I found and fixed a buffer overflow vulnerability in version 1.05's download feature. It shouldn't be too bad as it doesn't affect the stack (but static memory). Still it is strongly advised that you do upgrade.
This version also introduces some fixes to the renderer by marky and features an implementation of the raw input interface that is only experimental. Don't expect things to work necessarily the way they should. In that regard feedback is much welcome. Try setting in_mouse to 3.
Download links:
Full changelog: http://q3defrag.org/files/dfengine/history.txt
Comments
good work ;)
and i working on graphical mod for q3. We chandge only render, don't chandge other parts of game
http://files.mail.ru/K74JDI
last version of our mod + retextur ;)
press "скачать"
Show me some ingame screenshots. irc #2337 or icq 96512301
Q3Bers
ok, some screens here:
http://q3bers.nm.ru/screenshots/screenshots/
;)
our forum is broken now :(
http://forums.quakegate.ru/index.php/topic,429.364.html
I suppose this one is a bug
OK so I like having my gun drawn BUT adjusted to the low middle of the screen so I use these in my config (just an example):
seta cg_gunx "-1"
seta cg_guny "3"
seta cg_gunz "-9"
As a result I get this in the game:
http://s46.radikal.ru/i112/0905/fe/8d4bef1bb28f.jpg
cg_gunx, for instance, does actually work as shown here:
http://s51.radikal.ru/i134/0905/2a/883063296bf1.jpg (the difference from the picture above is obvious)
After using map_restart, gun* position is reseted to default and seems you can't change it since gun* will be "cheat protected" as shown here:
http://i075.radikal.ru/0905/cb/74d37510d0c5.jpg
Although one can always disable gun drawing I think I'm not the only one who likes gun position customized a bit :>
Is there any way to fix it?
Thanks in advance.
Map_restart has no effect at all for me. It'll reset when sv_cheats are set from 1 to 0.
The cg_ cvars are modspecific anyways and have nothing to do with the dfengine but defrag itself.
Removing the cheat protection in the mod itself takes not more than 10 seconds of coding time for cgg =)
cg_drawgun 0 is the best anyways =)
Works for me, it's done.
Works for me, it's done.
would be graet !
would be graet !
64bit binary
Can you please make a 64bit binary for linux?
I can launch it only after typing ~10 commands at bash, because i have no generic x86 support in kernel.
When I try to compile it i get lot of errors.
would be graet!
would be graet!
Cheat protection for +left and +right
+left and +right aren't used for anything except scripting. Why not make them cheat protected? It would reduce most of the cheating in Defrag, and the accusations of innocent players like me :)
It's not bannable because it's not mod specific.
Also it's not a cheat at all. It becomes a gain when used with varmath stuff e.g. which is why I suggested to delete the whole varmath, varcycle, etc stuff more than a half year ago.
True, those commands are part
True, those commands are part of the engine - such restrictions cannot be enforced from within defrag. I had a talk recently about cheat prevention and scripting. I don't intend to add restrictions on the varmath features because it would simply deport the problem onto people making custom code for the engine - while I believe those features are still generally helpful to players and movie makers. Reproducing varmath-like behaviors with the engine code is dead easy for anyone willing. Cheat prevention has to be enforced on the server level - people cannot be refrained from making custom versions of the client.
I have a simple idea to
I have a simple idea to detect cheaters, it would increase traffic slightly so perhaps it should be an admin cmd.
pm me @ irc
time not saved
I had the error that my new best time wasn't saved when I restarted the map right after finishing it. I got this only once. DFE 1.05 and newest DFversion.
Demo-saving stuff isn't (df)engine related.
fine
fine
And it might have occured due to a too high df_ar_stopdelay
Its set to 1800. What is the
Its set to 1800. What is the value standing for? milliseconds?
Yes
Demo startup
Why are demos started with a single click? It's really annoying, since you start a demo by mistake almost every time.
it's fixed.
it's fixed.
defrag demo playback
Hi, since i installed the latest patch, i cant watch dfwc2008 demos :(
what version of defrag do you
what version of defrag do you use?
what error message (if any) do you get?
When i change the volume with
When i change the volume with my keyboard (logitech g15) the weapon changes to gauntlet. I had this bug since ages but it just came to my mind and its not a big deal but anyway.
Well if you mind providing
Well if you mind providing the hardware I don't mind going for a debug :)
I can send you a photo ;D
I can send you a photo
;D
Had no problems with in_mouse
Had no problems with in_mouse 3 on vista 64, it's working great, cheers for all yoru work so far cgg! Since updating from 1.03 though (missed out the last couple of updates I think) half the menu buttons don't work, e.g. I can only change game physics from the console, not in the menu, and the back button doesn't work (have to use escape).
Sounds weird, too. I don't
Sounds weird, too. I don't see what in the changelog could cause this. Have you tried loading another mod and playing with the menus?
