Alloc Bloc Ful?
Confused
Wait. What? Join Date: 2003-01-28 Member: 12904Members, Constellation, NS2 Playtester, Squad Five Blue, Subnautica Playtester
![Confused](https://forumsdata.unknownworlds.com/uploads/userpics/293/nCQLKC0X9OW22.png)
<div class="IPBDescription">ah yes confusion reigns</div> on my most recent compile of my map after much structural change i had a leak and decideed to run teh map to find it but i get the error message alloc bloc full and get kick from the program. i have thought about using the optplanes tool on the map but havent yet because i am more concerened with findin said leak first.
i have compiled using wc a bacth compiler and tried using quarks find hole in map but cant find ****.
i think that was complex and rambiling enough to generate mas histeria so what the heck.
i have compiled using wc a bacth compiler and tried using quarks find hole in map but cant find ****.
i think that was complex and rambiling enough to generate mas histeria so what the heck.
Comments
Try the block method (putting hollow boxes around areas of the map and repeatedly compiling until it stops leaking, then mannually checking for leaks in the area you just boxed off).
your map (and hull file) gets too complex because of the leak.
the outer hull surface just needs too much memory.
when you find the leak your allockblockfull should be solved too.
<b>using pointfile when a leak causes an allockblockfull:</b>
you need to have an earlier "backup" .bsp file of your map that has no leak.
use this not leaked .bsp with the pointfile of the leaked .bsp (with renaming and copypasting to right directories)
You now have no allockblockfull because its the non leaked backup, but the pointfile shows a leak location of the actual leaked map.
use sv_cheats and noclip for sure.
<b>finding leaks when a leak causes an allockblockfull:</b>
Big block method ownz!
make a solid block around one half of your map and make binary trial and error search for the leak, big-blocking always one half of the map you dont know.
[b]most reasons for dificult leaks:[B]
- func_wall entities in outer hull (solved when you switch from "textured" mode to "filled" (in VHE) , brush entities should have other fill color than solid groups.
- point entities in the void. (use VHE entitie report wisely to check each point entity/ group entities, hide them and compile visible only)
- convex/concarve/non-flat surface. (this is an evil leak, caused by vertex manipulation-> one srface becomes non-flat and the compiler fixes this by moving vertexeds kinda randomly -> leak.)
- accidently moved a group of Solids. (I hate this leak. you just accidently moved a large group of brushes a tiny unit and cant find the leak because you search in the middle of the group and not at its border)
advantage: can be faster than big block, can find multiple leaks.
problems: little leaks hard to spot. useless for other problems (big block can be used to find almost any.)