BUG listing of STOCK bugs by the Mage.
These are some stock ROM2.4b4 and b6 bugs I have found. Basic descriptions
of
fixes are included some OLC 1.81 bugs are also included:
Note: All these fixes are in the Sacred Codebase and some in QuickMUD I
noticed.
--------------------------------------------------------------------------
Attacking while AFK: Made sure you cannot attack while AFK. Modified
is_safe() for this.
--------------------------------------------------------------------------
act_new() with wiznet: If someone typed a $t or $p with a delete
command the mud would crash. This happened in the LOG wiznet with
interp.c. Modified and made a wiznet_special just for this
log.. since it seems the only place a imm/user could send through
those codes.
--------------------------------------------------------------------------
magic.c - Ventrilloquate Does not properly show victims names if the
spell was not cast correctly. Fixed by changing the spell to have
these 4 lines replace the ones similiar to it.
sprintf(buf1, "%s says '%s'.\n\r",
IS_NPC(vch) ? vch->short_descr : vch->name, target_name);
sprintf(buf2, "Someone makes %s say '%s'.\n\r",
IS_NPC(vch) ? vch->short_descr : vch->name, target_name);
--------------------------------------------------------------------------
do_order - Changed to prevent a mob from going AFK.
--------------------------------------------------------------------------
do_steal - If the victim cannot SEE you then the yell will NOW reflect
that. It was shouting name no matter what.
--------------------------------------------------------------------------
magic.c - In some spells and skills Damage was called before affects.
--------------------------------------------------------------------------
fight.c What if they died? Was an affect leak. Ray of truth, dirt
kick, and demonfire were changed.
--------------------------------------------------------------------------
do_steal - Somehow could steal from a SAFE room. Fixed.
--------------------------------------------------------------------------
handler.c - If you made a spell that used TO_RESIST and the person
already had it via a race option it would remove that race
option. Did an exclusive OR afterwards to always reset the
RACE flags.
--------------------------------------------------------------------------
const.c - Identify and another spell had same slot. I reordered all
my slots. No biggie since slots are not used
--------------------------------------------------------------------------
act_obj.c - Steal cannot steal NODROP items now.
--------------------------------------------------------------------------
handler.c
fight.c - If you lose an item via alignment zapping. It will now
trigger the melt_drop flag.
--------------------------------------------------------------------------
act_obj.c - Haggle can gain you much money. Changed it so.. that you
can only sell up to 75% of value and haggle down to 76% of
value. also.. If you entered.. by 10000000000 diamonds.. it
would crash the system due to bit overflow.
--------------------------------------------------------------------------
special.c - Spec guard will attack imms with flags. Changed this.
--------------------------------------------------------------------------
fight.c - Lore plain does not exist or work. Yet it is listed.
basically made a modified spell_identify for lore.
--------------------------------------------------------------------------
act_obj.c - Envenom will multienvenom a food item. A pie. You can
make your skill 100% real easy this way. Check to see if
the pie was already envenomed and stopped check_improve if
so.
- 4th spell on potion/scroll is not work. Added the lines
in quaff and recite to make them work. oversite?
--------------------------------------------------------------------------
skill.c - In creation. You basics are sword and you add
weaponsmaster.. then decide you do not want it. It will
remove the rom basic field. Fixes so that all basic skills
for that class CANNOT be removed.
--------------------------------------------------------------------------
act_info.c - Race skills would be excluded if the class did not have
the skill. Fixed it so race skills could always be
practiced regardless of class.
--------------------------------------------------------------------------
act_obj.c - When you WEAR and object.. if it was of type WEAR_BODY it
would say torso instead.
--------------------------------------------------------------------------
magic.c - heat_metal did not trigger melt_drop flag.
I made melt_drop a function and placed it in all needed
places.
--------------------------------------------------------------------------
act_wiz.c - If you lower an IMM that had Holylight on.. to a
mortal.. that mortal will not have holylight. In do_advance
just remove the bit if it was set.
