1 2016-03-01 12:02:58	0|GitHub148|[13bitcoin] 15crowning- opened pull request #7624: [doc] Missing credit added (060.12...06patch-2) 02https://github.com/bitcoin/bitcoin/pull/7624
  2 2016-03-01 12:47:12	0|GitHub140|13bitcoin/06master 14f5ecd07 15Wladimir J. van der Laan: doc: Add missing credit to 0.12.0 release notes...
  3 2016-03-01 12:47:12	0|GitHub140|[13bitcoin] 15laanwj pushed 1 new commit to 06master: 02https://github.com/bitcoin/bitcoin/commit/f5ecd0737130eed8daf9d76c5232dce7e40b7150
  4 2016-03-01 12:47:22	0|GitHub104|[13bitcoin] 15laanwj closed pull request #7624: [doc] Missing credit added (060.12...06patch-2) 02https://github.com/bitcoin/bitcoin/pull/7624
  5 2016-03-01 12:52:33	0|GitHub142|13bitcoin/060.12 1435af157 15Wladimir J. van der Laan: doc: Clean out release notes...
  6 2016-03-01 12:52:33	0|GitHub142|[13bitcoin] 15laanwj pushed 1 new commit to 060.12: 02https://github.com/bitcoin/bitcoin/commit/35af157641ddbf6090e86edff7533d45ee4fb990
  7 2016-03-01 13:37:37	0|MarcoFalke|wumpus, I think 7607 needs backport to .11 and .12 as well
  8 2016-03-01 13:37:52	0|MarcoFalke|Do you prefer a pull or just do it directly?
  9 2016-03-01 13:39:40	0|btcdrak|if we have to change these we should also take the opportunity to change the fallback URL as well since it's until the curl change,it's already broken
 10 2016-03-01 13:39:52	0|btcdrak|and link directly to http://dev.bitcoincore.org/
 11 2016-03-01 13:41:08	0|wumpus|btcdrak: the redirect works fine IMO, dev.bitcoincore.org is a temporary solution
 12 2016-03-01 13:42:11	0|wumpus|MarcoFalke: backporting a 0.10 pull is confusing me, is this in master already?
 13 2016-03-01 13:42:21	0|MarcoFalke|jup
 14 2016-03-01 13:42:35	0|MarcoFalke|I linked to the backport, so it's easier which commits I mean
 15 2016-03-01 13:42:43	0|MarcoFalke|(it was 3 separate pulls)
 16 2016-03-01 13:43:06	0|wumpus|I'll just cherry-pick to the other branches then
 17 2016-03-01 13:43:24	0|btcdrak|wumpus: the redirects are very awkward creating tight coupling that is more liable to break someday. This would make a nice transition without breaking anything. At the end of the day, there will always be a URL, better if it's a separate subdomain that's also being used for something else (website)
 18 2016-03-01 13:43:39	0|wumpus|btcdrak: dev.bitcoin.org may go away any time
 19 2016-03-01 13:43:55	0|btcdrak|you mean the hosting server?
 20 2016-03-01 13:43:57	0|wumpus|yes
 21 2016-03-01 13:44:08	0|btcdrak|then what would be used as the fallback?
 22 2016-03-01 13:44:11	0|wumpus|I don't have commitment that it will stay up, or anyone will pay for the hosting, etc
 23 2016-03-01 13:44:28	0|btcdrak|dev.bitcoincore.org can always be pointed at any server, it's just a matter of changing DNS records.
 24 2016-03-01 13:44:48	0|wumpus|so I prefer ahving the redirects so we can move the fallbacks anywhere we want later on without any source change
 25 2016-03-01 13:45:27	0|wumpus|otherwise you're just moving the problem, you'd have to have a faux dev.bitcoincore.org with redirects to the actual location
 26 2016-03-01 13:45:30	0|btcdrak|right now, it's broken anyway until the wget is changed to curl, so it;s the perfect time to change the name. It isnt fixed to the server, it's just a dns record.
 27 2016-03-01 13:46:09	0|btcdrak|I'm already using another server to do the redirects..
