SmaugMuds.org - Viewing Topic: Bug: Switch

Bug: Switch

< Newer Topic :: Older Topic > can cause crashes.

Pages:<< prev 1 next >>

#1 Aug 5, 2006, 10:08 am

ToadVile

GroupMembers
Posts47
JoinedApr 1, 2006

Alright, i know this was talked about long time ago, but it's still around, so since it is.. i'll post a copy of the core from the crash, i don't know how it can be fixed myself..

But, you can switch into a mobile, then do a command, mud will crash.. Here is the core dump.

Core was generated by `afkmud 6500 hotboot 3 -1 -1 4'.
Program terminated with signal 11, Segmentation fault.
warning: current_sos: Can't read pathname for load map: Input/output error

Reading symbols from /lib/libdl.so.2...done.
Loaded symbols for /lib/libdl.so.2
Reading symbols from /usr/lib/libstdc++.so.6...done.
Loaded symbols for /usr/lib/libstdc++.so.6
Reading symbols from /lib/tls/libm.so.6...done.
Loaded symbols for /lib/tls/libm.so.6
Reading symbols from /lib/libgcc_s.so.1...done.
Loaded symbols for /lib/libgcc_s.so.1
Reading symbols from /lib/tls/libc.so.6...done.
Loaded symbols for /lib/tls/libc.so.6
Reading symbols from /lib/ld-linux.so.2...done.
Loaded symbols for /lib/ld-linux.so.2
#0 0x080c4302 in colorcode (src=0x822349c "&D\n\r", dst=0xbfffb2e0 "\033[0m", d=0x0, dstlen=20, vislen=0x0)
at color.c:918
918 mudstrlcat( dst, color_str( ch->desc->pagecolor, ch ), dstlen );
Current language: auto; currently c++
(gdb) bt
#0 0x080c4302 in colorcode (src=0x822349c "&D\n\r", dst=0xbfffb2e0 "\033[0m", d=0x0, dstlen=20, vislen=0x0)
at color.c:918
#1 0x080c4ce9 in colorize (txt=0xf <Address 0xf out of bounds>, d=0x8718d08) at color.c:1330
#2 0x080c5009 in color_send_to_desc (txt=0x0, d=0x8718d08) at color.c:1441
#3 0x080c50d5 in send_to_char (txt=0x875ec40 "", ch=0x0) at color.c:1457
#4 0x0807f787 in do_look (ch=0x86abea8, argument=0x817a1f3 "\n\r"

at act_info.c:1586
#5 0x080feae0 in interpret (ch=0x86abea8, argument=0xbfffe3d4 ""

at interp.c:987
#6 0x080cce6d in game_loop () at comm.c:4263
#7 0x080cd1b5 in main (argc=7, argv=0xbfffec84) at comm.c:4773

Thats all i got, never touched color.c other then for the fix that was on the page, but thats all.

#2 Aug 5, 2006, 10:53 am

Samson
Black Hand

GroupAdministrators
Posts3,643
JoinedJan 1, 2002

Happen to have the crash file that was generated from this one? A bit more detail is needed for frames 1, 2, and 3 since I see some disturbing things, like a NULL ch that should have been caught, and NULL txt that should have been caught. If any of those are truly NULL then I would have expected the safety checks to stop it from getting as far as it did.

#3 Aug 6, 2006, 6:36 pm

ToadVile

GroupMembers
Posts47
JoinedApr 1, 2006

Alright, Narrowing it down. switch don't like socials with colors, thats what causes it hehe

Here is a core dump, frame 1, 2 from the dump. (Happend when i switched into puff, did wdwon.

