#selenium

oooh

but that's a firefox driver (well, geckodriver) implementation detail.

Never considered that actually

b4hand: are you waiting for the new tabs to open?

whimboo: yeah I just used a sleep(5) to wait.

whimboo: I am an idiot. I think I put the sleep in the wrong place

So bummed that I cannot get the url paste to work after opening a new tab. Actions Ctrl+v seems to paste double.

jimevans: in his case he is testing the chrome scope of Firefox. So not really a topic for this channel

whimboo: i was more interested in superfreak3 than what b4hand was talking about.

oh, k

bad overlap then

whimboo: thanks though!

yeah, it's just that there's a common thread that people seem to want to open a new tab, but without thinking through *why* they think they need that.

we have an app that copys the url of a certain file in our app. to verify that the copy worked properly we were opening a new tab, pasting the url and waiting for the element with the file name to be displayed

worked smooth until upgrade to v3.5 (and now .6)

seems an odd approach. there are APIs to get the system clipboard contents.

as to the why, there really isn't any other way to check that the file's url copy worked. I don't think I can fire up another driver and just goto that url because there are login issues

there are api's - System.Windows.Forms

not sure if I wanted to include those though, unless there are others I can use

superfreak3: how does your app track login in the browser?

sorry

let me rephrase.

there is a token that is stored in local storage I believe

there are a couple of approaches.

if the login part of it gets too heavy duty, I can chat with our licensing/login/token counter

I can understand that some things changed in the latest firefox to allow the opening of the tabs, which was worked around by adding the proper .tabs profile preferences, but just can't get why the past isn't working

first, you could simply look at the URL, which you can get from the system clipboard. yes, you might need to know a little system programming depending on the OS you're on, but that's not that tough. on windows, i think the API call is GetClipboardData

yep, I think you're right - from system.windows.forms I believe

the funny thing is that if I paste to notepad after the copy, the url is fine

or, if I debug the test, it's fine

it's only release runtime that doubles the paste.

superfreak3: no, i'm going lower level than that. not depending on any framework but the windows API. but if you want to use the .NET wrapper, that's fine too.

i can look into wins api, no problem. Not sure what I'm going to do with that though, as, like I said, the clipboard contents are fine.

okay, so clearly, i'm not going to get to finish my thought, so let me just ask this question: if the URL is properly on the clipboard, are you confident that the feature is correct.

you mean buggy, it doesn't seem so

if I manually run through the test steps, the paste is good and the file is navigated to without issue

no, my question is more fundamental than that. do you need to actually navigate to the URL to be confident the feature is working?

ah, good question, I can check on that, but I believe that is the purpose of the whole url copy function. I will check to see what my limitations are there. Unfortunately I \'ve been called away so hopefully you'll be on later if there are any other questions. Thanks Jim!

Hello! A noob selenium question. Has anybody worked with React + Selenium? One particular question - How do you guys handle table components, and toggle components? Specially while getting the values?

(given that everything is react is in `div`s)

On a real yak shaving expedition today

jimevans: aren't the methods from ISpecificationCompliant public on DesiredCapabilities?

simonstewart: they're not accessible. not marked public.

Ohhh... interesting

Added an image: https://i.imgur.com/CMa13NA.png

because i don't want people toggling that bit.

it's an implementation detail

Indeed

i just got the implementation wrong

I didn't realise C# let you have non-public methods on interfaces

if the interface is non-public

Clever

The problem wasn't particularly obvious.

no, it wasn't. i was once again a victim of my own overzealousness

Hahaha

The New Session stuff has been an utter pathway for misery

i want to drop all OSS dialect of the protocol so badly, but we just. aren't. there. yet.

jimevans: the java tree will continue to support them

as it should.

At some point, we'll get there

Years from now

But by Christmas

hi

> how to get the xpath with firefox ?

hi all, just want to ask whether touch actions can work with chromedriver?

Automation: try them, maybe?

i tried and it does not work so I wonder if anyone has experience using Touch Actions in chromedriver before. Normally, I used it in Android/IOS web driver and it works fine but chrome driver is different story

Hahaha

Those end points were designed for mobile testing. I'd be a little surprised to find them in desktop chromedriver

kereliuk might know better

yes, because I tried to automate a mobile webpage run on desktop browser?

and the event is touch event instead of mouse event so that is why i am thinking of using touch actions

hey!

jimevans: it secretly does already....

Just Actions missing

ChromeDriver action chains will be in a future release soon!

lies, all lies I tell ya

Sorry taking a bit longer than expected ...

kereliuk, Automation was wondering if it was possible to use chromedriver with the Touch actions from the json wire protocol

simonstewart: well, that's the major bit, now isn't it? :)

kereliuk turned round the new session stuff in record time

But currently it won't work :(

It'll be okay

:P

i feel like there was a bot command added recently of relevance..

i wonder how quickly our friends in redmond and cupertino can get spec-compliant protocol handling done…

thanks

:tias

Try it and see - you'll be able to find out much faster if your proposal works by trying it than by asking us if it will work.

jimevans: for the most part ,the prtotocols are similar, particularly once you've sorted out the payload packaging

simonstewart: i'm well aware, having done a protocol conversion already.

lucast: I had no idea selbot2 could do that

:)

Same

WhereIsMySpoon_: added it recently :)

the hard parts are new session and actions (oddly enough, the major places where the protocols differ greatly)

New Session is making me weep

!m kereliuk

You're doing good work, kereliuk!

Actions was kind of fun to write in the java side of things

simonstewart: that's because you're following postel's law.

Indeed.

Given the number of people who are still on 2.x, I kind of feel compelled to be nice

your life would be much easier if you simply limited your backward compatibility matrix. of course, your users' lives would be harder...

Hey: we go all the way back to the original RC protocol :)

So you could connect a 1.0.3 client to a 3.6.0 server and have everything work

Once we get the translation stuff in place, everything tends to remain stable