Files
libakgl/examples/sidescroller/player.c

451 lines
16 KiB
C
Raw Normal View History

Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
/**
* @file player.c
* @brief The player's two hooks and its control bindings.
*
* Two of the six behaviour hooks on `akgl_Actor` are replaced here, and the
* split between them is the frame's own order. `akgl_game_update` calls every
* actor's `updatefunc`, then steps the physics, then draws -- so per-frame game
* logic that is not movement (picking a coin up, falling in the pit) goes in
* `updatefunc`, and anything that has to happen inside the physics step goes in
* `movementlogicfunc`, which is the only hook the step calls.
*
* Two of the library's documented gaps are worked around here rather than
Move outstanding work from TODO.md into the issue tracker TODO.md carried two records in one file: what had been done, with the measurements behind it, and what was left. The second half is what a tracker is for, and keeping it here has already cost something -- AGENTS.md records a round where eleven entries described code that had already changed, and this file admitted to three more. Every open item is now an issue on source.starfort.tech/andrew/libakgl, labelled by kind and blast radius and milestoned by what it can land in: 0.9.x for anything that breaks no ABI, 0.10.0 for new or changed public symbols, 1.0.0 for the design work. Four are epics: the performance plan (#60), coverage (#61), actor rotation (#62), and the false header comments (#63). Verified against the tree before filing rather than transcribed. Three entries were already fixed and were not filed: the akgl_path_relative context leak, the akgl_draw_background test extension, and the SDL enumeration audit -- keyboards, gamepads and mappings are all freed in CLEANUP today. Two were reworded because the code had moved: the fonts item is a missing teardown entry point rather than a missing API, since akgl_text_unloadallfonts exists, and draw_world's tilemap call is already bounded by numlayers, so only the per-layer actor rescan remains. TODO.md keeps the part a tracker has no place for: why a decision went the way it did, what the measurement was, and which arguments turned out to be wrong. TODO.txt is deleted. Four of its eight entries had shipped -- actor-to-actor collision, actor-to-world collision, automatic facing, image layers -- and the four that had not are #74 through #77, with the GPU renderer's research links kept because that is the part that took the time. Every reference that named an item number or a moved section is repointed, in the manual, the headers, the tests and the examples. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 18:47:34 -04:00
* papered over. Both are issues #29 through #32.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*/
#include <math.h>
#include <akstdlib.h>
#include <akgl/controller.h>
#include <akgl/heap.h>
#include <akgl/physics.h>
#include <akgl/registry.h>
#include <akgl/util.h>
#include "sidescroller.h"
/** @brief The player's collision box, as an offset into its 32x32 sprite frame. */
static SDL_FRect ss_player_body = {
.x = SS_PLAYER_BOX_X,
.y = SS_PLAYER_BOX_Y,
.w = SS_PLAYER_BOX_W,
.h = SS_PLAYER_BOX_H
};
/** @brief Where the map put the player, so a death can put it back. */
static ss_ActorData ss_player_data;
/**
* @brief The rectangle an actor is tested against, given a 32x32 sprite frame.
*
* Smaller than the frame on purpose. Sprite art does not reach the edges, and a
* hazard box the full size of the frame kills a player who is visibly nowhere
* near it.
*/
static akerr_ErrorContext *hitbox(akgl_Actor *obj, float32_t inset, SDL_FRect *dest)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, dest, AKERR_NULLPOINTER, "dest");
dest->x = obj->x + inset;
dest->y = obj->y + inset;
dest->w = 32.0f - (inset * 2.0f);
dest->h = 32.0f - (inset * 2.0f);
SUCCEED_RETURN(errctx);
}
/**
* @brief Put the player back where the map placed it, at rest.
*
* Everything the simulation carries between steps has to be cleared, not just
* the position: `ey` is where gravity has been accumulating, and a player who
* respawns still holding a full-speed fall lands dead again immediately.
