sdb behavior

[Mario T DeFazio]

In article <1991Apr30.150441.12993 at mccc.edu>, pjh at mccc.edu (Pete Holsberg) writes:
> In article <11537 at uwm.edu> adk at csd4.csd.uwm.edu (Andrew D Kailhofer) writes:
> =In article <1991Apr25.201206.12430 at mccc.edu> pjh at mccc.edu (Peter J. Holsberg) writes:
> =>I have a program that dumps core but when I invoke sdb with the name of
> =>the executable, sdb says "no source file."  The command line I used is
> =>simply "sdb program".  Program was built from "cc -g -o program program.c".
> =
> =Being an habitual System V person, I've seen this sort of thing
> =before.  There are several ways that this can be caused, the best two
> =contenders being (1) you aren't in the directory containing the source
> =of your code, or (2) [and I believe this is your problem] the error
> =actually occurred someplace inside a library call or some such like
> =that.
> 
> It turns out that the error was in the kernel's floating point driver.
> 
> =You might want to try doing a 't' and seeing what the stack
> =trace says, if it says anything.  Sdb isn't really very smart, and
> =gets pretty confrused if pointers get screwed up, which is also one of
> =the things that causes library calls to blow up...
> 
> But why does it say "no source file"???  I'm in the directory that
> contains core, program and program.c.
> 
> Pete

You just said the reason above, i.e. 
> It turns out that the error was in the kernel's floating point driver.

sdb is saying that it can't find a source file for the routine
where the bug was.  You don't have source for the floating point driver.
You probably don't have source for things like libc.a and libm.a either,
unless you have a source license.
I admit, though, that sdb's message could have been more explicit.


Mario T DeFazio			AT&T EasyLink Services, Lincroft, New Jersey
AT&T Mail: !mdefazio		
Internet:  mdefazio at attmail.com
           mtd at pegasus.att.com