Nickname SVN e nomi Git sono unificati per persona.
- Format CON crash output in VM_ScriptInfo() much more nicely.
- Execute VM_ScriptInfo() when DNDEBUG is used.
Comment ldcty[12] array from kpegrend() to remove last of set-but-unused warnings. I don't see any reason to leave this alone.
a.m32 light hotkeys: clamp hitag (range) to 0..32767 instead of 0..16000
make the 'palette' player member actually return the player's palette ID
instead of -1. Setting is still only through 'setgamepalette'.
Besides the performance issues, there's another reason why changing the palette
'frequently' isn't such a good idea. Assume an actor constantly changes the palette to some value and you're attacked by a newbeast at the same time. Then the pain tinting would be reset each time the actor's code is run. This commit adds a new flag 16 to setbrightness(), meaning "don't reset the fade tinting". Its only use right now is from CON's setgamepalette command (and there, always). The change is done in the engine and thus affects all platforms.
Trivial tweaks of setbrightness. Add an assertion that bits 1 and 4 are clear,
mainly. Also remove setvgapalette() from source.
The earlier changes introduced bugs when palettes were not set correctly in
certain situations (mostly cutscenes etc). This commit removes bit 1 from all flags that make it to setbrightness, the meaning of which is "don't actually update the palette". I have no idea what it was for and since the corresponding P_SetGamePalette() calls were from places like the mentioned cutscenes, I don't think it matters performance-wise.
SDL layer: don't call SDL_SetGammaRamp() if there would be no change.
This is in the similar vein as the set-palette deferring and it's this that fixes the low FPS in the radioactively contaminated area in AMC TC: Megabase (since setgamma() is called from setbrightness()). Again, update issues might be expected, but changing focus between EDuke32 and the desktop and back seems to restore the gamma settings properly on Kubuntu. Looks like X (or whatever above handles this stuff) maintains it on a per-application basis.
SDL layer: in 8-bit mode, defer setting the palette until showing the next frame.
CON code sometimes uses 'setgamepalette' from withing actors, which is a bit too frequent. Deferring the actual palette updating means that multiple requests in between two screen redraws are coalesced into one, which is desirable for performance reasons and for the reduction of 'tearing' artefacts (but not prevention, since the drawing isn't synced to the screen refresh rate with vsync off). (Did-not-)update issues might be introduced. Windows isn't touched because - I expect it to be more brittle, especially in light of the ATI HW gamma problems. - I haven't tested whether the particular performance penalty that lead to this change ('gas' in AMC TC) exists there too
make smoststart[] (which saves starting "smostcnt"s) an array of int32's
instead of shorts. This fixes their values' overflow in AMC TC megabase and probably a couple of other large maps.
fix bug in Net_ParseServerPacket introduced with SAMESIZE_ACTOR_T.
Fix sintable access with raw angle value (which might be outside [0..2047])
in overhead map drawing code.
patch up another out-of-bounds access in the voxel drawing code, though
I'm not sure whether the increased array size is enough.
savegame.c: fix initial setting of compression threshold on savegame load;
64-bit fixes
start CON code at script[3] instead of script[1]. The 'move' command permits
the constants 0 and 1, and moveptrs may reference the script up to index 2 then (though I'm not sure whether the code is reachable with moveofs either 0 or 1), so make sure it's nulled instead of whatever happened to compile at the first two positions. Move 1 is only used once in the original cons, when frozen and being shattered.
fix displaying the screenshot even for old-version savegames.
send BYTEVERSION as two bytes, high byte first.
Now we don't need to worry about the numbers running out soon. Bump BYTEVERSION by three to celebrate this, too.
Enable SAMESIZE_ACTOR_T described earlier.
Rewrite the savegame format in terms of the generic saving/loading system.
This makes savegames practically the same as the initial snapshot of a demo. Saves are now named 'dukesavX.esv' (demos: 'edemoX.edm'). Additionally, many changes that couldn't/needn't be cleanly separated are added with this commit: - make spriteext_t have the same size across 32/64 bit platforms (actor_t partially) - prevent saving/loading in MP games (it certainly didn't work and still doesn't) - it's time we start using assertions! Define NDEBUG for releasse builds. - reset savegame major and minor versions (we have a new magic string, so no conflict)
minor cosmetic tweaks in sector.[ch] and a off-by-one limit correction there
Fix out-of-bounds write when passing a map filename to Mapster32.
actor. (Needs SAMESIZE_ACTOR_T, which is not yet enabled)
misc. beautifying some game code...
some more clearbuf -> memset in gamedef.c plus add a forgotten keystatus
reset in the editor
Add new actor[].flags flag 2048 (SPRITE_NOCLIP), for 'spriteflags' CON usage.
