Hm.

First enable logging in autoexec.cfg with 'log on'.
At start time, do you see this?

L 09/11/2001 - 03:07:58: [META] Metamod v1.09.00, 2001/08/01
L 09/11/2001 - 03:07:58: [META] by Will Day <willday@...>
L 09/11/2001 - 03:07:58: [META]    http://www.metamod.org/
L 09/11/2001 - 03:07:58: [META] compiled: Aug  2 2001, 02:10:34
Eastern (optimized)
L 09/11/2001 - 03:07:58: [META] Recognized game 'cstrike'; using
dllfile 'cs_i386.so'
L 09/11/2001 - 03:07:59: [META] Game DLL for 'Counter-Strike' loaded
successfully
L 09/11/2001 - 03:07:59: [META] ini: Begin reading plugins list:
metamod.ini
L 09/11/2001 - 03:07:59: [META] ini: Finished reading plugins list:
metamod.ini; Found 0 plugins to load
L 09/11/2001 - 03:07:59: [META] ERROR: ini: Warning; no plugins found
to load?
L 09/11/2001 - 03:07:59: [META] dll: Loading plugins...
L 09/11/2001 - 03:07:59: [META] dll: Finished loading 0 plugins

In this case, I had no plugins configured, but metamod did indeed load
into memory.  Cheers.

-a

--- In metamod@y..., beruriah@m... wrote:
> Hi all...
> I'm having problems getting metamod to work, sort of...
> I'm running a dedicated, linux-based CS server, and using admin
> mod.  The admin mod is working fine, and, when the server runs,
> metamod appears to load:
>
> <OUTPUT>
>  Metamod version 1.09.00, Copyright (c) 2001 Will Day
> <willday@m...>
>    Metamod comes with ABSOLUTELY NO WARRANTY; for
> details type `meta gpl'.
>    This is free software, and you are welcome to redistribute it
>    under certain conditions; type `meta gpl' for details.
> </OUTPUT>
>
> However, whenever I try to use meta commands, or another plugin
> in addition to admin mod (e.g. LogD), nothing happens.  No crash,
> no error message, no apparent problems, but also no success.
> Typing meta list (or meta anything) results in silence from the
> server.
> I have not yet ruled out that I made some kind of an idiotic
> oversight in one of the .cfg files, but if I did, I can't find it.
Has this
> happened to anyone else?  Is there a solution?
> Thanks.
> --Jordan

Hi!

I have downloaded the metamod source and the full halflife sdk 2.2 from
metamod.org.
I'm using linux and the compiling and installtion works fine.
When I start hlds I get this message:
LoadLibrary failed on /home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so:
/home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so: undefined symbol:
__ti18CSaveRestoreBuffer
Host_Error: Couldn't get DLL API from
/home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so!
Error Host_Error: Couldn't get DLL API from
/home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so!

I hope anyone out there can help me fix this problem.

Thanks in advance,
Thomas Lundqivst.

Thomas Lundqvist wrote:
>
> LoadLibrary failed on /home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so:
> /home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so: undefined symbol:
> __ti18CSaveRestoreBuffer

Try if it works if you don't compile it in debug mode but
optimized.

Florian.

Thomas Lundqvist wrote:
>
> LoadLibrary failed on /home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so:
> /home/obi1/hl/hlds_l/cstrike/dlls/metamod_i386.so: undefined symbol:
> __ti18CSaveRestoreBuffer

Try if it works if you don't compile it in debug mode but
optimized.

Florian.