 28 2016-03-01 13:46:12	0|wumpus|it's not broken due to the bitcoincore.org though
 29 2016-03-01 13:46:21	0|wumpus|the reason to move to curl is some other certificate failure
 30 2016-03-01 13:46:41	0|btcdrak|remember the fallback URL does not need to be https://
 31 2016-03-01 13:46:56	0|btcdrak|either way, you have to change it, but http:// is the right thing not https://
 32 2016-03-01 13:46:58	0|wumpus|which I didn't realy agree on either, but this was cheaper than fixing the certificate issue I guess... :p
 33 2016-03-01 13:47:11	0|btcdrak|http://dev fix is free
 34 2016-03-01 13:47:40	0|btcdrak|since we dont need SSL, since verification is done by hashes
 35 2016-03-01 13:47:49	0|wumpus|but the fallback URL is not beign changes in anything here
 36 2016-03-01 13:47:54	0|wumpus|being changed*
 37 2016-03-01 13:48:44	0|btcdrak|either way, it's broken until a change is made. it's a hard fork :-P
 38 2016-03-01 13:49:06	0|wumpus|<btcdrak> then what would be used as the fallback? <- that's a good question, what is the cheapest way to host http-downloadable files?
 39 2016-03-01 13:49:14	0|wumpus|(with reasonable bandwidth, and limits)
 40 2016-03-01 13:49:30	0|btcdrak|I would guess a $60/yr VPS
 41 2016-03-01 13:49:38	0|wumpus|I don't want to babysit a VPS
 42 2016-03-01 13:49:53	0|wumpus|just host files
 43 2016-03-01 13:49:54	0|btcdrak|well there is also github large file hosting.
 44 2016-03-01 13:51:15	0|wumpus|anyhow these are three separate pulls, we can do a fallback location change later if that's necessary, there's notneed to correlate it to this
 45 2016-03-01 13:51:20	0|btcdrak|we could probably create a repository and just upload the sources using github's releases feature
 46 2016-03-01 13:51:56	0|wumpus|nice
 47 2016-03-01 13:51:57	0|btcdrak|for example: https://github.com/btcdrak/bitcoin/releases
 48 2016-03-01 13:52:17	0|btcdrak|that would be my preferred option. the URLs are predictable too iirc
 49 2016-03-01 13:53:18	0|btcdrak|let me experiment in a repository and get back to you
 50 2016-03-01 13:53:21	0|wumpus|in any case, if we have a redirect then we can change them to point at any location
 51 2016-03-01 13:53:29	0|Luke-Jr|fwiw, I have a terribly ugly hack for https://github.com/luke-jr/cross-binpkgs that could potentially work
 52 2016-03-01 13:54:30	0|wumpus|fallback location broken? just change the redirects and all old versions depends download will just keep working without code changes
 53 2016-03-01 13:54:48	0|Luke-Jr|+1
 54 2016-03-01 13:55:15	0|MarcoFalke|0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
 55 2016-03-01 13:55:15	0|MarcoFalke|100   639  100   639    0     0   2017      0 --:--:-- --:--:-- --:--:--  2017
 56 2016-03-01 13:55:15	0|MarcoFalke|Dload  Upload   Total   Spent    Left  Speed
 57 2016-03-01 13:55:15	0|MarcoFalke|of which: Fetching boost...
 58 2016-03-01 13:55:15	0|MarcoFalke|sha256sum: WARNING: 1 computed checksum did NOT match
 59 2016-03-01 13:55:15	0|MarcoFalke|% Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
 60 2016-03-01 13:55:45	0|MarcoFalke|https://travis-ci.org/bitcoin/bitcoin/jobs/112855470#L831
 61 2016-03-01 13:58:25	0|btcdrak|wumpus: look  https://github.com/btcdrak/sources/releases/tag/src
 62 2016-03-01 13:58:29	0|wumpus|it doesn't actually say what URL it had downloaded
 63 2016-03-01 13:58:37	0|btcdrak|so for example https://github.com/btcdrak/sources/releases/download/src/boost_1_59_0.tar.bz2
 64 2016-03-01 13:58:50	0|btcdrak|the fallback URL would be https://github.com/btcdrak/sources/releases/download/src/
 65 2016-03-01 13:59:28	0|btcdrak|so this could just be https://github.com/bitcoin-core/depends-fallback/releases/download/src/boost_1_59_0.tar.bz2
 66 2016-03-01 13:59:41	0|wumpus|boost hash wrong sounds like a pretty serious issue
 67 2016-03-01 14:00:28	0|MarcoFalke|maybe it's just travis-will-randomly-fail-tuesday
 68 2016-03-01 14:03:47	0|Luke-Jr|:/
 69 2016-03-01 14:03:51	0|btcdrak|need a hosting server for gitian fallback
 70 2016-03-01 14:03:51	0|btcdrak|so we could change the fallback URL to https://github.com/bitcoin-core/depends-fallback/releases/download/src/ but the best would be to setup a permanent redirect from dev.bitcoincore.org -> https://github.com/bitcoin-core/depends-fallbacks/releases/download/src/ then if for any reason the github repo changed you can just update the redirect and now we dont
 71 2016-03-01 14:04:41	0|wumpus|#7136 is already on 0.12
 72 2016-03-01 14:05:09	0|wumpus|btcdrak: but then why set up a dev.bitcoincore.org at all and just change the current redirect from bitcoincore.org?
