VladimirMangos

ciphercom work at cmake support for mangos show his interest and will to complete so big project with many multi-platform specific problems to state that will can ready for adding to mangos repo. He already presence some time at #mangos channel and i think also show possibility work in team. I think he can good addition to mangos team

this meaning that you or (1) use wrong extractor (must be used provided in sources), or (2) extract DBC files from wrong client version. Also note: you must copy in server dbc dir _all files from dbc dir generated by extractor

As i think it dependent from time when thread created. Old threads missing feature. Maybe other conditions also exist.

Starting from [11167] versions MaNGOS Master version use new build system base at cmake usage. * If you build MaNGOS at Unix/Linux then you need have installed cmake package and only can build mangos with cmake use: as example normal way build in like case.* If you plan build MaNGOS at Windows then for this time prefered use old way with Visual Studio projects in win directory. So nothing really changes in this case. You _can_ use cmake way MaNGOS build at Windows but this experemental mode. For cmake using build you need have cmake tool installed at Windows (with adding cmake to PATH). And need run cmake (most simple way run createprojects.bat in MaNGOS root directory) and then load generated Visual Studio solltion file build/mangos.sln [added]Mangos One also use cmake from [s0827] revision. [added2]MaNGOS Zero also use cmake from [z1369] revision

Added into [11167]+1 as default build system for Unix/Linux and experemental for Windows. Thank you, ciphercom.

Can't say anything about first case. So about second: fread call return amount really read data from requested in args. For really 100% safe code work need check return amount by compare with expected and process this in some way. We not always check thsi when expected "right" file structure. Ok, this not safe, i agree, but low chance problems in most like cases. Maybe fixing can be good but not high prio.

In [11157]. Thank you

I revert commit in [11154]. Maybe for first part check you right if not attribute like way check. But posisble need instead skipp los check replace it for like spell by outdoor check. But adding to ignore check all 1200 spells with TARGET_ALL_ENEMY_IN_AREA_INSTANT i think not right. This need additional research i think before adding to repo. Thread moved to under review section

In [11153] added same checked for not set aggro mode that used for normal spell hit case. In original added version code look strange, it can aggro target at miss in cases when normal hit will not triggering aggro.

But why in addition allowed any TARGET_ALL_ENEMY_IN_AREA_INSTANT spell targets ignore los

What about SPELL_ATTR_EX3_NO_INITIAL_AGGRO...

Understand now. Maybe need add comment to loop code for clarify this for code reader Comment added in [11152]

it still ugly in any form. and wrong also as i think: target unit is any unit target spell. I thin this wrong check just make all idea los check useless in this case.

+ for (Unit::AuraList::const_iterator itr = stunAuras.begin(); itr != stunAuras.end(); ++itr) + if (!(*itr)->HasMechanic(MECHANIC_STUN)) + { + is_stun_mechanic = false; + break; Look strange: why we stop at first non mechanic stun instead check that at least one mechanic stun exist? Or for stun need have all stun auras mechanic?

This is sd2 code, so no point discuss it at mangos _core_ forum.

it fixed in [11134]

any problem with tool/patch not included in mangos repo must be discussed in related tool/patch threads. Find related thread and reask.

With current state sources, don't must be any differences in cmake test results for One/master. About PREFIX. as cmake noob I from start use PREFIX ^^ so only from you read about CMAKE_INSTALL_PREFIX I test patch (with last fixes) at FreeBSD and it good work now. I think patch ready for repo in only-for-Unix-by-default mode. It just missing small cleanup parts for drop now not working anyway (as i understand from ciphercom notes) makefile.am and related remains. Because use it for windows still have some problems mostly related dev. way work (testing from IDE/etc), hmmm, at least for VS100 by default better preserve current project files. This not prevent ppl from use cmake way project files generation and use.

bobber

We follow in past always policy: it will be done when it will be done. no ETA or any timelines. All what we ready public show you can see at our repos and its progress. I not think that it will be changes now or in future.

so move to rejected as outdated.

maybe for targets amount limited area spells it must also ignore totems as for chained cases. In different unlimited area affects.

already fixed in [11116]

I not agree, some points have only 0x80 set that expected as outdoor as i remember. Ofc, i can recheck with this mask for find specific examples.

Possible related asm cases guarded by some defines specific for Mac OS in g3dlite as i think. Possible its can be replaced by arch specific defines and then not dependent from Mac OS. Also possible platform detection see __linux__/etc and set Linux platform and expect all linux is x86/x64 In short g3dlite or tbb can missing linux at PPC support. For overwrite it you need add something like LINUX_PPC platform to g3dlite code and update related defines use it for selection not using asm way build. It can be not easy task... Don't known what need for tbb.