[Bug 753513] Review Request: minetest - Multiplayer infinite-world block sandbox with survival mode

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=753513

--- Comment #8 from Aleksandra Bookwar <alpha at bookwar.info> 2011-12-02 20:58:43 EST ---
(In reply to comment #6)

There are two ways of doing things:
1) start the service, then run the client, set Server field to 'localhost' and
connect. Then minetest will use the existing server.

2) stop the server, run the client only, leave the Server field blank, and
Connect. Then client will run its own minetestserver instance on the same port
30000.

If you run both the server and the client, and then leave the Server field
blank, then minetest tries to create its own instance. It fails and gives the
following error:
=========================================
INFO: Full run of init_mapnode with g_texturesource!=NULL
AuthManager: loading from /home/bookwar/.minetest/world/auth.txt
BanManager: loading from /home/bookwar/.minetest/world/ipban.txt
6: Bind failed: Address already in use

In thread 7f8723576700:
/builddir/build/BUILD/celeron55-minetest-bc0e5c0/src/connection.cpp:577:
virtual void* con::Connection::Thread(): Assertion '0' failed.
Debug stacks:
DEBUG STACK FOR THREAD 7f8722d75700:
#0  virtual void* ServerThread::Thread()
#1  void Server::Receive()
(Leftover data: #2  void Server::SendBlocks(float))
DEBUG STACK FOR THREAD 7f872e683740:
#0  int main(int, char**)
(Leftover data: #1  void Server::start(short unsigned int))
Aborted
=========================================

This is not the segfault, just normal error. Here the "6: Bind failed: Address
already in use" is the main error string, and last is the debug output of the
second thread.

So this is not really the internal bug, just interface design flaw, which is
handled in a rather bad way though. I filed the corresponding issue here
https://github.com/celeron55/minetest/issues/27

And I will add README.fedora file to the docs/ directory with detailed
explanation of the server behavior.

> However, on installation of minetest and minetest server, my X session
> restarted.
>
> (...)
>
> On running minetest, I got the connection screen, clicked connect, and it
> segfaulted.
>
> (...)
>
> *** glibc detected *** /usr/bin/minetest: corrupted double-linked list:
> 0x0000000000fb2e90 ***
> 
> Program received signal SIGABRT, Aborted.
> [Switching to Thread 0x7fffee985700 (LWP 16958)]

Regarding these X-restarting and segfaults, minetest package can't be the
reason for such errors. My guess is that this is the glibc issue. Check for
example https://bugzilla.redhat.com/show_bug.cgi?id=754434 I stepped into this
recently with glibc 2.14.90-16 and yum, and had to update glibc running yum in
valgrind environment.

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.