martysama0134

1) It's so """native""" that half of the ports will throw lots of errors when you compile them with clang. (default behavior) Native, in this case, means "installed by default", not "it makes instructions more clear blablabla". 2) For binaries compiled with gcc4.9, you should usually use "gdb-7.10_4" (the minor version could vary), otherwise you'll get the "dwarf instruction problem" error. (pretty normal) Mysql 5.6 is one of the most criticized versions. Since you like to be alternative, you can also try MariaDB (mysql fork). 3) As I stated before, one could even use both. Use clang's modernizr and then compile it with gcc. (which is pretty edgy) Kinda all ymir's code is made with raw new/delete operators. The only code wrapped in smart pointers is related to the queries (and not all the times). Server-side, most all the stuff are loaded once (regens, map attrs, etc), and there's no default way to free them without closing the process. (I would call them "cached" and not "memory leaking" anyway) Client-side, it's the same thing, except of the Outdoor loading, which clearly leaks few mb every time you warp somewhere. (on the WE, it's pretty clear) Note: If he installs gcc on freebsd 10.x, he will have to remap few things otherwise cryptopp will be compiled with clang, and the rest of the game with gcc. (which will end up in a linking failure)

Just an question. Why to use Depend? It's useless. If you edit a mapped (aka inside Depend) .h, all the .cpp files that include it will also be recompiled as it should be. If you delete it, you have to recompile the whole source code all the time you edit a .h file.

What is the output doing this? # cd your/path/Srcs/Server/game/src # gmake dep

# gmake dep clean default Read this and you will see. [Hidden Content] Vanilla yes. when she started to developing source she used gcc after changed to clang. ---------------- Xd If you really want evidence that clang is better than gcc I'll do a video. Live when upgrading MySQL. And you will see. : D Clang regornize better codes than gcc. With gcc code is not always stable. If you will not use dump.core clang of external causes. Clang is indeed a nice alternative of gcc, but nothing more. Talking about benchmarks between gcc and clang, they are either balanced (one is faster than the other on few instructions, and vice versa). I usually use both, since they are helpful in pointing out new errors.

FreeBSD 10.1's gcc4.9 pre-compiled package was incomplete (initially it was ok, but after few updates it got broken). I previously installed it from package on different FreeBSD versions successfully. Anyway, install it, and perform a check if it's ok: # pkg check -s <name_of_that_package> The package should be called "gcc49-blablabla". You'll get the real name from "pkg info". In here the explanation of "-s": DESCRIPTION pkg check -B or pkg check --shlibs is used to reanalyse shared libraries of installed packages. pkg check -d or pkg check --dependencies is used to check for and install missing dependencies. pkg check -r or pkg check --recompute is used to recompute sizes and checksums of installed packages. pkg check -s or pkg check --checksums is used to find invalid checksums for installed packages.

The recent gcc49 package is incomplete. (it may vary) All the FreeBSD's minor updates will always break something, so it's not always as easy as it should be ['cause you'll end up fixing FreeBSD's settings after updated]. Compiling from the ports it's somehow futile, and time-wasting. I'd do such a thing only if the pre-compiled package is not compatible. If you've installed gcc4.9 from pkg and it's broken/incomplete (most probably it will be), try to install gcc4.8 instead: ## 1st way # pkg search gcc48 gcc48-4.8.5 # pkg install gcc48-4.8.5 ## 2nd way # pkg install gcc48

Remember to copy the script from the "raw" button: Makefile (lil edit #2) OT: It looks like a trivial bad-coded way to detect "legit windows". He thinks you coded such a thing (no idea of where he took it from), which means people can look down on you.

Makefile (lil clean) (game/src/) version.cpp (lazy way) (if you don't mind having no ver.txt at every game start up) What does the compiler print after this? (whole text)

For the ones who are interested in proxing their skype contacts, (like for ddos prevention, or reduce the multiple skype connections to one) I've moved the guide to github (since my bbcode.txt doesn't work in here anymore), and added few other points (2.1 and 2.2).

That's the first time that someone would get such an error. (maybe bad python code) Does cython trigger any warnings when compiling the .py files?

#Updated (v24) removed the boring CTRL requirement (to move the camera) when editing daylight/attr fixed refresh texture imagebox when onKey-pressing the down/up keys (like when onClicking them) fixed TextureSet file creation if not existing fixed new wolfman motion event handling NEW_MAP_MAPNAME_PREFIX config option [set a default mapname prefix when creating new maps ("" to disable)] created new TextureSet field when creating new maps created new Change/Delete Texture buttons when double-clicking a texture

You can't "unpack" a cythonized root, since there are no ".py" files at all to extract, but pure compiled c++ code. You can disassemble the launcher with IDA, but it will be asm/pseudo-c code. Even though you can (vaguely) understand what those scripts could do, you won't be able in any ways to get the original files back.

Excluding bugs, it should be BD-free. (like the rest of the other branches) Previously, in 2012, the r2089 (not r2089m) game had the "temp gm privilege" backdoor. (even official servers, except gf servers, were affected by this)

Actually, the "private port vulnerability" is neither a bug nor a backdoor. (highly used by kiddies since 2010) That's how "game"s and "db" communicate with each others. (even though an authorized connection wouldn't be so bad to implement) Ymir clearly specified the necessity of protecting such ports behind firewall. (in the source files, you can find some ipfw rules in korean) There are no backdoors so far. (I'm not considering the /war or /cube issues since they are bugs, and it would be funny considering them as BD)

1. Could you also post the last syslog lines before the relative core crash? 2. Have you changed somehow the design and the settings/server_attr of the a1/b1/c1 maps?

.pyc files are compatible only with the same python version they were "compiled".tl;dr that lib folder will work only with python2.7, and not with python2.2.

You can proxy skype just using random proxies servers retrievable from [Hidden Content] similar sites.

Guide moved to GitHub! Old revision:

He replaced the ui.py with the one I sent to him, and told me "everything is now ok".

Try using the default Resource Monitor tool to verify that: NB: Since I'm using a proxy, Skype establishes all the connection to just that proxy.

Tried it several times in the past years, and right now. [grabbing an IP from a friend contact of yours with such option enabled]NB: I'm not referring to the mythical beta skype resolvers, which I care the less. It's also written in there: The people in your contact list will get your real IP anyway. If you're interested, I can make a proxy-how-to thread for skype. (via .reg obviously)

Kinda wrong: Everyone in your contact list could always retrieve your real IP any time they want. Even though they won't be able to resolve your skype contact through web-resolvers, they will easily be able to use netstat&co to figure it out. There's also a myth where "beta skype resolvers" exist, which they are able to bypass "Allow direct connections to your contacts only.". ["beta", in this case, means the first skype version that implemented such an option]

For uiscriptlib, you should use the ui.py file you can find in the ./Extra/root/ folder. (at least, the code you can find about __USE_EXTRA_CYTHON__)

You don't have mysql-server installed in your machine.