--------------------------------------------------------------------------
- Alias r goto 3014 does not work. This is due to
find_location() not being able to determine that the passed
in argument is a string due to extra spaces placed in the
substitute. I have a routine that strips front and back of
strings to clean them up. I places this in find_location
and it fixed this particular problem.
--------------------------------------------------------------------------
comm.c - Page_to_char() allocates strlen + 1 and showstring
deallocates strlen. You then lose the +1 in memory.
--------------------------------------------------------------------------
skills.c - In the drop section there is a loop using the loop control
variable i. Inside the loop it is never used and only gn is
used for group dropping. This is a waste if never intended.
the function is in parse_gen_groups()
--------------------------------------------------------------------------
fight.c - in one_hit.. skill hand_to_hand means NOTHING.
--------------------------------------------------------------------------
skills.c - In do_gain.. if you gain an ENTIRE
group.. say.. protection and it has spells you cannot use
based on your class. Gain will gain them for you and you can
use them even if they are not listed. In order to fix this I
went into gn_add() and made sure that it did NOT ADD any
spell/skill that was toggeled to level -1(or 99).
--------------------------------------------------------------------------
fight.c - VORPAL?? Nothing is made of it.. so I added some
functionality to add VORPAL to the system.
--------------------------------------------------------------------------
mob_prog.c - The CHK_MONEY is being calculated WRONG!. It says it is
silver and yet.. it is not.. correct line should read.
lval_char = lval_char->silver + (lval_char->gold * 100);
--------------------------------------------------------------------------
fight.c - check_assist() the Mobs can assist even if they are BLIND.
This is simple fix: add a can_see(rch,ch) to the if attack line.
--------------------------------------------------------------------------
skills.c - Your TNL can lose exp when you lower you CP. in do_gain right
at end.. where you train "points" you change to have a temp xp
var save the current exp in this level.. and add it AFTER the
new calc is made.
xp = ch->exp - (ch-level * exp_per_level(*ch,ch->pcdata->points);
^^^ before the points are changed..
and after the new ch->exp is calculated.. add this
ch->exp += xp;
-- you do have to declare XP.. :)
--------------------------------------------------------------------------
-- Swimming and Drowning are not implemented at all.!! bug :)
--------------------------------------------------------------------------
effects.c - ACID pitting would never be saved when a character logged out.
This enabled them to just reset the effect easy. The issue has to
do with the way the affect is setup in the ACID effect. It is WRONG.
--------------------------------------------------------------------------
magic.c - IF an immortal is WIZI you can cast mass spells and have it
revealed
to you that an IMM is in the room. MAss healing, calm, chain_lightning,
faerie_fog, mass_invis. Just add a check to see if they are wizi imms and
ignore them.
--------------------------------------------------------------------------
magic.c - earthquake would reveal the presence of WIZI Imms in the room.
--------------------------------------------------------------------------
comm.c - There are four places where free_char() is called. This is not, in
itself, a bug. The bug is, those characters could theoretically have pets
and if they do, those pets will not be freed. You will sometimes see people
login with "ghost" pets grouped with them - pets they didn't buy and not
that
are not actually in the room but are showing up on the group list. To fix
this, you need to use nuke_pets() before each and every call to free_char().
Use the same argument that is passed to free_char() in each case.
