- Converting the wave speed back to floats, since I'm using sin() instead
of the finesine table now.
- Pre-shift the quake intensities to avoid extra shifting in GetModIntensity.
- Loading old saves will shift the intensities to fixed point but will not
convert the fixed wave speeds to floats. The chances of this
being in a savegame where the speeds are non-zero is pretty much nil,
and int and floating point 0 are bitwise identical (not counting -0.0).
- Squashed commit of the following:
commit bc45fe3263d34ef5f746f524687999c19bf7b779
Author: Randy Heit <rheit@users.noreply.github.com>
Date: Sun Mar 1 18:51:05 2015 -0600
wave scale -> wave speed
commit ff96388b128c724c1198757bfa52f1935a263356
Author: Randy Heit <rheit@users.noreply.github.com>
Date: Sun Mar 1 18:45:32 2015 -0600
More sine quake fixes
commit 2a89749a6fe6d271b9fbdc218779f680afcf4cb6
Merge: 719dfbe 5456074
Author: MajorCooke <paul.growney22@gmail.com>
Date: Sat Feb 28 20:37:22 2015 -0600
Added QF_WAVE to A_QuakeEx.
- Changes the random quakes into a sine wave (see Shadow Warrior/Rise of the Triad reboots, Hard Reset, etc.)
- Added 3 properties to control waves per second along each individual axis. Only works with QF_WAVE.
- Intensity X/Y/Z property becomes the amplitude of the wave.
- Stacks with regular quakes, allowing shaking along the camera which must be called using A_QuakeEx WITHOUT the flag, or the other quaking functions.
- Uses the youngest quake's time for positioning.
commit 54560741581e8d15cc7060e8e068cf85e9a4b432
Author: MajorCooke <paul.growney22@gmail.com>
Date: Sat Feb 28 20:21:19 2015 -0600
Recommitted recommended changes by Randi, with some modifications. Now, we should be finished!
commit 6f4473013411686d88fc185bdc1cc58b1035b0f1
Author: MajorCooke <paul.growney22@gmail.com>
Date: Sat Feb 28 12:52:57 2015 -0600
Finish this revert.
commit 467e53f9400f588a2ada9b32e7634cb1f4ad5066
Author: MajorCooke <paul.growney22@gmail.com>
Date: Sat Feb 28 12:46:02 2015 -0600
Reverted back to what was working.
commit da9de56a67efda08036e481fd5fccd5392ce6810
Author: MajorCooke <paul.growney22@gmail.com>
Date: Thu Feb 26 18:53:20 2015 -0600
Forgot this bit, for testing.
commit c5093d9bb97caf8478cefc32abc56a036feeea58
Author: MajorCooke <paul.growney22@gmail.com>
Date: Thu Feb 26 18:52:46 2015 -0600
Some more progress, but...
- This did not solve anything. In fact, it did the opposite -- completely broke wave quakes. Now they only happen whenever a random quake is in progress.
- Left in the commented code on purpose so Randi can test it.
commit 7e526405d2127cbb279f66008c8f8e55a5d497f3
Author: MajorCooke <paul.growney22@gmail.com>
Date: Wed Feb 25 17:50:42 2015 -0600
- Use newest waveform timer, not oldest.
commit 1356443609dbc6c7f46e081d0846816dc0836124
Author: MajorCooke <paul.growney22@gmail.com>
Date: Wed Feb 25 17:32:09 2015 -0600
- Got regular quakes to multiply onto sine quakes, but the vice versa needs fixing too.
commit d95796c94c70cd0229d4a6d30f69e3a7568b9588
Author: MajorCooke <paul.growney22@gmail.com>
Date: Wed Feb 25 16:46:21 2015 -0600
- Last hurdle. Now just need to figure out how to properly scale up and down.
commit 4bc3458e689155ce72c09776604d9eb4fa73d8be
Author: MajorCooke <paul.growney22@gmail.com>
Date: Tue Feb 24 23:18:03 2015 -0600
- Fixed the quakes being unstackable.
commit b51012d6d4ea065bf7f6fc9c1a0472966491f7af
Author: MajorCooke <paul.growney22@gmail.com>
Date: Mon Feb 23 23:48:34 2015 -0600
QF_WAVE renamed from SINE.