 73 2016-03-01 14:05:16	0|wumpus|+not
 74 2016-03-01 14:05:43	0|GitHub127|13bitcoin/060.12 1400d57b4 15Luke Dashjr: Workaround Travis-side CI issues...
 75 2016-03-01 14:05:43	0|GitHub127|[13bitcoin] 15laanwj pushed 2 new commits to 060.12: 02https://github.com/bitcoin/bitcoin/compare/35af157641dd...a10da9aa4933
 76 2016-03-01 14:05:45	0|GitHub127|13bitcoin/060.12 14a10da9a 15MarcoFalke: [depends] builders: No need to set -L and --location for curl...
 77 2016-03-01 14:08:11	0|GitHub172|[13bitcoin] 15laanwj pushed 3 new commits to 06master: 02https://github.com/bitcoin/bitcoin/compare/f5ecd0737130...732c01089601
 78 2016-03-01 14:08:12	0|GitHub172|13bitcoin/06master 145c70a6d 15Luke Dashjr: Bugfix: gitian: Add curl to packages (now needed for depends)
 79 2016-03-01 14:08:12	0|GitHub172|13bitcoin/06master 14e5daa2e 15Luke Dashjr: Merge branch 'master' into depends_curl
 80 2016-03-01 14:08:13	0|GitHub172|13bitcoin/06master 14732c010 15Wladimir J. van der Laan: Merge #7614: Bugfix: gitian: Add curl to packages (now needed for depends)...
 81 2016-03-01 14:08:16	0|GitHub49|[13bitcoin] 15laanwj closed pull request #7614: Bugfix: gitian: Add curl to packages (now needed for depends) (06master...06depends_curl) 02https://github.com/bitcoin/bitcoin/pull/7614
 82 2016-03-01 14:10:43	0|GitHub64|13bitcoin/060.10 144e1134b 15Luke Dashjr: Bugfix: gitian: Add curl to packages (now needed for depends)...
 83 2016-03-01 14:10:43	0|GitHub64|[13bitcoin] 15laanwj pushed 1 new commit to 060.10: 02https://github.com/bitcoin/bitcoin/commit/4e1134bdf1acff669c0f489934ac5f919c634d69
 84 2016-03-01 14:13:44	0|GitHub168|13bitcoin/060.12 14ca8f160 15Luke Dashjr: Bugfix: gitian: Add curl to packages (now needed for depends)...
 85 2016-03-01 14:13:44	0|GitHub168|[13bitcoin] 15laanwj pushed 1 new commit to 060.12: 02https://github.com/bitcoin/bitcoin/commit/ca8f160af5a54d08f8dc73acd959b0a73a7b427c
 86 2016-03-01 14:16:18	0|GitHub130|[13bitcoin] 15luke-jr opened pull request #7625: Bugfix: Check for bench_bitcoin being enabled where needed, and skip UniValue dependency when unused (06master...06bugfix_bench_checks) 02https://github.com/bitcoin/bitcoin/pull/7625
 87 2016-03-01 14:17:33	0|GitHub73|[13bitcoin] 15laanwj pushed 4 new commits to 060.11: 02https://github.com/bitcoin/bitcoin/compare/c40ec1421048...a0cfe3a9e6c5
 88 2016-03-01 14:17:34	0|GitHub73|13bitcoin/060.11 147815cb6 15Luke Dashjr: Bugfix: gitian: Add curl to packages (now needed for depends)...
 89 2016-03-01 14:17:34	0|GitHub73|13bitcoin/060.11 14a0e13f0 15MarcoFalke: Fix url in .travis.yml...
 90 2016-03-01 14:17:35	0|GitHub73|13bitcoin/060.11 1477841d4 15Luke Dashjr: Workaround Travis-side CI issues...
 91 2016-03-01 15:51:06	0|wumpus|I don't think the fallback logic is working properly, if boost really changed their tarball (or something else goes wrong), shouldn't it be trying to get it from the alternate URL?