------------------------ Yahoo! Groups Sponsor ---------------------~-->
Pinpoint the right security solution for your company- Learn how to add 128- bit encryption and to authenticate your web site with VeriSign's FREE guide!
http://us.click.yahoo.com/yQix2C/33_CAA/yigFAA/hWFolB/TM
---------------------------------------------------------------------~->

To unsubscribe, send an email to: metamod-unsubscribe@yahoogroups.com

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

Hi there,

I got several problems when building/running on metamod-1.10.00+hlsdk-2.2
on windows (linux runs fine):

1. Building it myself with MSVC 6.0 results in some strange error popup
    when starting a server. It says:

      Assertion failed,
      Program: C:\GAMES\SIERRA\HALF-LIFE\HLDS.EXE
      File: setnewh.cpp
      Line: 52
      Expression: new_p == 0

This happens immediately after metamod prints the "compiled: <date>..."
message in the log. If i use the precompiled dll, it works fine. Also,
For some reason, my self-build dll is much smaller than the pre-compiled
one?! I also examined both dlls with the depency-checker from the MS
devstudio and noticed, that the pre-build one links against msvcrt while
the self-build does not. I grep'd for setnewh but didn't find anything.
I assume, that some compiler- or linker switch is wrong but i really don't
know which. This happens regardless of debug/optimized setting. Funny
effect also: If i ignore this error (The Dialog offers Abort, Retry,
Ignore), Then the game runs, but i can't use the mouse anymore. Moving
around with the keyboard still works.
Is the pre-build dll compiled with mingw or cygwin's gcc?

2. Problem:

Using the pre-build dll, calling CENTER_SAY results in a server crash
with a message "NUM_FOR_EDICT: Illegal pointer".
If i compare the code in metamod against the similar code of hlsdk-2.2
(UTIL_HudMessageAll) i noticed, that in hlsdk-2.2 they use
UTIL_PlayerByIndex which also checks for (!pPlayerEdict->free) and
uses CBaseEntity::Instance(pPlayerEdict) instead of a cast. However,
i can't verify if that's the reason due to Problem 1.

3. I there a public CVS for metamod? I looked on sourceforge, but the
CVS tree is empty there.

Thanks
  -Fritz

On Wed, 28 Nov 2001, myself wrote:

[...]
> 2. Problem:
>
> Using the pre-build dll, calling CENTER_SAY results in a server crash
[...]

Got this working (tested while ignoring error 1).
Patch is attached (Added check for pEntity->free).

Ciao
  -Fritz

--- mutil.cpp.orig Wed Nov 28 07:31:01 2001
+++ mutil.cpp Wed Nov 28 07:31:48 2001
@@ -144,7 +144,7 @@
 	 mutil_LogMessage(plid, "(centersay) %s", buf);
 	 for(n=1; n <= gpGlobals->maxClients; n++) {
 		 pEntity=INDEXENT(n);
-  if(FNullEnt(pEntity))
+  if(FNullEnt(pEntity) || pEntity->free)
 			 continue;
 		 pPlayer=(CBaseEntity*)GET_PRIVATE(pEntity);
 		 UTIL_HudMessage(pPlayer, tparms, buf);

Hi,

again answering my own stuff ;-)

The assertion was raised in the function set_new_handler() from the
debugging version of MSVCRT. After having a look into MSVCRT sources,
the reason was revealed:

set_new_handler() is just a stub which (according to comments in the
source) should never be used. It is just an ugly hack to make STL
compile. It does _not_ set the new handler but always calls
_set_new_handler(0) instead. _set_new_handler is the "real" function and
uses the "old" semantic [handler-type is: int newhandler(size_t)].
I changed the relevant stuff in osdep.{cpp,h} and ifdef'd it into
defined(_WIN32) && defined(_MSC_VER) to make shure this only happens
if building wit MSC++ ...

Patch is attached...

  -Fritz

Index: osdep.cpp
===================================================================
RCS file: /home/cvs/cvsroot/metamod/metamod/osdep.cpp,v
retrieving revision 1.1
diff -u -w -r1.1 osdep.cpp
--- osdep.cpp 2001/11/28 08:39:24 1.1
+++ osdep.cpp 2001/11/28 08:57:01
@@ -166,8 +166,19 @@

