freghar

Just my quote few posts above. And the number N is getting bigger, you're ~27th person who I replied on this paticular question.

another one - http://gitready.com/

use SEARCH

He probably meant _real_ daemon, not just script handler. Well, I'd add that to the core once it's stable. Because a server, which is able to crash after few hours (days) without warning isn't a good candidate for daemon. Restarting such server would cause more pain than it's necessary. Not to mention that you can actually write a small standalone prog to daemonize mangos.

I heard TortoiseGit is now a lot more usable in v0.2 .. including patch creation. Otherwise .. omit the new repo creation + pull, you can use just local branch, no need for svn-style.

Are you sure? He fixed all 4 matches in current master # grep Ineract src/ -R -n src/game/Creature.h:431: bool isCanIneractWithBattleMaster(Player* player, bool msg) const; src/game/Creature.cpp:653:bool Creature::isCanIneractWithBattleMaster(Player* pPlayer, bool msg) const src/game/Creature.cpp:758: if(!isCanIneractWithBattleMaster(pPlayer,false)) src/game/Creature.cpp:806: isCanIneractWithBattleMaster(pPlayer,true); // output error message if need

I would guess doing a separate branch in git would be better .. doing a "git pull --ff balrok the_branch" or "git fetch -n balrok && git cherry-pick balrok/the_branch" is easier than downloading a patch file, applying it, writing a commit message, ...

You mean the diff? Just click the commit on github, gitk or use console "git log -1 -p <ref>" command.

Instead of sleep()ing I suggest using `expect' during login. It comes from tcl, but has support in bash as well: reply="blah" expect -c " spawn what_to_launch expect { string_to_expect: { send \\"${reply}\\n\\"; exp_continue } } exit" you can use wildcards in string_to_expect, such as string_to_* to match string_to_expect , string_to_something, string_to_somethingelse, etc. starttelnet.sh could be replaced with a simple alias or possibly even deleted. btw .. I never used MaNGOS RA, if it just prints nothing during login, sleep()ing is a pretty good way, but a one echo for both name and pass should be enough.

The same way as mangos-0.12, read 303 announcement or any of the N threads where it has been answered.

Generally, the easiest way is to use brain. You might eventually come to a conclusion that somebody actually had to ask this question in the past. So you use the seach button and type "patch apply" in that textbox and then you press the "search now" button and it finds (what a surprise!) several topics with such question already answered.

I agree, HG lacks some advanced abilities of Git, but it's a nice and easy-to-use "expansion" to the SVN.

I don't know how much this thread has been moderated or so, but I have something to add; Mangos uses so-called rolling updates. That's a development style. If you want some references, see Gentoo's latest (well, "latest") news, it isn't something bad or so, it just means that nobody can get a tested, fixed and fully stable release without regressions. But it's faster, the code gets tested faster and such ... it has nothing to do with (im)possibility of learning "how to code". If I'm ever going to have my own 112 public-wide repo, I'll probably use the Linux-kernel-similar style of development (with "stable" releases, release candidates, ...), because of the static client.

I just would like to recommend you a gtkmm (http://en.wikipedia.org/wiki/Gtkmm , http://www.gtkmm.org/docs/gtkmm-2.4/docs/) API for the GUI. It's a good start.

Simply read git-remote's manpage: git remote prune upstream

Hello, so you already searched around a bit? Good, you passed my personal IQ test! Doesn't make sense for me, perhaps you should describe what you mean in another way. "git clone some_remote_url local_dir" 1) git clone initializes a new local git repository in local_dir/.git/ and download all the object files (in a pack) to local_dir/.git/objects/ directory 2) then it creates ("checkouts") the working tree (= all the files you normally see, README, src/, sql/, ...) outside the .git/ dir (ie. to local_dir/), using the object files (which stores, in fact, almost everything in the repo) "git checkout --track -b my_branch origin/mangos-0.12" 1) it creates the my_branch branch (= a 41-byte txt file in .git/heads/my_branch), using the topmost commit from origin/mangos-0.12 remote branch 2) then it sets the tracking (see my "Chapter 2" for more about it) in .git/config 3) finally, it checkouts the whole working tree from my_branch, nothing 303-related left in there Thanks to that tracking, you can use "git pull" from inside the my_branch to fetch + merge new changes from mangos-0.12. If it prints an error like "failed to set tracking", then you have to use "git pull origin mangos-0.12" in there. Also note that git clone downloads everything, ie. you already have origin/mangos-0.12 in your local repo when you clone. And the compiling process should be the same as it used to be on 2.4.3. footnote: I'm closing this topic, if anyone has a new error (check for typos in command before saying that), feel free to post it in another topic. I hope there won't be any "HOW TO GET 2.4.3????" topics anymore. They will get locked, because nobody can be so stupid, it has been answered about 16 times, only from what I've seen.

So at first, read the stickies in the section you posted the topic in (can I call this an IQ test?). At second, there's that not-so-shiny search button on the menu-or-whatever bar. Good luck.

Well unless you have 50MB of disk space, you can always compile your own, I use a statically-linked executables of git 1.6.1 on a very old centos box (with old glibc) ..

Please take your time and read this short thread before posting any useless posts, thanks.

So perhaps you really don't have a mangos-0.12 ref. Try "git fetch origin" before that. and git for-each-ref refs/remotes/origin

Try git branch 243 --track origin/mangos-0.12 git checkout 243 This should work for Git version 1.4.x and lower (or at least I used it this way then).

If you have enough RAM, you can convert your DBs to InnoDB and tune it properly, including caches. From what I know, the DB slowdown can be caused by writing player saves as ~400-500KB data blocks per user. Wyk3d was working on some binary implementation of this, maybe it could result into splitting the large block to separate columns (at least to some point) later.

There was something called "writepdump" on mangos console. Not sure if that works, but it's probably better than exporting the data manually and rewriting item GUIDs later ..

Actually, IIRC the doxygen documentation is in the doc/ directory in the repo for quite some time ...

Just a small idea - you can change the packet display to something more readable .. For example a piece of tcpdump - http://paste2.org/p/124235 .. and I know about .. 5 opcode sniffers, so perhaps devs already have some, but it's always good to have a new one. edit: err, forgot to add - you can search for that formatting mechanism in tcpdump sources (from where some other projects took it), or perhaps dsniff, I remember such code there as well.