- Lots of ground covered, but still more to go.
- Still need to figure out how to make the camera properly shudder.
commit 427e4893193470bbf45415ffec70a0b69b8cccfd
Author: MajorCooke <paul.growney22@gmail.com>
Date: Sun Feb 22 16:52:30 2015 -0600
- Begin the groundworks for QF_SINE.
- Need to figure out how to rework and manipulate the sine wave to move faster, and to allow going below 0 without breaking it too much.
The 'Prologic' speakermode enumeration is removed since 4.40, not 4.44.
This makes possible to build it on linux x64, as the latest version of FMOD which has the dedicated linux64 download link is 4.42.
- The FMOD_DSP_TYPE_REVERB unit was removed in FMOD 4.36. It has been
replaced by a FMOD_DSP_TYPE_SFXREVERB unit, which is what should have
been used in the first place.
- Fixed: HasWeaponPiece assumed that once a piece was given, it would not
be removed. The easiest way to break this assumption is by loading an
earlier save where you had fewer pieces. (The catch being that if you
removed all pieces, it would notice that.)
- Quake scaling should be handled for each quake. The former
implementation was potentially taking flags from one quake and applying
them to another, which is broken.
- Remove all the int<->float conversions.
- This is RandomPick but for floats. Considering what RandomPick
translates into in the scripting branch, I hope I don't regret not using
type inference on RandomPick instead of creating a new keyword.
- When using both scaling flags, by default, the effect only reaches half peak going either way. This forces it to go all the way to the top (or bottom if using QF_MAX) before scaling back to its original height..
- QF_SCALEUP behaves like QF_SCALEDOWN: it gradually scales the tremors, only going upwards.
- QF_SCALEUP and QF_SCALEDOWN can be combined to make an earthquake that gradually smoothes in and out.
- QF_MAX can be used to invert this behavior, where it starts at the peak of the amplitude, fades out half way, and then grows back to maximum.
- This could cause problems in places like the GZDoom renderer in the odd circumstance, causing the camera to become stuck in the floor or ceiling until the quake expires.
- Relative quakes are different from other quakes; all quakes affecting
the camera do not become relative if one of them is relative.
- Use a single function call to get quake visual parameters instead of four.
- Thrust things in a psuedo-ellipse if they're inside a damaging quake whose
IntensityX != IntensityY.
- Don't break old savegames.
- Unlocks the full potential of using quakes, including the Z axis. Each intensity applies to X/Y/Z planes whenever a player is experiencing it.
- Flags:
- QF_RELATIVE - Adjusts the quaking of the camera to go in the direction its aiming (X: forward/backward. Y: Left/right.)
- Plans for including pitch will be implemented in the future for the Z axis with relativity.
Hexen uses the spawn height value from the mapthing_t structure to offset the item when floatbobbing. With proper gravity handling enabled this method doesn't really work so as a workaround ZDoom will now enable a hidden compatibility option when playing any map with a Hexen format MAPINFO so that any positive value in this field will make P_ZMovement revert to the original method of setting the items height (excluding the floatbob offset, of course.)
This also removes some code from P_NightmareRespawn that originate from the original Hexen method for floatbobbing which aren't needed anymore
- I'm assuming this check was here for a reason, but when both branches of
the if do the same thing and it's been this way since before recorded
history, it's not obvious what was intended here.
- "If it ain't broke, don't fix it."
- Some of the changes were downright wrong and some were pointless, so undo
everything that doesn't look like an actual improvement.
Added hud_showammo CVAR with three states:
* If value is 0, show ammo for current weapon only
* If value is 1, show ammo for available weapons
* If value is greater than 1, show ammo for all weapons
Default value is 2, so initial ammo display behavior isn't changed
- Fixed: If the menu cursor was on the topmost-displayed item, pressing up
would not scroll the view up. The check for scrolling only tested if the
newly selected item was the topmost one, since the menu code had
assumed the only time the cursor would be on the topmost visible line
was when it was the very first line of the menu. Using PgDn breaks this
assumption.
- ang_offset: the extra angle to apply to the actor after calculating the maximum turn.