For example:
+ nuke_pets( d->character );
free_char( d->character );
--------------------------------------------------------------------------
act_wiz.c - In do_outfit() it is possible for twits to crash the mud by
using
outfit, remove all, outfit, remove all, etc., and repeating that several
thousand times. Once they do it enough, the MUD will crash when they try to
look at their inventory. This happens because ROM tries to allocate too
much
memory to display the number of items in show_list_to_char() which is found
in act_info.c. Ideally, you should plug that possible crash also by
inserting
this little check:
act_info.c in the show_list_to_char() function:
count = 0;
for ( obj = list; obj != NULL; obj = obj->next_content )
count++;
+ if ( count > 5000 )
+ {
+ send_to_char( "That is WAY too much junk! Drop some of it!\n\r",
ch );
+ return;
+ }
prgpstrShow = alloc_mem( count * sizeof(char *) );
prgnShow = alloc_mem( count * sizeof(int) );
nShow = 0;
act_wiz.c in the do_outfit() function:
if (ch->level > 5 || IS_NPC(ch))
{
send_to_char("Find it yourself!\n\r",ch);
return;
}
+ if (ch->carry_number+4 > can_carry_n(ch))
+ {
+ send_to_char("You are carrying too much, try dropping some
items.\n\r",ch);
+ return;
+ }
if ( ( obj = get_eq_char( ch, WEAR_LIGHT ) ) == NULL )
{
A more elegant fix would be to also check for each item being in the
possession
of the character. This requires a simple recursion which goes through
inventory and all containers in inventory, counting the items. In any even,
the
above fix will stop twits from crashing the mud by attaining 10,000 items
with
the outfit command.
--------------------------------------------------------------------------
act_wiz.c - in do_mstat
sprintf( buf, "Hp: %d/%d Mana: %d/%d Move: %d/%d Practices: %d\n\r",
victim->hit, victim->max_hit,
victim->mana, victim->max_mana,
victim->move, victim->max_move,
IS_NPC(ch) ? 0 : victim->practice ); /* this line */
send_to_char( buf, ch );
Should be:
IS_NPC(victim) ? 0 : victim->practice );
--------------------------------------------------------------------------
Okay.. If you had an ALIAS with a " or a ' in it.. and saved it
it would cause players to login without EQ and spells/skills. and sometimes
crash the mud.
in do_alias do not allow the creation of an alias with those chars in them
simple.. yet effective for this strong bug.
--------------------------------------------------------------------------
Found a problem in the for-loop in fight.c/group_gain:
The loop-variable (gch) isn't used in the checking for alignment changes:
for ( gch = ch->in_room->people; gch != NULL; gch = gch->next_in_room )
{
OBJ_DATA *obj;
OBJ_DATA *obj_next;
if ( !is_same_group( gch, ch ) || IS_NPC(gch))
continue;
xp = xp_compute( gch, victim, group_levels );
sprintf( buf, "You receive %d experience points.\n\r", xp );
send_to_char( buf, gch );
gain_exp( gch, xp );
> for ( obj = ch->carrying; obj != NULL; obj = obj_next )
{
obj_next = obj->next_content;
if ( obj->wear_loc == WEAR_NONE )
continue;
> if ( ( IS_OBJ_STAT(obj, ITEM_ANTI_EVIL) && IS_EVIL(ch) )
> || ( IS_OBJ_STAT(obj, ITEM_ANTI_GOOD) && IS_GOOD(ch) )
> || ( IS_OBJ_STAT(obj, ITEM_ANTI_NEUTRAL) && IS_NEUTRAL(ch) ) )
{
> act( "You are zapped by $p.", ch, obj, NULL, TO_CHAR );
> act( "$n is zapped by $p.", ch, obj, NULL, TO_ROOM );
obj_from_char( obj );
> obj_to_room( obj, ch->in_room );
}
}
}
Fix: replace all the mentioned ch's with gch's.
--------------------------------------------------------------------------
In the same section there is the further problem of NODROP items being
dropped by an alignment shift. To remedy this, add the 3 lines marked by +.
