faramir118

I added debug configs to my branch. Has anybody run into trouble on Windows?

That part of Totem was changed in 7111, you might want to update and see if it helps?

Ok, try this: http://github.com/faramir118/mangos/tree/vmap_rewrite Amazingly, I didn't have to change a single line of code to get it to work. All I did was work with the includes. Changes: * Added bzip2 src to contrib/libmpq/dep/bzip2 * Added zlib src to contrib/libmpq/dep/zlib * Added test.cpp to contrib/libmpq/libmpq Tests mpq read and extraction. There's no error checking, it just assumes that you give it a valid filename (and optionally a file index) * Added VS90 solution to contrib/libmpq/win Windows binaries (libmpq.lib, test.exe) goes to contrib/libmpq/bin Happy testing! edit: forgot to mention that I didn't edit the debug configurations for VS, so they won't compile. Make sure you're compiling for Release. edit: fix path typos

Sorry, I guess I could have told you earlier that VS90 didn't like libmpq includes. There are three options: * use contrib/extractor/libmpq This library is ported to C++ This library seems to be a slightly modified version of rev5 of libmpq * use wowmodelviewer/src/libmpq This library is ported to C++ This library is heavily modified * attempt to make libmpq compatible with Windows on our own I'll try number 3, but don't count on me. If I do make progress, I promise it won't be super hacky! I'll probably also make some diffs to see what the differences between the three are.

Mangos revision: 9754 Glyph of Flash of Light changed, so this code doesn't do anything. From cbbcd3b08999b694592e8249b43e75bb04c2050b Mon Sep 17 00:00:00 2001 From: faramir118 <[email protected]> Date: Tue, 20 Apr 2010 00:24:57 -0500 Subject: [PATCH] remove old aura 54936 proc --- src/game/Unit.cpp | 7 ------- 1 files changed, 0 insertions(+), 7 deletions(-) diff --git a/src/game/Unit.cpp b/src/game/Unit.cpp index e13d4de..7a0f41d 100644 --- a/src/game/Unit.cpp +++ b/src/game/Unit.cpp @@ -6519,13 +6519,6 @@ bool Unit::HandleDummyAuraProc(Unit *pVictim, uint32 damage, Aura* triggeredByAu CastSpell(target,53739,true,NULL,triggeredByAura); break; } - // Glyph of Flash of Light - case 54936: - { - triggered_spell_id = 54957; - basepoints[0] = triggerAmount*damage/100; - break; - } // Glyph of Holy Light case 54937: { -- 1.6.5.1.1367.gcd48 For future reference, where should small changes like this go?

For future reference, it's easier to test if you post a diff. Simple diff containing your changes. git diff > my.patch OR Create full patch git commit -a -m "Short commit message" git format-patch -1 ./ Both methods create a new file in your repository, copy the contents and post that in the forum or paste2.org.

More than a few patches ago. http://www.wowwiki.com/Patch_2.0.3 It's there under the General section... Also, I think I read somewhere that if you tell the Pet to attack (ie, click the attack button on pet bar) a crowd controlled monster, it will do it. Is this still the case, or was that changed also? edit: I didn't make sense.

I got the lib compiled with MSYS. It does appear as if it contains the bz2 and zlib objects (open libmpq.a with a text editor and search for function names from bz2 or zlib). However, my libmpq.a is only 35.1kB I hacked around a while and finally got Visual Studio to attempt linking libmpq into vmap_extractor_v2, but I get: libmpq.lib(mpq.o) : error LNK2019: unresolved external symbol _fseeko referenced in function _libmpq__archive_open I went back and looked in MinGW's stdio.h, and all it has is fseeko64 With all of these problems, something in libmpq is going to have to be changed at some point.

I started with Cygwin. ./autogen.sh went smoothly ./configure throws error autom4te: invalid parameter '--trace=', along with a bunch of suggestions/warnings. I got stuck there, not experienced enough with *nix build tools. I didn't bother trying with MSYS after that, just figured it was hopeless. I'll try it tomorrow and let you know how it goes.

Did you add the realm to the realmlist table in realmd database?

