MC-Server on Raspberry Pi

[DJburner]

Hello,

I am trying to compile MC-Server on my Raspberry Pi (http://www.raspberrypi.org/).
I downloaded the source and made "make".
After a few minutes there is the error:"make: *** [build/debug/source/Caves.o] Error 1"

Whats Wrong?

DJburner

[xoft]

Hi,
welcome to the forum.
Some more info would come in handy. Since there is the Caves source, I suppose you're building from SVN? What revision? Does this happen with rev 698, too?
Also, the line you're mentioning is the last one in the process, but there should be an error line somewhere above it, with more specific information, such as the file and line number. If you run "make" again, it should print much less garbage and more about the error.
Alternatively, route the make's output into a file and attach the file to this forum thread.

I'm gonna warn you in advance - MCServer can run on the Pi, but it's incredibly slow there. People have reported that it takes 15 minutes just to generate the initial spawn area, and another 15 minutes for the player to join and download the world.

I should be getting my RasPi shortly, so I might look into that, but it's not probably gonna help too much - the Pi is just too slow and too limited to run any Minecraft server.

[DJburner]

Hmm,
Tomorrow, i will make it again.

And the pi isnt soo slow, luacraft server is running fine and fast on it, with flatland.

And Why is the last compiled revision 300? Not everybody will compile it.

DJburner

[xoft]

Flatland takes a big fat zero to generate, unlike our terrain, which we try to make as close to the vanilla server as possible. Of course you can set your world.ini to something that produces flatland, that should speed things up.
Still, the MC protocol (and all supported storage schemas, too) needs zlib compression and that takes some time, and that's probably gonna be the next bottleneck for the Pi. And if not that, then lighting will.

About rev 300, ask FakeTruth, really, he's the homepage maintainer And since I've set up Windows nightbuilds on my web, no-one really complains anymore

[DJburner]

Okey thanks for your answers

Should i forget it with mc-server on my pi?

On my root server i am using bukkit (because commandbook, worldedit and the other great plugins) but maybe i try mcserver also there

DJburner

[xoft]

By all means try it out, we're interested in more experience, especially on the Pi. Maybe try playing with world.ini so that the server can run nicely.

[DJburner]

okey, i build it again.



/usr/include/c++/4.6/bits/stl_list.h:1160:7: note: std::list<_Tp, _Alloc>::iterator std::list<_Tp, _Alloc>::erase(std::list<_Tp, _Alloc>::iterator, std::list<_Tp, _Alloc>::iterator) [with _Tp = cStructGenWormNestCaves::cCaveSystem*, _Alloc = std::allocator<cStructGenWormNestCaves::cCaveSystem*>, std::list<_Tp, _Alloc>::iterator = std::_List_iterator<cStructGenWormNestCaves::cCaveSystem*>]
/usr/include/c++/4.6/bits/stl_list.h:1160:7: note: no known conversion for argument 1 from 'std::list<cStructGenWormNestCaves::cCaveSystem*>::const_iterator {aka std::_List_const_iterator<cStructGenWormNestCaves::cCaveSystem*>}' to 'std::_List_iterator<cStructGenWormNestCaves::cCaveSystem*>'
make: *** [build/release/source/Caves.o] Error 1

Whats the Problem?

DJburner

[xoft]

Still missing a part of the error message, especially the location in MCS code. But that seems to be a general issue when using gcc, so I'll try compiling on gcc in a virtualbox, it'll probably produce the same error.

The Caves' code is very fresh, it has been committed only a few days ago, and since most of us build using MSVC, we don't catch gcc error immediately. I'll try to fix that asap. Thanks for the report.

Also, have a look at the GNUmakefile header, you might want to build in release mode for the Pi ("make release=1" instead of simple "make")

[DJburner]

Imtried both, make and make release=1
Same error
Im new in the linux world so a raspberry pi is got to learn using linux

Which is the stabelest source revision for 1.2.5?

DJburner

[xoft]

Well, you wouldn't believe what the issue was.

It seems that MSVC comes with a more universal STL than GCC. MSVC's STL has "const_iterator std::list::erase(const iterator)", while GCC's STL doesn't. I had no idea, and when MSVC compiled the code happily, I didn't even check the standard

---

I'd say rev 700 or 701 will be stable enough; depends on in which one I'll fix the iterators