razish / jasdk_modbase Goto Github PK

***DISCONTINUED, See Razish/OpenJK*** Yet another "fixed" or "clean" version of the Jedi Academy SDK, which people can use as a stable mod base

C++ 2.40% C 92.58% Objective-C 5.02%

jasdk_modbase's People

Contributors

Stargazers

Watchers

Forkers

redsaurus mrwonko blaenk stoiss wookiee- quartznets

jasdk_modbase's Issues

Convert cg_consolecmds.c, cg_servercmds.c, g_svcmds.c, ui_atoms.c to use bsearch for commands

Yep.

bg/pmove compat issues

apparently something related to weapon pickups and ammo count was different in base.

Client 0 Logic in CheckItemCanBePickedUpByNPC

        && item->activator != &g_entities[0] 
        && pickerupper->s.number

Plz 2 fix.

force regen bug: saber throw

related to the g_forceRegenTime framerate-independent fix

throw saber, takes away 50 force points
saber returns, all your force replenishes.

potentially uninitialised maxs in NAVNEW_TestNodeConnectionBlocked (g_navnew.c)

    if ( ignoreEnt )
    {
        VectorCopy( ignoreEnt->r.mins, mins );
        VectorCopy( ignoreEnt->r.maxs, maxs );
        ignoreEntNum = ignoreEnt->s.number;
    }
    else
    {
        VectorCopy( playerMins, mins );
        VectorCopy( playerMaxs, mins ); //bad, not sure what effect it has
        ignoreEntNum = ENTITYNUM_NONE;
    }

Ragnos crash

Infamous client-side Ragnos crash, related to G2 bones.

0x4DFAF9 related

Userinfo being set/changed twice when connecting

Cause seems to be saber1/saber2 fields not being properly read from userinfo in ClientUserinfoChanged.
Instead, ClientUserinfoChanged reads from session data (Which is not valid on the first connection) so saber1/saber2 will be treated as "Kyle" and "none" respectively
Saber-related session data is set in ClientBegin.

Recommended fix: Moving saber data out of session data and into client persistent data?
ClientUserinfoChanged should read from userinfo and set persistent data
ClientBegin should read values from client persistent data, not session.

Consolidate weapon data, general cleanup

damage/velocity defines in g_weapon.c -> (b?)g_weapons.h
could generalise weapons, e.g. fire(), alt_fire(), damage, velocity?
remove unused/commented code for: gauntlet, lightning gun, powerups, railgun..

Cmd_LevelShot_f is derp.

void Cmd_LevelShot_f(gentity_t *ent)
{
    if(!ent->client->pers.localClient)
    {
        trap_SendServerCommand(ent-g_entities,
            "print \"The levelshot command must be executed by a local client\n\"");
        return;
    }

    if(!CheatsOk(ent))
        return;

    // doesn't work in single player
    if(g_gametype.integer == GT_SINGLE_PLAYER)
    {
        trap_SendServerCommand(ent-g_entities,
            "print \"Must not be in singleplayer mode for levelshot\n\"" );
        return;
    }

    BeginIntermission();
    trap_SendServerCommand(ent-g_entities, "clientLevelShot");
}

Plenty of client 0 logic in the regular NPC code too.

Please investigate and make it more robust with multiple players / team allies?

ClientCommand rewrite

Remove leftover/broken debug commands?
Hash table for commands to drastically shrink code and speed up command execution

Q3_STATIC/*_HARD_LINKED is never referenced

plz remove

Add chat token support for /say cmds

Add chat token support for say cmds.