  // This used to be OS-dependent, as previous documentation indicated MSVC
  // used different typedef than gcc.  That appears to no longer be the case.
+// Well, not really ... If looking into MSVCRT debugging sources it turns out,
+// that set_new_handler is only a stub which always disables the handler
+// (and - even more confusing - raises an assert, if the argument is not NULL).
+// The _real_ function is _set_new_handler (_with_ leading underscore) which
still
+// uses the "old" semantic.
+#if defined(_WIN32) && defined(_MSC_VER)
+int meta_new_handler(size_t) {
+ return 0;
+}
+#else
  void meta_new_handler(void) {
 	 // This merely because we don't want the program to exit if new()
 	 // fails..
 	 return;
  }
+#endif
Index: osdep.h
===================================================================
RCS file: /home/cvs/cvsroot/metamod/metamod/osdep.h,v
retrieving revision 1.1
diff -u -w -r1.1 osdep.h
--- osdep.h 2001/11/28 08:39:24 1.1
+++ osdep.h 2001/11/28 08:57:01
@@ -69,6 +69,7 @@
  // DLL.
  #undef DLLEXPORT
  #ifdef _WIN32
+ #pragma warning(disable: 4390)
 	 #define DLLEXPORT __declspec(dllexport)
 	 // WINAPI should be provided in the windows compiler headers.
 	 // It's usually defined to something like "__stdcall".
@@ -160,6 +161,9 @@
 	 #define strcasecmp stricmp
 	 #define strncasecmp _strnicmp
 	 #define S_ISREG(m) ((m) & S_IFREG)
+ #ifdef _MSC_VER
+  #define set_new_handler _set_new_handler
+ #endif
  #endif /* _WIN32 */


@@ -168,8 +172,11 @@
  //    http://dragon.klte.hu/~kollarl/C++/node45.html
  // Although it appears MSVC++ is no longer different from gcc, according to:
  //   
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclang98/stdlib\
/info/NEW.asp
+#if defined(_WIN32) && defined(_MSC_VER)
+int meta_new_handler(size_t);
+#else
  void meta_new_handler(void);
-
+#endif

  // Thread handling...
  #ifdef linux

Fritz Elfert wrote:
>
Hi Fritz!

> Is the pre-build dll compiled with mingw or cygwin's gcc?

It is compiled with mingw under Linux, cross-compiled for Win32.

> 3. I there a public CVS for metamod? I looked on sourceforge, but the
> CVS tree is empty there.

Nope. Where did you look on SourceForge? I'm not aware of Metamod
being hosted as a SF project?

Florian.

Fritz Elfert wrote:
>
> Hi,
>
> again answering my own stuff ;-)

What a relief from the Admin Mod forums where nobody is able to help
himself and noone has a clue about what is actually going on. :)

> compile. It does _not_ set the new handler but always calls
> _set_new_handler(0) instead. _set_new_handler is the "real" function and
> uses the "old" semantic [handler-type is: int newhandler(size_t)].

Ugh, M$ and it's strange ways of prefixing standard clib functions
with an underscore. :/

> I changed the relevant stuff in osdep.{cpp,h} and ifdef'd it into
> defined(_WIN32) && defined(_MSC_VER) to make shure this only happens
> if building wit MSC++ ...

Will, any plans for an updated Metamod in the near future? (Just
curious because of our 2.51 plans).

Florian.

On Wed, 28 Nov 2001, Will Day wrote:

> Hmm, lovely.  I had included a url in osdep.h to the page in msdn where I
> had found the information, but it looks like msdn has (a) reorganized all
> their links (so I have to change all the urls in the source), and (b)
> removed that particular page, as I can't find it anymore.
I didn't find anything either ;-/ just in the source that comment ...

> I've made the changes back to the source, although not applying your
> patches directly.  My diff is attached; double-check it, if you could, to
[...]

The function header of int meta_new_handler(size_t size) in osdep.cpp
is wrong. In your patch, it reads:

+#elif defined(_MSC_VER)
+int meta_new_handler(size_t size);

while correctly it should read:

+#elif defined(_MSC_VER)
+int meta_new_handler(size_t size) {

Apart from that, everything is ok.

> make sure it covers everything properly.
>
> >+ #pragma warning(disable: 4390)
>
> What does this do?
[...]

