Jump to content

ensiform

Members
  • Posts

    1,626
  • Joined

  • Last visited

Everything posted by ensiform

  1. Look for UI_Alloc in ui_shared.cpp
  2. Try r_ext_compress_textures 0 in the console. And restart the game.
  3. Looks like you have a lot of other issues going on there too. Not sure what you are trying to do with "run" or what it is. cl_run is the auto run cvar. 1 being default to auto run afaik. +speed is the command to do the opposite (walk if auto run, or run if auto run is off) bind shift +speed
  4. You have r_ext_compress_textures enabled. Which is probably the cause.
  5. Hmmm, I don't think condump was added to SP by OpenJK... O.o I'll look. Edit: No it was not added by OpenJK, was in retail JASP exe. Did you try without using the '/' ?
  6. Because it just is the way it always was. Its not using the heap memory. This is a throwback to q3a which only used qvms and cannot allocate using malloc/calloc/alloca/new. Rather not go and change all of this for the time being.
  7. Well, something is using up a lot of the UI_Alloc memory to fill that 1 meg allocation pool. I'm going to bump it up to 2 megs to be the same as what MP currently is at for UI allocations. Maybe that will help. Update your repo to the master to see if bumping the limit to 2 megs helps you. Willing to bet its your species pack in combination with NPCSP that caused the issue.
  8. If you use the OpenJK engine, the console command which <filename> will tell you which search path it was loaded from.
  9. Whats the actual error? I don't have one that says ran out of ui memory. The one thing I notice when spawning NPCs is the console does say "unknown UI script ;" But still spawns anyway. @@therfiles
  10. No I want to see the full text log. Use the condump command. something like condump console.txt
  11. You'll have to wait until Monday unless I can get Cat to give me permissions to force build with upload again. The new build should be uploaded now but there seems to be an issue with the rsync process which I can do nothing about until Cat returns. @@Darth Futuza , perhaps you should upgrade your Atlantica as it was fixed like shortly after issue 271 was opened in 2013. http://jkhub.org/files/file/1561-atlantica-fix/ Which map is the first map?
  12. I never said HE personally had to do it =p. Someone from the community who maybe does, could, and release the hotfix pk3. And this was just a suggestion to avoid getting spammed in the future. Also, in theory they could/can use #s in their shadernames though which is one of two reasons why were against making the cop-out way. The other being someone decides to switch to *********** or ============= etc. Staying silent like most of the RP players have, is why none of their issues got fixed faster. FYI. I'm going to commit and push it for now since I haven't heard of any shaders that use # directly.
  13. Could you provide a console log from after the map loads? Use code tags or something.
  14. If they have different extensions they may not actually override. Files within pk3 namess that are at the bottom of the alphabet will have precedence over ones towards the top. Internally only one file of a specific name actually exists.
  15. Put all of the fixed shader files in a new pk3 that is like zzz-RPShaderFixes.pk3 The broken ones are no longer referenced. Its how the quake 3 virtual filesystem has always worked. Only one copy of a file is ever actually used when using the virtual filesystem. This is so you can release patches and so that assets3 contents replaces assets0 contents and 1 etc if 3 has something that was in any of the previous pk3s. -- http://slexy.org/view/s2Ft1VYYiz Here's the current patch, but I'm not sure if only checking for a single # is the best route.
  16. Rick, its primarily not a mod. Please stop calling it one. You personally wouldn't have to. In fact 1 pk3 with all the shader fixes could actually fix all of them. I would hardly call this lacking of backwards compatibility. Sure some _shaders_ are broken but the maps actually work, the game works. Regardless. Measures will be taken to allow your broken map shaders and skin/model shaders.
  17. If it did, it would be insignificant with todays computers. No FUD please.
  18. Vanilla never actually treated it as a comment, ever.
  19. They don't work in vanilla they just appear to do nothing, cluttering shader text. Which of course the users don't understand but ._.
  20. Pretty sure they just keep reusing someones old shader that has ########## in it, from Scziio or from cantina_mo.shader etc. Would rather leave it there, it says line ignored now as well. Scare people to pushing the map devs to fix it would be best. ###################### ////// Starport ////// ######################
  21. Yea but then you can end up with # randomly being skipped, at least in COM_Compress, no?
  22. How's this? Seems to work by not skipping the file anymore. http://slexy.org/view/s20xus2djt I need a map that for sure was still affected that isn't ord mantell rp to know for sure.
  23. But I still refuse to allow it on a global scale in all of COM_Parse.
  24. Technically it is a detriment because while it appears on the surface to do no harm it is allocating extra shaders(*2 for the hash) for potentially being there. Every time it hits one of those in your shader files. The warnings are were not put in place to protect against these types of incorrect shaders, they just happen to fall under the same category. There are a few types of shader syntax bugs that can cause the rest of the shader chain to literally break which is what this is there for. And why we really shouldn't revert totally back to archaic basic parsing with no syntax checking at all. The fact that it now displays the warning now shouldn't be that hard for someone to fix the shader files. Its not like you have to recompile each map. No that isn't hard just see above in this post as to why choosing not to for now.
×
×
  • Create New...