I like that; perhaps this idea could be extended to (some) other test
elements?
I can only think of Response assertions at present, for which there is
already an option, but there might be cases were a user wants to add a
certain test element to every sample group.
However, I'm not sure I understand how you are proposing to use the variable
in the timer.
Where would the variable value be defined? And how would the different
variable names be generated?
Or was the idea to copy the Timer element to the generated test plan, and
replace the variable name (if found) with the actual time delay?
I think this would be very flexible - the user could create a Timer under
the Proxy with one of:
- a special value, which would be replaced by the actual time
- anything else, which would not be replaced.
The latter could be a fixed value, or even a function or variable reference,
so long as it did not match the special "please replace me" string.
If the special string was replaced as a sub-string, it would even allow the
user to do something like:
Delay = ${__P(delay,REPLACE_ME_PLEASE)}
i.e. make the actual delay the default value for the "delay" property.
S.
-----Original Message-----
From: Jordi Salvat i Alabart [mailto:[EMAIL PROTECTED]
Sent: 02 December 2003 09:34
To: JMeter Developers List
Subject: Re: Another suggestion for the proxy
That's not what I was thinking about.
I was thinking about using a variable that would be replaced at
recording time -- that is: a timer placed in the HTTP Proxy would serve
as a template to create one timer inside the 1st request of each group.
Variable replacement for this particular variable would happen at
recording time, not at runtime.
This would give the user the choice between the trivial solution (a
Constant Timer for the interval from the previous request) or more
complex ones (such as a Gaussian Random Timer with the average set to
that interval and a deviation -- fixed or computed from it). The
Constant Timer would still be the most common usage, I guess.
--
Salut,
Jordi.
En/na [EMAIL PROTECTED] ha escrit:
> Your variable seems a sensible choice.
>
> Would this be the start of some pre-defined variables that would always
appear in the "User defined variables" section of the "Test Plan" Element?
>
> I've just scanned the all the timers. Would it be sensible to create a set
of default variables for:
> - Target Throughput (Constant Throughput timer)
> - Thread delay (Constant Timer)
> - Deviation (Gaussian Random Timer)
> - Constant Delay Offset (again, Gaussian Random Timer, and Uniform Random
Timer)
> - Random Delay Maximum (Uniform Random Timer)
>
> Paul
>
>
>
>> from: Jordi Salvat i Alabart <[EMAIL PROTECTED]>
>> date: Mon, 01 Dec 2003 12:44:10
>> to: [EMAIL PROTECTED]
>> subject: Re: Another suggestion for the proxy
>>
>>
>>
>>En/na Jordi Salvat i Alabart ha escrit:
>>
>>>... forgot about this one:
>>>
>>>Plus optionally add a timer to each sampler that is not the first in its
>>>group. We could use a constant timer with the time passed since the
>>>completion of the last sampling in the previous group.
>>>
>>>Or, if a sampler were found within the proxy, we could just stick that
one.
>>>
>>>Makes sense?
>>>
>>
>>Or even: we could have a variable that could be used in the "timer
>>template" and would be replaced with the time passed before sticking it
>>into the generated sampler.
>>
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: [EMAIL PROTECTED]
>>For additional commands, e-mail: [EMAIL PROTECTED]
>>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
