Vegar

Can you try reproducing it with a more recent checkout? Preferably compiled with debugging symbols (-ggdb if you're using gcc).

Ok, I've reverted the fix.

If nobody have any objections, I would like to revert the commit referenced in the first post until a better solution is found. The fix is causing more problems than it's resolving, making the CVS client a pain to use.

Yes, this seems to be the source of the problems people have reported with cluster insides enabled. In portland storage it's the water tiles from the potion school that's causing the crash. Maybe someone that know the reflections a bit better than me can look further into it, as I'll most likely be busy all next week.

Schmurk suggested to disable CLUSTER_INSIDES, which fixed the crash. Now I see that at the same angle as the client would crash, the map background turns blue, like when water tiles enter the frustum. Edit: Update: Commenting the CLUSTER_INSIDES code in reflection.c seems to fix the crash. I'll dig further into it.

Try to disable new selection in the config window under the video tab.

Known bug, revert this change to fix it (remove the -1): http://cvs.berlios.de/cgi-bin/viewcvs.cgi/....72&r2=1.73 Edit: Oops, florian beat me to it

I'm still not able to reproduce it, even with sky enabled. Someone with windows have to look into it.

I'm not able to reproduce this. Do you only have DEBUG enabled? no other debug options? How do you close the client? I tried with both ctrl+q and clicking the X in the top right corner.

Is this the issue you're talking about? (taken at beam) Edit: I've found that reverting the following patch, which is extracted from the commit "Fixed some bugs and changed some values in the sky code.", cvs date "1 Apr 2008 21:20:32 -0000", fixes the above issue. Those that are having issues with latest CVS can patch -R this patch for now. http://developer.berlios.de/patch/download.php?id=2429

To get back to the topic - the reason an array is used is because the text is displayed in several text field widgets, and the use of an array simplifies things a lot (you just pass the array pointer to the widget). Using a queue, a vector, a linked list or anything else either requires a change to the text field widget (affects a lot of code) or a new widget that would be so similar to the text field widget that it just doesn't make any sense to do it. There is nothing wrong with using an array, it's just done wrong here. As it is now, everything is a big mess, I don't think anyone can honestly say they understand all of the logic around the text buffer. I suggest to keep the array, but rewrite most of the code around it. Instead of using an integer to track which message is the most recent (I think that's how it's done now), it would make a lot more sense to use memmove() to move old messages backwards in the buffer (similar to what labrat suggested).

This is a different bug (we had no water shaders enabled).

Add this to the include list: #include "../misc.h" I'll commit it soon.

I'm seeing this too with latest CVS. (Water shader is 0, and no framebuffer). I investigated this a bit further. Steps to reproduce: - It must be day ingame. Use DEBUG_TIME to achieve this if it's night at the server. - Go to Irinveron without any warm clothes, stand by the sea. - When hurt by the cold and you lose some health (red number over character's head), the flickering occurs. Reverting this commit (25-03-2008 14:53): "Added a stencil test for reflections when not using the framebuffer and updated the skybox code according to that." by Schmurk seems to fix the flickering. Unfortunately I don't have time right now to look further into what the commit does. Video card: GeForce4 Ti 4200/AGP/SSE/3DNOW! OpenGL Version: 1.5.3 NVIDIA 71.74 Edit: fix typo

Fixed in latest CVS.

Can you indent your code with tabs? There's a line in the patch that's indented with spaces.

I find the category order in the storage window to be unusable in its current, random, state. I would like to propose that the categories are ordered alphabetically. What do you think? (How do I make this a poll?)

error_log.txt Log started at 2008-03-06 19:11:01 localtime (CET) [19:11:01] Using the server profile: main [19:11:01] EXTENDED EXCEPTION(5:opengl_error): Error: 'line 190, column 1: error: too many instructions ' in file 'shaders/anim.vert' in load_vertex_program at actor_init.cpp (line 159) [19:11:01] Error opening sound configuration file [19:11:01] unknown shield property "missile" Last message repeated 9 times [19:11:01] unknown weapon property "CAL_range_in" [19:11:01] unknown weapon property "CAL_range_out" .... lots of lines similar to the above .... [19:11:04] * server www.eternal-lands.com filename http://www.eternal-lands.com/updates150/files.lst [19:11:04] Downloading http://www.eternal-lands.com/updates150/files.lst from www.eternal-lands.com [19:11:04] Finished downloading http://www.eternal-lands.com/updates150/files.lst [19:11:05] * server www.eternal-lands.com filename http://www.eternal-lands.com/updates/custom_files.lst [19:11:05] Downloading http://www.eternal-lands.com/updates/custom_files.lst from www.eternal-lands.com [19:11:05] Finished downloading http://www.eternal-lands.com/updates/custom_files.lst [19:11:10] Client closed infos.log Log started at 2008-03-06 19:11:01 localtime (CET) Init extensions. GL_MAX_TEXTURE_UNITS_ARB: 4 GL_MAX_TEXTURE_MAX_ANISOTROPY_EXT: 8.000000 Init extensions done Init eyecandy Init eyecandy done Loading vertex programs Max bones per mesh: 17 Init actor defs Init actor defs done Init lights Init done Init done! Video card: GeForce4 Ti 4200/AGP/SSE/3DNOW! Vendor ID: NVIDIA Corporation OpenGL Version: 1.5.3 NVIDIA 71.74

UVP: 0 Video card: GeForce4 Ti 4200/AGP/SSE/3DNOW! Vendor ID: NVIDIA Corporation OpenGL Version: 1.5.3 NVIDIA 71.74

I've seen this one too.

That is awesome bluap. Could it be used for copying/pasting text as well (in the text field widget)?

Further investigations using strace gives me this: socket(PF_FILE, SOCK_STREAM, 0) connect(6, {sa_family=AF_FILE, path="/tmp/.X11-unix/X0"}, 19) That's the socket the excessive select() calls are checking. Alia: It's not SDL_net that's causing the select()s in question. I've recompiled SDL_net with some extra printfs, and the timeout is always 100 (as passed in multiplayer.c).

This is from the SDL_net source, inside SDLNet_CheckSockets() (the file I found this in hasn't changed since 2004): /* Set up the timeout */ tv.tv_sec = timeout/1000; tv.tv_usec = (timeout%1000)*1000; /* Look! */ retval = select(maxfd+1, &mask, NULL, NULL, &tv); This is certainly not passing {0, 0}, and I have no idea what else could be calling select().