Topic: [COLLECTED] 100 BTC Bounty For Windows Build Instructions (Read 18229 times)

In case anyone is still interested, I posted a simple guide about building the latest bitcoind on windows some time ago: https://bitcointalksearch.org/topic/building-headless-bitcoin-and-bitcoin-qt-on-windows-149479 

Don't try to use such an old Boost version, you can use 1.53 if you want and AFAIK the official client is using 1.50 currently.

Dia

after copying C:\boost_1_43_0-mgw\boost to C:\bitcoin-nftf-master\src

Building the dependencies with MinGW takes an eternity.  Not sure where the blame lies (as if it matters), but it is really painful.

Not only that, but they are also huge in terms of file size.  After compiling, they are nearly 4 GB.  Yes, disk space is cheap, but it's definitely not all necessary.

I think I'll put together a set of minimal binaries for the dependencies for Windows + MinGW.  No reason to go through all that unnecessarily.


As for missing .h files, double check the path names in the include path (-I on the command line) and double check that they exactly match the names of the directories on your file system.  Watch for subtle differences like dots instead of underscores or dashes.

Welcome to Linux development.  And people wonder why it isn't more widely adopted.

Say what you want about Microsoft, the lower barrier to entry has alot to do with its continued wide use.

I ran it, and it found my real-life wallet. I tried sending a tip to Matt at 1JBMattRztKDF2KRS3vhjJXA7h47NEsn2c, and the UI froze up. Oh well! At least I can compile SOMETHING now . . .

So lets say I want to try running linux in a virtual machine just for compiling bitcoin. What flavor of linux should I install? I've had passing exposure to Ubuntu . . .

Can anyone recommend which virtual machine software would be best for running a linux image? I would probably try VMWare first unless someone knows better.

Can anybody post the super-easy steps to downloading/installing/compiling everything to build bitcoin on such a linux virtual machine?

Should I be thinking about cygwin instead of a virtual machine? I've had passing exposure to that too.

I haven't tried running all these steps on a brand new XP-mode virtual machine inside Win7. If anybody has time to try that, I would be interested in whether that works.

Great guide!  TY!

The makefile now appears to look for C:\openssl-1.0.0d-mgw instead of C:\openssl-1.0.0c-mgw and it was failing on that.  But just a minor speed bump.

Again, thanks!

net.cpp:23:32: fatal error: miniupnpc/miniwget.h: No such file or directory

- If you have never used command-line compile tools before, prepare to be baffled and befuddled. This is not a nice clean GUI where you hit the compile button, wait a few seconds, and go try your executable. Thankfully, I've used gcc before, so I wasn't completely out of my element

- Expect the whole process from learning to downloading to compiling to troubleshooting, to take HOURS. I'm not kidding. These makefiles build THOUSANDS of object files, of which bitcoin probably uses a tiny percentage. Boost is the worst of them all. Just start the make and let it run overnight.

- Before running the steps outlined in the earlier post, I recommend you get yourself a github account, and follow their instructions to download and install "git" and set up your own sandbox (fork) of the bitcoin source. This will involve yet even more baffling command lines that need to be copied and run exactly as written on the github guide to getting started

- Do not watch the scroll of status messages going by during the eternity it takes to compile everything. You will see countless frightening-sounding warnings which will only confuse you
 - Do not expect a nice little "all done, everything went ok!" message at the end of a long compilation. The nice ones just stop printing status messages, and you are supposed to know from the acres of text above that nothing bad happened. The others are less nice . . .
 - Boost finally finished compiling with this very comforting message: Great.

- At the end of the make for openssl, the linker returned a very fatal-sounding I wasted some time trying to "make clean" and try "make" again, but eventually I went on to the next step, and I still got the expected libeay32.dll without any complaints.
 - Finally, I tried compiling bitcoin (ha!) - as you might expect, it wasn't happy. The error message was:
I found buffer.h in the includes directory for openssl, so I copied that directory out and renamed it "openssl" which I then put everywhere I could think of. No luck.

Linux compile tools let you stand on the shoulders of giants, but the giants all have Asperger's.

Put it in more places?

In the file trying to include buffer,h, change the include to have teh exact path to buffer.h? :?

Welcome to Linux development.  And people wonder why it isn't more widely adopted.

Say what you want about Microsoft, the lower barrier to entry has alot to do with its continued wide use.

My notes from TRYING to do this:

 - If you have never used command-line compile tools before, prepare to be baffled and befuddled. This is not a nice clean GUI where you hit the compile button, wait a few seconds, and go try your executable. Thankfully, I've used gcc before, so I wasn't completely out of my element
 - Expect the whole process from learning to downloading to compiling to troubleshooting, to take HOURS. I'm not kidding. These makefiles build THOUSANDS of object files, of which bitcoin probably uses a tiny percentage. Boost is the worst of them all. Just start the make and let it run overnight.
 - Before running the steps outlined in the earlier post, I recommend you get yourself a github account, and follow their instructions to download and install "git" and set up your own sandbox (fork) of the bitcoin source. This will involve yet even more baffling command lines that need to be copied and run exactly as written on the github guide to getting started
 - Do not watch the scroll of status messages going by during the eternity it takes to compile everything. You will see countless frightening-sounding warnings which will only confuse you
 - Do not expect a nice little "all done, everything went ok!" message at the end of a long compilation. The nice ones just stop printing status messages, and you are supposed to know from the acres of text above that nothing bad happened. The others are less nice . . .

...

Linux compile tools let you stand on the shoulders of giants, but the giants all have Asperger's.

Any help would be appreciated.