(This is the same section Edwin suggests changing each ch to a gch.)
for ( obj = ch->carrying; obj != NULL; obj = obj_next )
{
obj_next = obj->next_content;
if ( obj->wear_loc == WEAR_NONE )
continue;
if ( ( IS_OBJ_STAT(obj, ITEM_ANTI_EVIL) && IS_EVIL(ch) )
|| ( IS_OBJ_STAT(obj, ITEM_ANTI_GOOD) && IS_GOOD(ch) )
|| ( IS_OBJ_STAT(obj, ITEM_ANTI_NEUTRAL) && IS_NEUTRAL(ch) ) )
{
act( "|i|yYou are zapped by $p.|n", ch, obj, NULL, TO_CHAR );
act( "|i|y$n is zapped by $p.|n", ch, obj, NULL, TO_ROOM );
obj_from_char( obj );
+ if ( IS_OBJ_STAT(obj,ITEM_NODROP) )
+ obj_to_char( obj, ch );
+ else
obj_to_room( obj, ch->in_room );
}
}
--------------------------------------------------------------------------
In olc_save.c check this out:
char *fix_string( const char *str )
{
static char strfix[MAX_STRING_LENGTH * 2];
int i;
int o;
if ( str == NULL )
return '\0';
for ( o = i = 0; str[i+o] != '\0'; i++ )
{
if (str[i+o] == '\r' || str[i+o] == '~')
o++;
- strfix[i] = str[i+o];
+ if ((strfix[i] = str[i+o]) == '\0')
+ break;
}
strfix[i] = '\0';
return strfix;
}
Before, if this thing hit something that ended in '\n\r\0', it bypassed the
'\0' completely and went on into who knows what. The only good thing about
the code I had, I added a bounds check on this a long time ago because I
didn't trust it... it's a good damn thing too. It won't screw up on
saving, because it catches the '\0' in strfix... BUT, it will keep going
until it hits another '\0' which could be very soon, or it could be quite
a while.
What I added checks when it sets the strfix[i] and if it is setting a '\0'
the string is done, so you break out... you don't increment i, bypassing
the '\0', and going off into the wild binary yonder.
--------------------------------------------------------------------------
do_oset ignore.
One of my staff pointed this bug out...when they tried to set an extra
flag using set object it wouldn't set all it did was remove all the
extra flags from the object. Having Ivan's OLC I quickly found a quick
fix for this and here it is.
if ( !str_prefix( arg2, "extra" ) )
{
if ( ( value = flag_value( extra_flags, arg3 ) ) != NO_FLAG)
{
TOGGLE_BIT(obj->extra_flags, value);
}
return;
}
also add an #include "olc.h" with the rest of the includes at the top of
act_wiz.c
--------------------------------------------------------------------------
In act_obj.c do_put in this section
/* 'put all container' or 'put all.obj container' */
look for
&& get_obj_weight(obj) < (container->value[3] * 10))
^
^
change this to <=
--------------------------------------------------------------------------
Ok, I've found some more stock ROM memory leaks... damn.
These are all in save.c.
In fread_char():
KEY( "Clan", ch->clan, clan_lookup(fread_string(fp)));
The problem with this is, fread_string() is going to str_dup() whatever
it reads in... and this never gets freed. To fix this, you have to
keep a pointer to the string you read in:
if ( !str_cmp( word, "Clan" ) )
{
char *tmp = fread_string(fp);
ch->clan = clan_lookup(tmp);
free_string(tmp);
fMatch = TRUE;
break;
}
In that same function, the same thing is done for race:
KEY( "Race", ch->race, race_lookup(fread_string(fp)));
This needs to be fixed in the same way.
And also, fread_pet() contains duplicates of these errors reading race
and clan values.
Can't believe I missed that one for all this time, bleah.
--------------------------------------------------------------------------
you can purchase IMMORTALS and MORTALS.
in act_obj in do_list and do_buy for pets.. it does not check for
IS_NPC. It should otherwise a immortal with a nocancel flag on it
will...... crash system by being bought.
--------------------------------------------------------------------------
in fight.c in do_bash. the += GET_AC() line should be -=
check it out.. you will find out why
--------------------------------------------------------------------------
PETS will disapear when someone goes linkdead. Basically, the person
is transfered to LIMBO and then forced to quit. If you quit and your PET
is not in the same room.. well.. he go bye bye.
changed update.c to trans pet with char.. and idle_handling to make pet
go back.
