1 2014-07-10 00:07:31 <userguy243> Luke-Jr  - so theoretically there's no way to know for sure?  Right now, with 7 tps hard limit set various places, you still have 60 GB...and hardly anyone transfers it to another person, it's quite rare.  (to buy things iwth BTC and use them as currency.  it happens but it's rare.  it's a Pretty Big Deal for smeone to transfer BTC today, they make special arrangments.  it's not an off-hand thing)
  2 2014-07-10 00:08:03 <Luke-Jr> userguy243: O.o what? People buy things with bitcoins all the time.
  3 2014-07-10 00:08:45 <userguy243> Luke-Jr I mean as compared with charges on credit cards, for example.  most people who have btc in their wallet, still are about 50x more likely to charge something to their card.
  4 2014-07-10 00:09:06 <jcorgan> userguy243: those numbers sound made up
  5 2014-07-10 00:09:40 <userguy243> jcorgan - how many charges have you had on your credit/debit cards in the past 30 days, versus how many things have you bought with BTC?  That's my point.
  6 2014-07-10 00:09:56 <userguy243> when you're in a physical store(shp) you can't even spend your btc there.\
  7 2014-07-10 00:10:02 <userguy243> shp=shop
  8 2014-07-10 00:10:10 <userguy243> i missed hte o
  9 2014-07-10 00:11:13 <userguy243> And that's how the blockchain is 60 GB today.  And I'm talking YOU, who are here in #bitcoin-dev.  How many people won credit cards/debit cards worldwide who have 0 walleters, with a total of 0 btc in them?
 10 2014-07-10 00:11:25 <userguy243> won=own
 11 2014-07-10 00:11:33 <jcorgan> this is pretty off-topic for -dev
 12 2014-07-10 00:11:37 <userguy243> walleters = wallets
 13 2014-07-10 00:11:41 <userguy243> jcorgan point taken
 14 2014-07-10 00:12:12 <userguy243> I was just asking how the blockchain could be pruned, for once a billion people use it.  it'llget pretty gnarly.  luke-jr (and earlier someone else) said that it can't get pruned
 15 2014-07-10 00:17:11 <WeCluster> 2014-07-10 00:13:04   nActualTimespan = 941328  before bounds 2014-07-10 00:13:04 GetNextWorkRequired RETARGET 2014-07-10 00:13:04 nTargetTimespan = 1209600    nActualTimespan = 941328 2014-07-10 00:13:04 Before: 1c2a1115  000000002a111500000000000000000000000000000000000000000000000000 2014-07-10 00:13:04 After:  1c20bca7  0000000020bca74a83a83a83a83a83a83a83a83a83a83a83a83a83a83a83a83a
 16 2014-07-10 00:20:16 <userguy243> there are semiformal plans to ease the TPS restrictions too, once there is more adoption.  And semiformal plans for more adoption :)
 17 2014-07-10 00:21:06 <WeCluster> bitcoind working as expected in headless build -- seems like mining could now occur on this box. I will be doing development in this area specifically, as well as core protocol improvements. any pointers for where to find some bugs that need fixing and other useful work that I can do while still learning?
 18 2014-07-10 00:22:19 <kazcw> WeCluster: if you could find the commit that breaks windows build and include that in a bug report, it'd help fix it
 19 2014-07-10 00:22:57 <kazcw> WeCluster: do you know git bisect? it's pretty cool, you can do a binary search through commits to find the one that breaks it
 20 2014-07-10 00:23:49 <kazcw> WeCluster: but you should use a release version (like v0.9.2.1) for mining
 21 2014-07-10 00:23:58 <kazcw> and wallets for that matter
 22 2014-07-10 00:24:15 <WeCluster> understood. will do.
 23 2014-07-10 00:24:20 <userguy243> cu guys later
 24 2014-07-10 01:00:43 <btc111> are there any android devlopers here? pls msg me . have a question will pay bounty
 25 2014-07-10 01:27:58 <ttgg> What compiler/IDE is best for working with the bitcoin source on Windows?
 26 2014-07-10 01:29:48 <kazcw> ttgg: mingw. see doc/build-msw.md
 27 2014-07-10 01:30:59 <ttgg> Thanks kaz
 28 2014-07-10 01:49:03 <ttgg> "this is a security-critical project where any mistake might cost people lots of money."