*/
static akerr_ErrorContext *respawn(akgl_Actor *obj)
{
ss_ActorData *data = NULL;
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, obj->actorData, AKERR_NULLPOINTER, "obj->actorData");
data = (ss_ActorData *)obj->actorData;
obj->x = data->home_x;
obj->y = data->home_y;
obj->ex = 0.0f;
obj->ey = 0.0f;
obj->tx = 0.0f;
obj->ty = 0.0f;
obj->vx = 0.0f;
obj->vy = 0.0f;
ss_game.deaths += 1;
SDL_Log("Player died (%d) and respawned at %f, %f", ss_game.deaths, obj->x, obj->y);
SUCCEED_RETURN(errctx);
}
/**
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
* @brief The player's `movementlogicfunc`: friction and the jump.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*
* Called by `akgl_physics_simulate` once per actor per step, *before* gravity,
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
* drag, the velocity recomputation and `move`. Terrain is no longer this
* function's problem -- the library resolves it after the move, which is the
* only place it can be resolved without predicting the arithmetic above.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*/
static akerr_ErrorContext *ss_player_movement(akgl_Actor *obj, float32_t dt)
{
float32_t friction = 0.0f;
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
/* The default logic still has to run: it is what copies the character's
* speeds onto the actor and signs its acceleration by the movement bits. */
PASS(errctx, akgl_actor_logic_movement(obj, dt));
/*
* Workaround 1: there is no friction anywhere in the arcade backend.
*
* akgl_actor_cmhf_left_off zeroes `tx` outright, so releasing a direction
* stops the actor dead inside one frame -- correct for a top-down Zelda,
* wrong for a sidescroller. This game binds its own `_off` handlers that
* leave `tx` alone (see below) and decays it here instead, faster on the
* ground than in the air. Below a pixel per second it is snapped to zero,
* because an exponential decay never actually arrives.
*/
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
/* Where the last step left the player, one pixel down. This runs before the
* step it is deciding about, so it is one frame stale by construction --
* which is one frame of coyote time, and not a thing a player can feel. */
PASS(errctx, ss_grounded(&obj->shape, obj->x, obj->y, &ss_game.grounded));
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
if ( AKGL_BITMASK_HASNOT(obj->state, AKGL_ACTOR_STATE_MOVING_LEFT) &&
AKGL_BITMASK_HASNOT(obj->state, AKGL_ACTOR_STATE_MOVING_RIGHT) ) {
friction = SS_FRICTION_AIR;
if ( ss_game.grounded == true ) {
friction = SS_FRICTION_GROUND;
}
obj->tx -= obj->tx * friction * dt;
if ( fabsf(obj->tx) < 1.0f ) {
obj->tx = 0.0f;
}
}
/*
* A jump is an impulse straight into the environmental term, because `ey` is
* the axis gravity accumulates on and the two have to cancel for the arc to
* come back down. Thrust would not: `ty` is capped against the character's
* `speed_y`, which is 0 for this character, so a jump written as thrust is
* scaled to nothing by the ellipse cap in akgl_physics_simulate.
*
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
* `ss_game.grounded` is the probe above.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*/
if ( (ss_game.jump_requested == true) && (ss_game.grounded == true) ) {
obj->ey = -SS_JUMP_SPEED;
}
ss_game.jump_requested = false;
/*
* The character maps MOVING_UP to the jump sprites, so the state word says
* "in the air" whether the actor is rising or falling. Nothing else reads
* the bit here: `speed_y` and `acceleration_y` are both 0 in
* character_ss_player.json, so the thrust it would authorise is capped to
* nothing.
*/
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
if ( ss_game.grounded == true ) {
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
AKGL_BITMASK_DEL(obj->state, AKGL_ACTOR_STATE_MOVING_UP);
} else {
AKGL_BITMASK_ADD(obj->state, AKGL_ACTOR_STATE_MOVING_UP);
}
SUCCEED_RETURN(errctx);
}
/**
* @brief The player's `updatefunc`: the animation, then what it is touching.
*
* Runs in `akgl_game_update`'s actor sweep, before the physics step and before
* anything is drawn.
*/
static akerr_ErrorContext *ss_player_update(akgl_Actor *obj)
{
SDL_FRect player;
SDL_FRect other;
bool hit = false;
int i = 0;
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
/* Facing, then the animation frame. Replacing a hook does not mean
* reimplementing it. */
PASS(errctx, akgl_actor_update(obj));
/* Off the bottom of the world. Nothing in the library stops an actor
* leaving the map -- akgl_physics_arcade_move does not clamp -- so falling
* out of the level is a thing the game has to notice for itself. */
if ( obj->y > (float32_t)(akgl_gamemap->height * akgl_gamemap->tileheight) ) {
PASS(errctx, respawn(obj));
SUCCEED_RETURN(errctx);
}
PASS(errctx, hitbox(obj, 8.0f, &player));
for ( i = 0; i < SS_COIN_COUNT; i++ ) {
if ( ss_game.coins[i] == NULL ) {
continue;
}
PASS(errctx, hitbox(ss_game.coins[i], 8.0f, &other));
PASS(errctx, akgl_collide_rectangles(&player, &other, &hit));
if ( hit == true ) {
/*
* Giving the pool slot back is what unregisters the actor and stops
* it being drawn; there is no "despawn" call. Releasing another
* actor from inside this sweep is safe -- akgl_game_update re-reads
* `refcount` at the top of every iteration and skips a slot that has
* gone free.