This disables a stupid warning about empty statements. There are loads of
them. The appear at code like this:

   if (plid); // satisfy gcc -Wunused

or at most of the RETURN_... macros

BTW: You don't need the above code to make gcc shut up. Instead, simply
don't name the parameter in the arguments (use an anonymous parameter).
E.g.: Instead of writing

void mutil_logConsole(plid_t plid, const char *fmt, ...) {
   // ... code, not using plid ...

simply write:

void mutil_logConsole(plid_t, const char *fmt, ...) {
    ...


> >> 3. I there a public CVS for metamod? I looked on sourceforge, but the
> >> CVS tree is empty there.
>
> Nope, I've never really gotten into CVS.

I really would like that. If you are not familiar with CVS, just give me
dev-access on sourceforge (my account there: felfert@...)
and i can do the initial setup (i have _lots_ of experience with CVS).


> Yes, there's a couple mod/ent support additions, and these fixes, so I'll
> probably push something out this week.
>
I just found that metamod'ized HPB_Bot at your Site. I'm currently doing
something similar. I'm writeing a Bot based on TeamBot (which is based on
HPB_Bot). Obviously, my bot is called "MetaBot" ;-). However, since
botman's  (and TeamBot's) code is a lot of difficultly readable
spaghetti-code, i decided to convert it into clean C++ (a base bot-class
with mod-specific bot-variants derived from it. The plain stuff is done
and now i'm designing a waypoint-class.

Because of that:
In the notes, you write:

    One thing I'm planning to look at is to have metamod manage the
    RegUserMsg calls and then provide the necessary information to plugins

Can i expect something like that, or shall i start hacking metamod myself?

Ah, and also: Something like MDLL_* but for the engine functions would
be really useful since engine functions are used a lot.

Ciao
  -Fritz

On Wed, 28 Nov 2001, Will Day wrote:

[...]
> >void mutil_logConsole(plid_t, const char *fmt, ...) {
>
> Gotcha.  I usually don't like to use anonymous parameters, as I prefer to
[...]
In those cases, i usually write it in comments:
void mutil_logConsole(plid_t /* thePLID */, const char *fmt, ...) {


> Well, it's been awhile since I last looked at that.  I think there were two
> things I wanted to do:
>
>  - Have metamod cache the Name to ID mapping, from RegUserMsg, so that bots
>    wouldn't have to catch RegUserMsg, and thus wouldn't have to be
>    restricted to loading only at startup.  And of course provide an
>    interface for the plugin/bot to query the mapping, something like:
>
>     int         (*pfnGetUserMsgID)      (plid_t plid, char *name);
>     char*       (*pfnGetUserMsgName)    (plid_t plid, int *msgid);
>
>    That doesn't include the iSize from RegUserMsg, though.  Wonder if
>    that's needed; it might be nice to have in any case.
At least in HPB's code it's not needed.
>
>  - Have metamod cache the entire _msg_, from MessageBegin through various
>    Write's until MessageEnd, and then be able to provide the message as a
>    whole to the plugin/bot, so the bot doesn't have to catch all the Write
>    functions.
Probably not worth the effort. When i look into the bot's code, in most
cases either there's a single element of interest while the rest of the
message is ignored, and also the message varies from mod to mod. From my
point of view, it would only make sense, if this forces kind of strong
type-checking but with so much message variations thats probably
impossible. You always have to cast something.

>
> The first should be pretty easy, and I could probably do that real quick.
> Any preference for interface names/arguments?
Not really, your suggestion above looks nice to me.


> You should already be able to call engine functions from a plugin, in the
> exact same manner as you would from a gamedll, using the macros in the SDK
> dlls/enginecallback.h where available, ie:
>    char *cp = (char *) LOAD_FILE_FOR_ME("myfile", &len);
>
> assuming of course you're using the same naming convention for your globals
> that Valve does. :)
That's the point. It needs that global g_engfuncs. I try to avoid globals
where i can cause they break the concept of instances in C++. Ok, in this
case it's academic ;)

-Fritz
PS: Again: If you need help with setting up some CVS, just drop me a note.

