NotAShelf/rogged
1
1
Fork 3
Code Issues Pull requests 1 Projects Releases Packages Wiki Activity Actions

movement: generalize; use vectors #16

Merged
NotAShelf merged 6 commits from SquirrelModeler/rogged:generalize-movement into main 2026-04-09 14:11:47 +00:00
Conversation 4 Commits 6 Files changed 9 +5 -5
1 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files
Showing only changes of commit 4044543fe1 - Show all commits

fix: move guard check

Prev Next 
I am stupid. I added the guard on the wrong check. Ffs...
Squirrel Modeller 2026-04-09 15:28:57 +02:00
commit 4044543fe1
No known key found for this signature in database
GPG key ID: C9FBA7B8C387BF70

10
src/main.c
View file

@ -446,14 +446,14 @@ static int handle_movement_input(GameState *gs) {
MoveResult result =
try_move_entity(&gs->player.position, direction, &gs->map, &gs->player, gs->enemies, gs->enemy_count, true);
if (result == MOVE_RESULT_MOVED) {
if (target != NULL) {
player_on_move(&gs->player);
}
player_on_move(&gs->player);
action = 1;
} else if (result == MOVE_RESULT_BLOCKED_ENEMY) {
SquirrelModeler marked this conversation as resolved Outdated
NotAShelf commented 2026-04-09 13:00:33 +00:00
Outdated
Copy link

You should probably guard against potential null target here before attacking.

Because if try_move_entity returns MOVE_RESULT_BLOCKED_ENEMY but player_find_enemy_at fails to find the enemy (be it due to differing lookup logic, or if the enemy was killed by effects earlier in the turn), target would be NULL, and thus lead to an ub in player_attack. Something like:

   } else if (result == MOVE_RESULT_BLOCKED_ENEMY) {
     target = player_find_enemy_at(gs->enemies, gs->enemy_count, new_x, new_y);
-    player_attack(&gs->player, target);
-    action = 1;
+    if (target != NULL) {
+      player_attack(&gs->player, target);
+      action = 1;
+    }
   }
You should probably guard against potential null target here before attacking. Because if `try_move_entity` returns `MOVE_RESULT_BLOCKED_ENEMY` but `player_find_enemy_at` fails to find the enemy (be it due to differing lookup logic, or if the enemy was killed by effects earlier in the turn), target would be `NULL`, and thus lead to an ub in `player_attack`. Something like: ```diff } else if (result == MOVE_RESULT_BLOCKED_ENEMY) { target = player_find_enemy_at(gs->enemies, gs->enemy_count, new_x, new_y); - player_attack(&gs->player, target); - action = 1; + if (target != NULL) { + player_attack(&gs->player, target); + action = 1; + } } ```
SquirrelModeler commented 2026-04-09 13:20:06 +00:00
Outdated
Copy link

If we move the action = 1 into the null guarded if statement, then enemy AI breaks. We can do:

if (target != NULL) {
  player_attack(&gs->player, target);
}
action = 1;

Which fixes that bug.

Ignore this. I added an incorrect guard.

~~If we move the action = 1 into the null guarded if statement, then enemy AI breaks. We can do:~~ ``` if (target != NULL) { player_attack(&gs->player, target); } action = 1; ``` ~~Which fixes that bug.~~ Ignore this. I added an incorrect guard.
target = player_find_enemy_at(gs->enemies, gs->enemy_count, new_x, new_y);
player_attack(&gs->player, target);
action = 1;
if (target != NULL) {
player_attack(&gs->player, target);
action = 1;
}
}
if (action)