 29 2014-07-10 01:49:40 <Emcy> sure is
 30 2014-07-10 01:51:39 <ttgg> bitcoin-qt is this still considered like the default wallet? Will it continue to 'basically' function if I compile my own bitcoin source, but I make patches that add functionality? I'm assuming I'd have to also make changes to the wallet code to use my changes from a wallet...
 31 2014-07-10 01:51:48 <ttgg> I know that might be a stupid question, but I'm just starting with this
 32 2014-07-10 01:59:43 <Emcy> how can anyone say how it might function if you add your own patches
 33 2014-07-10 02:09:29 <ttgg> This is just my own internal testing and stuff
 34 2014-07-10 02:09:46 <ttgg> I'm not considering putting in any pull requests or anything like that until I have a fundamental understanding of the code
 35 2014-07-10 02:09:52 <WeCluster> ttgg: we're using mingw with msys and in that case you should follow doc/build-unix.md instead of doc/build-msw.md
 36 2014-07-10 02:09:57 <ttgg> I plan on playing with it extensively
 37 2014-07-10 02:10:08 <ttgg> Thanks WeCluster
 38 2014-07-10 02:10:24 <PRab> ttgg: If you are testing, make sure you use testnet coins instead of real bitcoin in case something goes wrong.
 39 2014-07-10 02:10:29 <ttgg> I'm playing with mingw and msys right now and I just started eyeballing the bitcoin.org developer guide
 40 2014-07-10 02:11:02 <WeCluster> ttgg: also note that you will need to add -D_WIN32_WINNT=0x0601 -DWINVER=0x0601 (for Windows 7+) to the Makefiles and probably also -lz (for including libz at link time) to get the source to build under windows if you're using gcc under msys/mingw
 41 2014-07-10 02:11:18 <ttgg> thank you WeCluster
 42 2014-07-10 02:12:01 <ttgg> I've been looking around and I can't find any good newbie guides for this stuff, I'm coming to the table here with just a basic understanding of C++, a layman's understanding of the Bitcoin protocol, and the desire to help
 43 2014-07-10 02:12:30 <WeCluster> we've just gone through the process of building the source in this way on Windows 7 -- kazcw helped earlier today with this, we started out using the latest beta version of the bitcoin source but found it would not build. v0.9.2.1 does build, and if you can't get it built we will be glad to help.
 44 2014-07-10 02:13:02 <ttgg> Thanks WeCluster, should I just try to compile and build it as my first mission?
 45 2014-07-10 02:13:09 <ttgg> I'm on win7
 46 2014-07-10 02:13:46 <WeCluster> start here https://bitcointalk.org/index.php?topic=149479.0
 47 2014-07-10 02:14:55 <WeCluster> we were only interested in building a headless installation, which is simpler, but do not look at step #3 and think to yourself "the newer beta source would be more fun"
 48 2014-07-10 02:15:31 <ttgg> "fun"
 49 2014-07-10 02:15:58 <ttgg> Are all of the library links in this thread up to date/functionable still?
 50 2014-07-10 02:16:04 <ttgg> I guess I'll find out...
 51 2014-07-10 02:16:58 <WeCluster> you can use any version of boost and any version of openssl ... we used boost 50 and openssl 1.0.1h
 52 2014-07-10 02:18:09 <WeCluster> building the source as a headless bitcoind installation made it simpler. no need for Berkeley DB, and none of the qt libs are required. to build headless you specify --disable-wallet
 53 2014-07-10 02:18:34 <ttgg> Should I go headless my first go? I kind of want to play with the whole enchilada
 54 2014-07-10 02:18:51 <WeCluster> also --disable-upnp-default
 55 2014-07-10 02:19:07 <ttgg> Also, why hasn't someone made a unified library installer for this?
 56 2014-07-10 02:19:21 <ttgg> Step #2 is entirely too long imo
 57 2014-07-10 02:19:30 <ttgg> If people are using this as a reference
 58 2014-07-10 02:20:10 <WeCluster> we were unable to get boost 55 to build under mingw/msys
 59 2014-07-10 02:20:22 <ttgg> If no one is doing it, I'm going to make a single download and installer for all of these, to simplify step 2
 60 2014-07-10 02:24:13 <WeCluster> the most difficult part of getting it built under mingw/msys on win7 was ensuring the build environment/toolchain and paths required were present. when you make your first build attempt, let us know.
