Hey James,

>       Should the middle case be failing? The users take on this was
> "I said zero probes matched was okay, so its fine if that doesn't  
> match"

We did this intentionally because what the user typed there is always
incorrect. While it's true that -Z means that it's fine that we matched
no probes, in this case the user has specified a probe that isn't just
absent now, but will _never_ be created. The intention of the -Z option
is to catch probes that aren't yet created, but could be potentially.
We don't want to allow the user to create a script that is necessarily
incorrect.

Adam

On Nov 24, 2008, at 12:58 PM, James McIlree wrote:

>
>       I think this is a bug, but I wanted to double check...
>
>       A user reported that the -Z flag wasn't working correctly,
> and I have narrowed it down to this test case (on Solaris):
>
> /*
>  * Fully specified, no errors
>  */
> /home/ovrskeek# dtrace -Z -n 'pid$target:libc.so.1:malloc:entry' -c ./
> simpledtrace: description 'pid$target:libc.so.1:malloc:entry' matched
> 1 probe
>
> /*
>  * Fully specified, but function name is wrong
>  */
> /home/ovrskeek# dtrace -Z -n 'pid$target:libc.so.1:NoMatch:entry' - 
> c ./
> simpledtrace: invalid probe specifier pid$target:libc.so.
> 1:NoMatch:entry: failed to lookup 'NoMatch' in module 'libc.so.1'
>
> /*
>  * Fully specified, function name and library name is wrong
>  */
> /home/ovrskeek# dtrace -Z -n 'pid$target:libc.NoMatch:NoMatch:entry' -
> c ./simple
> dtrace: description 'pid$target:libc.NoMatch:NoMatch:entry' matched 0
> probes
>
>       Should the middle case be failing? The users take on this was
> "I said zero probes matched was okay, so its fine if that doesn't  
> match"
>
>       I originally thought it was an error due to being fully specified
> (and wrong), but the final case is also fully specified and wrong, and
> does
> not issue an error.
>
>       James M
>
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss@opensolaris.org


--
Adam Leventhal, Fishworks                        http://blogs.sun.com/ahl

_______________________________________________
dtrace-discuss mailing list
dtrace-discuss@opensolaris.org

Reply via email to