Lynx3d

I bet hiding this option a bit saves them tons of traffic from people that refuse to learn a few basic git commands...

You can install pre-build ACE and TBB (e.g. with apt-get etc., and don't forget development files), and configure with ACE_USE_EXTERNAL=1 and TBB_USE_EXTERNAL=1. PCH means precompiled headers, that's nothing you can install, it's just a compilation option. That should reduce build times considerably...

Can't tell you about mmaps, but for vmaps (it's version 3 actually, but don't ask me what v1 was...), it's more or less still a dump of the data structures, i.e. some section identifier ("magic bytes"), size to read and then just raw values. It just avoids pointers and padding bytes in contrast to the previous version. So yea, it pretty much means "read the code"...there's no standard associated to it, just reflects the actual memory content.

I'm not aware of any tbb concurrent hash map usage, just plain STL (well, still TR1 for most released compilers) hash maps. The main workload still is not multithreaded anyway, and there's only plans for map level parallelism, which won't interfere in writeable data much.

There's no direct way to do this. And i'm not aware of anything that has this behaviour retail, "world drops" are always limited to certain level ranges, continents/instances etc. and always excludes mobs that have no other loot (guards typically cannot be looted from what i know). So the only way to do this currently is adding the loot(-template) to the loot of each mob in the world... Though note that there's plenty of dummy/trigger mobs that can't be directly killed or can't be seen, or both, and giving them loot can lead to weird behaviour too.

If you can tell me a damage spell that takes longer than 5s to cast... ^^ And yes, you do enter combat, and leave about 5s after you stop attacking, just tried retail. Offtopic: Actually i think most "normal" mobs have timer based combat reset, they don't seem to unconditionally reset once you crossed a certain radius from their spawn location...

Actually, you can, through various virtualization techniques...but "dual boot" typically refers to having both installed and choosing which one to boot from a boot loader. And as i understood, mangos will not run on his new 64bit PC but another dedicated machine with 32bit Windows.

If you don't have *.vmtree, *.vmtile and *.vmo files you did something very wrong... You obviously pulled out a really old vmap extractor/assembler from somewhere. You should do what the text say then "redo vmaps with latest vmap_assembler.exe program" Note to self: remember that .exe only applies to windows, as the tools build fine on linux now too ^^

Hm i'm still (or again?) getting the forum-inside-forum thing when trying to rename a thread, which consequently fails Have to hope that editing the first post gives me the option to edit the topic, which apparently depends on the age?

That's very simple, because "REPLACE" does not exist in the SQL standard. It's just a custom extension in several SQL-compatible server implementations.

Hm not exactly much going on lately, will any dev besides me actually care to comment? Don't make it feel like the Captain is leaving the sinking ship... About faramir118 and Schmoozered i have no complaints, if they are interested of course. From what i can tell Schmoozered tries very hard to things less hacky than in stuff you find out in the wild. faramir has for sure good core knowledge by now, and i'm not sure anyone can even review the whole mmap project without his direct involvement, and it certainly adds a lot of quality to game play when it is stable. To Vladimir, thanks for all the work you put into this project, and i can fully understand that after such a long time, and many people come and leave, it gets time to do something different. Since you want to stay in #mangos, I guess we'll get to know when you found a new cool project, whatever it may be

I have to correct myself, the mac clients just aren't stripped, which means all function names are available. So the few leaked internal builds are the only source so far that reveals the original DBC struct member names.

No idea what is inside, there doesn't seem to be any good information, from what i heard it's basically everything RSA has been putting in his branch and then some. But RSA already never really cared to outline his modifications (at least not here in english forums), so hardly anyone would ever try to make some useful feature patch for mangos. There may be many nice things inside, but basically, they're rather useless to us because they are mixed with so many other things, making a port arguably as complicated as re-implementing it. Many unfished patches from our core modification section seem to go into such forks and get fixed/hacked there some more, but usually nothing comes back from it. Sometimes someone drops a big pile of code and expects us to review it without their help, but that's just infeasible for most of us time-wise.

Assembly normally doesn't contain any variable names either... AFAIK some 4.x mac clients accidently (?) were released with debug info, which allowed to extract named structs. No idea if it's still the case...i'm not the binary dissection guru here

This "mangosR2" is not mangos! It's some custom fork with god knows what mods and hacks. So please everyone stop flooding the forum with problems with it...could aswell ask for arcemu or trinity support :rolleyes:

Tedious research. basically. The files themselves don't contain any info about it.

Stupid question, but did you try to actually read the error? Since you didn't give any details about the Linux distro you're using, i can't tell you what package you need, in case of Ubuntu it would be libmysqlclient-dev for example.

Of course that didn't help, it just screws the chances for selecting the type of mineral that will spawn, which happens after that location has already been selected.

Install a 64bit compiler first?

What do you mean with conflict? Source compiles just fine on Linux if you didn't mess around with it...I just don't see a reason to undo a set of fixes without a real reason...

So in other words you suggest to revert the code to rev. [10924]? And for what reason?

It's a bad idea for a very simple reason, SD2 database is configured independently from the other DBs, so assuming you have access to other DBs from SD2 DB connection is just false. And the name of the character DB depends on the configuration too, so this simply cannot work in any generic way. So better listen to what zergtmn says

"/bin/sh^M" According to Wikipedia, "^M" stands for Carriage return...i guess the script somehow got converted to windows line endings, and your shell doesn't like that...

I don't think enabling the vmaps is the problem here, core demands them by default now anyway... The problem is not really vmaps related at all, rather a logic problem how to select a proper height on very uneven terrain without screwing up in some other way. Best way to solve it would be proper path finding of course, but that's not available yet, though you may find faramir's mmaps repo useful.

They always required an actual item that counts as fire/water/earth/wind totem, either the annoying inventory-stuffing quest totems or a relic acting as such (see flavour texts), much like gathering or crafting abilities that need tools. The problem probably has to do with the fact that they are cast as triggered by the "Call of ..." spells. Not sure how it works (or rather, worked) retail though, they tend to change their mind more quickly than anyone can react