 61 2014-07-10 02:25:29 <ttgg> I'll be doing it within the hour
 62 2014-07-10 02:28:21 <ttgg> within the mingw and msys shell "cd /c/deps/"
 63 2014-07-10 02:28:38 <ttgg> I don't quite understand, is this a directory I need or is it some pseudo directory in the shell?
 64 2014-07-10 02:32:26 <ttgg> I feel like an idiot because I'm getting stuck at things that should be simple
 65 2014-07-10 02:32:46 <WeCluster> your entire hard drive is available to you as /c (in case you're new to msys) and deps is the name of the directory those instructions recommend for unpacking the dependencies like boost
 66 2014-07-10 02:33:11 <ttgg> Woo!
 67 2014-07-10 02:33:16 <ttgg> Thanks, that's the explanation I needed
 68 2014-07-10 02:35:50 <WeCluster> ttgg: FWIW, we did not follow the instructions you're attempting to follow... never issued any command from command prompt, instead we did everything within msys
 69 2014-07-10 02:36:53 <ttgg> ah
 70 2014-07-10 02:37:03 <ttgg> Well fwiw, it's working so far
 71 2014-07-10 02:37:08 <ttgg> :)
 72 2014-07-10 02:38:16 <ttgg> making stuff right now in step 2
 73 2014-07-10 02:38:18 <ttgg> no hangups yet
 74 2014-07-10 02:38:56 <ttgg> Did Satoshi use all these libraries in the original? I'm sure we obfuscated the hell out of the code by now... Maybe he used boost?
 75 2014-07-10 02:44:44 <sipa> jcorgan: yes
 76 2014-07-10 02:45:15 <sipa> jcorgan: after calling the init function, and before calling the free function, you can call multiple functions in parallel
 77 2014-07-10 02:46:59 <jcorgan> thanks.  it looked that way from the calling sequence but i wasn't sure
 78 2014-07-10 02:53:16 <sipa> ttgg: which libraries in particular?
 79 2014-07-10 02:53:24 <sipa> he certainly used wx, bdb, openssl
 80 2014-07-10 02:53:38 <sipa> boost i don't remember
 81 2014-07-10 02:55:15 <ttgg> I'm going through a library list for 0.9.2.1 compile guide for Win7
 82 2014-07-10 02:55:42 <sipa> not at a computer now
 83 2014-07-10 02:56:03 <sipa> qt was only introduced in 0.5
 84 2014-07-10 02:56:18 <sipa> qrencode probably too
 85 2014-07-10 02:56:41 <sipa> you canfetch an early version from git and look :)
 86 2014-07-10 03:11:33 <ttgg> bootstrap.bat mingw
 87 2014-07-10 03:11:38 <ttgg> I got stuck here
 88 2014-07-10 03:11:48 <ttgg> is the mingw supposed to be my mingw shell path?
 89 2014-07-10 03:12:02 <ttgg> when trying to configure/install boost
 90 2014-07-10 03:12:53 <WeCluster> which version of boost have you selected?
 91 2014-07-10 03:13:01 <ttgg> 1_55_0
 92 2014-07-10 03:15:09 <ttgg> vcvars32.bat is not recognized
 93 2014-07-10 03:15:21 <ttgg> i get a string of errors like that when trying to do bootstrap.bat mingw
 94 2014-07-10 03:15:46 <WeCluster> tried to build 1_55_0 under mingw/msys but failed and gave up
 95 2014-07-10 03:16:32 <ttgg> you said you used boost 1_50_0?
 96 2014-07-10 03:21:25 <WeCluster> you should probably be using bootstrap.sh rather than bootstrap.bat
 97 2014-07-10 03:21:50 <ttgg> can i run the .sh in windows?
 98 2014-07-10 03:21:59 <WeCluster> no, run it in msys
 99 2014-07-10 03:23:01 <WeCluster> one of the reasons we did not follow, precisely, those instructions that you're trying to follow was that we had no intention of using msvc -- wanted to build using gcc toolchain instead.
100 2014-07-10 03:24:03 <ttgg> and... broken
101 2014-07-10 03:24:04 <ttgg> execunix.c:20:26: fatal error: sys/resource.h: No such file or directory
102 2014-07-10 03:24:30 <WeCluster> because your CPATH is not set properly
103 2014-07-10 03:26:39 <WeCluster> we added the following to /etc/profile to set CPATH, BOOST_ROOT, and OPENSSL_CONF in the shell:
104 2014-07-10 03:27:03 <WeCluster> OPENSSL_CONF="/c/MinGW/msys/1.0/local/ssl/openssl.cnf" CPATH="/c/MinGW/include" BOOST_ROOT="/c/MinGW/msys/1.0/dev/boost_1_50_0" export OPENSSL_CONF CPATH BOOST_ROOT
105 2014-07-10 03:27:24 <ttgg> the msys /etc/profile ?