Fritz Elfert wrote:
>
> That's the point. It needs that global g_engfuncs. I try to avoid globals
> where i can cause they break the concept of instances in C++.

In how far do they break the concept of instances?

Florian.

i am writing a small plugin to analyze traffic...

and i am having some seg faulting problems with meta mod under linux... if
i try any of:

meta reload traffic
meta unload traffic, meta load traffic

it will seg fault every time IF the library has been updated, regardless
of whether anything was actually changed.  I tested this in the following
way:

start hlds,
meta load traffic (loads fine)
i log in to the server and run around, test out features, save some data,
etc.
i disconnect,
touch dllapi.cpp
make
make install
then in hlds console:
meta reload traffic
-- segfault here.

I repeat all of the above but with:
meta unload traffic
meta load traffic
-- segfault here


the first meta load traffic always works, so if it seg faults i can always
restart the server and load the plugin, but what a pain... i intend to use
this to do traffic analysis on a live server, and i cant have it seg
faulting all the time... any one seen this before?

much thanks...

jesse.

I don't have such problems here, but you probably want to run the server
in gdb for finding out _where_ it segfaults. Here's how to do that:

1. Compile both metamod and your traffic.so with -g to have debugging
    info.
2. Create a file .gdbinit which sets the server's commandline in your
    server's toplevel directory. Here is an example from one of my
    debugging sessions:

    set args "-game cstrike -nomaster +sv_lan 1 +exec autoexec.cfg"

3. Run the server in gdb manually with the following cmdline: (asuming
    libhlwon.so and hlds_run are in the same dir)

    LD_LIBRARY_PATH=`pwd` gdb ./hlds_run

    (At gdb's prompt, type "run<CR>"

4. Trigger the segfault (it returns to gdb's prompt)
5. Type "bt<CR>" to get a backtrace.

If you don't know how to interpret the backtrace, post it here ...

-Fritz

On Fri, 30 Nov 2001, jesse l. dailey wrote:

>
> i am writing a small plugin to analyze traffic...
>
> and i am having some seg faulting problems with meta mod under linux... if
> i try any of:
>
> meta reload traffic
> meta unload traffic, meta load traffic
>
> it will seg fault every time IF the library has been updated, regardless
> of whether anything was actually changed.  I tested this in the following
> way:
[...]

well, i ran the backtrace, and here it is:
none of these functions appear to be mine...

#0  0x419805a2 in Q_strncasecmp () at stdexcept:53
#1  0x41980612 in Q_strcasecmp () at stdexcept:53
#2  0x4197f7e0 in Cmd_ExecuteString () at stdexcept:53
#3  0x4197ea86 in Cbuf_Execute () at stdexcept:53
#4  0x4198f6c5 in _Host_Frame () at stdexcept:53
#5  0x4198f858 in Host_Frame () at stdexcept:53
#6  0x804a828 in Eng_Frame ()
#7  0x804a4d2 in main ()

i have a suspicion that it could be a lack on clean up on my part... since
this seg fault happens anytime i do a load->unload->load... but i dont
know what kinds of things i need to clean up, i mean global structures
that i use should be destroyed when the library is unloaded right? and if
they aren't they will be memory leaks not seg faults, no?


On Mon, 3 Dec 2001, Fritz Elfert wrote:

