @aguycalled 😄 [01:53:34] nice [01:53:42] lol at that mis-tag [01:53:54] i know im pretty slack, but calling me a procrastinator is a bit of a burn [01:53:58] Yeah, 59 versions in the future 😄 [01:54:10] no, the username tag [01:54:21] Hahaha [01:54:27] LOL, tots missed that [01:54:34] I type pro->tab [01:54:35] Hahaha [01:54:39] yah [01:56:53] It's pretty fun coding in c++ again [01:57:29] It's been ~5 years since I last worked on a c++ project, getting rusty [01:58:51] Anyways, I'm gonna catch some 💤 [01:58:55] TTYL [02:03:20] See ya! glad you're having fun [07:44:53] i built osx again and i have the same hashes i had bbefore (different from you @mxaddict ) [07:45:19] have you built @aguycalled ? it would be great to get this stitched up and published [10:43:50] https://build.nav.community/master/master.SHA256SUM.asc [11:08:15] Looks like our hashes for windows and OSX don't match [11:08:36] Maybe osx has same issue as the one @prole mentioned about windows? [11:38:49] before it was only win with unmatched hashes [13:27:31] Hmmm, I'll see if I can build OSX and get the same hashes again then. [13:28:11] BTW, is there a way to run all the gitian-*.yml? I ran them individually. [13:31:19] @aguycalled @prole what is lacking for relase of 4.6.0? [13:35:00] i have a script to run all of them but any should be a different run of gitian-builder [13:35:46] hash verification and then release the binaries [13:40:01] Ahh, I see [13:40:19] I just recreated a temp.sh and pasted the calls to gbuild 😄 [13:42:04] BTW, If you saw in my PR for the currency changes, seems there was an issue with QT Json support [13:42:17] It would hang when decoding the JSON from the API that I used. [13:42:25] So I had to switch it to boost json [15:35:51] @aguycalled I have 2 wallets running locally with price update code, one compile against QT 5.7.1 and one compiled against Qt 5.9.8 [15:35:57] Waiting to see if they loose connection [15:36:20] I've compiled them with 1 sec interval for price update to speed up the process [15:37:10] great i was trying to find the original issue description [15:38:53] If you find it, let me know. [15:39:49] @aguycalled Your hashes for OSX build did not match @prole's as well right? What do you think the cause was? [15:40:09] I've rebuilt on my gitian, and the hashes are the same as my first build [15:40:21] 4b85898fbb98e492fc1d1a64152bbddeb73137f3eb6b01103b806110bb243c35 navcoin-4.6.0-osx.dmg 72bceae82fdf36397ebc0787a2529dafd7ae81f259afab371a0b40270d16fcf6 navcoin-4.6.0-osx.tar.gz 5987daea2b9a25f65f8fb2596d290cc7370a664e28c188c18efb69733cae5a4f navcoin-4.6.0-osx64.tar.gz e8c56e47e99af1d41eb05ee961258c8ca4c1822b7a5e72c7961cb32650f2c72f src/navcoin-4.6.0.tar.gz 571bf86264463e561aefc06acec74e5182b7efd0e96054ab307db6d7103c0a49 [15:40:22] navcoin-osx-4.0.0-res.yml [15:40:39] So i'm stumped on why they don't match [15:40:51] But our arm and linux builds seem to match [15:41:59] Scratch that, your hashes sent did not match my linux hashes either [15:42:27] we had before osx incongruences with @paul and he solved it by rebuilding his gitian. never managed to figure out why the windows build was not deterministic [15:42:50] I think the windows build might be an issue with mingw [15:43:15] im going to restart my gitian now [15:43:19] building master [15:43:21] I've seen this issue before with builds using mingw [15:43:32] My build was not from master [15:43:38] It was from 4.6.0 tag [15:43:49] Is master head the same as 4.6.0 right now? [15:44:11] --navcoin-core=4.6.0 is what I ran [15:45:32] 4.6.0 tag is at bb51359 [15:45:39] same as master right now [15:47:38] Ok [15:48:18] It's the only difference in our build though, maybe gbuild takes the --navcoin-core param into acount when creating the hashes [15:49:24] ill use the tag just to eliminate possible causes [15:49:50] ACK [15:50:35] I need to get a new PC, my i7-4790k is not cutting it 😄 [15:50:56] Not enough threads to run multiple builds at a time [16:25:48] @aguycalled do you know how long it would take for the issue to come up with the price update code? [16:26:00] I know the old code was set to 120 sec interval [16:26:06] with previous value it was a few days [16:26:15] 2-3 days? [16:26:30] yep [16:26:49] lets say with the old version it took a conservative estimate of 5 days to cause the issue [16:27:04] I'm running 2 nodes @3 sec interval [16:28:03] so you should see within a few hours [16:29:29] So it should pop up in about 1.5 hours [16:29:38] If my math is correct [16:30:16] I'll run these nodes for a day and see what happens [16:31:31] I'm also building PR 470 now to test changes against boost 1.64 [17:27:58] @aguycalled I got the build with PR for prices to crash, but for the wrong reason 😄 [17:28:10] It crashed because of json parsing error [17:28:26] I added a try/catch to fix the issue and pushed to the PR branch [17:43:01] @aguycalled I've updated the PR with another commit to fix the ->deleteLater call [17:43:28] The original code used for updating the price used the QJson* classes [17:43:39] And I noticed that sometimes it would not parse the JSON and would hang [17:43:51] Which meant it would never get to the end of the callback function [17:44:02] And would not run ->deleteLater [17:45:13] And since my PR now uses the boost json functions this no longer happens in my testing, hence this comment in the code: cpp // Parse json // NOTE: Had to use boost json as Q5's json support would not work with // the json data that I was getting from the API, IDK why ¯\_(ツ)_/¯ [17:46:21] I'm running the nodes with the latest patch on PR branch, and will let it run for the rest of the day, I'll let you know how it goes later [18:11:37] @aguycalled @prole I do have permission to merge PR's with 2 approvals? I see the merge button on PR: https://github.com/NAVCoin/navcoin-core/pull/470 [19:04:01] @aguycalled , I got this error: 2019-05-06 19:00:48 ERROR: ReadBlockFromDisk: OpenBlockFile failed for CBlockDiskPos(nFile=0, nPos=915812) [19:04:07] Any idea what could cause this? [19:05:07] -reindex? [19:08:58] @mntyfrsh I got the error on a node that was already synced to the network [19:09:12] After restart it worked fine again [19:37:29] r/w error id say [19:39:59] Ok, so not related to the price auto update changes? I commented on the issue with an error that I think might be the issue that you were talking about. [19:40:12] 2019-05-06 19:16:18 socket send error Bad file descriptor (9) 2019-05-06 19:16:19 INFO: Updating prices 2019-05-06 19:16:19 ERROR: Could not parse price data json 'boost::property_tree::json_parser::json_parser_error' 2019-05-06 19:16:24 INFO: Updating prices [19:40:34] i meant about 2019-05-06 19:00:48 ERROR: ReadBlockFromDisk: OpenBlockFile failed for CBlockDiskPos(nFile=0, nPos=915812) [19:40:36] After those error logs, the wallet closed/crashed [19:40:57] the wallet was not closing back on the days, just losing connectivity iirc [19:41:28] it was showing some message like too many descriptors open [19:41:46] Ahh, [19:42:00] This one is bad file descriptor [19:42:04] So it could be related [19:42:14] Were the errors that you got on a linux or a mac wallet? [19:43:27] Right after that bad descriptor error, the wallet crash with segmentation fault (core dumped) [19:44:09] mac and windows for sure [19:44:55] Ok, I will do more testing and try to find out the cause [19:45:26] I've restarted the node that crashed (node built against QT 5.7.1) [19:45:38] Node that was built agains QT 5.9.8 is still running [19:56:20] @aguycalled your build is still running right? https://build.nav.community/4.6.0/ [19:56:31] yes [19:56:33] all those hashes match me so far (except windows exxs) [19:56:56] if your OSX matches mine, i think i will go ahead and attach the binaries to github and update the website etc [19:58:42] Is there a bounty for fixing windows build hashes? 😃 [20:00:28] there should be [20:01:48] @aguycalled I think I found the issue [20:02:03] I update the code to fetch json from my localhost [20:02:08] And set the interval to 100ms [20:02:22] QT 5.7.1 build crashes every time with: QThreadPipe: Unable to create pipe: Too many open files [20:02:39] Build with QT 5.9.8 does not crash even after 10 minutes of running [20:05:46] NVM node with QT 5.9.8 crashes after 16 minutes [20:06:01] Node with QT 5.7.1 crashes in about 10 seconds [20:06:12] Same error QThreadPipe: Unable to create pipe: Too many open files [20:06:24] Was this the issue that you were having? [20:31:10] very likely was the same msg, does it crash or lose connectivity? [20:31:12] I added some more details to the PR [20:31:20] ill check after the roundtable [20:31:59] It crashes (I think it looses connectivity before it crashes) cause It also logs error connecting to node before the crash [20:32:17] It might only crash because I have it set to such a fast interval [20:32:31] Thanks @aguycalled [21:53:31] @mxaddict https://github.com/NAVCoin/navcoin-core/issues/471 [21:54:21] gave it a high bounty as this is critical to have a trust-less release process [21:55:17] @mxaddict if github allows you to merge, then go for it. [21:55:27] 😉 [22:17:24] @navcoin core @community dev Some reviews would be greatly appreciated : https://github.com/NAVCoin/navcoin-core/pulls/mxaddict 😄 [22:18:16] will review tomorrow what i havent yet. patch-5 gitian build is running now. will approve if it passes [22:20:48] I didn't mean you, some of my PRs have 1 approval (From you) I'm looking for the second approvals 😄 [22:36:16] i will try to get to them over the next couple of days. I have a few other things on my plate at the moment. We will get through them! [23:05:39] @aguycalled your build has finished and the hashes match mine [23:05:51] ill publish them to github and start on my release checklist [23:17:02] Nice! [23:17:28] I'm working on the windows hashes issue now. [23:17:49] Will create PR once i get it working on my gitian VM