Re: [SCM] systemtap: system-wide probe/trace tool branch, master, updated. release-0.9.8-214-g9ace370

[Josh Stone <jistone at redhat dot com>]

On 07/30/2009 07:59 AM, Mark Wielaard wrote:
> Hi Josh,
> 
> On Wed, 2009-07-29 at 21:40 +0000, jistone@sourceware.org wrote:
>> commit 0a98fd42f87ef9217917931bd8a47e272a5664ce
>> Author: Josh Stone <jistone@redhat.com>
>> Date:   Wed Jul 29 14:07:37 2009 -0700
>>
>>     Enable variable listing (-L) for dwarf probes
>>     
>>     All $target variables and their C-types are now printed in -L mode.
> 
> Very neat!
> Please do add a NEWS entry for this also. This is really useful.

Thanks!  I'll try to write something up about it.  There's only one
thing in NEWS since 0.9.8 though -- we should really search for other
things too that need to be announced.

> It did break one testcase though.
> FAIL: cmd_parse13: unexpected EOF
> The regexp doesn't expect all the new type information:
> -re {(syscall\.a[_a-zA-Z0-9]*(\ [_a-zA-Z0-9\$]+:(string|long|unknown|
> stats))*\r\n)+}
> 
> Is it so that the $vars always come after to other variable?
> Then maybe we need to split this and only expect something like
> (\ $[_a-zA-Z0-9]+:[^$])* (untested) last?

The $vars are currently always last, but I don't think we should rely on
that.  I pushed a change to the testcase that makes it very relaxed
about matching types, since it's not really doing anything with what it
finds.

Josh