> I don't have such problems here, but you probably want to run the server
> in gdb for finding out _where_ it segfaults. Here's how to do that:
>
> 1. Compile both metamod and your traffic.so with -g to have debugging
>    info.
> 2. Create a file .gdbinit which sets the server's commandline in your
>    server's toplevel directory. Here is an example from one of my
>    debugging sessions:
>
>    set args "-game cstrike -nomaster +sv_lan 1 +exec autoexec.cfg"
>
> 3. Run the server in gdb manually with the following cmdline: (asuming
>    libhlwon.so and hlds_run are in the same dir)
>
>    LD_LIBRARY_PATH=`pwd` gdb ./hlds_run
>
>    (At gdb's prompt, type "run<CR>"
>
> 4. Trigger the segfault (it returns to gdb's prompt)
> 5. Type "bt<CR>" to get a backtrace.
>
> If you don't know how to interpret the backtrace, post it here ...
>
> -Fritz
>
> On Fri, 30 Nov 2001, jesse l. dailey wrote:
>
> >
> > i am writing a small plugin to analyze traffic...
> >
> > and i am having some seg faulting problems with meta mod under linux... if
> > i try any of:
> >
> > meta reload traffic
> > meta unload traffic, meta load traffic
> >
> > it will seg fault every time IF the library has been updated, regardless
> > of whether anything was actually changed.  I tested this in the following
> > way:
> [...]
>
>
>
>
> To unsubscribe, send an email to: metamod-unsubscribe@yahoogroups.com
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

On Mon, 3 Dec 2001, Will Day wrote:

>
> I trust your "make install" doesn't simply overwrite the currently open
> file (ie with 'cp'), but either renames or deletes the current file before
> copying the new one in place, or else uses /usr/bin/install to take care of
> that.


> When are these in relation to the 'make install'?  Does it still segfault
> if you do:
>    meta unload traffic
>    make install
>    meta load traffic
>


well i do just use cp, but it even crashes if i do:

meta load ...
meta unload ...
meta load ...

with no 'make install' at all...

ill change things over to /usr/bin/install maybe there is more than one
source of the crash.



> >well, i ran the backtrace, and here it is:
> >none of these functions appear to be mine...
> >
> >#0  0x419805a2 in Q_strncasecmp () at stdexcept:53
> >#1  0x41980612 in Q_strcasecmp () at stdexcept:53
> >#2  0x4197f7e0 in Cmd_ExecuteString () at stdexcept:53
> >#3  0x4197ea86 in Cbuf_Execute () at stdexcept:53
> >#4  0x4198f6c5 in _Host_Frame () at stdexcept:53
> >#5  0x4198f858 in Host_Frame () at stdexcept:53
> >#6  0x804a828 in Eng_Frame ()
> >#7  0x804a4d2 in main ()
>
> I think this the normal stacktrace when HLDS is looking for a function to
> call that corresponds to the command you typed at the HLDS console
> (iterating through the list of REG_SVR_COMMAND strings).

hmmm... perhaps it is still thinking that a server command is registered
and trying to call a function pointer that has been yanked from memory?


>
> Does it segfault if you just unload the plugin, without loading it again?

on a full server yes, it never has on my test server (which is usually
empty or has one person (me) in it )...


jesse.

>
> >
> >#0  0x419805a2 in Q_strncasecmp () at stdexcept:53
> >#1  0x41980612 in Q_strcasecmp () at stdexcept:53
> >#2  0x4197f7e0 in Cmd_ExecuteString () at stdexcept:53
> >#3  0x4197ea86 in Cbuf_Execute () at stdexcept:53
> >#4  0x4198f6c5 in _Host_Frame () at stdexcept:53
> >#5  0x4198f858 in Host_Frame () at stdexcept:53
> >#6  0x804a828 in Eng_Frame ()
> >#7  0x804a4d2 in main ()
>
> I think this the normal stacktrace when HLDS is looking for a function to
> call that corresponds to the command you typed at the HLDS console
> (iterating through the list of REG_SVR_COMMAND strings).

well, with this hint in mind i poked around actually any command following
an unload dies... i tried status, quit, meta list, meta cmds, meta
cvars... all seg faulted... do i need to un-register commands in
Meta_Detach?

here's how i register the one command i use, is this correct?

(*g_engfuncs.pfnAddServerCommand) ("traffic", &traffic_handler);

thanks for your help...

jesse.

the only engine functions i use are AddServerCommand and Cmd_Argv.... its
a very simple plugin to gather traffic data... so it implements
GameDLLInit to register its one command and PlayerPreThink to do the data
logging of a player's position...

there's really very little going on (like a hundred lines or
something)... i could send you dllapi.cpp if you are interested...

