Well, here's the deal. I would like to view a stack trace of everything that's being called in my web application, starting from when the web page is loaded to browsing through the different pages. Starting with the -X is certainly not an option in my case since the Apache server I'm using is maintained by another dept and all i have is access to start, stop or restart the server though a script.

I could certainly have Apache, mod_perl etc running on my own development box and have root access to it, but the above question stems more out of curiousity and an intention to experiment given the circumstances, rather than having to solve some immediate problem.

Now, I could certainly use Carp::Cluck to obtain a stack trace but that also means that I would have to 'use Carp' in every module that could ever get called, right?

Also, since this is just an experiment, I certainly don't mind having to write my own XS or C code. Just curious.

Thanks again,

Gokul

Perrin Harkins <[EMAIL PROTECTED]> wrote:
On Mon, 2006-03-06 at 13:54 -0600, Chris Werner wrote:
> I don't think that is what he was asking... He asked if there was a
way to attache the
> DEBUGGER to a specific pid; he did not address the issue of getting
the correct input to
> that pid.

The one isn't much good without the other.

In theory, something elaborate could be done with signals in order to
tell the debugger to activate (you'd have to write some XS code), or you
could put something in the URL that triggers it in a handler, but you
still have to send a request to that specific process.

In short, it's a lot of trouble. What's the problem you're trying to
solve, Gokul? Restarting with -X is usually not a big deal for a
developer.

- Perrin



Yahoo! Mail
Use Photomail to share photos without annoying attachments.

Reply via email to