[p4] Trigger scripts (Update)

[Xianzhou Guo]

Update:
It's only slow when I use external IP and triggers
tuened on.

If no triggers turn on, it's still fast!

Not wure why?

===================
I guess it's partially my fault.

I found if I use an internal IP, it's fast. If I use
the external IP (even I am on lan), it's slow.
As I work on my laptop a lot, I want to make sure it's
always accessible, so I set it as external IP.


Not sure why the external IP/internal IP make such a
big difference.


Maybe security setup issue.

=========================================

Something else is going on, and you might want to look
at the
"p4 triggers" specification and ask tech support about
it.     They'll
probably help you turn on debugging logs to track it.

4x to 5x slower makes no sense. Something is going on.

	-Jeff Bowles

On Oct 7, 2004, at 9:39 AM, Xianzhou Guo wrote:

> I have a simple trigger written in perl to ensure a
> changelist has a job associate to it.
>
> It has 2 p4 calls though, one to get the joblist,
> another one to describe the change list.
>
> one user. :-)
>
> The speed is not very slow, but compare the checkin
> process without a trigger, the actual speed is 4 - 5
> times slower.
>
> So I decided a trigger may not be very scalable at
> all. I've heard p4 has a c++ API, wondering will
that
> be faster.
>
> I was suprised by the fact that trigger is done by
OS
> commands anyway. I understand the overhead. I am not
> saying perforce doing a poor job, but the
architecture
> of the trigger is doomed to be slow, I just didn't
> realize it can be SO slow. :-)
>
> When I run the trigger in commandline outside of
> perforce, it's fast though.
>



		
__________________________________
Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.
http://promotions.yahoo.com/new_mail