Actors are always moved such that _they_ clip against blocking sprites, even if their blocking bit is clear. Setting the new bit make them not clip against anything. The promary use of this is for decorative moving sprites that are spawned in masses, such as rain or snow.
engine.c, in engine_addtsprite: if we can't add more tsprites because we
already have MAXSPRITESONSCREEN of them, return 1 to signal that the caller should break out of the loop.
Tweaks to compile the release build with clang. Now if it only would not
generate curious code like this one (Ubuntu 11.10's clang 2.9 on i386): 0x080a57ea <+538>: xor %al,%al 0x080a57ec <+540>: test %al,%al 0x080a57ee <+542>: jnz 0x80a57f2 <initgroupfile+546> => 0x080a57f0 <+544>: jmp 0x80a57f0 <initgroupfile+544> 0x080a57f2 <+546>: shl $0x4,%edi
Patch up out-of-bounds access in some voxel drawing code.
Don't spawn weapon tiles (WEAPONx_SPAWN) if any such value has been set
to negative numbers. Previously, only checks for being zero were performed, but CON code in the wild also has WEAPONx_SPAWN -1, which would propagate to the sprite picnum...
a.m32: new state 'print_parallel_midpts'
Rudimentary support for auto-aligning along TROR nextwalls.
Usage: press the ['] (quote) modifier together with the PERIOD key. Aligns only the immediate upper and lower neighbors of each wall that got aligned by following point2s or nextwalls. Doesn't play well with Alt (carry over xrepeat) yet, might need two subsequent alignment passes.
Fix out-of-bounds access in tile selector code in astub.c
checkdefs.sh script: hackish support for CON files, only 'include *.con' and
'definesound ... *.(voc|wav|ogg)' commands (missing: 'music', something else?).
Fix two yax_drawrooms() calls with too large sectnums (copy&paste error).
Fix a number of oob accesses with "switch switches".
Various code checked for switch tiles in the following manner, for (ii=0; ii<2; ii++) switch (DynamicTileMap[sprite[i].picnum-1+ii]) { case SOME_SWITCH: ... } which blows for picnum 0. Now, we simply disallow it.
Remove {eduke32,mapster32}.memmap files on clean/veryclean targets
(bioman request).
Fix negative array access in premap.c
Fix potential out-of-bounds array access in P_DisplayTip (happened with
WGR2 SVN, so not sure where there's a tipping animation there).
Fix sintable[] access with out-of-bounds values in core engine functions
(getzrange, hitscan, clipmove, drawmapview). They're reproducible with DNE1.3D; Since these angle values most likely come from CON code, my suggestion would be to bit-and all values representing angles with 2047 when setting the from CON, but I'd have to think about potential side-effects of doing this.
- Fix invalid polymer light array access originating from game's cleanup
callback 'G_Polymer_UnInit'. - In astub.c, reset spritelightid to -1 when nulling the spritelightptr of a sprite (no crashes observed, just precaution) - In sdlayer.c, catch SIGILLs with our signal handler too, since illegal instructions are what the instrumented debug builds will execute when attempting to carry out an undefined (per C) operation.
engine.c: fix shift by negative value in parascan().
Simple one-liner from Robin Green to fix a problem where the menu background tint would sometimes draw over the menu contents
a.m32: New function save_restore_hlsectors, also accessible from the Quote-F
menu. If some sectors are highlighted, save their indices. The highlighting of those particular sectors can then be restored at a later time (but careful across structural modifications that tweak sector indices) by running the same function. Shares a 'gamearray' with function 'collect_teleporting_sectors', so the latter will reset this one when run.
when failing to compile a state with a menu name, don't keep the
menu entry.
fix filling all sectors of a bunch with Ctrl-U. Previously, only
one sector was filled and the 'clear bunch' message was displayed even if the operation was canceled.
CON parser cleanup: make labelcode save offsets to CON code instead of
pointers and make it into an int32_t*.
CON parser cleanup: make labeltype be an int32_t* instead of an intptr_t*.
CON parser cleanup: when getting a value from labelcode[], assume that it
represents a pointer into the script for label types other than LABEL_DEFINE instead of checking inclusion of the value in the script bounds. This assertion was tested with WGR2 SVN.
Add the following commands to those which shouldn't be optimized away
with empty branches: ifcanseetarget, ifpdistl, ifpdistg, ifgotweapononce.
more Makefile tweaks; spriteext
Fix issue reported at http://forums.duke4.net/topic/5068-bug-introduced-in-old-versions-of-eduke32
When failing to extend a collection of sectors because one of
them has a different height, print the sectnums of two differing sectors to so that the mapper can find and examine them with Ctrl-J.
new cfg var 'glusememcache', the same as r_memcache cvar.
when having some sectors highlighted, also show how many walls
they contain in total.
Makefile bits for Windows.