--------------------------------------------------------------------------
olc_act.c in MEDIT(add_mprog) for the trigger..
the one_argument(argument,phrase) line takes out all capital letters
change it to strcpy(phrase,argument);
--------------------------------------------------------------------------
in bust_a_prompt make the %e option not work when you are blind. A Prompt
can show exits when you are blind.. if this is not done.
--------------------------------------------------------------------------
magic.c - bless spell.
if (FIGHTING or AFFECTED)
send_to_char( You are already blessed) ??
Bad message at best..
If try to cast bless during fight get You are already Blessed? ack
split it to two seperate if statements.
--------------------------------------------------------------------------
Found another stock ROM bug here in do_at() in act_wiz.c:
original = ch->in_room;
on = ch->on;
char_from_room( ch );
char_to_room( ch, location );
interpret( ch, argument );
/*
* See if 'ch' still exists before continuing!
* Handles 'at XXXX quit' case.
*/
for ( wch = char_list; wch != NULL; wch = wch->next )
{
if ( wch == ch )
{
char_from_room( ch );
char_to_room( ch, original );
ch->on = on;
break;
}
}
The problem with this section of code is what happens if 'on' is somehow
extracted, picked up by the character, or moved from it's original room?
With the above code, ch->on is set back to the original furniture,
regardless of what happened to it - no check is made.
I changed the last section of the code to make sure that the 'on' object is
still present in the original room before ch->on is set back to it:
/*
* See if 'ch' still exists before continuing!
* Handles 'at XXXX quit' case.
*/
for ( wch = char_list; wch != NULL; wch = wch->next )
{
if ( wch == ch )
{
OBJ_DATA *obj;
char_from_room( ch );
char_to_room( ch, original );
/* See if on still exists before setting ch->on back to it. */
for ( obj = original->contents; obj; obj = obj->next_content )
{
if ( obj == on )
{
ch->on = on;
break;
}
}
break;
}
}
Note: If you have mob programs you need to fix do_mpat() also!
--------------------------------------------------------------------------
in save.c fread_char.. KEY("Prom") should be KEYS(Prom)
use KEYS to free up STRINGS.. otherwise memory leaks.
--------------------------------------------------------------------------
in do_switch minor string overwriting at end. Look at the prompt and
free_string(victim->prompt) first.
--------------------------------------------------------------------------
Drop a corpse and have one in inventory. examine corpse, which one
does it examine? both?? I had this, not sure if Stock.
--------------------------------------------------------------------------
Change_sex. Needs to have the affect applied AFTER the ACT statement
talking about the change sex.
--------------------------------------------------------------------------
We had a recursive crash today on Abysmal. I can't believe it has never
happened before either, but here is how it works:
A character gets to a "Hit return to continue" prompt.
The character then lags terribly such that their output buffer grows to
as big as it can and then overflows:
if (d->outsize >= 32000)
{
bug("Buffer overflow. Closing.\n\r",0);
close_socket(d);
return;
}
Ok, in close_socket(), there is a section that looks like:
if ( dclose->outtop > 0 )
process_output( dclose, FALSE );
That calls process_output() again. In process output, it goes in and if
they have d->showstr_point (meaning they are in the pager). It calls
write_to_buffer()... that function when called on a descriptor that is
already overflowed the buffer just tries to close_socket() again.
Result? A recursive loop that crashes the MUD.
Here is an easy way to fix it, in process_output(). All this does is
prevent ANY write_to_buffer() to the descriptor if fPrompt is FALSE. The
only time that should be FALSE is in close_socket().
/*
* Low level output function.
*/
bool process_output( DESCRIPTOR_DATA *d, bool fPrompt )
{
extern bool merc_down;
/*
* Bust a prompt.
