Webrecorder NB questions

jhanasik · Post by **jhanasik** » Fri Feb 08, 2008 5:29 pm

I have been using Macro Scheduler for years but my users have given me a request for which I have questions. The users have a new browser based accounting system and want me to assist them with automating some of the reporting requirements.

I downloaded the trial version of Webrecorder and it appears that this may be of assistance but I have questions.

1. I went through the steps to record the processing all the way to the point of submitting the report request. When I run the process while in the editor I get the "trial version" warning, which of course I expected. I click OK and it runs fine. If I compile the script and try to run the resulting EXE the automation appears to be waiting. Is this the "trial version" warning (which I do not see) holding it up? If so, that's ok, I just wanted to see how it would function when run from the user's perspective.

2. Once the URL loads the user is presented with the logon page. My security people would not be happy if they knew that I was hard coding user and password in these scripts or even providing the userid and passwoard outside of the script and plugging it in when needed. So, I would like the automation to hook into the IE instance AFTER the user has signed on and sitting on the main page. Is this possible. I searched the forum and found 1 reference to doing something like this and this answer was that Webrecorder had to create the IE instance in order to maintain the connection to the brower page, that hooking in after the fact is not possible. If that is true that will probably be a showstopper for me. Security breaches is cause for termination.

3. If I am able to work around #2 above and purchase the required licensing to place the IEAuto.dll on the users computer, I am assuming that I can still just create a Macro Scheduler exe file and give it to the user for them to run without having to purchase the full blown Macro Scheduler product for each user. Webrecorder webpage states that MS version... is required in able to use Webrecorder but I am assuming that this means the developers computer, not the user. Is this correct?

Thanks in advance for any feedback.

Bob Hansen · Post by **Bob Hansen** » Sat Feb 09, 2008 1:28 am

You could have the script open the browser, open the application, and pause untl name and password are entered, and then continue. The pause may be using something like WaitWindowOpen or WaitWindowClosed.

Post by **Marcus Tettmar** » Sat Feb 09, 2008 9:52 am

If you compile to an exe you need to distribue the IEAuto library with the EXE If you are using the Imports method (i.e. you have functions that look like IE_Navigate) then you need an Imports folder beneath your EXE containing IEAuto.DLL and IEAuto.INI. If you are using the straight LibFunc method then you only need to ship the DLL and make sure the script knows where to find it (I usually refer to %SCRIPT_DIR% so it finds it in the same folder as the EXE).

jhanasik · Post by **jhanasik** » Wed Feb 13, 2008 5:14 pm

OK...So now that I have worked around the sign on issue, I discovered that I am dealing with IFRAME which according to live chat support IFRAME is bad as webrecorder does not "see" the elements on the page within the frame. They suggested using send character/text to fill the text fields which appears to work fine. When I have finished with the current request I must click the OK button (which I do using key presses since I can't "see" the button) to register the entered information. Works fine. When done I must refresh the page to setup for the next entry. I do this by sending the following

Let>FrameName={""}
Let>FormName={""}
Let>TagValue={"Charge back Report"}
IE_ClickTag>%IE[0]%,str:FrameName,str:FormName,A,TEXT,str:TagValue,r

IE_Wait>%IE[0]%,r
Wait>delay

This appears to work OK as the IFRAME appears again on the webpage. The problem is when I do the next send character/text the first element does not get populated. It is like it is lost or lost focus or something like that. If I let it run up to the point where it needs to populate the first text field on the second pass and then manually enter the keystrokes everything is fine so I know that the webpage is focused correctly from what I can tell. I have tried everything to get it to start filling in the first field on the second pass but can't seem to get it to work. Anyone have any ideas. Thanks.

Post by **Marcus Tettmar** » Wed Feb 13, 2008 5:32 pm

You probably just need to press TAB a few times to get the focus back to that first field. Or use image recognition to find the field and click in it.

Just a point of clarification - WebRecorder can work with some IFRAMEs, but not all.

jhanasik · Post by **jhanasik** » Wed Feb 13, 2008 5:46 pm

I have tried tabbing even tabbing forward then back to get to the field. Will try the image recognition you suggested. About the IFRAME, when I initially did the recording and was watching the code building in the left pane as I navigated, when I got to the point of clicking the report request button to fill the page with the text fields that is when the recording appeared to stop. When using the advanced tools to try to tell me where I was that was when I discovered that every field was reporting back as IFRAME with index 7.