Lift raises as expected, then is triggered by a separate linedef to descend to the next lowest floor. It does this successfully and stops but then immediately descends to the next lowest floor again which breaks subsequent uses of that lift.

Sometimes you have to assign a new key to a keybind before Backspace allows you to clear out the current binding. Will see if this is somehow related to the SDL1->SDL2 transition, or if other factors are at play

So clang on -std=c++17 does not like the below:

p_user.cc:988:72: error: cannot pass object of non-trivial type 'epi::strent_c' through variadic function; call will abort at runtime [-Wnon-pod-varargs]
                I_Warning("WEAPON %s has no sprites and will not be added!\n", info->name);

For myself, I stubbed it as it isn't anything critical. Something to look at I suppose.

https://imgur.com/7lb2jX7

runs okay in v1.21

As support for other games (like Heretic) improves, it might be worth it to move the effect of various cheats outside of the C++ realm (except for the responder which needs to match the input string). This will allow for some of the more complex or unique cheats to be done via scripting in keeping with the spirit of customization/flexibility that EC has.

As reported by @atsb , the sky texture seems to be clamped starting at a vertical line in the middle of the player's viewpoint. I've attached a version of r_sky.cc with explicit GL_REPEAT flags added for testing.
r_sky.zip

Would also be nice if this could produce ES2 desktop artifacts

This would help reduce GL overhead in a similar manner as our IMAGE and TRUETYPE fonts

There is an initial degree of DDF parsing support for this option, which is essentially a DDF-ized method of doing something like TEXTURE1/2/PNAMES. Need to put it through its paces to see what remains to be done.

This is an example from DBP59. The entire area in black should also be flooded with the wood floor flat.

Stretch, Zoom, etc

[  6%] Building CXX object source_files/ddf/CMakeFiles/edge_ddf.dir/anim.cc.o
In file included from /Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:22:
In file included from /Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/local.h:22:
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/../epi/epi.h:48:37: error: expected function body after function declarator
void I_Error(const char *error,...) GCCATTR((format(printf, 1, 2)));
                                    ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/../epi/epi.h:49:41: error: expected function body after function declarator
void I_Warning(const char *warning,...) GCCATTR((format(printf, 1, 2)));
                                        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/../epi/epi.h:50:40: error: expected function body after function declarator
void I_Printf(const char *message,...) GCCATTR((format(printf, 1, 2)));
                                       ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/../epi/epi.h:51:40: error: expected function body after function declarator
void I_Debugf(const char *message,...) GCCATTR((format(printf, 1, 2)));
                                       ^
In file included from /Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:22:
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/local.h:242:42: error: expected function body after function declarator
void DDF_Error    (const char *err, ...) GCCATTR((format (printf,1,2)));
                                         ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/local.h:243:42: error: expected function body after function declarator
void DDF_Warning  (const char *err, ...) GCCATTR((format (printf,1,2)));
                                         ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/local.h:244:42: error: expected function body after function declarator
void DDF_WarnError(const char *err, ...) GCCATTR((format (printf,1,2)));
                                         ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:82:3: error: use of undeclared identifier 'DDF_WarnError'
                DDF_WarnError("New anim entry is missing a name!");
                ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:91:4: error: use of undeclared identifier 'DDF_Error'
                        DDF_Error("Unknown animdef to extend: %s\n", name);
                        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:120:2: error: use of undeclared identifier 'DDF_WarnError'
        DDF_WarnError("Unknown anims.ddf command: %s\n", field);
        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:127:3: error: use of undeclared identifier 'DDF_WarnError'
                DDF_WarnError("Bad TICS value for anim: %d\n", dynamic_anim->speed);
                ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:136:4: error: use of undeclared identifier 'DDF_Error'
                        DDF_Error("Missing animation sequence.\n");
                        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:140:4: error: use of undeclared identifier 'DDF_Error'
                        DDF_Error("TYPE=GRAPHIC animations must use the SEQUENCE command.\n");
                        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:203:2: error: use of undeclared identifier 'I_Error'
        SYS_ASSERT(storage);
        ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/../epi/asserts.h:41:20: note: expanded from macro 'SYS_ASSERT'
        ((cond)? (void)0 :I_Error("Assertion '%s' failed in %s (%s:%d).\n",  \
                          ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:215:3: error: use of undeclared identifier 'DDF_WarnError'
                DDF_WarnError("Unknown animation type: %s\n", info);
                ^
/Users/ats/Downloads/EDGE-classic-Preview-1/source_files/ddf/anim.cc:244:3: error: use of undeclared identifier 'I_Debugf'
                I_Debugf("- ANIMATED LUMP: start '%s' : end '%s'\n", first, last);
                ^
16 errors generated.