#0  0x080c4302 in colorcode (src=0x824393c "&D gain &YW&Rorld &YD&Romination&D today!\n\r", dst=0xbfffbbb4 "\033[0m",
    d=0x0, dstlen=20, vislen=0x0) at color.c:915
915                          mudstrlcat( dst, color_str( ch->desc->pagecolor, ch ), dstlen );
Current language:  auto; currently c++
(gdb) bt
#0  0x080c4302 in colorcode (src=0x824393c "&D gain &YW&Rorld &YD&Romination&D today!\n\r", dst=0xbfffbbb4 "\033[0m",
    d=0x0, dstlen=20, vislen=0x0) at color.c:915
#1  0x080c4ce9 in colorize (txt=0xf <Address 0xf out of bounds>, d=0x872eac8) at color.c:1327
#2  0x080c5009 in color_send_to_desc (txt=0x0, d=0x872eac8) at color.c:1438
#3  0x080c50d5 in send_to_char (txt=0x872cf88 "", ch=0x0) at color.c:1454
#4  0x080c9516 in act (AType=48,
    format=0x85c51b0 "You proclaim that you &WWILL&D gain &YW&Rorld &YD&Romination&D today!", ch=0x8704938, arg1=0x0,
    arg2=0x0, type=3) at comm.c:5179
#5  0x080fdcad in check_social (ch=0x8704938, command=0xbfffd4b8 "wdwon", argument=0xbfffdcf5 "" at interp.c:402
#6  0x080fe984 in interpret (ch=0x8704938, argument=0xbfffdcf5 "" at interp.c:903
#7  0x080cce6d in game_loop () at comm.c:4263
#8  0x080cd1b5 in main (argc=2, argv=0xbfffe5a4) at comm.c:4773
(gdb) frame 1
#1  0x080c4ce9 in colorize (txt=0xf <Address 0xf out of bounds>, d=0x872eac8) at color.c:1327
1327             ln = colorcode( colstr, colbuf, d, 20, NULL );
(gdb) frame 2
#2  0x080c5009 in color_send_to_desc (txt=0x0, d=0x872eac8) at color.c:1438
1438       write_to_buffer( d, colorize( txt, d ), 0 );

hope that helps out a bit

#4 Aug 6, 2006, 6:40 pm

Zeno
Sorcerer

GroupMembers
Posts723
JoinedMar 5, 2005

I actually get this too. Meant to fix it, but haven't yet.

I'm pretty sure it's due to the fact it's trying to send color a mob, which shouldn't happen. Either that or it's trying to check the custom colors on the mob. Just a guess offhand.

#5 Aug 6, 2006, 7:12 pm

Remcon
Geomancer

GroupAdministrators
Posts1,873
JoinedJul 26, 2005

You are right zeno,
in color.c function colorcode
find

            case 'D':  /* Reset to custom color for whatever is being displayed */
               if( ansi )
               {
                  /* Yes, this reset here is quite necessary to cancel out other things */
                  mudstrlcpy( dst, ANSI_RESET, dstlen );
                  if( ch )
                     mudstrlcat( dst, color_str( ch->desc->pagecolor, ch ), dstlen );
               }
               break;

change it to this

            case 'D':  /* Reset to custom color for whatever is being displayed */
               if( ansi )
               {
                  /* Yes, this reset here is quite necessary to cancel out other things */
                  mudstrlcpy( dst, ANSI_RESET, dstlen );
                  if( ch && ch->desc )
                     mudstrlcat( dst, color_str( ch->desc->pagecolor, ch ), dstlen );
               }
               break;

It seems that if your switched into a mob and its sent an &D ch is valid while ch->desc is NULL so when it trys to check it, it just crashes.
So its basically just adding an && ch->desc to the if( ch ) line

#6 Aug 7, 2006, 2:04 am Last edited Aug 7, 2006, 2:05 am by ToadVile

ToadVile

GroupMembers
Posts47
JoinedApr 1, 2006

Alright, another problem found when switched

If switched, you can use attrib (Which is actually dale_score), it'd crash the mud.

Here is a fix for it

in idale.c, in dale_score

Find:

   if( IS_IMMORTAL( ch ) )
   {
      send_to_char( "----------------------------------------------------------------------------\n\r", ch );

      ch_printf( ch, "%sIMMORTAL DATA:  Wizinvis [%s]  Wizlevel (%d)\n\r", s1,
                 IS_PLR_FLAG( ch, PLR_WIZINVIS ) ? "X" : " ", ch->pcdata->wizinvis );

and replace that with

    if( !IS_NPC(ch) && IS_IMMORTAL( ch ) )
   {
      send_to_char( "----------------------------------------------------------------------------\n\r", ch );

      ch_printf( ch, "%sIMMORTAL DATA:  Wizinvis [%s]  Wizlevel (%d)\n\r", s1,
                 IS_PLR_FLAG( ch, PLR_WIZINVIS ) ? "X" : " ", ch->pcdata->wizinvis );

That'll solve that problem for that