Yeah, using the latest
Yeah, using the latest dfengine most menu buttons don't work in any mod. Just tried with 1.05 though, everything works fine, so the bug has come in with the most recent changes.
...while typing just thought of something, tested and I was right - this problem only occurs when in_mouse 3 is set, any other mouse mode and menus work fine, so something for you to fix in next update =). In game in_mouse 3 is working fine for me though.
ah I misread your original
ah I misread your original post. Yeah nice catch - happens consistently on my setup as well. That one escaped me.
- fixed demo playback failing
- fixed demo playback failing to start when a local server is up
Woot! :D
In regards to testing raw input, in windowed mode the xhair can escape the window.
Some .h files seems to be
Some .h files seems to be missing in the source since I'm getting a bunch of undeclared identifiers errors etc when trying to compile it
mind to make a dump of said
mind to make a dump of said errors?
Make sure you use the msvc project files that are provided in the package.
Sure...
1>------ Build started: Project: quake3, Configuration: Release Win32 ------
1>Compiling...
1>win_wndproc.c
1>.\win32\win_wndproc.c(425) : error C2065: 'WM_INPUT' : undeclared identifier
1>.\win32\win_wndproc.c(425) : error C2051: case expression not constant
1>.\win32\win_wndproc.c(426) : error C2065: 'RAWINPUT' : undeclared identifier
1>.\win32\win_wndproc.c(426) : error C2146: syntax error : missing ';' before identifier 'ri'
1>.\win32\win_wndproc.c(426) : error C2065: 'ri' : undeclared identifier
1>.\win32\win_wndproc.c(427) : error C2143: syntax error : missing ';' before 'type'
1>.\win32\win_wndproc.c(431) : warning C4013: 'GetRawInputData' undefined; assuming extern returning int
1>.\win32\win_wndproc.c(431) : error C2065: 'HRAWINPUT' : undeclared identifier
1>.\win32\win_wndproc.c(431) : error C2146: syntax error : missing ')' before identifier 'lParam'
1>.\win32\win_wndproc.c(433) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(435) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(439) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(439) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(440) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(440) : error C2224: left of '.data' must have struct/union type
1>.\win32\win_wndproc.c(440) : warning C4047: 'function' : 'int' differs in levels of indirection from 'void *'
1>.\win32\win_wndproc.c(440) : warning C4024: 'Sys_QueEvent' : different types for formal and actual parameter 4
1>.\win32\win_wndproc.c(440) : error C2198: 'Sys_QueEvent' : too few arguments for call
1>win_input.c
1>.\win32\win_input.c(198) : error C2065: 'RAWINPUTDEVICE' : undeclared identifier
1>.\win32\win_input.c(198) : error C2146: syntax error : missing ';' before identifier 'dev'
1>.\win32\win_input.c(198) : error C2065: 'dev' : undeclared identifier
1>.\win32\win_input.c(198) : error C2059: syntax error : '{'
1>.\win32\win_input.c(204) : warning C4013: 'RegisterRawInputDevices' undefined; assuming extern returning int
1>.\win32\win_input.c(214) : error C2146: syntax error : missing ';' before identifier 'dev'
1>.\win32\win_input.c(214) : error C2059: syntax error : '{'
1>Generating Code...
1>Build log was saved at "file://d:\q3\dfengine_1.06_src\code\Release\BuildLog.htm"
1>quake3 - 21 error(s), 4 warning(s)
========== Build: 0 succeeded, 1 failed, 3 up-to-date, 0 skipped ==========
whoa. Yeah I guess that makes
whoa. Yeah I guess that makes sense if you're using a version of windows prior to xp. Is it the case ?
_WIN32_WINNT should be set to
_WIN32_WINNT should be set to at least 0x0501 (xp)
Hey,
I'm using XP actually (SP3)... :)
greetz
That's weird. The input code
That's weird. The input code calls SetCapture(), which as far as I can make out should prevent the cursor from escaping the window - at least it works that way on my setup.
Which version of windows do you have ?
I'm on WinXP SP 3 (32 bit).
I'm on WinXP SP 3 (32 bit). It's somewhat strange since I have the same kind of bug in my code but it tends to escape if I only move upwards and do +attack at the same time. Honestly, I figured a call to IN_ActivateWin32Mouse would cause it to stop escaping in Windowed mode but it seems I was wrong.
I remember you mentioned this
I remember you mentioned this when we talked about raw input.
No luck in trying to reproduce the issue. What sounds weird is that SetCapture() should be a standard, setup independent win32 call - or at least so I figure. I wonder if the issue is going to show up again. There are workarounds to this.
Maybe check if X and Y are
Maybe check if X and Y are out of bounds and just clamp it to the edge?
or maybe just center the
or maybe just center the cursor every frame
Actually did something
Actually did something similar to this and got it working, calling Activatewin32Mouse multiple times however, heh
Sorry for double post
I should clarify that its not actually holding +attack and moving the mouse up but rather moving crosshair upwards and then +attack'ing that causes it.