 92 2016-03-01 15:52:07	0|wumpus|this was the same for libqrcode in the original problem, it was failing the download on ssl cert, but it didn't try the fallback
 93 2016-03-01 15:52:50	0|arlisk|opaa
 94 2016-03-01 15:54:04	0|wumpus|and if the fallback logic isn't working, the last thing we should be worrying about is where its files are hosted :p
 95 2016-03-01 15:58:46	0|wumpus|ok just downloaded http://sourceforge.net/projects/boost/files/boost/1.59.0/boost_1_59_0.tar.bz2 and it still has the same sha256sum, 727a932322d94287b62abb1bd2d41723eec4356a7728909e38adb65ca25241ca. It maybe that the file is corrupted on one of SF's mirrors, ofc.
 96 2016-03-01 15:59:12	0|btcdrak|or tampered with...
 97 2016-03-01 15:59:32	0|btcdrak|it wouldnt be the first time sf filedownloads have been compromised
 98 2016-03-01 16:01:44	0|wumpus|could be, though I think it's unlikely someone would mess with an old boost download
 99 2016-03-01 16:01:57	0|wumpus|not impossible ofcourse
100 2016-03-01 16:30:16	0|GitHub162|13bitcoin/06master 14fafe446 15MarcoFalke: [depends] Delete unused patches...
101 2016-03-01 16:30:16	0|GitHub162|[13bitcoin] 15laanwj pushed 2 new commits to 06master: 02https://github.com/bitcoin/bitcoin/compare/732c01089601...639ec582d0f3
102 2016-03-01 16:30:17	0|GitHub162|13bitcoin/06master 14639ec58 15Wladimir J. van der Laan: Merge #7616:  [depends] Delete unused patches...
103 2016-03-01 16:30:26	0|GitHub177|[13bitcoin] 15laanwj closed pull request #7616:  [depends] Delete unused patches  (06master...06Mf1602-boost155) 02https://github.com/bitcoin/bitcoin/pull/7616
104 2016-03-01 16:58:28	0|gmaxwell|wumpus: well an old boost download we use?
105 2016-03-01 16:58:51	0|wumpus|yes but as the download process detects it, doing it just for us is pointless
106 2016-03-01 17:00:10	0|wumpus|tho would be kind of funny if someone invested a lot in an attack that they didn't even test locally :p
107 2016-03-01 17:33:27	0|Giszmo|Looking into bip142 I wonder if there is a schema to optionally allow segWit that would be compatible with bip21? some bitcoin:1....?amount=12&segWitAllowed=true kind of downwards compatible style we could use in mycelium to leave it to the sender if he wants to use segwit?
108 2016-03-01 17:34:00	0|Luke-Jr|Giszmo: uh, you can't use segwit like that
109 2016-03-01 17:34:12	0|Luke-Jr|segwit is tied to the address
110 2016-03-01 17:34:22	0|Luke-Jr|a 1… address can never be segwit
111 2016-03-01 17:34:37	0|Luke-Jr|only 3…
112 2016-03-01 17:39:40	0|GitHub111|[13bitcoin] 15laanwj pushed 4 new commits to 06master: 02https://github.com/bitcoin/bitcoin/compare/639ec582d0f3...e5121eb951c4
113 2016-03-01 17:39:41	0|GitHub111|13bitcoin/06master 14fa06ce0 15MarcoFalke: Fix doxygen comment for payTxFee
114 2016-03-01 17:39:41	0|GitHub111|13bitcoin/06master 14fa97f95 15MarcoFalke: [doc] Fix markdown
115 2016-03-01 17:39:42	0|GitHub111|13bitcoin/06master 14fa26652 15MarcoFalke: Make sure LogPrintf strings are line-terminated
116 2016-03-01 17:39:53	0|GitHub37|[13bitcoin] 15laanwj closed pull request #7617: [doc/log] Fix markdown syntax and line terminate LogPrint (06master...06Mf1602-trivial9) 02https://github.com/bitcoin/bitcoin/pull/7617
117 2016-03-01 17:40:02	0|Giszmo|Thanks Luke. Back to study the bips ...
118 2016-03-01 19:17:13	0|michagogo|AIUI increasing the prune limit just pauses it until the new threshold is hit. Why not go back and fetch the old data?
119 2016-03-01 19:17:31	0|michagogo|I mean, we still have the header chain, right?
120 2016-03-01 19:17:40	0|sdaftuar|michagogo: download old blocks, just to store them?
121 2016-03-01 19:17:48	0|michagogo|Oh. Wait.
122 2016-03-01 19:18:02	0|michagogo|Forgot about the undo data.
