Subversion Repositories eduke32

Rev

Show changed files | Details | Compare with Previous | Blame | RSS feed

Filtering Options

Rev Age Author Path Log message Diff
5059 2127d 21h hendricks266 /polymer/eduke32/build/src/cache1d.c cache1d: Allow loading GRP files from within other GRP files.  
5058 2127d 21h hendricks266 /polymer/eduke32/build/src/cache1d.c cache1d: Enumify the magic constants for "file is in a zip" and "file is in the filesystem".  
4905 2198d 2h terminx /polymer/eduke32/build/src/cache1d.c Crash fix  
4886 2201d 12h hendricks266 /polymer/eduke32/build/src/cache1d.c Clean up autodetected game install search paths on all platforms that use them.

This should prevent Lunar Apocalypse from becoming Nuclear Winter with Megaton on Linux and Mac.
 
4838 2214d 10h helixhorned /polymer/eduke32/build/src/cache1d.c cache1d.c: Revert changes in r4665 unrelated to unrolling, fixing oob write.

Specifically, we must make sure that dasizeof <= LZWSIZE (or so) in
c1d_{read,write}_compressed() (what the assertion in the last commit was about,
and is now redundant). This fixes saving of e.g. large gamearrays with the C-CON
build. Bump BYTEVERSION.

An alternative for setting dasizeof to 1 would be to factor it and keep dividing
as long as it's > LZWSIZE.
 
4837 2214d 10h helixhorned /polymer/eduke32/build/src/cache1d.c cache1d.c: in (de)compr. funcs, add an bound-checking assert that fails w/ AMC TC.

Also, formatting.
 
4680 2277d 0h terminx /polymer/eduke32/build/src/cache1d.c Another 5000 lines of pain and tears. DONT_BUILD.  
4665 2277d 0h terminx /polymer/eduke32/build/src/cache1d.c Minor optimizations (loop unrolling) to LZW functions in cache1d, improves throughput ~20% on my i7  
4661 2277d 0h terminx /polymer/eduke32/build/src/cache1d.c Additional engine cleanups and optimizations. About as generic as it sounds. Also changes internal dummytile/tilefromtexture behavior so that the LZ4 compressed version of a texture is kept and the cache1d entries associated with the tiles can expire, like any other tile.  
4658 2277d 0h terminx /polymer/eduke32/build/src/cache1d.c Various additional optimizations: add CLASSIC_SLICE_BY_4 mode to unroll some of the loops in a-c, replace integer divisions by a divisor unknown at compile time with usage of libdivide, clean up pragmas further by removing more old stuff that wasn't used anywhere. This is another one of those nasty commits that make people cry. DONT_BUILD.  
4637 2301d 23h terminx /polymer/eduke32/build/src/cache1d.c Re-use temporary buffer for tilefromtexture true color -> paletted conversion and free when finished with the .def instead of freeing and re-allocating for every tile  
4543 2371d 20h hendricks266 /polymer/eduke32/build/src/cache1d.c Ball-busting Makefile restructure.

DONT_BUILD.
 
4502 2410d 19h hendricks266 /polymer/eduke32/build/src/cache1d.c Wrap most calls to exit() with a macro that will print the exit status, file name, line number, and function name to the log when using a debug build.  
4491 2425d 4h helixhorned /polymer/eduke32/build/src/cache1d.c Bye, bye, B*alloc(), all hail X*alloc()! Replace large portion of the calls...

... and cull code that is dead with the X*alloc() versions since they never
return NULL on requesting memory.

Use something like
git grep '[^Xx]\(m\|c\|re\)alloc *('
and
git grep '[^Xx]strdup *('
to see places where I left the B*alloc() calls intact.

BUILD_LUNATIC.
 
4298 2535d 13h helixhorned /polymer/eduke32/build/src/cache1d.c Add Makefile.common switch ALLOCACHE_AS_MALLOC, remove DEBUG_WITH_VALGRIND.  
4255 2562d 13h helixhorned /polymer/eduke32/build/src/cache1d.c Stylistic/formatting tweaks in cache1d code.  
4188 2604d 9h helixhorned /polymer/eduke32/build/src/cache1d.c cache1d: add FIND_NOCURDIR option to klistpath().  
4187 2604d 9h helixhorned /polymer/eduke32/build/src/cache1d.c Fix building stand-alone libcache1d.so.  
4175 2613d 8h helixhorned /polymer/eduke32/build/src/cache1d.c Give rotatesprite 'orientation' bits enum names, too.  
3880 2775d 10h helixhorned /polymer/eduke32/build/src/cache1d.c cache1d.c: cover up preceding assertion failure by bumping size of lzwrawbuf[].

It's hard to call this a fix, since ideally, buffer sizes like these would be
only as large as needed in the worst case.
 

Show All