[sudo-users] Has anyone successfully modified the default "mailto" string in sudoers file

[Stoddard, Kevin M]

For Christ sake!!  I've been in this I.T. industry over twenty years.
I've done every task in this industry with the exception of management.
This is simply a problem with visudo finding an error on line 41 and
reporting it as an error on line 40.  Is that because the programmer
that wrote the code started off with line numbers at 0 (zero), or 
maybe because the error caused the line number NOT to get updated 
properly because that comes at the end of the line-syntax-error checking
loop and since it failed, the original programmer forgot to make a 
call to Increase_Line_Number:?

Who cares?!!  The point of this dialog was intended to let other users
be aware of the "possible" error in visudo syntax checking.  You know,
save them some time "chatting" with "experts" on the WEB.

Get a Life.  I have work to do with FISMA issues and interfaces across
dissimilar platforms.  I don't have time to dissect minutia.


-----Original Message-----
From: Russell Van Tassell [mailto:russell+sudo-users at loosenut.com]
Sent: Thursday, January 19, 2006 10:26
To: Stoddard, Kevin M
Cc: sudo-users at sudo.ws
Subject: Re: [sudo-users] Has anyone successfully modified the default
"mailto" string in sudoers file


On Thu, Jan 19, 2006 at 08:11:34AM -0700, Stoddard, Kevin M wrote:

> After commenting out (#) the 'mailto' line, visudo still complained about 
> a syntax error on this line.  Now, that's just wrong.
>
>  [...]
> 
> SUMMARY:  visudo points to the wrong line number when using the -s flag.

As someone else pointed out, because of the way files are tokenized,
you're getting the error at the point the tokenizer file went 'huh?'
Depending on the lines, or what code you're writing, this can often bear
little resemblence to the line where the actual error occurs (things
like quotes and squirrley brackets or other logical delimiters can be
excellent demonstrations of this idea).

The oftenways the simplest way to troubleshoot where the actual fault
may lie is to take a "comment out and test" approach with the tokenizer.
That is, comment out segments of the block(s) you were just working with
and then attempt to compile, test, etc.  Usually the error will become
obvious or, well, you'll finally narrow the offending line dow to the
real line using typical binary sort-like logic to eliminate candidates.

Hope that helps!
Russell

-- 
Russell M. Van Tassell
russell at loosenut.com

"Can't Sleep.  Clowns Will Eat Me."