[7428] New .map file format and some fixes

[NewGotham]

... Yeah, that worked. Apparently, GIT wasn't updating the source despite the fact it SAID it was...

Got it sorted, works fine!

[lenst]

I'm running mangos on Linux and I still have problem extracting the new map files. I have recompiled ad according to instructions. Running it generates 1206 maps files, but the lowest numbered file is 2692527.map. No 0004331.map gets generated.

[hongchan]

thanks.

nice work.

[henhouse]

Awesome!

However can this be pushed to 0.12 like the previous one was too?

[jorooo]

@henhouse:

Back port to 2.4.3

DOWNLOAD

i don`t have this client, but must work. If all ok i add it.

[Laizerox]

I'm running mangos on Linux and I still have problem extracting the new map files. I have recompiled ad according to instructions. Running it generates 1206 maps files, but the lowest numbered file is 2692527.map. No 0004331.map gets generated.

I ran small tests and did come to result that linux refuses to read 3 .MPQ files...

These files are common, common-2 and new lichking one.

In other way I tried to open them in text editor and text editor fails to open them as well where as it can open others... So I have no idea what causes the real problem...

[lenst]

I ran small tests and did come to result that linux refuses to read 3 .MPQ files...

These files are common, common-2 and new lichking one.

In other way I tried to open them in text editor and text editor fails to open them as well where as it can open others... So I have no idea what causes the real problem...

These are large files. That might explain why the text editor can't open them. They are all larger than 2Gb. I think they are the only files that large.

I put in some printfs and found that fopen fails with "Value too large for defined data type". But why now? What changed in the patch that could have an effect on this? Any change in libraries or compiler settings?

[Lynx3d]

Hm for me "ad" refuses to extract any DBC files and maps since the new map format when compiled on 64bit Ubuntu.

Looks somehow like file extraction is screwed somewhere.

I chose "Release" in cmake wizzard. I could swear it worked fine with the "old" version, a few commits before the new format. Now the output is merely:

Map & DBC Extractor
===================

Detected locale: enGB
Opening ./Data/enGB/locale-enGB.MPQ
Opening ./Data/enGB/patch-enGB.MPQ
Opening ./Data/enGB/patch-enGB-2.MPQ
Extracting dbc files...
Extracted 0 DBC files

Detected locale: deDE
Opening ./Data/deDE/locale-deDE.MPQ
Opening ./Data/deDE/patch-deDE.MPQ
Opening ./Data/deDE/patch-deDE-2.MPQ
Extracting dbc files...
Extracted 0 DBC files

Using locale: enGB
Opening ./Data/enGB/locale-enGB.MPQ
Opening ./Data/enGB/patch-enGB.MPQ
Opening ./Data/enGB/patch-enGB-2.MPQ
Opening ./Data/common.MPQ
Opening ./Data/common-2.MPQ
Opening ./Data/lichking.MPQ
Opening ./Data/expansion.MPQ
Opening ./Data/patch.MPQ
Opening ./Data/patch-2.MPQ
Extracting maps...
Read Map.dbc file... Fatal error: Invalid Map.dbc file format!

However, the precompiled ad.exe still worked fine (don't have 32bit libmpq shared lib for the precompiled linux binary, so back to wine...)

[lenst]

On 32bit linux fopen can't open large files (>2Gb i think). System.cpp uses fopen to test if a file exists. The code that reads the files uses open (not fopen) with flag O_LARGEFILE. Before the map patch these files where just read, not tested for existens. One solution could be to change the exitens test to use open. The following patch works for me:

--- a/contrib/extractor/System.cpp
+++ b/contrib/extractor/System.cpp
@@ -1,6 +1,9 @@
#define _CRT_SECURE_NO_DEPRECATE

#include <stdio.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
#include <deque>
#include <set>
#include <cstdlib>
@@ -81,9 +84,10 @@ void CreateDir( const std::string& Path )

bool FileExists( const char* FileName )
{
-    if(FILE* fp = fopen( FileName, "rb" ))
+    int fd = open( FileName, O_RDONLY | O_LARGEFILE );
+    if(fd >= 0)
    {
-        fclose(fp);
+        close(fd);
        return true;
    }

[VladimirMangos]

O_LARGEFILE not portable, not exist like definition in win32 api.