*/
PASS(errctx, akgl_heap_release_actor(ss_game.coins[i]));
ss_game.coins[i] = NULL;
ss_game.coins_taken += 1;
SDL_Log("Collected coin %d of %d", ss_game.coins_taken, SS_COIN_COUNT);
}
}
for ( i = 0; i < SS_HAZARD_COUNT; i++ ) {
if ( ss_game.hazards[i] == NULL ) {
continue;
}
PASS(errctx, hitbox(ss_game.hazards[i], 6.0f, &other));
PASS(errctx, akgl_collide_rectangles(&player, &other, &hit));
if ( hit == true ) {
PASS(errctx, respawn(obj));
SUCCEED_RETURN(errctx);
}
}
SUCCEED_RETURN(errctx);
}
/*
* Workaround 2, the release half.
*
* akgl_actor_cmhf_left_off and _right_off clear the movement bit, zero `ax`
* *and* zero `tx`. That last one is the "stops dead" behaviour; these two do
* everything else it does and leave `tx` for ss_player_movement to decay.
*/
static akerr_ErrorContext *ss_control_left_off(akgl_Actor *obj, SDL_Event *event)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, event, AKERR_NULLPOINTER, "event");
obj->ax = 0.0f;
AKGL_BITMASK_DEL(obj->state, AKGL_ACTOR_STATE_MOVING_LEFT);
SUCCEED_RETURN(errctx);
}
static akerr_ErrorContext *ss_control_right_off(akgl_Actor *obj, SDL_Event *event)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, event, AKERR_NULLPOINTER, "event");
obj->ax = 0.0f;
AKGL_BITMASK_DEL(obj->state, AKGL_ACTOR_STATE_MOVING_RIGHT);
SUCCEED_RETURN(errctx);
}
/** @brief Ask for a jump. Whether one is allowed is ss_player_movement's call. */
static akerr_ErrorContext *ss_control_jump_on(akgl_Actor *obj, SDL_Event *event)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, event, AKERR_NULLPOINTER, "event");
ss_game.jump_requested = true;
SUCCEED_RETURN(errctx);
}
/**
* @brief Cut a rising jump short when the button is let go.
*
* Variable jump height for four lines, and it works because `ey` is an ordinary
* field that nothing else owns between steps.
*/
static akerr_ErrorContext *ss_control_jump_off(akgl_Actor *obj, SDL_Event *event)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
FAIL_ZERO_RETURN(errctx, event, AKERR_NULLPOINTER, "event");
if ( obj->ey < 0.0f ) {
obj->ey *= 0.4f;
}
SUCCEED_RETURN(errctx);
}
akerr_ErrorContext *ss_player_bind(akgl_Actor *obj)
{
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, obj, AKERR_NULLPOINTER, "obj");
ss_game.player = obj;
Delete the sidescroller's collision, and let the library do it collision.c was 383 lines that re-implemented akgl_physics_simulate's own arithmetic, because movementlogicfunc runs before gravity, drag and the move and so had to predict where the actor would end up. The whole file goes. What replaces it is three lines in main.c, a shape on the player and the blob, and a `collidable` property on the map's terrain layer -- which also retires SS_TERRAIN_LAYER_ID, a compiled-in layer id that broke whenever somebody reordered layers in Tiled. grounded stays a probe rather than becoming a contact test. A contact says something pushed back this step; grounded asks whether there is a floor to push off, and that stays true through the frame after landing. The blob keeps its own wall and ledge probes for the same reason: the resolver does not say which side of the blob it was pushed from, and says nothing at all about a floor that is missing. The summary line now carries the player's resting position, because exiting 0 was not evidence that collision ran. It reports 136.0,160.0 grounded after 240 autoplay frames -- feet at y=192, flush on the surface of terrain row 12. The tutorial's collision section is rewritten around the library's, keeping the quarter-of-a-pixel story as what a predicting resolver costs. The docs harness learns `json excerpt=`, so the map property the chapter quotes is checked against the map. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KzBDV2fqgnUAcqCKqKvc71 Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 07:35:08 -04:00
/* The library's default response only stops an actor *entering* geometry;
* one that starts inside it is simply stuck, so lift the spawn point clear
* of the step it overlaps *before* recording it. */
PASS(errctx, akgl_collision_shape_box(&obj->shape, &ss_player_body, 0.0f));
obj->shape_override = true;
PASS(errctx, akgl_collision_settle(&ss_collision, &obj->shape, &obj->x, &obj->y, 0));
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
ss_player_data.home_x = obj->x;
ss_player_data.home_y = obj->y;
obj->actorData = (void *)&ss_player_data;
/*
* The default `facefunc` clears every facing bit and then sets one from the
* movement bits -- so an actor that stops moving is left facing nowhere, its
* state drops to bare ALIVE, and character_ss_player.json has no sprite for
* that. The actor becomes invisible while standing still.