106 2014-07-10 03:27:33 <WeCluster> yes
107 2014-07-10 03:27:47 <WeCluster> after you modify it, use "source /etc/profile" to reset your current environment in your shell
108 2014-07-10 03:29:19 <ttgg> same error :\
109 2014-07-10 03:30:05 <ttgg> do the line breaks matter for the code you posted?
110 2014-07-10 03:30:34 <WeCluster> yes
111 2014-07-10 03:31:05 <WeCluster> but there's another include directory under msys rather than MinGW ... what is the name of the file your build is searching for again? (my IRC history got cleared)
112 2014-07-10 03:31:26 <ttgg> execunix.c:20:26: fatal error: sys/resource.h: No such file or directory
113 2014-07-10 03:31:29 <ttgg> sys/resource.h
114 2014-07-10 03:33:17 <WeCluster> yes, so, the correct CPATH for this is /c/MinGW/msys/1.0/include
115 2014-07-10 03:33:49 <WeCluster> we simply switched back and forth to get the dependencies to build
116 2014-07-10 03:34:14 <WeCluster> it wasn't pretty, but it worked
117 2014-07-10 03:36:39 <WeCluster> seems like you should be able to use a CPATH that references both the msys and mingw include directories ... but not being 100% sure that's a valid thing to do, we didn't do it.
118 2014-07-10 03:37:00 <ttgg> Can I have multiple CPATH's and live?
119 2014-07-10 03:38:07 <WeCluster> you can have multiple paths in the CPATH, separated by either : or ;
120 2014-07-10 03:38:38 <WeCluster> but not multiple CPATH lines, or the last one will be the only CPATH
121 2014-07-10 03:39:49 <WeCluster> since mingw/msys is a combination of Windows environment and msys environment variables, it isn't clear whether the separator character for a compound CPATH path would be : or ;
122 2014-07-10 03:48:40 <earlz> Was getwork removed from master?
123 2014-07-10 03:54:04 <ttgg> CPATH="/c/MinGW/include;/c/MinGW/msys/1.0/include"
124 2014-07-10 03:54:07 <ttgg> tried, didn't help
125 2014-07-10 03:55:24 <earlz> ugh. I can't get master to work at all
126 2014-07-10 03:55:31 <earlz> I guess go to 9.2.1
127 2014-07-10 03:57:08 <WeCluster> try CPATH="/c/MinGW/msys/1.0/include"
128 2014-07-10 03:57:33 <ttgg> figured it out
129 2014-07-10 03:57:35 <ttgg> missing an option
130 2014-07-10 03:57:37 <WeCluster> you can see that resource.h is present here: /c/MinGW/msys/1.0/include/sys/
131 2014-07-10 03:58:00 <ttgg> had to do bootstrap.sh --with-toolset=mingw
132 2014-07-10 03:58:11 <ttgg> found answer on stackoverflow
133 2014-07-10 03:59:15 <WeCluster> good. when you get to the ./configure step for building bitcoin src, here's the command we used: CPPFLAGS="-I/dev/boost_1_50_0 \ -I/dev/openssl-1.0.1h/include" \ LDFLAGS="-L/dev/boost_1_50_0/stage/lib \ -L/dev/openssl-1.0.1h" \ ./configure \ --disable-wallet \ --disable-upnp-default \ --with-boost-libdir=/dev/boost_1_50_0/stage/lib \ --with-boost-system=mgw48-mt-1_50 \ --with-boost-filesystem=mgw48-mt-1_50 \ --with-boost-progr
134 2014-07-10 03:59:58 <WeCluster> similar to the one shown here in step #3: https://bitcointalk.org/index.php?topic=149479.0
135 2014-07-10 04:00:38 <sabfer> hi! has anyone else seen this error while running the deploy script for OS X? https://gist.github.com/anonymous/07dd73647a3e2d484ddf
136 2014-07-10 04:00:45 <WeCluster> the configure script does test-compilations using boost, so it turned out to be very important that the configure script get those paths right.