Perhaps something more like:
$h - health
$a - armor
$f - force pool
$g - ammo in current weapon
$n - name
$s - force side
$l (L not I) - location
$d - last death location
$c - team color (red,blue,free,spectator)
$t - team name (^1Red Team, ^4Blue Team, ^3Free, Spectator)
$e - enemy team color(blue,red,free,spectator)
$w - current weapon name
$i - player id
$h - siege class name
$o - current held flag in ctf

Open for other styles, we used this method in ETF, even allowed $r to show reader name. I used a different method in EnsiMod though. More along the lines of:
[h] [a] [f] [g] [n] [s] [l] [d] [c] [t] [e] [w] [id] [cl] [o]

EF_DISINTEGRATION rendering bug

When EF_DISINTEGRATION (e.g. hit with a fully charged disruptor shot) is set on a client performing a special move (notably staff saber butterfly) their disintegrating body will move quickly in a (random?) direction.

UI cvar table cleanup

akin to cgame/game, might not need a cvar->update() but doesn't hurt..

Fraglimit warning message is not played to blue team players in TFFA

In cg_playerstate.c after the following line:

highScore = cgs.scores1;

add:

if ( cgs.gametype == GT_TEAM && cgs.scores2 > highScore ) {
    highScore = cgs.scores2;
}

Fix entitylist command to not skip client-0 and add missing types.

/*
===================
Svcmd_EntityList_f
===================
*/
void    Svcmd_EntityList_f (void) {
    int         e;
    gentity_t       *check;

    check = g_entities;
    for (e = 0; e < level.num_entities ; e++, check++) {
        if ( !check->inuse ) {
            continue;
        }
        G_Printf("%3i:", e);
        switch ( check->s.eType ) {
        case ET_GENERAL:
            G_Printf("ET_GENERAL          ");
            break;
        case ET_PLAYER:
            G_Printf("ET_PLAYER           ");
            break;
        case ET_ITEM:
            G_Printf("ET_ITEM             ");
            break;
        case ET_MISSILE:
            G_Printf("ET_MISSILE          ");
            break;
        case ET_SPECIAL:
            G_Printf("ET_SPECIAL          ");
            break;
        case ET_HOLOCRON:
            G_Printf("ET_HOLOCRON         ");
            break;
        case ET_MOVER:
            G_Printf("ET_MOVER            ");
            break;
        case ET_BEAM:
            G_Printf("ET_BEAM             ");
            break;
        case ET_PORTAL:
            G_Printf("ET_PORTAL           ");
            break;
        case ET_SPEAKER:
            G_Printf("ET_SPEAKER          ");
            break;
        case ET_PUSH_TRIGGER:
            G_Printf("ET_PUSH_TRIGGER     ");
            break;
        case ET_TELEPORT_TRIGGER:
            G_Printf("ET_TELEPORT_TRIGGER ");
            break;
        case ET_INVISIBLE:
            G_Printf("ET_INVISIBLE        ");
            break;
        case ET_NPC:
            G_Printf("ET_NPC              ");
            break;
        case ET_BODY:
            G_Printf("ET_BODY             ");
            break;
        case ET_TERRAIN:
            G_Printf("ET_TERRAIN          ");
            break;
        case ET_FX:
            G_Printf("ET_FX               ");
            break;
        default:
            G_Printf("%-3i                ", check->s.eType);
            break;
        }

        if ( check->classname ) {
            G_Printf("%s", check->classname);
        }
        G_Printf("\n");
    }
}

Also correctly left justifies the default case instead of right justify.

Missing functions/files cleanup

missing functions: PrecacheItem, TossClientCubes, SetRespawn, ArmorIndex, Think_Weapon, UseHoldableItem, CheckGauntletAttack, [...]
missile files: p_hud.c, g_pweapon.c, [...]
unnecessary extern keyword usage everywhere

UI still references g_spskill

The UI dll still references g_spskill cvar instead of the changed g_npcspskill.

Move weapon defines from g_weapon.c to a separate header to clean the file up a bit.

Move the weapon defines to a separate header and clean the comments.

make smooth first-person anims optional

or figure out why it causes animevents (specifically saber swing noises) a lot louder/play more often

Block ability to /team scoreboard...