*
* Clearing this is the documented way out: akgl_actor_automatic_face leaves
* an actor alone entirely when it is clear, and the facing bits stay
* wherever the control handlers last put them.
*/
obj->movement_controls_face = false;
/* Replace the hooks after akgl_actor_initialize, never before -- it
* overwrites all six. The tilemap loader has already run it here. */
obj->movementlogicfunc = &ss_player_movement;
obj->updatefunc = &ss_player_update;
SUCCEED_RETURN(errctx);
}
akerr_ErrorContext *ss_player_controls(int controlmapid, char *actorname)
{
akgl_ControlMap *controlmap = NULL;
akgl_Control control;
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, actorname, AKERR_NULLPOINTER, "actorname");
/* akgl_controller_pushmap checks the upper bound and not the lower one, so
* a negative id indexes before the start of akgl_controlmaps. TODO.md,
Move outstanding work from TODO.md into the issue tracker TODO.md carried two records in one file: what had been done, with the measurements behind it, and what was left. The second half is what a tracker is for, and keeping it here has already cost something -- AGENTS.md records a round where eleven entries described code that had already changed, and this file admitted to three more. Every open item is now an issue on source.starfort.tech/andrew/libakgl, labelled by kind and blast radius and milestoned by what it can land in: 0.9.x for anything that breaks no ABI, 0.10.0 for new or changed public symbols, 1.0.0 for the design work. Four are epics: the performance plan (#60), coverage (#61), actor rotation (#62), and the false header comments (#63). Verified against the tree before filing rather than transcribed. Three entries were already fixed and were not filed: the akgl_path_relative context leak, the akgl_draw_background test extension, and the SDL enumeration audit -- keyboards, gamepads and mappings are all freed in CLEANUP today. Two were reworded because the code had moved: the fonts item is a missing teardown entry point rather than a missing API, since akgl_text_unloadallfonts exists, and draw_world's tilemap call is already bounded by numlayers, so only the per-layer actor rescan remains. TODO.md keeps the part a tracker has no place for: why a decision went the way it did, what the measurement was, and which arguments turned out to be wrong. TODO.txt is deleted. Four of its eight entries had shipped -- actor-to-actor collision, actor-to-world collision, automatic facing, image layers -- and the four that had not are #74 through #77, with the GPU renderer's research links kept because that is the part that took the time. Every reference that named an item number or a moved section is repointed, in the manual, the headers, the tests and the examples. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-02 18:47:34 -04:00
* "Defects". */
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
FAIL_NONZERO_RETURN(
errctx,
((controlmapid < 0) || (controlmapid >= AKGL_MAX_CONTROL_MAPS)),
AKERR_OUTOFBOUNDS,
"Control map id %d is outside 0..%d",
controlmapid,
(AKGL_MAX_CONTROL_MAPS - 1)
);
PASS(errctx, aksl_memset((void *)&control, 0x00, sizeof(akgl_Control)));
controlmap = &akgl_controlmaps[controlmapid];
controlmap->target = SDL_GetPointerProperty(AKGL_REGISTRY_ACTOR, actorname, NULL);
FAIL_ZERO_RETURN(
errctx,
controlmap->target,
AKGL_ERR_REGISTRY,
"Actor %s is not in AKGL_REGISTRY_ACTOR; bind controls after the map is loaded",
actorname
);
/*
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
* A control map listens to one keyboard and one gamepad, matched by id, and
* 0 means "any". That is what a one-player game wants, and it is the only
* thing that works: SDL_GetKeyboards() reports what is attached, but the id
* a key event carries is chosen by the video backend, and on X11 it is
* SDL_GLOBAL_KEYBOARD_ID (0) without XInput2 and the physical device's
* sourceid with it -- never the SDL_DEFAULT_KEYBOARD_ID of 1 that SDL
* registers when XInput2 is missing. This function used to bind
* SDL_GetKeyboards()[0] and the arrow keys did nothing at all.
