thanks, didnt think of the listeners thing , but probably too much work for my liking :). I normally write custom stylesheets to get the data as I want which is reasonably ok (and similar to your second option), as in most cases the phb needs reports he can understand. regards deepak
On Thu, Jul 23, 2009 at 1:51 AM, Adrian Speteanu <[email protected]>wrote: > There is a workaround to this: > > -> depending on the way you organize your test plan tree, your > listeners record only samples from there level down (inclusive) - I > know I read this somewhere, but could not find a reference link for > you. > example: > 1. login > |-[ simple controller ]- > |- 2. user profile page > |- 3. change settings > |-[ loop controller ]- > |- 4. some other action > |- 5. even another action > |- listener(s) #1 -> this listener(s) only records samples > 4. and 5. but not above > |- listener(s) #2 -> this listener(s) records 2. 3. but also > 4. and 5. but not 1. login > this one very simplist example, i get usually into weird situations > because of this method and is not elegant, but thought to mention it > just in case it helps. > > -> if you write your results to a file it gets even simpler to do > this (i do this even if i have to run in gui mode, to have some backup > results just in case - i prefer csv files if i am only interested in > response times) > you just edit the results file and remove unwanted results for one > type a listeners, save the file and then load it into the listener in > gui mode and the listener will only use what you wanted. > > again, not elegant, but effective. > > > On Wed, Jul 22, 2009 at 8:27 PM, sebb<[email protected]> wrote: > > On 22/07/2009, Deepak Shetty <[email protected]> wrote: > >> Is there any feature request to eliminate samples from the response time > >> calculations/listeners or is there any way to achieve this? This is > useful > >> for initialisation sort of samplers (e.g. beanshell) or in cases where > a > >> frame / iframe makes a request to say blank.html which is needed for > >> fidelity of the test , but is not really needed to measure response > times .. > > > > Listeners only see samples in their scope, but otherwise there is no > > provision for ignoring samples. > > > > AFAIK there's no outstanding feature request for this; feel free to > > raise one via Bugzilla. > > > >> (currently I do this after the fact using custom stylesheets on the jtl > >> file). > >> regards > >> > >> deepak > >> > > > > --------------------------------------------------------------------- > > 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] > >
