freghar

You probably don't have "fresh source", see git status.

There's more than just this condition, I already have such dismounts (bootybay, ships, ..) removed in my custom repo and it has pretty much nothing to do with 3.0.3 .. Or has Blizz removed the dismount ?

http://xkcd.com/138/

As for Mangos-related code to begin with, try src/game/* , perhaps SD2 guys can tell you where to begin with scripts.PS: nice to see that at least someone remembers the original meaning of that word

thyphoon: without the @ .. it's a standard URI - http://en.wikipedia.org/wiki/URI protocol://username:Password@server_address:Port/path/to/file.txt it's minimal form is a protocol+server_address. laflare: use the shiny search button at the top of this page TheLuda: Make the search button blink, please.

There are few displayIds, like ships, small houses, etc... but not the big ones like stormind cathedral and objects that are tied to some specific environment ("houses" in OG / IF), ...

Whether the multiprocessing or multithreading is slower, that depends on style of communication. For example multiprocessed Apache web server - it uses many forks, each one for one connection, it handles it separately, without much communication with the parent server, so the performance it's pretty good. However if the map server would need to exchange a lot of data with other processes, it would be pain. Passing structures to threads is simplier (and in this case faster) rather than sending them through UNIX sockets, TCP/IP conn, SIGPIPE, ... ... and multipocessing isn't hard to use even for noobs, if they have the right tutorial ..

Yes, such diff can be applied with git-apply or using GNU patch utility for example.

Please, search around. Plenty of such stuff here.

So you want a commit (state) of the repository at 2008/10/31 13:00 hrs ? Well, open up gitk and scroll down till you see that date on the right side, copy the hash (or write somewhere first 7 characters) and do git reset --hard abcde12 (replace with that hash).

You obviously haven't been near the Linux community since ~1993 .. Slackware was the first GNU/Linux distribution can I remember. Led by Patrick Volkerding and his wife for all those years, "Debian" was something in the far future. Well I can't recommend Slackware for any of you, unless you want to build a super-stable router, database server or similar thing. It is also very useful when you want to learn something about the system (and kernel) internals. It's the same as with CentOS - those are really stable distros, but with VERY OLD software, which is the cost of being stable. I recommend you using something like Debian testing (Ubuntu/Mandriva/Fedora/wtf for newbies), perhaps Archlinux (sometimes _really_ "unstable") or Gentoo. GCC 4 (which isn't probably in Slack and CentOS) really made things faster by another optimizations, newer glibc, mysql, etc.

Perhaps nowhere, since 3.0.3 still isn't supported. 303-willcrashforsure: # grep -i 'spell power' src/game/* src/game/ItemPrototype.h: if(ScalingStatValue & 0x00008000) // spell power

You probably selected a "bare repository" when cloning, ie. repo without a working tree (= without source files and such things). Such repository contain all you need, just set "bare = true" to false in config and do that hard reset. It's possible to "convert" a bare repo to normal one this way. Note that Git recognizes a bare repo by the name of your git data directory, the "bare = true" is just to disable commands like git-checkout, ... That means if your data dir is called ".git", it's normal repo and source files should be outside it, with any other name, it's a bare repo. Bare repositories are used for hosting (where you just push & pull, no need for working tree (at least in Git)) for example.

your .git/objects/pack/ directory? Go to your working tree (outside .git/) and do a "git reset --hard".

It hasn't been changed - hourly.

Do a "git clean -d -f" after that reset to remove uncommited, newly added files and directories.

An old joke? That's the reason of this post:D Sorry, no Vista here to try out ..

1) There are tons of examples, use search 2) You need to understand path specification (like ".." is one directory back, "directory1/directory2/file.txt" and so on ... Windows Explorer uses silimar style, with backslashes however. If you get this, there's pretty no way to don't understand what to do.

First, do a fetch (download new objects from a remote repository), then simply view the remote branch. For example: git fetch origin git log -3 origin/master to view latest 3 commits.

I develop on a clean mangos, without any full DB .. and the world server starts in ~3 seconds, it shows some errors though, so those gates aren't probably in the base map. [Gameobject template 179830 not found in database! BattleGround not created! BatteGroundWS: Failed to spawn some object BattleGround not created! [Gameobject template 183978 not found in database! BattleGround not created! BatteGroundNA: Failed to spawn some object! [Gameobject template 180087 not found in database! BattleGround not created! BatteGroundAB: Failed to spawn some object BattleGround not created! [Gameobject template 183971 not found in database! BattleGround not created! BatteGroundBE: Failed to spawn some object! [Gameobject template 184719 not found in database! BattleGround not created! BatteGroundEY: Failed to spawn some object BattleGround not created! [Gameobject template 185918 not found in database! BattleGround not created! BatteGroundRL: Failed to spawn some object! IIRC they're even called by gameobject ID from the BattleGround*.cpp ...

I'd say "RTFM", unfortunatelly, there's no official Mangos manual. Well the start was hard for me as well, but as I searched the thing around, starting from custom commands, through packet handling "hacks", ..., I got the basic idea of what is where and how to do it. ontopic: Most of the spells is from DBC, only spell effects are generally handled (AFAIK).

Things are pretty different than you might imagine, even lot of Linux kernel developers haven't much idea of how the memory management is handled in there and I'm not expert in it either, but know a few bits 1) You can limit number of players and so on, generally, it isn't a good idea to set a strict memory limit for a process, because after exceeding that limit, mmap() and other functions will return error and the whole process segfaults (in most cases, I really doubt MaNGOS has some protection for that), however you can set it with ulimit, see "ulimit -m". 2) Yes, it is .. and it's done that by default (MS Windows IIRC doesn't support this), on every read(). Cached files gets deleted whenever the memory is needed for some process. If you want to make a storage in ram, use ramdisk (pure kernel one or simple tmpfs, disadvantage for kernel ramdisk is that it's size can be set only at boot, tmpfs may, on the other hand, get swapped, but it's more easy to use). 3) "free" , "top", "cat /proc/meminfo", ... The thing is that you don't want memory to be free. If you don't need all the space, you want the disk cache in it (which is a very nice feature in fact), so you might see something like "1G used, 80MB free, 6G cache", which is good in fact, you have ~6G of "free" memory (which can be used when needed). Another pretty cool thing for running multiple almost identical servers is virtual memory page sharing, google/wiki for it if you want PS: Google for keywords from /proc/sys/vm/

I really doubt someone will lead you here step-by-step on every simple thing. I'm also interested in the spell system, since I don't know too much about it (spell_affect table connection to c++ code for example), but for example armor can be easily found out in the source, so search in src/game/* ...

So it works this way as well? I thought you have to specify optional refspec like "<remote> <refspec>" without the slash ... (even if your previous impconfig/impconfig had downloaded only impconfig branch, this fetch will download everything:)) Making it this way isn't the best. Yes, you fetch also origin, but you do only one rebase, so your branch isn't updated from origin, but from impconfig (thus depending on it's author to merge mangos/master regullary, which don't have to be necessarily true as merge compares two tree objects). So (imho) the better way when updating things would be to hard reset the current branch and then re-applying the rest with rebase. Like (I suppose you don't have your custom changes on current branch) git reset --hard origin/master (hard-reset your branch to the remote master) git pull --rebase impconfig impconfig (fetch + rebase the remote one into your current) Good luck.

And by Google I guess 'ON DUPLICATE KEY' is also mysql-only ...