*/
- if (!merc_down && d->showstr_point)
+ if (fPrompt && !merc_down && d->showstr_point)
write_to_buffer(d,"[Hit Return to continue]\n\r",0);
else if (fPrompt && !merc_down && d->connected == CON_PLAYING)
{
CHAR_DATA *ch;
CHAR_DATA *victim;
--------------------------------------------------------------------------
From: Edwin Groothuis <[EMAIL PROTECTED]>
To: Rom Mailing-list <[email protected]>
Subject: Overflow error in substitute alias / prefix
Date: Thursday, October 14, 1999 11:01 AM
Text is also available from URL below.
Prefix
Problem
When having a prefix and the command entered is too long, the mud will
crash.
Cause
The length of the prefix and the length of the command are incorrectly
checked against the length of the string they have to form.
Checking
Set your prefix to : and enter a command too long. First a function in
comm.c will complain, then your mud will crash.
Patch
In alias.c/substitute_alias(), fix the next lines:
/* check for prefix */
if (ch->prefix[0] != '\0' && str_prefix("prefix",argument)) {
- if (strlen(ch->prefix) + strlen(argument) > MAX_INPUT_LENGTH)
+ if (strlen(ch->prefix) + strlen(argument) > MAX_INPUT_LENGTH-2)
send_to_char("Line to long, prefix not processed.\r\n",ch);
else {
sprintf(prefix,"%s %s",ch->prefix,argument);
argument = prefix;
}
}
--------------------------------------------------------------------------
Problem: Can gate to a guild room that is NOT YOURS.
Fix: Make a routine called check_gate and put all the checks for
the gate/portal/nexus spells in it. Then put the check_gate function
in those spells. Basically, this will help out.
Note: Clan rooms, Is Safe rooms, etc. Think about WHAT ELSE you would do.
--------------------------------------------------------------------------
You can order a Pet to bash every round, rescue every round. This is
because
there is no wait state for mobs. This is solved by a couple things:
putting a check for IS_NPC(ch) && ch->wait) return in interpret_cmd. That
will fix the things. Then you make sure that you can subtract a wait
properly
Do this in violence_update. if (IS_NPC && ch-wait) ch->wait-1; I am not
giving you exact code, but you should be able to fix that. Make sure
WAIT_STATE
is adding the wait to a npc as well as a PC.
--------------------------------------------------------------------------
Cursed Pets can recall with you. I have done a really modified recall but
basically you do a IF_PET() and if cursed, dont recall. Let your master
know
with a send_to_char or something.
--------------------------------------------------------------------------
in spell_recharge, it is check obj->value[3]/ This is the numberical
representation of your spell. I dont know why this is there... but to do a
level check on it does not make sense, so either remove it or make it a real
level check.
--------------------------------------------------------------------------
in db.c, in load_old_object, it allocates an affect manually. replace with
new_affect().
this is ALSO the case with load_threads in note.c, use new_note().
--------------------------------------------------------------------------
Pets with SANC and HASTE (and some other spells) will duplicate affects.
This is
due to fread_pet loading the original pet and the original affects, then
loading
the affects on top of that. The fix has to do with save.c
in fread_pet.. move the int vnum to a global to the function instead of
local
to the Vnum read.
Next in the Affc section.
if (!check_pet_affected(vnum,paf))
{
paf->next = pet->affected;
pet->affected = paf;
} else{
free_affect(paf);
}
in place of the paf->next and pet->affected lines. This will enable the
check function
to determine if they should PUT the affect on the pet.
In db.c (where I put it) put the check function:
/* remove duplicate affects on pets */
bool check_pet_affected(int vnum, AFFECT_DATA *paf)
{
MOB_INDEX_DATA *petIndex;
petIndex = get_mob_index(vnum);
if (petIndex == NULL)
return FALSE;
if (paf->where == TO_AFFECTS)
if (IS_AFFECTED(petIndex, paf->bitvector))
return TRUE;
return FALSE;
}
put the function prototype in merc.h and you are done!!!
