[next] line.sidenum[i] doesn't return the correct sidedef number compared to 2.2.13

2.2.13:

next:

These two screenshots display the numbers of the sidedefs facing the player in the beginning sector of GFZ1. For some reason, next is returning 0 in some cases when it shouldn't.

Example WAD: SidenumTesting.wad

assigned to @LoganAir

after doing git bisect, I traced down when this issue first shows up, it started happen on commit 8fd9a82a in !2163 (merged)

I seem that 0xFFFF was replaced with 0xFFFFFFFF and sidenum was changed from UINT16 to UINT32 in src/r_defs.h

maybe that number is too big for Lua and gets converted to 0?

added Bug Lua labels

or you get back a zero if there is no backside?

see from 06b5c9d5/src/lua_maplib.c#L1121

    case line_frontside: // frontside
        LUA_PushUserdata(L, &sides[line->sidenum[0]], META_SIDE);
        return 1;
    case line_backside: // backside
        if (line->sidenum[1] == NO_SIDEDEF)
            return 0;
        LUA_PushUserdata(L, &sides[line->sidenum[1]], META_SIDE);
    return 1;

assigned to @Lactozilla and unassigned @LoganAir

assigned to @LoganAir and unassigned @Lactozilla

added Changed behavior label

removed Bug label

added Bug label

removed Changed behavior label

The problem here is actually because sidenum_get in lua_maplib.c wasn't updated to use UINT32 instead of UINT16, so line.sidenum[1] spits out 0 instead of the back sidedef number like it should be doing. #line.backside works just fine.

added Regression label

mentioned in merge request !2252 (merged)

closed

reopened

fixed by Monster Iestyn's merge request !2252 (merged)

closed

assigned to @MonsterIestyn and unassigned @LoganAir