- pitch_offset: Just like ang_offset, with pitch instead.
- FAF_BOTTOM: Aim straight for the actor's z coordinate, not with +32 units.
- FAF_MIDDLE: Aims for the direct middle of the actor.
- FAF_TOP: Aims for the very top of the actor.
- FAF_NODISTFACTOR: Use the raw pitch offset without calculating it into the distance of the aimed actor.
This reverts commit a4c07a9ee3.
- This commit touched every line of thingdef_codeptr.cpp, which would make
merging it into the scripting branch a pain in the ass.
This happened when the polydoor was open. If a mobj blocked the poly door, such that the door could not move from its open position, the poly door could rotate a little bit more than needed, making the door partially closing when the door managed to close later.
There are a few quite specific steps to reproduce this issue:
* 640x480 video resolution
* -iwad ... -warp ... command line parameters
* OS X 10.4 or 10.5 PowerPC, maybe performance related
When all these requirements are met, content view doesn't show up sometimes
The simplest solution for this issue is to set initial window size to non-existent video resolution
- The error "You cannot pass parameters to..." used the most recent token,
which was always ( and not the function name. (Note that this was
already fixed in the scripting branch, so this is probably going to be a
conflict. Meh.)
- Added A_Warp flags:
- WARPF_BOB: Gets the bob offset of actor pointer with the FLOATBOB flag.
- WARPF_MOVEPTR: The function is inversed to move the pointed actor with applied flags, but only the original caller will make the success/jump.
This feature helps a lot with buggy gamepads that constantly generate events from "sticky" hats/sticks
Polling is enabled by default, use joy_axespolling CVAR to turn it on/off
- Fixed: TicSpecial::CheckSpace() never thought it ran out of space due to
unsigned math.
- Fixed: TicSpecial::GetMoreSpace() assumed only relatively small amounts
of data would be written at a time so made no effort to ensure it
actually got enough space. This assumption could be violated by writing
a very long string, which can happen when replicating a string cvar.
- Rippers will rip through anything with an equivalent ripper level, or if their level is between or on the min and max ranges.
- If no min or max is defined, it simply checks if the monster's ripper level is lower than the missiles.
- Functions: A_SetRipperLevel(int level), A_SetRipMin(int min), A_SetRipMax(int max)
- Properties: RipperLevel, RipLevelMin, and RipLevelMax.
- RipperLevel: Applicable to monsters and projectiles.
- RipLevelMin and RipLevelMax are only useful on monsters.
- By default, all are 0.
- Explicitly size and position the text cursor in
FListMenuItem::DrawSelector() the same way that the options menu does it
using M_DrawConText(): By sizing it to a scaled 8x8 cell.
- Take advantage of the new _Pragma operator to hide the printf warning
suppression inside of macros instead of needing to litter the code
around Printfs with a bunch of junk.
- Fixed: When DSBarInfo::DrawGraphic() is used scaled, without fullscreen
offset, if one of the top and left clip locations was 0 and the other
was non-0, the 0 one would be clipped to the edge of a 4:3 box centered
on the screen instead of the edge of the screen.
- Fixed: Using +set cvarname and +cvarname on the command line would not
work if cvarname was defined in CVARINFO. This should be the proper way to fix
it. Rather than move all command line execution after loading CVARINFO,
keep command line execution before wads are loaded. If an attempt is
made to set an unknown cvar or to run an unknown command (which could
potentially be shorthand for setting an unknown cvar), save it and try
running it again after all CVARINFOs have been handled.
- Fixed: Script arrays didn't work in named scripts because the loader
read the script number as an unsigned word, hence it would never find
named scripts, since they are stored with negative numbers.
- Do a filter and species check first to save time.
- Added DMSS/KILS/RMVF_EITHER, which means if the actor is of type or species, it counts.
- A_DamageTarget(20,"Normal",DMSS_EITHER,"DoomImp","CyberdemonSpecies")
- This affects actor DoomImp, and anything that's of species CyberdemonSpecies.
- Added a little more documentation via comments.
- Added two more flags for each of the functions, EXFILTER and EXSPECIES.
