Analog vs object test automation

[CarlFerrer 0]

Hello there!

I have started to automate an app (mostly windows/desktop) and am wondering what is the best way to build test scripts avoiding analog code. That is, avoiding mouse movements, coordinates, etc. I really would want to go for the object recognition way of automating!

Here is a snippet of code I recorded with the Au3record program, which I like to avoid as much as possible.

Thanks in advance for your time.

CarlF

---

_WinWaitActivate("Citrix Receiver - Internet Explorer","Address Combo Contro")

;Maximize
WinSetState("Citrix Receiver - Internet Explorer","",@SW_MAXIMIZE)
;_WinWaitActivate("ABS-Core TEST <on DTMLA93> - \\Remote","")
;MouseClick("left",744,14,1)

MouseClick("left",542,89,1)
MouseClick("left",388,282,1)
_WinWaitActivate("Login - \\Remote","")
MouseClick("left",34,115,1)
MouseClick("left",413,467,1)
_WinWaitActivate("ABS-Core TEST <on DTMLA93> - \\Remote","")
MouseClick("left",52,123,1)
 

[Earthshine 287]

I would start here for automating Citrix with AutoIT and branch out as you see fit

Unfortunately, for you, the way Citrix offers the session to users (a big updating image) you can only do keyboard/mouse/pixelsearch stuff, no object clicks as I understand from these threads

 

Edited June 14 by Earthshine

[junkew 299]

Citrix is mainly bitmap finding see for example findbmp thread in examples with some other image finding solution. But even if bmp is found you have to move mouse and click. Maybe if yo understand the vnc citrix protocol you can do smarter things but so far never seen a solution for that.