[Lynx3d]

Hm...not sure if it's a good idea to go that low-level for that anyway. POSIX way to check file existence/permissions would be access() from unistd.h...that would work nearly everywhere except Windows of course :rolleyes:

#ifndef WIN32
bool FileExists( const char* FileName )
{
   if( access(FileName, F_OK) || access(FileName, R_OK) ) return false;
   return true;
}
#else
bool FileExists( const char* FileName )
{...windows code...}

That would be all to check for existence and read access. The actual reading seems to be part of libmpq, and if it fails to handle the file it seems to be caught in MPQArchive::MPQArchive() anyway...?

[dardennf]

There are few problems with portals, hearthstones, battle grounds and arenas ...

With some players, those problems occur when they are on a mount or when they jump before the teleportation : they die.

With some others players, when they take a portal (...) without a mount : they die.

Some ideas ?

[hikikomori18]

There are few problems with portals, hearthstones, battle grounds and arenas ...

With some players, those problems occur when they are on a mount or when they jump before the teleportation : they die.

With some others players, when they take a portal (...) without a mount : they die.

Some ideas ?

i believe thats unrelated, it was posted in another thread about the portal to the mages in SW. It apparently has something to do with them taking fall damage after the port.

[Sephiroth1983]

The changes done by this patch break Blink even more than before

at least in Ruins of Lordaeron arena.

[VladimirMangos]

It apparently has something to do with them taking fall damage after the port.

This related to recieve outdated movement packets from clients from past after server side teleport that considered by fall code as fall. But bug not in fall code, but in not implementing some way to skip outdated packets in case near (same map) teleport. Maybe need rememeber time of teleport and skill any packets with in packet time < remembered. Or after send teleport packet to client skill all movement pakcets until some client confirmation about teleport packet recieve. Anyway this optopic for thread.

[VladimirMangos]

/root/mangos-old/extractor_v2/loadlib/loadlib.cpp:8:27: error: ..\\mpq_libmpq.h: No such file or directory

I think you use not recent sources.

In recent sources:

#include "../mpq_libmpq.h"

Hmmm, or modified. This fiel has been added and not chnage in this line from creating.

I think you must get proper not modified soueces instead attempt fix manually your existed version.

[go2wolfie]

i did change that Vladimir, if you read my post i put that line on bold and after that error code i put that i corrected that part and then i got another error.

and i'm using the http://depositfiles.com/files/y524hv1wr which i found from this Threat post #67 where the applying maps.patch is

[VladimirMangos]

As i write: you don;t must have fixed line. It just not exist in not-modified sources.

[VladimirMangos]

i ran all those, then i used the ad.exe from the files above

Where in _above_ you found ad.exe

_Only_ _single_ appropriate ad.exe for your version if you build from own sources in contrib/extractor directory.

BUT this can not work if you have some stranged way modifed source. And you already prove by quoted lines that you have its in modified state.

[go2wolfie]

Ohhh ok, thank you very much Hiki, i went to the git site and got the ad.exe file and guess what, IT WORKED xDDDDDDD

Now, i'm about to extract the vmaps jopefully it wil work also because the last one i got was crashing when assembling the vmaps.

So far so good, at least i can start the server with those maps i created.

And Vladi thanks for your releases. but maybe for a future refence can you add on the ad.exe a comment with the version extrated of the file and which one you need for the server you are runnnig.

[Sephiroth1983]

The changes done by this patch break Blink even more than before

at least in Ruins of Lordaeron arena.

To this I can add, while blink started to work a bit better, now it launch us sideways (a bit off of the straight trajectory) or even backwards. Sometimes for no apparent reason.

[Lascax]

To this I can add, while blink started to work a bit better, now it launch us sideways (a bit off of the straight trajectory) or even backwards. Sometimes for no apparent reason.

I am teleported backward when there is an obstacle in front of me. Try to Blink in front of a closed door.