Lucien (Jun 13) [60522] successfull
Lucien (Jun 13) [60526] thes -> this

ERROR 1064 (42000) at line 2: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'IF EXISTS 'AwakeMUD'@'localhost'' at line 1

mysql> select @@Version;
5.5.55-0ubuntu0.14.04.1

https://dev.mysql.com/doc/refman/5.7/en/drop-user.html
As of MySQL 5.7.8, the IF EXISTS clause can be used, which causes the statement to produce a warning for each named account that does not exist, rather than an error.

...shes have been exposed just by sitting in a vehicle and performing various commands. It would be prudent to search for all calls to world[ch->in_room] and other in_room variables and verify that the room is not NOWHERE before we execute.

Filed by Lucien (id #1) from the rear of a vehicle of vnum 1309, located in room 10000.

Room descriptions should probably cleared of specific references to awakenedworlds. A generic placeholder maybe?

The documentation in docs/staffdocs needs to be added to the wiki for easy browsing. (Thanks for writing it up for us!)

Trainer needs update

As the first step of coded playergroups, the ability to create and disband a group should be instated. This will also include the backend data structures to enable playergroups to exist.

Member management includes the following commands:

PROMOTE / DEMOTE
GRANT / REVOKE
OUTCAST

GRANT / REVOKE is required for testing the majority of playergroup commands.

Really, anything that forces large volumes of text on the client should be toggled.

It's hard to tell what the effect of 'set x auth on/off' is; it controls the authorize flag, but unless you know what the authorize flag does, you may think that setting it to on authorizes the character fully. We should either add an authorize command or better messaging to the set command to clarify this use case.

For RP purposes players should be able to record messages using pen and paper.

Precondition: have an object with a PEN flag and an object with WRITEBAL in inventory.

Restrictions: length

Could use a dumbed down implementation of iedit. Abuse possible?

A new character with 1 quickness and 0/0 armor got an armor-overload (encumbrance) warning when wearing a non-armored set of chains. This isn't the expected behavior unless any clothing encumbers a qui-1 char by canon.

This issue is a catch-all for any issues with 'set'

set <initdice|initroll> does nothing. (act.wizard.cpp:3655-3664).

set <hunger|thirst> removes from 'score' output but doesn't gag hunger/thirst messages. Output in 'score' does not seem to return.

Setting a player's level to 0 removes them from wholist. Intended behavior?

nowizlist doesn't seem to be implemented (act.wizard.cpp:3770, commented out). I wonder why.

idnum doesn't seem to to anything and would probably screw up the DB if it did. Remove it?

passwd doesn't seem to to anything. Bugged?

newbie seems bugged. PLR: output from stat doesn't get updated properly.

pg does not seem implemented (missing case in act.wizard.cpp). Remove it?

help 'hunger%' OR '%status

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff6acc319 in _IO_vfprintf_internal (s=, format=, ap=) at vfprintf.c:2047
2047 vfprintf.c: No such file or directory.
(gdb) bt
#0 0x00007ffff6acc319 in _IO_vfprintf_internal (s=, format=, ap=) at vfprintf.c:2047
#1 0x00007ffff6acee71 in buffered_vfprintf (s=s@entry=0x7ffff6e421c0 <IO_2_1_stderr>, format=format@entry=0x7fffffff5a90 "Executing query: SELECT * FROM help_topic WHERE name LIKE '%\'hunger%\' OR \'%status%' ORDER BY name ASC", args=args@entry=0x7fffffff5998) at vfprintf.c:2356
#2 0x00007ffff6ac9dfe in _IO_vfprintf_internal (s=0x7ffff6e421c0 <IO_2_1_stderr>, format=0x7fffffff5a90 "Executing query: SELECT * FROM help_topic WHERE name LIKE '%\'hunger%\' OR \'%status%' ORDER BY name ASC", ap=0x7fffffff5998) at vfprintf.c:1313
#3 0x00000000005bacd9 in log (format=0x7fffffff5a90 "Executing query: SELECT * FROM help_topic WHERE name LIKE '%\'hunger%\' OR \'%status%' ORDER BY name ASC") at utils.cpp:633
#4 0x000000000053891e in mysql_wrapper (mysql=0x89c580, query=0x7fffffff7ae0 "SELECT * FROM help_topic WHERE name LIKE '%\'hunger%\' OR \'%status%' ORDER BY name ASC") at newdb.cpp:54
#5 0x0000000000433381 in display_help (help=0x7fffffff9b30 "\200\241\377\377\377\177", arg=0x7fffffffddb5 "'hunger%' OR '%status") at act.informative.cpp:2239
#6 0x00000000004336bb in do_help (ch=0xb41630, argument=0x7fffffffddb5 "'hunger%' OR '%status", cmd=200, subcmd=0) at act.informative.cpp:2278
#7 0x0000000000518a1d in command_interpreter (ch=0xb41630, argument=0x7fffffffddb0 "help 'hunger%' OR '%status", tcname=0xb41ea0 "Finster") at interpreter.cpp:1333
#8 0x00000000004aca2a in game_loop (mother_desc=3) at comm.cpp:660
#9 0x00000000004abbfe in init_game (port=4000) at comm.cpp:351
#10 0x00000000004ab669 in main (argc=1, argv=0x7fffffffe708) at comm.cpp:235
(gdb)