123 2016-03-01 19:18:08	0|michagogo|Never mind…
124 2016-03-01 19:18:38	0|GitHub82|[13bitcoin] 15ericshawlinux opened pull request #7628: QT: Add 'copy full transaction details' option (06master...06master) 02https://github.com/bitcoin/bitcoin/pull/7628
125 2016-03-01 22:09:51	0|gmaxwell|Is it just me or has openssl missed their announced window?
126 2016-03-01 22:10:11	0|gmaxwell|oh they haven't just nothing to their announce list. :(
127 2016-03-01 22:10:23	0|GitHub106|[13bitcoin] 15pstratem opened pull request #7629: Order CTxMemPool::queryHashes result by feerate including descendents. (06master...062016-03-01-queryhashes) 02https://github.com/bitcoin/bitcoin/pull/7629
128 2016-03-01 22:10:59	0|sdaftuar|https://www.openssl.org/news/secadv/20160301.txt
129 2016-03-01 22:20:41	0|phantomcircuit|sdaftuar, do i have that correct that using the descendent modified fee index in reverse order will produce a stream of transactions with no orphans
130 2016-03-01 22:20:42	0|phantomcircuit|?
131 2016-03-01 22:20:51	0|sdaftuar|phantomcircuit: was just going to comment on the PR
132 2016-03-01 22:20:58	0|sdaftuar|phantomcircuit: no, you need to walk the ancestors
133 2016-03-01 22:21:20	0|phantomcircuit|descendants are the things which depend on the transaction right?
134 2016-03-01 22:21:20	0|sdaftuar|this is slightly tricky to write efficiently actually
135 2016-03-01 22:21:28	0|phantomcircuit|ie children not the parent
136 2016-03-01 22:21:35	0|sdaftuar|yes that's right
137 2016-03-01 22:21:43	0|sdaftuar|however the highest scoring thing might have ancestors with low fee
138 2016-03-01 22:22:34	0|sdaftuar|there's no direct relationship between any of the feerate scores and a "no orphan" order
139 2016-03-01 22:22:51	0|phantomcircuit|but for any given tree of transactions the parent transaction(s) are going to have a higher descendent fee total ... oh right feerate
140 2016-03-01 22:23:10	0|sdaftuar|see my PR for a "CPFP" mining algorithm,
141 2016-03-01 22:23:17	0|sdaftuar|i think i solve the exact problem you're trying to solve
142 2016-03-01 22:23:28	0|sdaftuar|#7600
143 2016-03-01 22:24:50	0|sdaftuar|is there a reason to prefer descendant fee rate over ancestor fee rate here?
144 2016-03-01 22:25:21	0|sdaftuar|the latter is much closer to being in sync with what is valuable to miners
145 2016-03-01 22:26:09	0|sdaftuar|(though at the low end, descendant fee rate score is more accurate for what gets evicted from the mempool)
146 2016-03-01 22:27:52	0|phantomcircuit|sdaftuar, the descendant tracking essentially answer the question, if i removed this transaction what would the net effect be while ancestor tracking answers the question if i added this transaction and all of it's parents what would the net effect be
147 2016-03-01 22:28:42	0|phantomcircuit|the descendant tracking can get you the same result as ancestor tracking if you start out with the full mempool and simulate reducing the mempool limit to the size of block you're trying to build
148 2016-03-01 22:29:05	0|phantomcircuit|(i think)
149 2016-03-01 22:29:31	0|sdaftuar|i don't disagree with what you wrote (well not sure i fully follow but what you wrote sounds right).  if you're trying to communicate the most valuable transactions, ancestor feerate should capture that
150 2016-03-01 22:29:45	0|sdaftuar|if you're trying to communicate the least valuable transactions, i think the worst descendant fee rate captures that
151 2016-03-01 22:30:08	0|phantomcircuit|yeah that's what i was saying basically
152 2016-03-01 22:30:43	0|sdaftuar|ok, so i'm still not sure exactly how you plan to use this but you might want to check out the SortForBlock code in #7600
153 2016-03-01 22:31:01	0|sdaftuar|basically if you want to take a tx and communicate it in a no-orphan way
154 2016-03-01 22:31:11	0|sdaftuar|you call CalculateMemPoolAncestors on it
155 2016-03-01 22:31:26	0|sdaftuar|and then sort by nCountWithAncestors
156 2016-03-01 22:31:58	0|sdaftuar|(this will be trickier if you are looking at descendant fee rate, in which case you presumably want to communicate the descendants of a tx)
157 2016-03-01 22:32:18	0|phantomcircuit|sdaftuar, i think you're right that using ancestor feerate is better
158 2016-03-01 22:32:30	0|sdaftuar|i have to run now, family time.  back in a couple hours...