137 2014-07-10 04:02:29 <sabfer> I found something similar on the logs, but it referred to another library http://bitcoinstats.com/irc/bitcoin-dev/logs/2013/09/15
138 2014-07-10 04:03:59 <ttgg> I had to jerry-rig the b2 build of boost
139 2014-07-10 04:04:07 <ttgg> missing a .jam file but I found a way to force it
140 2014-07-10 04:04:33 <WeCluster> are you taking detailed notes so you can fix all of this and make a single build/installer program?
141 2014-07-10 04:04:46 <ttgg> Yes, well enough WeCluster
142 2014-07-10 04:04:54 <ttgg> I'll be making the all-in-one this weekend probably
143 2014-07-10 04:05:11 <ttgg> And I'm also working on a bitcoin source map when I learn the bitcoin source code well enough
144 2014-07-10 04:05:26 <ttgg> I want to make this as easy for a newbie as possible
145 2014-07-10 04:12:56 <WeCluster> understood. but newbies download and install binaries, not source code. there isn't as much room for introducing ad hoc changes to bitcoin as there might be in, say, the Linux Kernel, because no bitcoind is an island.
146 2014-07-10 04:16:00 <ttgg> This is for newbies who want to work with the code but don't know their ass from a hot rock
147 2014-07-10 04:41:43 <michagogo> WeCluster, ttgg: thanks for working on getting windows builds working
148 2014-07-10 04:42:06 <michagogo> Right now windows release binaries are built with mingw on linux
149 2014-07-10 04:42:07 <WeCluster> michagogo: thanks for the thanks!
150 2014-07-10 04:42:32 <michagogo> Just one tip: look at build-unix.md's history, I think improvements were made to it
151 2014-07-10 04:42:47 <michagogo> And I don't think build-msw got the same treatment
152 2014-07-10 04:43:36 <michagogo> (Note that I'm unfamiliar with both docs, so I can't specifically endorse or [the opposite, I just woke up] them)
153 2014-07-10 05:05:54 <WeCluster> ttgg: have you built everything successfully yet? did boost 1_55_0 work for you?
154 2014-07-10 05:14:12 <wumpus> michagogo: indeed, I'm even going to remove the build_msw.md if no one is going to pick it up and maintain it
155 2014-07-10 05:15:00 <michagogo> wumpus: what's in that file, anyway?
156 2014-07-10 05:15:19 <michagogo> (I'm on my phone atm, or I'd just go and look at it)
157 2014-07-10 05:15:38 <phantomcircuit> michagogo, instructions which probably dont work for building the windows binaries on linux
158 2014-07-10 05:15:48 <michagogo> Heh
159 2014-07-10 05:15:49 <wumpus> michagogo: windows build instructions from eh, three versions ago or so?
160 2014-07-10 05:16:01 <wumpus> phantomcircuit: no it's instruction for building windows binaries on windows
161 2014-07-10 05:16:08 <michagogo> Wait, three versions ago?
162 2014-07-10 05:16:16 <michagogo> 0.9.0?
163 2014-07-10 05:16:21 <michagogo> That doesn't sound right
164 2014-07-10 05:16:24 <wumpus> michagogo: more like three major versions lol
165 2014-07-10 05:16:34 <michagogo> Ah.
166 2014-07-10 05:16:38 <michagogo> So, 0.6?
167 2014-07-10 05:16:38 <wumpus> it's never updated, no one cares about it, it should go
168 2014-07-10 05:16:55 <michagogo> Well, apparently a couple people do!
169 2014-07-10 05:16:56 <phantomcircuit> wumpus, oh really?
170 2014-07-10 05:17:02 <phantomcircuit> iirc that never worked
171 2014-07-10 05:17:05 <phantomcircuit> like ever
172 2014-07-10 05:17:14 <wumpus> well there are a few people that build on windows
173 2014-07-10 05:17:23 <wumpus> but they keep their tactics veeery secret :p
174 2014-07-10 05:17:51 <wumpus> they wouldn't contribute anything back to git or something like that
175 2014-07-10 05:18:18 <michagogo> 
ACTION is not one of those people

176 2014-07-10 05:18:27 <wumpus> anyhow, the gitian build works best for building windows executables
177 2014-07-10 05:18:45 <michagogo> I think the one time I ran experimental code on Windows I gbuilt it
178 2014-07-10 05:20:21 <wumpus> this issue is now more than a year old, and the problem existed before that, https://github.com/bitcoin/bitcoin/issues/1401
179 2014-07-10 05:23:43 <WeCluster> we are curious about building with visual studio 2013 on Windows ... but presume that will be such a big undertaking that we started with mingw/msys building with gcc toolchain instead, so we didn't use build-msw.md
180 2014-07-10 05:24:09 <wumpus> build-msw has only msys/mingw build steps
181 2014-07-10 05:24:25 <WeCluster> following the build-linux instructions under msys/mingw worked well enough. it was decipherable.