*
* Bind a specific id when two players share a machine on two keyboards.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*/
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
controlmap->kbid = 0;
controlmap->jsid = 0;
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
/* ---- keyboard ---- */
control.event_on = SDL_EVENT_KEY_DOWN;
control.event_off = SDL_EVENT_KEY_UP;
control.key = SDLK_LEFT;
control.handler_on = &akgl_actor_cmhf_left_on;
control.handler_off = &ss_control_left_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
control.key = SDLK_RIGHT;
control.handler_on = &akgl_actor_cmhf_right_on;
control.handler_off = &ss_control_right_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
control.key = SDLK_SPACE;
control.handler_on = &ss_control_jump_on;
control.handler_off = &ss_control_jump_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
/* ---- gamepad ---- */
/* Clear the keycode first: a keyboard event is matched on `key` whatever
* else the binding carries, and 0 is a keycode like any other. */
control.key = 0;
control.event_on = SDL_EVENT_GAMEPAD_BUTTON_DOWN;
control.event_off = SDL_EVENT_GAMEPAD_BUTTON_UP;
control.button = SDL_GAMEPAD_BUTTON_DPAD_LEFT;
control.handler_on = &akgl_actor_cmhf_left_on;
control.handler_off = &ss_control_left_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
control.button = SDL_GAMEPAD_BUTTON_DPAD_RIGHT;
control.handler_on = &akgl_actor_cmhf_right_on;
control.handler_off = &ss_control_right_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
control.button = SDL_GAMEPAD_BUTTON_SOUTH;
control.handler_on = &ss_control_jump_on;
control.handler_off = &ss_control_jump_off;
PASS(errctx, akgl_controller_pushmap(controlmapid, &control));
SDL_Log("Bound %s to keyboard %d and gamepad %d", actorname, controlmap->kbid, controlmap->jsid);
SUCCEED_RETURN(errctx);
}
akerr_ErrorContext *ss_player_autoplay(int frame)
{
SDL_Event synthetic;
PREPARE_ERROR(errctx);
FAIL_ZERO_RETURN(errctx, ss_game.player, AKERR_NULLPOINTER, "ss_game.player");
PASS(errctx, aksl_memset((void *)&synthetic, 0x00, sizeof(SDL_Event)));
/*
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
* Dispatch through akgl_controller_handle_event rather than calling the
* handlers directly, so a scripted run exercises the binding table the way a
* player does. The first version of this called the handlers, which meant
* the smoke test passed while the control map matched nothing at all and the
* arrow keys did nothing -- a test that cannot fail is not a test.
*
* SS_AUTOPLAY_KBID is deliberately not 0: the map binds keyboard 0 meaning
* "any", and driving it from a non-zero device id is what proves that.