- Stands for "exclude filter/species" and makes the function not take them into account.
- Cleaned up the code and placed all the checking in their own subfunctions.
- The wiki said the minimum distance to teleport defaults to 0, actor.txt on the other hand said otherwise. I was wondering why it was still broken somewhat...
- Prevent stickiness from happening, a.k.a. getting stuck in ceiling or floor and letting the engine unstick the actor. This caused velocity loss.
- TF_KEEPVELOCITY: Keep the velocity after teleporting.
- TF_KEEPANGLE: Don't use the special spot's angle.
- TF_USESPOTZ: Normally, this function
- TF_NOSRCFOG: Don't leave fog at the previous location.
- TF_NODESTFOG: Don't leave fog at the arriving location.
- TF_USEACTORFOG: Use the actor's TeleFogSourceType and TeleFogDestType properties.
- TF_NOJUMP: Don't require or cause a jump. In this case, put 0 (or "") in for the jump destination when using CustomInventory items to teleport actors around.
- Instead of reverting the teleport fog back to defaults, if there wasn't a class or if the class failed to be found, set it to NULL.
- P_SpawnTeleportFog will not spawn anything if it's NULL.
- Added "" so it can be used to mean 'don't change anything' for A_SetTeleFog.
- New properties include TeleFogSourceType and TeleFogDestType.
- TeleFogSourceType is the fog left behind where the actor teleported away from.
- TeleFogDestType is the fog the actor sees when it arrives at its destination.
- Added A_SetTeleFog(<oldpos>,<newpos>) -- oldpos sets TeleFogSourceType, newpos sets TeleFogDestType.
- CLOFF_SETTARGET | CLOFF_SETMASTER | CLOFF_SETTRACER
- An actor that causes A_CheckLOF (and only an actor) to succeed will set the intercepting actor as its target, master, and/or tracer, respectively.
- Added more resolutions to the hard coded table (up to 5K).
- Since we're using scaling for fullscreen, we should probably just use the hard coded table for fullscreen resolutions as well.
- Fixed: Resolution menu used fake aspect ratio to determine which aspect to file a resolution under.
- Added a way to use SDL surface blitting instead of accelerated textures.
1) Don't show OSX_COCOA_BACKEND option if the host is not OSX;
2) Don't use the '-msse' compiler flag in dumb/ if the architecture does not support it.
- Improved international keyboard support by falling back to scan codes if we don't recognize a keycode.
- Clear out any residual mouse movement when entering relative mouse mode.
- HITTARGET, HITMASTER, and HITTRACER.
- A missile that dies hitting an actor will set this impacted actor as the new target/master/tracer, depending on specifications.
- Added THRUREFLECT, MIRRORREFLECT, and AIMREFLECT. All require REFLECTIVE to work.
- THRUREFLECT prevents missiles from changing course if reflected, and takes precedence over all reflective flags.
- MIRRORREFLECT causes missiles to perform a direct 180 turn-around.
- AIMREFLECT turns the missile back to the original shooter, and does not slow the missile down.
Of course, just changing the name would break many old ZDoom mods so now Boom's name is explicitly being translated into ZDoom's before setting the string.
- FTF_REMOVE: Removes the actor when the alpha hits a certain level.
- - A_FadeIn - 1.0
- - A_FadeOut - 0.0
- - A_FadeTo - Alpha target level reached
- FTF_CLAMP: Automatically fixes the alpha so it won't leave the range [0.0, 1.0].
- Avoid doing the division if the result would be outside the ]0,1<<24[ range:
-- if the numerator is nonpositive, ie <=0, truncate the result to 0,
-- if the numerator is greater or equal than the denominator, the result will be outside the allowed range, hence truncate the result to 1<<24.
-- otherwise, the result will be inside the range. Knowing that the denominator is greater than the numerator, if the numerator has the last 24 bits non zero, the denominator can't be less than 1<<24, hence the denominator won't be truncated to 0.
- Add comment details to help who doesn't know math. Big deal!
-Fixed: ALLOWPAIN should not trigger pain states if the damage is 0, but still allow for infighting to occur.
-Fixed: an unneeded logic call was processing damage modification when it was 0 to begin with.
