GitLab

The first is something that was missed when the hardcoded World Tour stuff was added, and the second looks like it was just a typo.

Fixes terminx/eduke32!40

Jordon Moss authored Aug 16, 2020 and Evan Ramos committed Aug 17, 2020

mapster32/build: Make AlignWalls_ not a partially broken, cryptic mess. Correctly aligns textures of any size now.

Jordon Moss authored Aug 11, 2020 and Evan Ramos committed Aug 17, 2020

engine: Implement tileflags member of picanm, and added support for new flag: TILEFLAGS_TRUENPOT, that allows you to enable true non-power-of-two texture support per-tile.

(Fixes classic's buggy tiling and panning behaviour and bypasses the emulation of it in Polymost)

Jordon Moss authored Aug 11, 2020 and Evan Ramos committed Aug 17, 2020

engine: Enable non-power-of-two wall support in classic renderer. (Doesn't do anything yet without either a new map format, or upcoming TILEFLAGS_TRUENPOT).

I needed to change the weird VS workaround in order for the EDUKE32_STATIC_ASSERT on line 850 to succeed.

Things seemed OK when I tested them, but I'd like more information on the suspected compiler bug in question and why this was necessary in the first place.

Also moves a bunch of VM_ASSERT statements up next to the variables they check.

When this macro was first introduced, all of an instruction's operands needed to be read from the bytecode before aborting due to failure. Since this isn't the case anymore, all of the VM_ASSERTs can be moved up.

These are actually fairly commonly used, so I'm not sure why I bothered with having them the other way before.

This is only ever used for refreshing pointers to certain special gamevars and for checking if variables in savegames exist in the currently running script--it doesn't need to print errors to the console.

Duke3d: get rid of separate check for common gamevar types before the switches in getvar__ and setvar__

It wasn't any faster.

I've heard several times that functions with a single exit point are faster, but profiling showed that removing the intermediate variable and returning directly from the switch wins here.

Also lowers MICROPROFILE_MAX_COUNTERS to 256.

At some point years ago I switched the input_t struct inside the player struct to use dynamic memory allocation for some reason. This changes that back--input_t is only 20 bytes and dynamically allocating it saves at most a couple hundred bytes of memory. Pointless.

Like DukePlayer_t, I do not intend to modify this structure again in a way that breaks backwards compatibility, unless 100% absolutely necessary.

Backported from NBlood.

Backported from NBlood.

Backported from Rednukem.

It was caused by the fix for crashing when firing outside the map.

Also port said fix to the bunny rockets.

NY00123 authored Aug 06, 2020 and Richard Gobeille committed Aug 10, 2020

the list as returned by ALSADrv_MIDI_ListPorts. Opening a device with
ALSA_PortID == 0 might technically be successful even if not present.