I want to join

[sirherrbatka]

libbullet 2.80 (or 2.77 but nothing in between (bug in bullet))
libogre 1.8rc1
libmygui 3.2.0

Wiki is outdated (as usual).

[ezzetabi]

@ezzetabi: I also prefer Linux to Windows, but even with Wine, a fair amount of games does not work reasonably well. Granted, it has been years since I tried it out... so many things, so little time...

I understand, do not worry.

Changing topic: Sorry if I insist. But you still need web space as stated at the end of your movie presentation? Are you interested in a shell account in the virtual machine of mine? ``Not really...'' is a perfectly valid answer, just I would like to know.

[mark76]

@sirherrbatka: Thanks for the update on the versions required for building OpenMW. I am just waiting for finishing the DirectX SDK download, required to build OIS, required to build Ogre 1.8.0rc1, only 15 hours left... quite a lot of work for one stupid missing library (dxguid.lib). The server over at MS is so terribly slow right now . And I already had to switch to Visual Studio 2010 (Express), as 2008 is no longer able to build OpenMW (it seems to use some methods introduced in the latest revision of the language standard - C++0x - and the 2008 version does not support this).

Anyway, I have already looked into the OpenMW source code, and I do have a roadmap for adding multiple esX support (which I'll post soon in the task description). Fortunately, the existing code looks quite reusable, and the first steps should be finally ready in no time... when this stupid download finally finishes. I wish I could have better news right now . Anyway, at least some of the stuff should be ready for 0.15.

@ezzetabi: Thanks for your offer of web space. I might actually accept it, but right now, I just have not enough time to to pursue my "move homepage" task further. Waiting for GB-sized downloads @10kb/s has consumed too much spare time in the past weeks (I don't have a good web access during the week, only during the weekends - just when the other server's bandwidth is loaded out.)

So, in a nutshell, could we try again this in a month or two? (If you find a different use for your web space in the meantime, I can live with it.)

[Zini]

And I already had to switch to Visual Studio 2010 (Express), as 2008 is no longer able to build OpenMW (it seems to use some methods introduced in the latest revision of the language standard - C++0x - and the 2008 version does not support this).

This would be rather surprising, because we do not use C++11 yet. My compiler is pretty old too (gcc 4.4.3) and it does not have any problems with the current master branch. Could you give me a few pointers where MSVC 2008 has problems with our code base?

[mark76]

Finally... I got everything set up and actually working. I can leave the starter cave and enter the outside world. What a beautiful water shader.

Anyway. Now it is time to start hacking and I already got two questions:

1) Is there some way to hack a starting position in the outside world, so I don't have to move all way to Solstheim from the starter cave each time I want to test landscape merging?

2) How is pushing a commit supposed to work? Do I make an account and push myself, or do I submit a patch which is then pushed by somebody else?

@Zini: After switching to VC++10, all things that didn't work mysteriously vanished (not too bad per se), but I managed to reconstruct the following example. In openal_output.cpp, there is the following snippet:

Code: Select all

void OpenAL_SoundStream::play()
{
    std::vector<char> data(mBufferSize);

    // ... snip ...
    for(ALuint i = 0;i < sNumBuffers;i++)
    {
        size_t got;

        //
        // the following line results in an "undefined symbol: data.data()" error!
        //
        got = mDecoder->read(data.data(), data.size());

        alBufferData(mBuffers[i], mFormat, data.data(), got, mSampleRate);
    }
    // ... snip ...
}

I did have a look into the "vector" header included with VS9, and this method is indeed missing. Some googling turned up that this seems to be a new feature added in the C++0x draft, and that Microsoft only starts supporting this feature set with VS10. Somewhere in the "vector" header shipped with VS10, there is the following snippet, which supports my interpretation that VS10 is indeed required:

Code: Select all

 #if _HAS_CPP0X
	pointer data()
		{	// return address of first element
		return (this->_Myfirst);
		}

	const_pointer data() const
		{	// return address of first element
		return (this->_Myfirst);
		}
 #endif /* _HAS_CPP0X */

Now, as to why this hasn't turned up before... I have no idea. Concerning the GCC, there is supposed to be another snipped somewhere in it's libraries,

Code: Select all

  // _GLIBCXX_RESOLVE_LIB_DEFECTS
  // DR 464. Suggestion for new member functions in standard containers.
  // data access

but I have to admit I haven't looked into it further. C++ standard history isn't quite my strong side, so I don't know how long this has been floating around. But on the other hand, GCC is known to adopt new additions to the standard quite fast, while MS is known for the opposite (i.e. it must becone an official standard before it is even considered).

Anyway, using VS10 magically fixed everything. (If you want a more detailed list of what does not work, you'll need to wait for the next weekend - I have a VS9 setup on another computer I don't have access to right now.)

[ezzetabi]

1) Is there some way to hack a starting position in the outside world, so I don't have to move all way to Solstheim from the starter cave each time I want to test landscape merging?

coc "Thirsk" in the console might help?

[Zini]

You commit your changes to your local repository and push them to your repository on github and when you are done with a branch, you send a pull request to me. I think there is a checklist on the wiki that describes the initial setup and github has also tons more of documentation on the process.

[sirherrbatka]

yes you can use coc in console and that's what we are doing all the time.

Is the setpos command, actually working?

[Zini]

Nope. Not implemented yet, because we can't move objects around (except for the player). We first need to implement cell-border crossing properly.

[Okulo]

We first need to implement cell-border crossing properly.

Is that why the game screeches to a halt and we have to wait for more cells to be loaded when we enter a new cell?