Circle's wizhelp guide states on page 5-6:

Sets the message seen by players when you use GOTO to move to a different
room. POOFIN with no argument resets your poofin to the default "appears with
an ear-splitting bang"; POOFOUT with no argument defaults to "disappears in
a puff of smoke."

However, Awake does not set a default message for both commands if called without arguments.

As (null) is not a nice thing for a player to see this should be changed.

Rooms track peace through the .peaceful variable, but have the PEACEFUL flag settable through redit. On removal of this flag, it should decrement the peaceful variable appropriately, but currently does not appear to do this.

The 'policy' command currently returns empty.

Having a radio set to 8 is critical to being able to communicate with the game. New characters should be auto-loaded with a radio set to 8, perhaps with a room along the way that tells them that the yellow text they've been seeing is coming from their radio. The goal is to prevent any player from leaving chargen without a radio.

If an object was created directly from proto and has its pointers still pointing to the proto obj, don't delete the pointers. Same for all other things too.

Not sure whether I should open an issue for an unmerged branch but I think I'm missing something here.
gcc 4.8.4 on Ubuntu 14.04

g++ -DDEBUG -DSELFADVANCE -Dlinux -Wall -I /usr/local/include/ -ggdb -c -o playergroup_classes.o playergroup_classes.cpp
playergroup_classes.cpp: In member function ‘void Playergroup::audit_log_vfprintf(const char*, ...)’:
playergroup_classes.cpp:215:24: error: ‘va_start’ was not declared in this scope
va_start(args, format);
^
playergroup_classes.cpp:217:14: error: ‘va_end’ was not declared in this scope
va_end(args);
^
make: *** [playergroup_classes.o] Error 1

• pocketsec (discussable, of course)
• survival gear vendor only offer trauma patches
• Missing completely: armor/clothing vendor
• melee vendor has a very sparse supply
• firearms vendor has a very sparse supply / room description does not fit stock
• ammo shop needs update (substitute magazines with generic ones)
• talismonger doesn't have enough variety and is lacking enchanting mats and libraries

After a copyover, your character is forced to LOOK the moment your descriptor is brought online. This creates a potential knowledge gap where another player in the room could load after you look and thus appear without you knowing they're there. Solutions for this include either doing a second loop and forcing a look after everyone has loaded, or alternatively printing a message when someone restores from copyover (potentially spammy).

Also, screenreader players probably do not appreciate being forced to look at the full description. Investigate ways to opt out of this behavior.

see issue #62. Output from INFO should be generic.

I'd like to make the following changes to vehicles:

• If you're strong enough, you can pick up a vehicle, transferring it from the world into your inventory and encumbering you appropriately.
• If you're strong enough, you can drag a vehicle, provided it meets the towing requirements (destroyed or unlocked, not moving, etc).
• Destroyed vehicles are not immediately purged on crash/copyover, but are instead transferred to a junkyard so that players have the chance to recover their vehicle before it's purged out. Destroyed vehicles in the junkyard are purged on next crash/copyover.
• Trucks should be towable provided your vehicle is powerful enough.

While standing in a room with a shopkeeper, 'info' with no arguments acts like 'info 1' rather than showing you the MUD info string. It should either display an error message telling you to select an item, or should display the MUD info string.

Filed by Lucien (id 1) from the front of a vehicle of vnum 1317, located inside another vehicle of vnum 1309.

Implement a toggle that lets the player indicate they're using a screenreader.

It appears that item 121 (A UCAS Army Rangers Beret) is an ammo box in the base game; it's referenced by number in ammo creation, and you end up with a beret full of ammo at the end of it. While this is an entertaining mental image, either the item should be rewritten into an ammo container, or the vnum should be reassigned to something that is an ammo container, preferably between the 100-200 vnums.

long shirts[5] = { 38072, 16218, 1024, 28462, 22680 }, pants[5] = { 1023, 39828, 17954, 38009, 22683 };

Intent is to keep radio spam down during creation.

Not sure if it's really an issue. I created a mob (5001) with a weight of 2.00 kg. Looking at it displays the weight as 0kg.

relevant source: act.informative.cpp:500

Commands such as 'mute', 'freeze', etc do not have any timeouts and are set-and-forget commands. Contesting or appealing these commands is difficult, especially if the staffer who set them is not present or has left the game. Forcing the staffer to set a maximum timeout on these commands ensures that problem players have a chance to mend their ways.