182 2014-07-10 05:24:49 <WeCluster> build-msw appears to depend on msvc
183 2014-07-10 05:24:53 <wumpus> no, it does not
184 2014-07-10 05:25:25 <wumpus> not that it matters, it's hopelessly outdated
185 2014-07-10 05:26:04 <wumpus> see, this mentions msys about a zillion times, https://github.com/bitcoin/bitcoin/blob/master/doc/build-msw.md
186 2014-07-10 05:26:10 <WeCluster> hmm... just looked at it again a couple hours ago. the instructions refer to at least one .bat file which can only use msvc
187 2014-07-10 05:27:19 <wumpus> anyhow, if you succeed in doing this it would be great if you updated the build-msw.md  with the new information
188 2014-07-10 05:28:04 <wumpus> they are meant to be msys build instructions, if something about msvc sneaked into there that's possible but wrong
189 2014-07-10 05:29:12 <WeCluster> you're right, it's all MSYS-based without msvc ... was thinking about these instructions instead: https://bitcointalk.org/index.php?topic=149479.0
190 2014-07-10 05:29:55 <WeCluster> build-msw.md references build-unix.md so it's not clear why build-msw.md even exists any longer. vote to remove it.
191 2014-07-10 05:30:37 <wumpus> oh the forum instructions... right, I don't know about those at all, they are a brave attempt at bulding with msvc, but it was never contributed back to git
192 2014-07-10 05:31:02 <wumpus> it exists to provide easy steps for building the dependencies on windows
193 2014-07-10 05:31:05 <WeCluster> this seems to be the only thing of potential importance: change 'MAKE' env. variable from 'C:\MinGW32\bin\mingw32-make.exe' to '/c/MinGW32/bin/mingw32-make.exe'
194 2014-07-10 05:31:13 <wumpus> which is *not* the same as on linux
195 2014-07-10 05:31:30 <wumpus> and on linux you can usually install distro packages
196 2014-07-10 05:31:31 <WeCluster> but there's no need to build from the windows command prompt, when msys shell works so much better anyway.
197 2014-07-10 05:31:58 <wumpus> ... but it mentions only the msys shell
198 2014-07-10 05:32:35 <WeCluster> we built the source just fine without using mingw32-make.exe
199 2014-07-10 05:32:42 <WeCluster> on mingw/msys
200 2014-07-10 05:32:46 <wumpus> that's fine, update the document if you can
201 2014-07-10 05:33:19 <wumpus> the build instructions are not supposed to be the only, holy way to do things, there are alway alternative ways and they may work better...
202 2014-07-10 05:33:28 <WeCluster> ttgg is working on a revision and a single build script, we'll see what we can come up with together
203 2014-07-10 05:34:08 <wumpus> ok great
204 2014-07-10 05:36:53 <WeCluster> building the dependencies on windows by attempting to follow windows build instructions instead of using msys shell seems to be voodoo art that depends on random chance and sunspots or something, whereas following the straight-forward instructions for building under unix seemed to make sense at each step.
205 2014-07-10 05:37:20 <wumpus> then put those in the document
206 2014-07-10 05:37:30 <WeCluster> if people are interested in building using mingw WITHOUT msys then that would be another matter and would probably justify keeping a separate build-msw.md
207 2014-07-10 05:37:53 <wumpus> no, people are interested in having one build guide for windows that works :)
208 2014-07-10 05:38:00 <wumpus> if that's mingw with msys that's great
209 2014-07-10 05:38:52 <wumpus> ... and as I said two times already, build-msw.md references MSYS in every step, why do you think it is about building without MSYS?
