Re: Dup Bug Revisited

From: Cris Jacobin (jacobin@BELLATLANTIC.NET)
Date: 01/22/98

>On Thu, 22 Jan 1998, Cris Jacobin wrote:
>>[...] if (GET_NAME(dt->character) && !str_cmp(GET_NAME(dt->character),newn
>>(gdb) print *dt
>>$1 = {descriptor = 30, host = "", '\000' <repeats 15 times>,
>>  large_outbuf = 0x0, input = {head = 0x0, tail = 0x0}, character = 0x0,
>The problem is 'Valid_Name' is called before you have a character. Note the
>'0x0'.  I see you have the ANSI color snippet installed.  Check your
>nanny.c and make sure 'Valid_Name' is called *only* at the GET_NAME step.
>The patch may not have applied entirely correctly either, did you see 'fuzz
>factor'? That almost always means something went wrong. :)
>Anyone without that color snippet seeing the problem?
>(perhaps I should add a dt->character check...but at that point, you
>*should* have one.)

        Hmm.  valid_name is called twice in nanny().  Once when it checks
for the false positive due to deleted players.  Once again when it falls
through the rest of the case statement and creates a new 'unknown' char.

        I guess tommorow when the eyestrain hurts less, I'll look at the
code and try and determin if there's a return/break missing causing the
fallthrough else block to be executed after a conditional is processed.

        As far as the color snippet, never had a problem with it and it
patched clean.  The dup.diff patch however was added by hand.

        Since the problem is sporatic and evidently not reproducable, and
shared with another RH 5.0 user is it possible that it might be yet another
problem with the glib?  There -is- a libso bugfix, but almost positive that
RPM is installed.

-jac 'Thanks for the assistance'

     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     | |

This archive was generated by hypermail 2b30 : 12/15/00 PST