It looks to me like the code in contrib/extractor/libmpq is already cross-platform. Why not use that? Here's an alternative: Built using Eclipse CDT, MinGW 5.1.6, and MSYS 1.0.10 for Win32: http://rapidshare.com/files/377488543/libmpq-win32.zip.html The whole eclipse project is in the zip, along with a Win32-Release binary. I had to make similar changes as the cross-platform version in contrib/extractor/libmpq, because I wasn't able to get config.h. //#include "config.h" //const char *libmpq__version(void) { // // /* return version information. */ // return VERSION; //}

I'll add that sort of thing to the todo list, but it will be a while before I get to it.

It's possible, but I don't think it's worth the time it would take to process all of the maps: The recast parameters need some tweaking - you'll run into issues where NPC don't 'fit' through doors, or a hill is too steep for them, etc The generator also crashes on certain maps. Looking into it. Thanks, I'll push that in a bit. Only if you generate all of the mmaps yourself. Read above why I think it's a waste of time.

This was on my list of things to do with pathfinding, but I think that there are other, faster ways to fix it similar to what NoFantasy posted above. However, there is a problem with checking JUST the height. Some monsters can throw you into the air high enough that they need to follow to keep attacking. (see http://www.wowhead.com/npc=5474#abilities) If the target isn't in range, that ogre will just evade and reset even though the player wasn't really out of range. So, you need to see if the player is actually flying, versus if he is falling, knocked back, etc.

Pushed small fix for linux compile. Sorry I can't test these things myself

Hmm, that's ugly. I'll look at it. try generator 0 48 33 Not anymore! I changed it so you can just put true or false in the config. The obj file goes to the same location as the mmaps if the setting is true.

I just pushed the generator to my repo. * I made a few minor code changes * mmap.conf is simpler, and has documentation * a smidge of compile fixes (VS90 and earlier don't come with a stdint.h) * some code cleanup (more needed) * I cut out a lot of the things that were left over from the previous mmap project There's plenty left though, just need some time to figure out what is safe to delete * removed some things from the make configs - need linux testers * need VS80 and VS100 solutions (accidentally wiped the VS80 stuff out, oops!) Updated first post with directions.

Took me a while to sort through this, but you speak truth. When I add it to my repo, it won't contain any of the code left over from the mmap project Derex started. It will be just the recast/detour parts.

There is already inhabit type somewhere (DBC or SQL DB?), which keeps creatures from entering water if they can't swim, etc. Getting the navmesh to contain the terrain info is the part we don't know. I haven't done any of the map generation, so someone will have to correct me if I'm wrong: You need to specify coordinates for the first mmap as a command line argument - read the readme.rtf. That being said, I'm hoping to redo the generator. Recast doesn't need coordinates to generate a navmesh, it just needs the map mesh. I may be putting my foot in my mouth here, but I think it's possible to get the map mesh data from the vmap without needing the start/end positions. edit: No, don't need start/end coords. Gotisch had taken care of that.

Not a bug, but if you want to change it you need to alter the source. Search for and change the last parameter of the following in World.cpp: setConfigMinMax(CONFIG_UINT32_MAX_PLAYER_LEVEL, "MaxPlayerLevel", DEFAULT_MAX_LEVEL, 1, DEFAULT_MAX_LEVEL); But also pay attention to following from DBCEnums.h: // Client expected level limitation, like as used in DBC item max levels for "until max player level" // use as default max player level, must be fit max level for used client // also see MAX_LEVEL and STRONG_MAX_LEVEL define #define DEFAULT_MAX_LEVEL 80 // client supported max level for player/pets/etc. Avoid overflow or client stability affected. // also see GT_MAX_LEVEL define #define MAX_LEVEL 100 I haven't tried it, so you might crash your client trying it. Have fun!

True, that is better than nothing. Paladin can Judgement of Light + Beacon of Light, then sit back and watch party members do the healing for him.

Ok, good. I'll work on adding the generator to the branch on my repo. Will probably put them under contrib with the vmap tools.

That link is for the non-recast generator. Thyros, could you edit your post and remove it, just so people don't get confused? The one modified by Gotisch is here: http://rs806.rapidshare.com/files/368138336/recastmmaps.tar.gz I will try to add it to my mmaps branch. Gotisch was using linux, so I'll have to make his changes work in visual studio.

updated in original post

Updated for 9750: http://paste2.org/p/775871