Clang Target:

Apple clang version 13.0.0 (clang-1300.0.29.3)
Target: x86_64-apple-darwin20.6.0
Thread model: posix

SDL2/GLEW etc.. are all correctly resolved and identified via CMake.

file_c::ReadText allocates a local std::string which is returned and then copied to a new string in assignment. It would be better to pass a std::string reference to receive the value

Hi, congrats for adding support for Boom displacement and accelerative scrollers.

However, with Boom wall scrollers 249, 218 and 254 the walls won't always move to the direction where they should.

249 scroll tagged wall when control sector height changes
218 accelerate tagged wall when control sector height changes
254 scroll tagged wall, the action line sets the speed and direction

Here's a test wad for Doom2 map01 with 249 scrollers. There are four switches to scroll the opposite walls. All sets are otherways identical except roteted by 45/90 degrees. All the walls should have the same scrolling result. The arrows indicate in which direction each part of the wall should move.

boom_249_scroller.zip

An other concern is that the walls desync. Although in this case the 249 lines are in a lift which will return to it's original height the walls do gain a small offset change in the process.

DOOMWADDIR and DOOMWADPATH support is not working, since the results of FS_ReadDir() are relative to the path it is given, but you don't do a FS_Join() to get the full path of those filenames.

I suggest updating FS_ReadDir() to always join the given path onto the results, since most usages of FS_ReadDir() need it anyway. The one case I noticed to fixed is CMD_Dir (in console code) would need a PATH_GetFilename() on the results to keep the output nice.

Another thing I suggest is treating both these environment vars in a uniform way, namely that once you convert DOOMWADPATH into a list of directories, simply add DOOMWADDIR to this list (assuming it exists).

Easier to understand by seeing in-action, but teleporting things appear to be visible as they are moved to their new location, causing it to have a "zipping" effect

Opening first door in Aliens Stranded get a reproducible crash in fpe_glDrawArrays, have not yet run under address sanitizing to see if can catch first bad access.

Callstack

libGLESv2.dll!00007ff845d88313() (Unknown Source:0)
libGLESv2.dll!00007ff845d5dc6a() (Unknown Source:0)
libGLESv2.dll!00007ff845af6add() (Unknown Source:0)
edge-classic.exe!fpe_glDrawArrays(unsigned int mode, int first, int count) Line 803 (c:\Dev\EDGE-classic\libraries\gl4es\src\gl\fpe.c:803)
edge-classic.exe!draw_renderlist(_renderlist_t * list) Line 801 (c:\Dev\EDGE-classic\libraries\gl4es\src\gl\listdraw.c:801)
edge-classic.exe!gl4es_flush(...) Line 1013 (c:\Dev\EDGE-classic\libraries\gl4es\src\gl\gl4es.c:1013)
edge-classic.exe!gl4es_glColorMask(unsigned char red, unsigned char green, unsigned char blue, unsigned char alpha) Line 1128 (c:\Dev\EDGE-classic\libraries\gl4es\src\gl\gl4es.c:1128)
edge-classic.exe!RGL_FinishSky() Line 715 (c:\Dev\EDGE-classic\source_files\edge\r_sky.cc:715)
edge-classic.exe!RGL_RenderTrueBSP() Line 3183 (c:\Dev\EDGE-classic\source_files\edge\r_render.cc:3183)
edge-classic.exe!R_Render(int x, int y, int w, int h, mobj_s * camera, bool full_height, float expand_w) Line 3379 (c:\Dev\EDGE-classic\source_files\edge\r_render.cc:3379)
edge-classic.exe!HUD_RenderWorld(float x, float y, float w, float h, mobj_s * camera, int flags) Line 1330 (c:\Dev\EDGE-classic\source_files\edge\hu_draw.cc:1330)
edge-classic.exe!HD_render_world(coal::vm_c * vm, int argc) Line 606 (c:\Dev\EDGE-classic\source_files\edge\vm_hud.cc:606)
edge-classic.exe!coal::real_vm_c::EnterNative(int func, int argc) Line 334 (c:\Dev\EDGE-classic\source_files\coal\c_execute.cc:334)```

So I've gotten another one and I'm starting to recompile this on macOS.

