Hi,
I'm trying to set up a dev environment on Windows 10 so I can build OpenMW (and then, if I can find an issue simple enough, actually contribute to the project while I have free time over Christmas). However, I'm at a point where it looks to me like I should be able to build the latest master-branch revision, but Visual Studio spits out a bunch of errors, both due to linking and code it can't compile in the first place. I've never directly used tools like Git or CMake, or, in fact, worked on any C/C++ code with external dependencies before, so I've been prodding at this for a few days now.
A list of what I've already done to get to where I am:
-
Spoiler: Show
- Installed Visual Studio 2015 and 2013, and got them working for basic projects.
Installed MSYSGit (via VS 2015's installer)
Installed CMake
Start of things that shouldn't have made any difference in the end.
'Installed' Boost and Bullet with unknown levels of success using guesswork and this: https://wiki.openmw.org/index.php?title ... #Libraries
Followed this guide as best I could, in the hope it would make me familiar with the tools: https://gist.github.com/ace13/a25557c096c80a37ddcb
Where links were dead, I used web archived versions. Even though OSG is now used, I still got Ogre3D, as the instructions were easier to follow that way (the MyGUI ones referred back to those for Ogre). I had issues with Qt, as I ended up with a 32-bit only variant at one point, but I'm pretty sure I put that right.
Installed and (mostly) built Scrawl's fork of OSG. I had issues at first, but eventually everything was fixed except the DAE plugin, which didn't like an implicit type cast from double to value_type in lines 44 and 159 of daeWTransforms.cpp, so wouldn't compile.
Tried to adapt ace13's guide for configuring CMake for OpenMW itself to work for OSG instead of OGRE, but even after pointing every thing I could think of at what I perceived to be the right place, it still didn't detect OSG properly, and I couldn't generate VS solutions.
End of things that shouldn't have made any difference in the end
After that had taken up a few evenings, and not given me a build environment, I decided to create a new folder and take the 'cheat' route and use this: https://gist.github.com/ace13/de7f30454e99b3952405
I successfully cloned Ace's version of the repo, checked out the right branch, and started the script, which ran and broke, as I'd not actually got CMake on the path.
Having fixed that, I ran it again, and it complained that QMake wasn't listing resources that actually existed.
I guessed that it was supposed to run QtBinPatcher and wasn't for whatever reason, so I added a pause to the script and ran the patcher manually, which made CMake succeed.
I attempted to clone my (up-to-date) fork of the official repo over the top of Ace's so I could actually get something done at some point, and accidentally put it in a subfolder, which wasn't so bad, as it turns out I don't know enough about how to use git to swap out a local copy of one repo for another.
I finally opened OpenMW in Visual Studio 2013, and built the BUILD_ALL project, and everything failed horribly with more errors than would be reasonable to post in a forum without someone asking me to, even in a spoiler tag.
There might be a really, really simple thing I have to do to fix this, even if it requires deleting things and starting again, but I have no idea what's actually broken, so can't make any more headway with fixing this myself.
I
have received all the C/C++ teaching Cambridge University offers its Computer Science students (which is surprisingly little, but 'enough'), so I
should be able to find something I can do once I have a codebase that can just be built and run, but unfortunately I'm shooting in the dark when it comes to anything that isn't actually the code itself.