On Mon, 3 Dec 2001, Will Day wrote:

> A short time ago, at a computer terminal far, far away, jesse l. dailey wrote:
> >> I think this the normal stacktrace when HLDS is looking for a function to
> >> call that corresponds to the command you typed at the HLDS console
> >> (iterating through the list of REG_SVR_COMMAND strings).
> >
> >hmmm... perhaps it is still thinking that a server command is registered
> >and trying to call a function pointer that has been yanked from memory?
>
> That particular case should be handled (REG_SVR_COMMAND); the command
> string and function pointer that the HL engine sees are actually from
> metamod, rather than from the plugin.
>
> >> Does it segfault if you just unload the plugin, without loading it again?
> >
> >on a full server yes, it never has on my test server (which is usually
> >empty or has one person (me) in it )...
>
> Hmm.  What engine functions are you calling in the plugin?  There may be
> other functions that need to be similarly handled.  Right now, it handles
> REG_SVR_COMMAND, CVAR_REGISTER, and REG_USER_MSG.
>
> --
> Will Day                  Those who would give up essential Liberty, to
> willday@...       purchase a little temporary Safety, deserve
> UNIX System Programmer    neither Liberty nor Safety.  - Benjamin Franklin,
> Atlanta, GA                  Pennsylvania Assembly, Nov. 11, 1755
>

works perfectly... thanks a million! :)

On Mon, 3 Dec 2001, Will Day wrote:

> A short time ago, at a computer terminal far, far away, I wrote:
> >>hmmm... perhaps it is still thinking that a server command is registered
> >>and trying to call a function pointer that has been yanked from memory?
> >
> >That particular case should be handled (REG_SVR_COMMAND); the command
> >string and function pointer that the HL engine sees are actually from
> >metamod, rather than from the plugin.
>
> Okay, although it _should_ be handled, it in fact isn't.  It's still
> actually sending the original command string from the plugin to the engine,
> instead of the locally allocated copy of the string as was intended.  Dunno
> how I never ran into this.  If you're set up to compile metamod, try the
> attached patch and see if that fixes the problem.
>
> --
> Will Day                  Those who would give up essential Liberty, to
> willday@...       purchase a little temporary Safety, deserve
> UNIX System Programmer    neither Liberty nor Safety.  - Benjamin Franklin,
> Atlanta, GA                  Pennsylvania Assembly, Nov. 11, 1755
>

I wanna check a file for consistency on a player. I run ForceUnmodified in
DispatchThink, I tried to implement it in a ClientCommand but that didn't
work out. I made a corrupt file and tested it on a server and I never got
InconsistentFile back from the engine. Can I only check .mdl files or can I
check every file in halflife/cstrike directory?

Thanks in advance
/[DING]OBi1

Thomas Lundqvist wrote:
>
> I wanna check a file for consistency on a player. I run ForceUnmodified in
> DispatchThink, I tried to implement it in a ClientCommand but that didn't
> work out. I made a corrupt file and tested it on a server and I never got
> InconsistentFile back from the engine. Can I only check .mdl files or can I
> check every file in halflife/cstrike directory?

You can check any file but you will have to prechache it first.
Use PRECACHE_GENERIC() and then FORCE_UNMODIFIED() in
DispatchSpawn() (pfnSpawn()) to set it up and provide a
InconsistenFile() function in your plugin. But be aware that this
will download the file to the client if it does not yet exist on
the client.

Florian.

On Thu, 6 Dec 2001, Florian Zschocke wrote:

>
> You can check any file but you will have to prechache it first.
> Use PRECACHE_GENERIC() and then FORCE_UNMODIFIED() in
> DispatchSpawn() (pfnSpawn()) to set it up and provide a
> InconsistenFile() function in your plugin. But be aware that this
> will download the file to the client if it does not yet exist on
> the client.
>
> Florian.

Is it possible to check .exe and .dll files also? If an exe file
doesn't exist on the client will it then be sent to the client?

Regards,
Thomas Lundqvist

Thomas Lundqvist wrote:
>
> Is it possible to check .exe and .dll files also? If an exe file
> doesn't exist on the client will it then be sent to the client?