A few things:

• i_sdlinc.h - the current SDL2/SDL.h headers fail if using MacPorts instead of Homebrew. For MacPorts, I had to modify it like so:

#ifdef _MSC_VER
#include "SDL.h"
#elif __APPLE__
#include <SDL.h>
#else
#include <SDL2/SDL.h>
#endif

Since MacPorts points to the SDL2 headers inside the SDL2/ directory unless the -I parameter is modified.

• the dependent files (edge-defs, config files etc..) currently point either to a directory using -ewad etc.. or to a self-defined directory in the current location. For macOS bundles, this poses a problem as users cannot easily set the path via the terminal (unless they are technical users) and that the macOS bundle see's itself as 'the current root directory'.

To solve this, I suggest adding APPLE directives for SDL_GetPrefPath - this would place all these files into /Users/user/Library/Application Support/edge directory and would read/write them from here, allowing the bundle to be self-contained.

This could be used for various effects (pulsing smoke, etc) without needing to create specific images or DDFIMAGE entries with the sizes in question. Could also help in transformations that change the size of the thing that culminate in it BECOME()-ing something else.

[img]https://i.imgur.com/BAte3oP.png[/img]

when I try to run ec from UDB

resolved: IWAD was set to read-only
but then the question arises why the IWAD does not work when it is ready-onle
none of the other ports care, they just work

The screenshots show the problem and how it should look like as shown in a screenshot from doomretro

pic 2 is when standing on risen stair sector 43.

debug.txt

I have attached the debug log for a user who is encountering issues launching the game. The user had a clean install and confirmed that the files reported missing in the log were indeed present. Will update as progress is made.

If I set any non-integer mouse sensitivity in EDGE config, it rounds it to an integer. This makes it impossible to get anywhere near proper sens.

Check if this also needs to be done for JPG/TGA

This would allow for alternate firing/reloading/etc animations based on whether or not the weapon is zoomed in versus our very much static current zoom state.

Edge writes files and directories into the current binary folder, which is problematic for a number of reasons.

Using SDL_GetPrefPath could be a solution to this, with org and app names being configurable, preferably with some app metadata, though could be at compile time.

I am trying EC for the first time and built it on Linux Mint 21; then copied the files and directories mentioned in the wiki into /usr/games as /usr/games/edge-classic:

drwxr-xr-x 13 root root    4096 Aug 30 23:23 autoload
drwxr-xr-x  2 root root    4096 Aug 30 23:23 edge_base
-rwxr-xr-x  1 root root 4058704 Aug 30 23:23 edge-classic
-rw-r--r--  1 root root 1505324 Aug 30 23:23 edge-defs.wad
drwxr-xr-x  2 root root    4096 Aug 30 23:23 edge_fixes
drwxr-xr-x  2 root root    4096 Aug 30 23:23 soundfont

If I run the game from within that directory, specifying the iwad on the command line, it works fine; however, if I run the game from outside the directory, it fails to find edge-defs.wad and other files; it's as if it expects the game to be run from the game files directory and nowhere else; there seems to be no clean separation between where the engine and support files are located and where it might be run from.

Surely one should not, with any modern program, have to cd into the directory the executable is located in, in order to run it; especially as one might be loading wad files from a directory where they are kept?

This is not a major thing, but I think when searching for an IWAD to load, certain ones should be preferred, in particular DOOM2.WAD since most mods tend to use that IWAD as their base and may fail with DOOM.WAD (etc).

So instead of stopping at the first usable IWAD, give each one a score based on their name and pick the highest score. I suggest "DOOM2" should have the highest score, then maybe "FREEDOOM2" since it is highly compatible with it, then "TNT" and "PLUTONIA", etc...

This would help with liquid splashes, fireballs, etc, where the flat or sprite in question has changed in appearance from its vanilla form.

For example, if imp fireballs are recolored to green, the dynamic light would still be red/orange unless the DDF for the fireball attack were updated as well. If the dynamic light could be 'told' to use the fireball sprite's average color as a reference, then in theory the light color should be updated automagically.

Hi, the Boom displacement and accelerative scrollers (scroll/accelerate when control sector's height changes) do not always work right.

