#selenium

Made some headway with the PHP webdriver, It seems that standalone version creates an actual session and returns valid capabilities

where the docker version for some reason fails to return valid capabilities as well as fails to start a new firefox instance

I bet the same thing happens with the plan jar files, having a hub and nodes

I would post the returned capabilities by both if that helps? (might be spammy) as well as the capabilities that were passed in, they were identical

if I am not wrong, the standalone and hub/nodes have different logics, which could be the cause

Mantas: :gist stuff like that

Please paste >3 lines of code or text to https://gist.github.com and share the resulting URL

can you reproduce by starting a hub/nodes with the jar file?

Does that work? ^^

There's the what's been sent and what's been received in both

sent are both same, whereas the received are completely different

I'll see if I can reproduce with hub/node jar next

WhereIsMySpoon_ & diemol-sl starting the selenium-server-standalone.jar in hub and node configuration also fail and produce the same returned capabilities, however, this time the browser did start

if you check inside the container with the debug images, the browser also opens

Found the issue

It's to do with the way session id is returned

give me a few

That's the line that passes in the sessionID into the driver

I need help guys, is there a way too retrieve the responses after executing an action on a webpage? for instance "response 200" or "statusCode: 200" etc. p.s i'm using python

Now my question, is that the PHP remote webdriver implementation that's incorrect, or does that come from selenium?

Mussojack not through selenium per se, you would have to use a proxy, like browser mob proxy to inspect all the requests a web page makes.

as far as I know, those are the only bindings with that issue Mantas all the other ones (at least based on the official ones), work normally

dude-x: could you direct me to a tutorial of sorts, could not find any usefull tips on google

not at the moment, in the middle of a meeting

google "python browser mob proxy"

Alright, I'll keep digging, thank you for the information and pointing me in the right direction with hub/node

i think it can retrieve HAR files

sure Mantas, I hope the PHP bindings work out that issue, because lots of people rely on them

dude-x: i've been trying to use scrapy to retrive the httpRespons but unfortunatlly i only get 200 :P

you should use a proxy server to intercept all the requests

i think Chrome with chromedriver can do this without a proxy

chromedriver page from google will tell you how to retreive this info

under performance monitoring

@hazmeister ping :)

pong

hello all

I'm having trouble starting Firefox using selenium

Hi all, has anyone automated something like this with standard java/selenium => https://www.ag-grid.com/example.php I'm looking on some advice on how best to approach it

geckodriver won't start (exit status 1) and provides a Ruby-based error log

in this case, I'm not even using the Ruby webdriver bindings

jackofspaces-sl what functionality are you trying to automate, I haven’t used AG grid, but the source doesn’t look too scary

joshin4colour-sl you’re getting a ruby driver from geckodriver or is that from something else?

it's from something else

relatively simple, using that scenario something like writing a method that can take a person name say, and return the value from column 4 in their row

even calling geckodriver --version produces that error

a ruby error?

I guess so

very weird, I haven’t seen that before

:geckodriver

Info: https://goo.gl/Xbtrrm | Download: https://goo.gl/iMvo1d | Status: https://goo.gl/N8DiZU

can’t see any issues in the codebase mentioning “no implicit conversion”

smells like something environment based, can you try your code somewhere else?

I suspect it is environmental

I'll boot up my other machine and try it

jackofspaces-sl `[[role='row']>[col-id='name']`

you’d have to loop through all the matches

or you can do `[role='row'][row-business-key="Bas Rahman"]`

hazmeister-sl if you remember my issue from last night, i have investigated a little further. it appears that even if i pass `options.addArguments("-profile", "/home/seluser/.mozilla/firefox/unauthorized-profile")`, Firefox shows in `about:profiles` that the other profile (`/home/seluser/.mozilla/firefox/owner-profile`) is active

gettign a little closer

i tried setting `Default=1` in `profiles.ini`, didn't make a different

don't get why it's ignoring the arguments

that is the right question to be asking

:)

can you go inside the docker container and look at the profile on the filesystem after the browser starts?

let me try

both folders are there

hazmeister-sl looks like on an envirorment without ruby/rbenv installed, geckodriver starts up fine

which is...kind of weird

FYI, for whoever cares, created this: https://github.com/facebook/php-webdriver/issue...

joshin4colour-sl: do you have titus's webdrivers gem installed?

which geckodriver is it calling?

the global webdriver I have installed at /usr/local/bin/geckodrivrer

I don't think I have that gem installed

but I'm also not using Ruby at all for this

if you ever installed the gem it still uses it without using ruby

hmmm ok

is /usr/local/bin/geckodriver a symlink?

if it was using the gem, it'd be in a different location so you're good there then

oh actually from your stacktrace above, it was using the gem from /Users/joshua.grant/.rbenv/versions/2.4.1/lib/ruby/gems/2.4.0/gems/webdrivers-2.4.0/bin/geckodriver

ah ha, yes, this looks like the case

I'm not using the geckodriver I think I'm using

yes, definitely looks like a PATH issue

thanks everyone!

a while ago, someone told me about a Chrome extension that worked like the Firebug plugin FireChange where it can record DOM changes in Chrome.

DomListener seems to be one the one