#selenium

Need some guidance to investigate further

bindings are 3.3.3

python 2.7 using latest pytest and tox

barancev good morning

any chance you know this failure? https://ci.appveyor.com/project/SeleniumHQ/sele...

p0deje: outdated JNI library in the classpath?

*JNA

ussher_: none of the selenium drivers do a wait after sendKeys

I see that randomly failing in master

ussher_: its not that "Firefox is doing wrong" or "GeckoDriver is doing wrong" but we dont support that feature

p0deje: google-search said -Djna.nosys=true might help

barancev thank you, I'll try that!

but I don't know how to pass it through buck

ah

I'll try adding to .buckjavaargs

lucast: dwt is back 😒

p0deje: do you work on capybara?

AutomatedTester no, I avoid it :)

why?

p0deje: ^

p0deje: apparently capybara removes hidden elements?

let me check

p0deje: thanks!

AutomatedTester you're weclome

s/weclome/welcome

[13selenium] 15p0deje 04force-pushed 06rb-handshake from 14f6e9e7d to 14f45ca34: 02https://git.io/vSPOI

13selenium/06rb-handshake 14f45ca34 15Alex Rodionov: Support both OSS and W3C dialects in Ruby bindings...

[13selenium] 15p0deje created 06outdate-jna (+1 new commit): 02https://git.io/v9nva

13selenium/06outdate-jna 143b8a2b0 15Alex Rodionov: Attempt to fix failure on AppVeyor CI

AutomatedTester lucast who is dwt and what's so special about him?

p0deje: he has had a really bad attitude on the issue tracker'

looked through his issues

I see what you're saying

barancev looks like -Djna.nosys=true worked https://ci.appveyor.com/project/SeleniumHQ/sele...

I'll push to master

though I have no idea how this works

p0deje: -Djna.nosys=true forces JVM to use its own JNA implementation instead of any other available in classpath

I love it! https://bug1355890.bmoattachments.org/attachmen...

barancev buck actually had the same problem

[13selenium] 15p0deje 04force-pushed 06outdate-jna from 143b8a2b0 to 142b08036: 02https://git.io/v9nTG

13selenium/06outdate-jna 142b08036 15Alex Rodionov: Prevent JNA library conflicts on AppVeyor...

[13selenium] 15p0deje merged 06outdate-jna into 06master: 02https://git.io/v9nkg

[13selenium] 15p0deje 04force-pushed 06rb-handshake from 14f45ca34 to 1438efaf1: 02https://git.io/vSPOI

13selenium/06rb-handshake 1438efaf1 15Alex Rodionov: Support both OSS and W3C dialects in Ruby bindings...

[13selenium] 15p0deje 04deleted 06outdate-jna at 142b08036: 02https://git.io/v9nkw

[13selenium] 15barancev pushed 1 new commit to 06master: 02https://git.io/v9nIc

13selenium/06master 1491f9f2a 15Alexei Barantsev: Fixing key of pageLoad timeout

AutomatedTester: the issue I ran into was I was doing a wait->until-> expectedCondition( elementToBeClickable ). that passed, then I 'that same element'->click() which would fail because of 308. so needed to add a wait/sleep/pause in there.

have since added a waitForPageToLoad() class that checks the document.ready state and that seams to get around it as instructed in the comments in issue 308

ussher_: thats now fixed in Firefox 55. the discussion was also around sendKeys from the way I read it

so it was just a guess from me that if he added a wait/pause/sleep before the send keys maybe that would help.

great to know its fixed though. :)

p0deje: did you do this from HEAD https://github.com/SeleniumHQ/selenium/issues/3...

03barancev 37 minutes ago - Fixing key of pageLoad timeout | https://github.com/SeleniumHQ/selenium/commit/9...

lucast yes

k, it might be my fault, current released gem handles the error incorrectly https://github.com/SeleniumHQ/selenium/commit/6...

03lmtierney 6 days ago - [rb] fix naming convention for new w3c error classes | https://github.com/SeleniumHQ/selenium/commit/6...

and capybara maintainer pointed it out :-P

#3878

03twalpole closed New Errors added in selenium-webdriver 3.4 never get used due to naming error [rb] - https://github.com/SeleniumHQ/selenium/issues/3878 []

although i'd still expect a WebDriverError

lucast why? capybara doesn't intercept any of those errors

I can check on 3.4.0

k, i'm not sure how capybara does things

looks like he worked around my oops

looks hacky but I guess it should work

not sure why it's not working for dwt

yeah, i'm sure Thomas will figure it out

hi

anyone ever got selenium resulting into firefox hanging up?

I run it over ssh on a centos in headful state

and when a sound starts to play - firefox attempts to show a panel (part of it shown on the screenshot

and that's when it hangs

+ firefox consumes nearly 100% of CPU since that

that panel says 'To play audio, you may need to install the required PulseAudio software'

I do not need it, it's a VM for tests

Hi All, hope everyone is doing good?

I can't navigate to a URL, in Firefox. Just a blank window opens and that's all. The follow is my test https://gist.github.com/anonymous/20263009a338e...

do you get an error?

no just an# empty Firefox window

I got this org.openqa.selenium.WebDriverException: Not a Number

from which line?

wewlad: that looks like a gecko or firefox issue

maybe try on the mozilla channel/irc server

<lucast> its the version of the Selenium-Java.jar ?

or maybe someone from mozilla here can chime in :)

huh?

<lucast> the version of the Selenium-Java.jar in my POM file

how is that throwing a WebDriverException

<lucast> I'm using FF 48, Selenium-Java 3.4.0, GeckoDriver 16

that's not telling me what line in your code is causing the error

<lucast> sorry I will have a look I'm newbie

For the ones who are planning to come to Berlin, for the Selenium Conf, you can get an first idea of the city with this :) https://www.youtube.com/watch?v=ORroLqHB_XM

<lucast> This is the output I get https://gist.github.com/anonymous/2dc4a419db334...

PeterJones: it's because of the Firefox version you are using

can you update to 53?

if you can't, you'll need to downgrade selenium and geckodriver

<lucast> Thanks I will upgrade to FF53 now

<lucast> Thanks I will upgrade to FF52 is it?

i *think* 52 should work

<lucast> Thanks, you the MAN!

WhereIsMySpoon: yeah, you are right. Looks like I've found a way to disable those notifications and I hope that helps, but turns out that selenium creates a new firefox profile each time, how to push some changes that'll be hooked by all new profiles?

you need to make your own profile and load it, wewlad

there are classes for this

like online classes?

uh, I don't want to mess with that :c

I'd better try to add user.js to firefox dir

wewlad: there is FirefoxProfile class in the python bindings to do the work for you

AutomatedTester: nah, I've just added a pref to the defaults for all firefox profiles

this way it solves the problems with all firefox profiles on this vm, not just the ones ran by selenium

wewlad: and that worked?