210 2014-07-10 05:39:27 <wumpus> please at least read a document before criticizing it
211 2014-07-10 05:39:44 <WeCluster> we will be attempting to build with visual studio 2013 next, and may recommend modifications to build-unix that incorporate all of the mingw/msys knowledge and steps, while revisions to build-msw would become a how-to for visual c++ and/or mingw/win32
212 2014-07-10 05:40:13 <wumpus> modifications for windows building belong in build-msw.md, not not in build-unix.md
213 2014-07-10 05:40:18 <wumpus> build-unix is about unix only
214 2014-07-10 05:40:48 <WeCluster> agreed, the criticism of build-msw was a criticism of the forum instructions for building the dependencies with .bat files and command prompt environment settings. misplaced blame.
215 2014-07-10 05:41:22 <WeCluster> but the build-msw steps that reference msw (which is, as you noted, all of the steps) weren't as useful as the build-unix instructions during our build effort.
216 2014-07-10 05:41:53 <wumpus> sigh, build-msw stands for Build MS Windows... it does not refer to msvc at all
217 2014-07-10 05:41:58 <WeCluster> (oops, meant to write "msys" not "msw" above)
218 2014-07-10 05:42:06 <wumpus> it is a  guide for building on windows by any means necessary
219 2014-07-10 05:42:12 <wumpus> I don' t even care what compiler is used
220 2014-07-10 05:42:28 <wumpus> build-unix however is about building on unix and should stay about building on unix
221 2014-07-10 05:42:31 <WeCluster> building under msys isn't exactly building under windows.
222 2014-07-10 05:42:44 <wumpus> strictly it is - msys runs on windows
223 2014-07-10 05:43:51 <WeCluster> if you like it the way it is, that's okay. what we'd like to see is build-msw giving steps for msvc and mingw without msys, and for the unix build instructions to incorporate those msys steps.
224 2014-07-10 05:43:59 <wumpus> I'd agree with you if you used cygwin or another full unix wrapper on windows, but msys is only a very thin shell and not a unix-on-windows simulator
225 2014-07-10 05:44:30 <WeCluster> this would be consistent with how boost gives the same instructions: http://www.boost.org/doc/libs/1_55_0/more/getting_started/windows.html
226 2014-07-10 05:44:41 <WeCluster> A note to Cygwin and MinGW users  If you plan to use your tools from the Windows command prompt, you're in the right place. If you plan to build from the Cygwin bash shell, you're actually running on a POSIX platform and should follow the instructions for getting started on Unix variants. Other command shells, such as MinGW's MSYS, are not supported—they may or may not work.
227 2014-07-10 05:45:11 <wumpus> I do like it seperated per platform
228 2014-07-10 05:45:40 <WeCluster> if build-msw is strictly "from the command prompt" or "from visual studio" that would make more sense
229 2014-07-10 05:45:57 <wumpus> well visual studio building is not something we can recommend
230 2014-07-10 05:45:59 <WeCluster> MSYS steps are nearly identical to building under Unix/Linux
231 2014-07-10 05:46:16 <wumpus> and just using the command prompt is needless masochism, msys was designed to make this easier
232 2014-07-10 05:46:34 <WeCluster> haha we could rename build-msw build-masochism
233 2014-07-10 05:46:36 <wumpus> I don't care about ideological purity here it should just work
234 2014-07-10 05:46:46 <WeCluster> what about build-eclipse ?
235 2014-07-10 05:47:08 <WeCluster> build-msw could include both visual studio 2013 and eclipse
236 2014-07-10 05:47:15 <wumpus> it could
237 2014-07-10 05:47:46 <wumpus> but I prefer one solution that works and is maintained to two that don't work or are out of date in one version :)
238 2014-07-10 05:48:16 <wumpus> so instead of build-*every ide or variant*.md we just have one per platform
239 2014-07-10 05:49:47 <wumpus> whether that uses msvc, mingw, msys, eclipse, code::blocks, or turbo C 2.0 I don't mind
240 2014-07-10 05:50:56 <wumpus> and the official builds will continue to be done using gitian
241 2014-07-10 05:52:36 <wumpus> which uses mingw, as that's the only way to do deterministic cross-builds from linux
242 2014-07-10 06:13:47 <wumpus> Looking for code review of https://github.com/bitcoin/bitcoin/pull/4400 ; it's just a trivial code move, but as it affects the random number generation I'd like some people to check just for sure
243 2014-07-10 06:55:30 <wumpus> Luke-Jr: do you have any scripts to test longpolling? I'd like to work on #2834 to not use boost primitives directly but use our own threading ones - but it'd be nice to have something to test the functionality with