Eduke32 Changelog ViewerMake a couple of player functions take player indices instead of DukePlayer_t *.
All Entries
Guard retrieval of player index (.yvel) from APLAYER sprite.
It is unacceptable that yvel is on one hand modifiable without restriction from
scripting, but can be used as an array index without prior bound check in the C
code. Because that member has an overloaded meaning and is also used for
innocuous purposes such as the green color intensity of an SE light, it's
infeasible to restrict access from scripting. Consequently, we must add bound
checks on the C side. This is the first part of the effort to make .yvel safe,
adding two functions P_Get() and P_GetP(). There are a couple of other uses as
some kind of index.
Reserve 16 tiles from the end of MAXTILES, don't allow as user tiles from ART/DEF.
Lunatic: add example toggling forcefield wall to test.lua.
Don't consider W_FORCEFIELD+2 (surgery instruments) a forcefield picnum.
Clean up code related to forcefield player damage etc.
Rename 'camsprite' to 'g_curViewscreen', other related cleanup.
Add helper utility lunatic/util/cmpart.lua.
Lunatic: swap con.insetsprite's 4th and 5th positional arguments.
Lunatic: expose player[].gotweapon, document it and *_weapon() methods.
Minor stylistic tweaks of code related to weapon switching.
Factor some code in config.c into OSD_WriteAliases(), move symbol_t to osd.c.
Don't autoexec <mapname>.cfg files for clipshape maps.
Lunatic: don't allow starting demo recording, since gamevar diffing is NYI.
osd.c: patch up a possible oob access of osdtext[].
Lunatic: error on invalid static data read like 'sprite.picnum'.
Pass RETURN to EVENT_DAMAGEHPLANE in the 'dist' arg to allow chained callbacks.
Scripting: Add EVENT_DAMAGEHPLANE and document it in Lunatic doc Appendix B.
Also add test/damagehplane.lua with these demonstrations:
- custom breakable ceilings
- breakable TROR glass
Generalize Sect_DamageCeiling to Sect_DamageCeilingOrFloor, use appropriately.
- in the function itself: for floor, currently, do nothing (but this commit
is in preparation of the next one, again)
- In Proj_MaybeDamageCF(), Proj_MaybeDamageCF2() and the A_RadiusDamage() use
if the function, generalize to floors, but with the special case that
parallaxed floors keep blocking projectiles, as before (in constrast to
parallaxed ceilings). However, Sect_DamageCeilingOrFloor() is only
called for non-parallaxed ceilings *and* floors.
Minor no-functionality-changing tweaks.
- factor out: G_WallSpriteDist()
- factor out: Proj_MaybeDamageCF() and Proj_MaybeDamageCF2()
in preparation for the next commit
- Make PROJ_DECAYVELOCITY macro take an arg for readability's sake
sector.c: in ST9 handling code, prevent oob for improperly constructed doors.
Lunatic: fix a couple of spelling mistakes in the doc, found using Emacs ispell.
Lunatic: expose "con.shoot(tilenum, i [, zvel])". Document and add example.
Lunatic: add "for s, what in sectorsofbunch(bunchnum, gv.BOTH_CF)" iterator.
a.m32: don't clamp SE50 extra (=horiz) to [-500 500] with KP8/KP5.
LunaCON: statically check some member assignments with literal numbers.
And issue warnings whenever the assignment would fail (error) at game time.
Really fix the Pause key on Windows/SDL2.
This requires enabling the same workaround as for Linux. Maybe it's not
system-dependent at all.
Polymer: invalidate a wall when its sector's visibility has changed.
Do not consider system gamearrays for updating from demos.
This generally fixes demo playback. Before, tilesizx[]/tilesizy[] were written
into twice as many bytes as needed. Now, don't do that at all.
Lunatic: in fs.listpath() use FIND_NOCURDIR. Also document that function.
cache1d: add FIND_NOCURDIR option to klistpath().
Fix building stand-alone libcache1d.so.
defs.c: in check_tile(), pass 'tile' by value.
Guard a sprite/actor access with sector[].hitag as index from OOB.
Unlikely but possible for rogue CON code.
multivoc.c: use a local variable instead of a file-scope one used locally.
a-c.c: Make vector extensions also be compiled with Clang.
I was checking for GCC >= 4.7, while Clangs pretends to be GCC 4.2. Use a
feature test macro instead. The comment I made in r4161 regarding GCC vs.
Clang code was wrong. Now, Clang generates slightly faster code for these cases
(solid and masked 4-pixel wide vlines).
In P_HandleSharedKeys(), clean up weapon index bound checking.
In particular, (int32_t)j != -1 is replaced by (uint32_t)j < MAX_WEAPONS
to prevent a value >= MAX_WEAPONS to be assigned to p->wantweaponfire (it's
used as a weapon index in P_CheckWeapon().
Don't assign result of A_PlaySound() to p->scream_voice if the former is >127.
Because p->scream_voice is an int8_t.
This is an unlikely situation (126 voices ought to be playing simultaneously
while starting the "scream" sound), but still...
LunaCON: disable some player members, tweak decls of others.
The ones disabled are because they were found to not be used in any mod I
tested. Exposing members directly is rarely good API.
- bool toggle_key_flag -> const uint8_t ... (because value 2 is used)
- make scream_voice, cheat_phase read-only
Lunatic: slightly tweak the event/actor times printout when exiting.
LunaCON: fix assignment to projectile sound/bsound/isound.
Give rotatesprite 'orientation' bits enum names, too.
Use named enum constants for the HUD flags instead of literal numbers.
Polymodes: bump the allowed range for 'r_pr_hudangadd' to [-1024 .. 1024].
Polymost: handle HUD model 'fov' DEF token.
This "simply" requires an appropriate factor in the projection matrix.
Also, get rid of a special-case factor for >=1.6 aspects (making HUD models
wider then, I think).
A couple of trivial changes.
Classic: implement horiz. tex panning for sprites (spriteext[].xpanning).
Known as getactor/sprite[].xpanning from CON, values are 0-255. It should be
assumed that this is only valid for wall-aligned sprites (currently, face
sprites also pan).
A test is provided in lunatic/test/animatesprites.con. It should be run in E1L2
and only with the Lunatic build.
Amend r4142: make e.g. 'include a' (names shorter 4 chars) append ".m32".
DONT_BUILD.
Game: make underwater blocking floor sprites w/ xvel 1 NOT act like sprite bridge.
Likewise for blocking wall sprites. This makes it possible to construct
underwater sections containing complex spritework without losing the
"underwater-ness" when over such sprites, in a backward-compatible way.
Actually, the check is for bit 1 being set, but all other bits should be
considered reserved.
lunatic/util/foreachmap.lua: don't expand .xxx after ']'.
Add an example searching for all maps containing blocking floor-aligned
sprites in underwater sectors.