Block ability to /team scoreboard or remove its functionality entirely while retaining the flag for compatibility reasons

mins/maxs consistency

mainly for playing bounding boxes, lots of hardcoded values (-15 <-> 15)

Mover code is still susceptible to griefing

Movers are still very susceptible to griefing in various forms. Siege items aren't even handled at all. AFAIK no mods have actually even fixed those cases yet.

lchat in cg_servercmds.c is missing a space like ltchat

Title describes.

Com_sprintf(text, sizeof(text), "%s<%s>^%s%s", name, loc, color, message);

Should be

Com_sprintf(text, sizeof(text), "%s<%s> ^%s%s", name, loc, color, message);

s/trap_Printf/trap_Print/

No message is displayed if flag is returned by timer/death pit

as described here http://jkhub.org/topic/1550-no-message-is-displayed-if-flag-is-returned-by-timerdeath-pit/

charge_stick possible minor bug

self->s.pos.trDelta[0] += vNor[0]*(tN[0]*(((float)Q_irand(1, 10))*0.1));
self->s.pos.trDelta[1] += vNor[1]*(tN[1]*(((float)Q_irand(1, 10))*0.1));
self->s.pos.trDelta[2] += vNor[1]*(tN[2]*(((float)Q_irand(1, 10))*0.1));

specifically vNor[1] usage
need to investigate further

Capturelimit missing from master list of valid vote strings.

Capturelimit missing from master list of valid vote strings. (And from the "error message")

Usage info cleanup for g_svcmds.c

G_Printf("Usage: addip <ip-mask>\n");

G_Printf("Usage: removeip <ip-mask>\n");

Add this to forceteam:

    if ( trap_Argc() < 3 ) {
        G_Printf("Usage: forceteam <player> <team>\n");
        return;
    }

General flag returns print no message to anyone.

Return from nodrop/timeout/or whatever besides touch return doesn't print any kind of message but is supposed to. Seems to be issue with handling of the flag msg functions in cgame.

CMD_ALIVE doesn't check spec/tempspec like CheatsOk did

CMD_ALIVE is missing checks for not being spectator or a siege spectator like the CheatsOK function did. Plz2fix.

donedl exploit

Results in disappearing powerups and other client data (JA+ grapple etc)

Issue lies in ClientBegin being called without a chance for player_die to be called.
Has been hooked on windows and linux servers, but we've yet to determine a way to check if it is valid use of the command. Could just always block on some servers but this is not an ideal fix.

Could check if they have an active download (cl->download?), if they've already been in the game, etc.

Another "fix" is to call player_die from ClientBegin in under certain conditions.
This is not ideal, and may be error-prone.
Waiting on @TheSil for some experimenting

    //assign the pointer for bg entity access
    ent->playerState = &ent->client->ps;

    //fix for donedl command bug, that could cause powerup dissapearing
    if ( ent->health && ent->client && ent->client->sess.sessionTeam != TEAM_SPECTATOR
        && clientNum == ent->client->ps.clientNum) {
        // Kill him (makes sure he loses flags, etc)
        ent->flags &= ~FL_GODMODE;
        ent->client->ps.stats[STAT_HEALTH] = ent->health = 0;
        g_dontPenalizeTeam = qtrue;
        player_die( ent, ent, ent, 100000, MOD_TEAM_CHANGE );
        g_dontPenalizeTeam = qfalse;
    }

Alt Attack should cycle backwards through specs (if already following only)

Code to do this: (SpectatorThink g_active.c)

        /* the regular attack button check is here don't include this line */

        // alt attack button cycles through spectators
        if ( client->sess.spectatorState == SPECTATOR_FOLLOW ) {
            if ( ( client->buttons & BUTTON_ALT_ATTACK ) && ! ( client->oldbuttons & BUTTON_ALT_ATTACK ) ) {
                Cmd_FollowCycle_f( ent, -1 );
            }
        }

        /* the jump button check is here don't include this line */

We check for is already following because JKA has the speed mode when in free flight.

The bot triggering of the new message for clientdisconnect is becasue of G_AddBot is broken.

Compare this version with yours, you can drop some of it since I added IP and guid to bots.

http://slexy.org/view/s21nqIfRjg