Thanks guys. I was able to log in finally. Tried a lot of stuff, but what worked was unchecking "redirect automatically" on the HTTP samplers. I would like to figure out why unchecking "redirect automatically" made a difference. Insights? Opinions?
Thanks Sam On Tue, Jun 11, 2013 at 11:57 PM, Robin D. Wilson <[email protected]>wrote: > By the way, this is where recording a login using the HTTP Proxy Recorder > would help up you figure this problem out. > > -- > Robin D. Wilson > VOICE: 512-777-1861 > > > > On Jun 11, 2013, at 10:54 PM, "Robin D. Wilson" <[email protected]> > wrote: > > It appears to me that the way this works is to hide/unhide some elements > on the page using JavaScript. The way that would work is to hide an element > that says "you must enable JavaScript" and unhide an element that has the > login form on it. > > If that is the case, then you can just ignore the "you must enable > JavaScript" warning, and just submit the form anyway. The server has no way > to know whether the browser hid/un-hid anything, so if you submit the login > form it will assume you saw the login form. > > When you are looking in the tree listener, are you looking at the "text" > of the response, or are you looking at the rendered HTML? You really need > to look at the "text" since that's what JMeter actually sees. It may be a > "red herring" to assume that the JavaScript warning makes a difference > since you are looking at the response in a tool (the Tree Listener) that > doesn't execute the JavaScript, and never will. > > -- > Robin D. Wilson > VOICE: 512-777-1861 > > > > On Jun 11, 2013, at 9:41 PM, nmq <[email protected]> wrote: > > Take a look at this code snippet I found for the login page. > > <script type="text/javascript"> > // activate login feature if script is activated and browser is supported > if ($.browser.msie && $.browser.version < 8) { > $('#browser-redirection').css('display', ''); > } else { > $('.script-checking').css('display', ''); > $('#warnings').css('display', 'none'); > } > </script> > > So my understanding is that the login feature is not getting activated at > all as JMeter does not run javascript. > Is that correct? > Is there any way for me to simulate a user logging in with this situation? > > > > > > On Tue, Jun 11, 2013 at 5:10 PM, Deepak Shetty <[email protected]> wrote: > > >> . If the recorded requests have the same problems as your test plan did, > > which is fairly common when you have dynamic data and is not a good > > indicator. > > > > > > > > On Tue, Jun 11, 2013 at 1:52 PM, Robin D. Wilson <[email protected]> > > wrote: > > > >> If you use the Proxy setup, you can then just 'replay' the previous > >> requests and see if they have the same problem as you were > >> having. Basically, disable your test requests, and copy/paste the ones > >> from the Proxy recording in their place. Run the test using > >> the recorded requests, and watch the Tree Listener for the responses > from > >> the server. If the recorded requests have the same > >> problems as your test plan did, then you will definitely need to discuss > >> with your developers (maybe it's just a bug in their > >> code?). If the recorded proxy script works normally, then you have a > >> problem in your JMeter test plan setup - and you can use the > >> proxy requests to figure out what is missing from your test plan. > >> > >> -- > >> Robin D. Wilson > >> Sr. Director of Web Development > >> KingsIsle Entertainment, Inc. > >> VOICE: 512-777-1861 > >> http://www.kingsisle.com > >> > >> > >> -----Original Message----- > >> From: nmq [mailto:[email protected]] > >> Sent: Tuesday, June 11, 2013 3:31 PM > >> To: JMeter Users List > >> Subject: Re: Login failed - javascript > >> > >> I meant they're encoding the request using javascript. Should I have a > >> talk with the developers? > >> Problem is they're offshore *sigh*. > >> > >> > >> On Tue, Jun 11, 2013 at 4:27 PM, nmq <[email protected]> wrote: > >> > >>> Hi Deepak > >>> > >>> Thanks for all that info. I installed fiddler quickly. > >>> > >>> This is what I got in request header: > >>> /UpdateCheck.aspx?isBeta=True HTTP/1.1 which I don't think is > >>> significant OR I could be wrong. Correct me if I am. > >>> It also says "response is encoded and may need to be decoded before > >>> inspection" when I clicked on Inspectors tab. Do you think this might > >>> be the problem? They're encoding the password using javascript? If > >>> yes, what can I do to bypass this? > >>> > >>> > >>> Hey Robin, I've done all of that. I used a tool called badboy to > >>> capture the script, so didn't need to use the proxy. I've tried both > >>> Firefox and Chrome strings for the user-agent in HTTP Header Manager. > >>> Everything was working fine before they deployed the current build > >> yesterday. > >>> > >>> > >>> Regards > >>> Sam > >>> > >>> > >>> > >>> On Tue, Jun 11, 2013 at 4:18 PM, Robin D. Wilson <[email protected] > >>> wrote: > >>> > >>>> First, this isn't really a "limitation" of JMeter, it is an artifact > >>>> of the way web sites work. Keep in mind, JMeter is designed to test > >>>> the 'server' part of the web system, but web systems include the > >>>> 'browser' in the application logic (often times incorporating a lot > >>>> of logic in the JavaScript code that runs in the browser, or in other > >>>> coding systems such as Flash and Silverlight). You could call that a > >>>> 'limitation' of JMeter, but that would be like saying that a chainsaw > >>>> is limited because it can't be used as a good hammer. > >>>> > >>>> There are a couple of ways this is measured, depending on the site in > >>>> question. If it is coming from the server, it is probably looking at > >>>> a header in the request to figure out if you have JavaScript enabled. > >>>> Add an "HTTP Header Manager" element to your test plan, and set a > >>>> User-Agent value... > >>>> > >>>> We use the following User-Agent value: > >>>> > >>>> User-Agent: Mozilla/5.0 (compatible; MSIE 9.0; Windows NT > >>>> 6.1; WOW64; Trident/5.0) > >>>> > >>>> This essentially tells the server that you are making requests with > >>>> the > >>>> IE9.0 browser (which supports JavaScript by default). (NOTE: > >>>> we use this because it is still our most popular browser (actually, > >>>> that's not quite true - it is the most common version of the most > >>>> popular browser 'type' (IE)) - for users hitting our sites.) > >>>> > >>>> But if you have a different user population, you might prefer to use > >>>> Chrome or Firefox or Safari as your 'standard test' User-Agent. > >>>> You can look up their User-Agent strings here: > >>>> > >>>> http://www.useragentstring.com/pages/useragentstring.php > >>>> > >>>> If the HTTP Header Manager + User-Agent value configuration doesn't > >>>> work, you will need to figure out how the server is determining that > >>>> the browser supports JavaScript, and mimic that with your test. It is > >>>> usually easier to setup the 'HTTP Proxy Server', and just collect a > >>>> session from your browser than it is to try to figure it out manually > >>>> though. > >>>> > >>>> To setup the proxy and capture a session: > >>>> > >>>> 1) Create a new Test Plan. > >>>> 2) Right-Click on "Workbench" and select: > >>>> > >>>> Add->Non-Test Elements->HTTP Proxy Server > >>>> > >>>> 3) Make sure "Capture HTTP Headers" is checked > >>>> 4) Click "Start" on the HTTP Proxy Server configuration page (at the > >>>> bottom of the page) > >>>> 5) In your browser, set your Proxy Server address to "localhost", and > >>>> use the port specified in your HTTP Proxy Server configuration > >>>> (default is 8080). > >>>> 6) Visit your site, and perform some functions you want in your test. > >>>> > >>>> These should start to record your requests in the test plan, below > >>>> the workbench section. You can click on one of the requests and see > >>>> what the "HTTP Header Manager" looks like, and use that as your > >>>> default HTTP Header Manager for your tests. You can also see what > >>>> sort of interactions are taking place between the browser and the > >>>> server - some of which may be under-the-covers (hidden from the user) > >>>> and allowing the server to figure out whether the site supports > >>>> JavaScript. > >>>> > >>>> -- > >>>> Robin D. Wilson > >>>> Sr. Director of Web Development > >>>> KingsIsle Entertainment, Inc. > >>>> VOICE: 512-777-1861 > >>>> http://www.kingsisle.com > >>>> > >>>> > >>>> -----Original Message----- > >>>> From: nmq [mailto:[email protected]] > >>>> Sent: Tuesday, June 11, 2013 2:41 PM > >>>> To: JMeter Users List > >>>> Subject: Login failed - javascript > >>>> > >>>> Hi everyone > >>>> > >>>> I have run into an issue running my basic login script for the AUT. > >>>> It was working fine till we got a new build this week. > >>>> > >>>> Now, I have been a functional tester my whole career. My company > >>>> wanted me to do some performance test for them and I figured why the > >>>> heck not. I'll learn along the way, so basically I'm a newbie in this > >>>> area. > >>>> > >>>> Since JMeter is an open-source (translated: free of cost) tool that > >>>> is supposedly powerful, we decided to use it (stupidly, without > >>>> finding out its limitations). I've invested quite some time in > >>>> learning the tool so I'm not ready to give up or switch to another. > >>>> I'm also not a programmer and don't have much info on java or > >> javascript. > >>>> > >>>> Anyways, getting back to the point..... I looked at the response in > >>>> ResultsTree in HTML format and this is the message I'm getting on the > >>>> Login > >>>> page: > >>>> > >>>> "This website requires JavaScript > >>>> Please activate JavaScript and press F5" > >>>> > >>>> HELP!! > >>>> > >>>> Regards > >>>> Sam > >>>> > >>>> > >>>> --------------------------------------------------------------------- > >>>> 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] > >