Yes and yes. :)

Florian.

1. I opened up the metamod VC++ project file and compiled the metamod dll. When
I ran hlds with this metamod.dll I get the following assertion error:
file: setnewh.cpp
line: 52
expression: new_p == 0

2. When I compile the trace plugin, metamod isn't able
to load it. Under meta info the error is:
status: badfile any ideas?


Before answering the above, currently I'm using VC++ to do this, if there is
another compiler that is easy to get up and running I'll gladly switch to it.
I'm just looking to get productive without pissing about with tools :)

RefuX

Have a look at November's list archive via

http://groups.yahoo.com/group/metamod/messages/8

There's a thread with subject "Several problems/questions". Two of my
mails have patches attached.

Ciao
  -Fritz

On Thu, 20 Dec 2001, refux wrote:

> 1. I opened up the metamod VC++ project file and compiled the metamod dll.
When I ran hlds with this metamod.dll I get the following assertion error:
> file: setnewh.cpp
> line: 52
> expression: new_p == 0
>
> 2. When I compile the trace plugin, metamod isn't able
> to load it. Under meta info the error is:
> status: badfile any ideas?
>
>
> Before answering the above, currently I'm using VC++ to do this, if there is
another compiler that is easy to get up and running I'll gladly switch to it.
> I'm just looking to get productive without pissing about with tools :)
>
> RefuX
>
>
>
> To unsubscribe, send an email to: metamod-unsubscribe@yahoogroups.com
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>

--
Fritz Elfert <felfert@...>               Thinking Objects Software GmbH
Lilienthalstr. 2                                  Phone: +49 711 88770 400
70825 Stuttgart                                     FAX: +49 711 88770 449
--------------------------------------------------------------------------

> There's a thread with subject "Several problems/questions". Two of my
> mails have patches attached.

Thanks for the help Fritz :)

Your email mention CVS, or rather the lack of it.
Has anything changed? does metamod reside in CVS now?

RefuX

On Thu, 20 Dec 2001, refux wrote:

[...]
> Your email mention CVS, or rather the lack of it.
> Has anything changed? does metamod reside in CVS now?
[...]

No, afaik unfortunately not.

Ciao
  -Fritz

--
Fritz Elfert <felfert@...>               Thinking Objects Software GmbH
Lilienthalstr. 2                                  Phone: +49 711 88770 400
70825 Stuttgart                                     FAX: +49 711 88770 449
--------------------------------------------------------------------------

Here is a question I post to you'll in the group.  I am not a coder,
or admin or even pretend to understand the finer parts of all the new
anti hacks out.  However, I question this new one....It kicks people
who have deleted most of the valid ogc files...or at least the ones
they thought make it work(curiosity is bound in the heart of a
child).  Yet now one Rocketboy is getting banned and really not
cheating after seeing what all the hub bub is about....ever consider
keeping or making a list of what files you check for so people can
self disinfect their hard drives be they reformed cheaters or just
curious?

Rocketboy

Ok here is what you need to do: Format your hard drive and re-install your
OS. That is the only way to get ridof all traces of OGC.

Good Luck

----- Original Message -----
From: "trident96.rm" <Trident96@...>
To: <metamod@yahoogroups.com>
Sent: Sunday, January 13, 2002 10:03 PM
Subject: [metamod] Reformed


> Here is a question I post to you'll in the group.  I am not a coder,
> or admin or even pretend to understand the finer parts of all the new
> anti hacks out.  However, I question this new one....It kicks people
> who have deleted most of the valid ogc files...or at least the ones
> they thought make it work(curiosity is bound in the heart of a
> child).  Yet now one Rocketboy is getting banned and really not
> cheating after seeing what all the hub bub is about....ever consider
> keeping or making a list of what files you check for so people can
> self disinfect their hard drives be they reformed cheaters or just
> curious?
>
> Rocketboy
>
>
>
> To unsubscribe, send an email to: metamod-unsubscribe@yahoogroups.com
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>