Can I only send to controls that are visible?

[coghlan]

I want to automate the configuration of <n> objects in a 3rd-party application through the available drop-down menus, radio buttons and text fields, and repeat the procedure say, 100 times. The application is, I believe, COM, not Java or .NET

Is it possible I could be looking at working at the level of specifying mouse coordinates and button presses?

The reason I ask is because I tried the window information tool on other apps like Thunderbird and noticed that it often only identifies the parent object (combo box), and only provides mouse coordinates as I navigate the list. I was going to try creating a test script that would send 10 e-mails (not spam) using Thunderbird, but seem to be limited by the fact that I can only see mouse coordinates and not the class names or indices when inside a combo box. I don't see anything at all for the basic file/edit/etc. toolbar for the window itself.

I was expecting to be able to jump to a control, select the 3rd item in a pull-down menu, left click and then click the SAVE icon.

Are there some apps for which I either can only work with mouse coordinates or possibly not be able to automate input to them at all?

[PsaltyDS]

I want to automate the configuration of <n> objects in a 3rd-party application through the available drop-down menus, radio buttons and text fields, and repeat the procedure say, 100 times. The application is, I believe, COM, not Java or .NET

Is it possible I could be looking at working at the level of specifying mouse coordinates and button presses?

The reason I ask is because I tried the window information tool on other apps like Thunderbird and noticed that it often only identifies the parent object (combo box), and only provides mouse coordinates as I navigate the list. I was going to try creating a test script that would send 10 e-mails (not spam) using Thunderbird, but seem to be limited by the fact that I can only see mouse coordinates and not the class names or indices when inside a combo box. I don't see anything at all for the basic file/edit/etc. toolbar for the window itself.

I was expecting to be able to jump to a control, select the 3rd item in a pull-down menu, left click and then click the SAVE icon.

Are there some apps for which I either can only work with mouse coordinates or possibly not be able to automate input to them at all?

Thunderbird is a Mozilla application like Firefox and does not use standard Windows APIs to draw it's controls. Nor does it expose a COM automation interface the way Outlook or IE do.

AutoIt uses the standard APIs. Apps that draw their own graphical objects (Flash is another example) are opaque to AutoIt.

[James]

Look at the ControlSetText etc in the helpfile.

[coghlan]

Thunderbird is a Mozilla application like Firefox and does not use standard Windows APIs to draw it's controls. Nor does it expose a COM automation interface the way Outlook or IE do.

AutoIt uses the standard APIs. Apps that draw their own graphical objects (Flash is another example) are opaque to AutoIt.

Okay, but if I know that the list item that I want to select is at coordinates (x, y) after opening the drop-down menu, can I still send mouse input to the app and create the <n> objects, or should I not investigate any further?

[PsaltyDS]

Okay, but if I know that the list item that I want to select is at coordinates (x, y) after opening the drop-down menu, can I still send mouse input to the app and create the <n> objects, or should I not investigate any further?

There is a scripting language available INSIDE Mozilla apps. You can create XUL extensions and add them to do just about anything. If you want to automate from outside the app, then blind mouse/keyboard actions are all you got.