The examples from Boomedit.wad do work right. In those the control sector is always a lift; the scrolling or acceleration starts when the lift is lowered/scrolled back or deaccelerated when the lift returns back. But for other type of control sector movements the results are mixed.

Here's a test wad for Doom2 map01. It has ten scrolling tracks with different setups of control sectors. In all tracks when the A-switch is triggered the medikit should scroll away from the player. On displacement scrollers the srolling should stop when the control sector stops moving. One accelerative scrolers the scrolling should continue. When the B-switch is triggered the medikits should scroll back or the scrolling should deaccelerate and stop to a stand still.

https://www.dropbox.com/s/ckil8870o6m1jxp/scrollers.zip?dl=1

Not fully sure of this but it looks like that monsters spawned by the monster spawner ignore if the player has fired a weapon in the area.

Here's a test wad for Doom2 MAP01. The player is in a west side room and the monster spawner is in the east behind a window. The spawned monsters will appear in the middle and always face direction 0 to the east.
monsterspawner.zip

If the player shoots the pistol the monster spawner awakes. In vanilla Doom all the spawned monsters start chasing the player right from the start was player in sight or not. In Edge Classic the spawned monsters start to chase only if the player is in sight (in the moment of monster's appearance) or makes more noise. If the player is not in sight the spawned monsters just stand still.

If the monster spawner is awakened silently by player going to the window it works the same way both in vanilla Doom and in Edge Classic. When the player hides away from the window the spawned monsters stand still and won't wake up if the player comes back to the window.

With the Boom generalized floor and ceiling changers the numeric and trigger modes seem to be swapped. If the line trigger is set up to use the numeric mode it currently uses the trigger mode instead and vice versa. This happens with the Boom generalized actions only. Regular Doom/Boom actions such as in Doom E2M1 do use the correct modes.

However it also seems that all numeric mode changers take the sector behind the highest numbered 2s line as the target when it should be the sector behind the lowest numbered 2s line instead. Full explanation on lines 356-359 in here: https://github.com/fragglet/mbf/blob/master/boomref.txt?plain=1#L358

And if the tagged floor or ceiling is at the target height already the texture and type are not changed.

Here's a test wad for Doom2 map01.
floor+ceilingchangers.zip

It is assuming that it is a PWAD/-file parameter type of situation, so if another valid IWAD with a higher score is in one of the default search paths both IWADs will end up being loaded.

It would be nice for wads to be able to have a single lua script lump text lump, similar to COALHUDS

This is the last remaining item to be implemented to have full UMAPINFO v2.2 support.

Tracy is a profiler which works across platforms and rendering backends, without needing IDE integration

https://github.com/wolfpld/tracy

Hi,

I noticed you're using glatter (which is not compatible with Macs). Currently it is now impossible to build a Mac version.

I suppose this change is permanent?

So on Version 1.1, increasing the resolution slowly pushes the image off to the right side of the screen

Our current SDL_Joystick implementation for gamepads is getting a bit long in the tooth and is starting to show issues with more modern devices.

In addition to updating the functions behind the scenes, we will probably need a dedicated "live input mapping/test" subpage within the Options menu.

EC currently is not High-DPI/Scaling aware; resolution lists are either not created properly or display values inconsistent with what the user expects (i.e., if the OS scaling is set to 150%, the native resolution of the display will be listed as the native resolution divided by 1.5)

The 1.1 tag builds fine on Linux, but the master branch does not. I looked into why this is happening. Here's my make log for the failing git build: make.log

In the last few lines, I noticed that it said:

.../source_files/edge/umapinfo.cc:399:22: error: 'stricmp' was not declared in this scope; did you mean ‘strncmp’?

So, I just looked up "stricmp linux" (thinking maybe I need to install a dependency or something) and I see a lot of people with the same "not declared" problem. I think this is what is causing the issue. stricmp is apparently not POSIX.

Some obvious examples are Ancient Aliens MAP01 and Firerainbow MAP01. For those that preserve speed and the player is carried through by scrollers, they get an unwanted speed boost when exiting. The angle preservation is also weird.

See issue topic and example:

Leak in source_files/edge/p_spec.cc was line 1153 though probably different now post formatting
slope_plane_t *result = new slope_plane_t;

Would be fun but also in the pause betwen the intro graphic and the first title/music playing it would help avoid the impression that something is locked up or delayed