Rough design:
freeze <character> <minutes>
if minutes > 1 d * 24 h/d * 60 m/h (one day), display caution message (ex: "Freezing players for too long can provoke further antisocial behavior. Are you sure you want to continue?")
if minutes > arbitrary maximum, display error message and return.

The above will need discussion and calibration before it's designed into the code, and the timeout thresholds will need to be configurable on a per-codebase level.

I don't know why the iload restriction was put into place (my suspicion is that it's so an imm can't iload a credstick, put an arbitrary amount of nuyen on it, and hand it off to a mort without anyone tracking how much was given), but the restriction is not universal. It can be bypassed with LOAD OBJ , or the imm could just buy a credstick from one of the vendors.

If the goal is to make it so imms Cannot Give Out Credsticks Ever, restrict them from doing so in give/drop.

Otherwise, just add logging so that the value of a credstick is displayed in the cheatlog and remove this restriction.

...n't ring (likely due to send_to_room(buf, k->phone->in_room); instead of having any vehicle handling)

Filed by Lucien (id 1) from the front of a vehicle of vnum 1317, located inside another vehicle of vnum 1309.

After installation on a brand-new VirtualBox Ubuntu 16.04.2 LTS image, a new character was created and logged in, then a copyover performed shortly after. When the copyover finished, the character was in 0, paralyzed with 0 qui, and did not appear to be a staff member anymore (no room num, flags, etc). Console showed an idle_delete() took place and dropped 1 character. There was also an invalid val passed char-to-room.

Avenues of exploration:

• Does Ubuntu have special snowflake timestamps?
• How about MySQL on Ubuntu?
• What happens if a character's pfile is lost in the DB during copyover, but the descriptor stays live?
• How can we set it up to allow recovery of first-come admins without requiring direct DB manipulation?
• Why does 0 qui stop you from using wiz commands?

The following modes for the 'show' command are unimplemented (act.wizard.cpp):

library
lodge
rent

This is informational only. It should be discussed whether these modes should be implemented or removed.

constants.cpp lists 133 skills, yet the DB has 135 skill rows.

the pfiles_skills table is regularly updated for immortal characters:

May 26 09:29:13 :: Executing query: DELETE FROM pfiles_skills WHERE idnum=1
May 26 09:29:13 :: Executing query: INSERT INTO pfiles_skills (idnum, skillnum, rank) VALUES (1, 1, 100), [...], (1, 134, 100), (1, 135, 100);

Hence we get two "empty" skills:

< 10P 10M > skills
You know the following skills:
[...]
Blowgun (godly)
Pharmacuticals (godly)
(godly)
(godly)

On each healing tick, for each character, all the items in their room/vehicle are scanned for a workshop. This is probably the source of a lot of lag when you have hoarder players with thousands of items in their apartment while they program. This system should be updated such that a room has an attached list of deployed workshops that is updated when the workshop state changes (deploy, pack up, destruction, purge, world load, etc). This list will be much faster to scan than the list of all objects.

Flagged as both a bug an enhancement because it's 'working as intended' but the original design was flawed.

Reproduction steps:

iload 898
goto 60586
collect westwind
copyover
look

Car is gone.

Players should be warned not to spend money on a radio in chargen.

Players should have the freedom to choose their first language. Check canon rules on that. Skill level could be INT+2 or 3 if no rules exist.

Players need to be warned of difficulties in communicating with other players ICly.

The set command has a switch to modify stats of an offline player (act.wizard.cpp:3460).

set file Puppet Qui 6
A Male Human's qui set to 6.
Saved in file.

However, the value does not get committed to the DB.

I converted 10003 into a power focus and attempted to bond with it, but got the 'You can't bond with that' error (paraphrased). This was resolved when I removed the ImpInvis (Spell) flag from it, after which it bonded normally.

This phase focuses on audit logs (they need to be added to all commands, and viewing them should be doable for both privileged players and high-leveled immortals).

Key features of logs:

• Unless the group is secret, pretty much every action is logged (see command spreadsheet for complete details).
• Logs should persist after a group is disbanded / deleted.
• Log viewing should allow for sectioning by date.
• Log viewing should allow for searching by keyword.

Spellset needs the following:

• Have its TODO entries completed
• Figure out how to remove spells
• Revisit syntax for improved attribute spells

ACMD(do_tail) uses two system calls (ls, tail) to list available logs and tail them. As log/ doesn't exist the command does nothing at the moment. I'm not sure I like passing commands to a shell from the game. However, it'd be nice for a staffer without shell access to view and/or search logs.

Hilarious, but I'm not sure if this is intended behavior. It doesn't seem worth designing around since the cases of staff attacking mortals will be rare to begin with and the staff probably won't care about a fair fight at that time.

The only way I've found to purge vehicles is to purge the whole room, which is unappealing for a variety of reasons. The purge command should be expanded to allow targeting of vehicles.