Start a new topic

Better interaction with web pages for Key Sequences

This is specifically related to Web Page connections (Chrome and/or Edge). While it is possible to use the existing Key Sequence (on connection) syntax to automate login to certain basic web login pages, two things would make this much more powerful:


1) Add {WAIT:ONLOAD} (or maybe just {ONLOAD}) that would wait until the page has been loaded; it could be ignored for SSH/etc. connections, but for Web Page connections, it would simply use the browser's existing "load" event. This would allow more responsive auto login, rather than a fixed ms delay, especially on connections that have variations in load time due to link speed, etc.


2) Add {FOCUS:<name>} (e.g. {FOCUS:username}); this allows a specific input field to be focused explicitly, when {TAB} alone doesn't provide a deterministic way to get to the target field.



Hi Eric,


thanks for the feedback. Regarding your first suggestion: it's actually not really possible to wait until a page has been loaded. This is a common issue with web sites. I can wait until a certain URL has been downloaded (typical document completed event in various web browser APIs) but unfortunately in today's world, this doesn't mean that the web page has been loaded. SPA applciations or web sites which are using various javascript frameworks, do a lot of stuff afterwards - sometimes even create the DOM. So, for simple web sites this may work but from our experience the majority of login pages cannot really rely on such events.


Regarding point 2 and the mentioned issue for point 1, I like the idea to have some special tokens for this case. In our last couple of releases we invested quite a bit in our injected javascript to do auto fill. I'm more inclined to leverage the javascript/auto fill feature for special cases. We plan to allow you to include arbitrary custom javascript in an autofill step for special cases. We could also allow sending keyboard events in an auto fill step. I will discuss your suggestions with my colleagues.


Regards,
Stefan

Yes, that certainly makes sense. Most of the web pages I'm dealing with are of the "simple" variety, but I'm sure the ones you describe are becoming more common.


On a related note, what about the browser's built-in password autocomplete feature? For reasons that I can't quite pin down, this went from not working at all, to working sometimes, to working partially (like completing the user name but not the password). Is this a feature that should work? If so, what do I need to do to consistently "enable" it?

Btw, the "simple" case I was describing (waiting for a page to be downloaded) is already implemented. We kick off the key sequence task once this event has been triggered (even wait 200ms IIRC).


Regarding the built-in password dialog: there have been some changes in Chromium and the wrapper from EssentialObjects which caused this popup to show up. Since there are no APIs to fully take advantage of this, the vendor (EO) provided a way to completely disable this popup - which we implemented. I guess that's why you see a different behavior in the most recenet releases.


Login or Signup to post a comment