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
*/
if ( frame == 1 ) {
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
synthetic.type = SDL_EVENT_KEY_DOWN;
synthetic.key.which = SS_AUTOPLAY_KBID;
synthetic.key.key = SDLK_RIGHT;
PASS(errctx, akgl_controller_handle_event((void *)&akgl_game.state, &synthetic));
FAIL_ZERO_RETURN(
errctx,
AKGL_BITMASK_HAS(ss_game.player->state, AKGL_ACTOR_STATE_MOVING_RIGHT),
AKGL_ERR_BEHAVIOR,
"the right arrow did not reach the player; the control map matched nothing"
);
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
}
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
if ( (frame % SS_AUTOPLAY_JUMP_PERIOD) == 0 ) {
Let a control map listen to any keyboard, and fix the dead arrow keys The sidescroller's controls did nothing. `akgl_controller_handle_event` matched `event->key.which == curmap->kbid` exactly, with no way to say "whatever keyboard the player is typing on", so the game did the obvious thing and bound `SDL_GetKeyboards()[0]`. That cannot work. The id a key event *carries* is chosen by the video backend and is not required to be an id `SDL_GetKeyboards()` reports. On X11 without XInput2 every key event carries `SDL_GLOBAL_KEYBOARD_ID`, which is 0, while `SDL_AddKeyboard` registers the attached keyboard as `SDL_DEFAULT_KEYBOARD_ID`, which is 1; with XInput2 the events carry the physical slave device's `sourceid` while `SDL_GetKeyboards()[0]` may be the master. Either way the map matched nothing and every key press was dropped. A `kbid` or `jsid` of 0 now matches any device of that kind. 0 is safe to spend: SDL documents `which` as 0 when the source is unknown or virtual, and joystick ids start at 1, so no real device is 0. A non-zero id still matches only that device, which is what keeps two local players on two keyboards apart, and there is a test for that half too. `util/charviewer.c` already passed 0 and depended on the old behaviour by accident; it works under XInput2 now as well. Two reasons this shipped, both closed: - Every test in tests/controller.c dispatched an event whose id equalled the id the map was bound with, so none of them could see it. - The sidescroller's smoke test called the control handlers directly instead of dispatching events, so it passed against a control map that matched nothing. It now pushes synthetic events through akgl_controller_handle_event from a deliberately non-zero device id and fails if the press does not arrive. Verified by breaking the binding and watching the run exit non-zero. `test_controller_wildcard_device_ids` was written first and failed against the unfixed library with "a map bound to keyboard 0 ignored a key press from device 11", which is the whole reason to trust it now that it passes. The controls were documented, in one sentence. Chapter 19 has a proper table of them, and chapter 15 explains why not to reach for SDL_GetKeyboards(). The header said the match was exact and now says what it does. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 21:23:02 -04:00
synthetic.type = SDL_EVENT_KEY_DOWN;
synthetic.key.which = SS_AUTOPLAY_KBID;
synthetic.key.key = SDLK_SPACE;
PASS(errctx, akgl_controller_handle_event((void *)&akgl_game.state, &synthetic));
} else if ( (frame % SS_AUTOPLAY_JUMP_PERIOD) == SS_AUTOPLAY_JUMP_HOLD ) {
synthetic.type = SDL_EVENT_KEY_UP;
synthetic.key.which = SS_AUTOPLAY_KBID;
synthetic.key.key = SDLK_SPACE;
PASS(errctx, akgl_controller_handle_event((void *)&akgl_game.state, &synthetic));
Add two tutorial games that build, run in CI, and cannot drift examples/sidescroller and examples/jrpg are complete programs, built with the library and exercised headless by ctest. The chapters quote them with `c excerpt=examples/...` blocks rather than restating the code, so a chapter cannot drift from a program that compiles -- the excerpt check fails the moment the source moves. 34 excerpts in one chapter, 21 in the other. The two are complementary. The sidescroller is the physics tutorial: gravity, a jump, coins, hazards. The JRPG is the content-pipeline tutorial: a town map, NPCs spawned from map objects, four-way per-facing animation, a text box, a follower. Both smoke tests drive real SDL_Events through akgl_controller_handle_event and step the physics clock at a fixed 1/60s rather than sleeping, so a scripted run is deterministic and finishes in under five seconds. Writing them is what turned up most of the defects recorded in the next commit, because a game exercises paths a unit test does not. Each workaround says in the chapter which library gap forced it: - collision is written in a custom movementlogicfunc, because akgl_physics_arcade_collide raises AKERR_API and akgl_physics_simulate never calls collide at all; - the sidescroller cancels the step's own gravity when it blocks downward, because otherwise a quarter-pixel of penetration makes the *horizontal* sweep report blocked and the character walks backwards a tile at a time; - both clear movement_controls_face on every map-spawned actor, because the default facefunc leaves a stopped actor with no facing bit, no sprite, and no draw; - the JRPG's follower gets a renderfunc that nulls obj->parent for the duration of the draw, because a child's offset is counted twice. Assets are CC0 from three Kenney packs, vendored with per-pack licence text, per-file provenance, and the geometry contract in docs/tutorials/assets/README.md. CC0 specifically rather than merely free: a reader who copies a tutorial into their own game inherits no obligation. scripts/fetch_tutorial_assets.sh refreshes them in the shape mkcontrollermappings.sh was fixed into for 0.5.0 -- it checks curl's status, refuses a pack page that does not say CC0, verifies the archive and the staged dimensions, and leaves the tracked bytes untouched on any failure. Both failure paths were tested, and a no-op refresh is byte-identical. Co-Authored-By: Claude Code <noreply@anthropic.com> Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
2026-08-01 20:59:00 -04:00
}
SUCCEED_RETURN(errctx);
}