#selenium

titusfortner-sl: do you have the standalone output when you got that last exception? #2856

03titusfortner open Window Size returning error - https://github.com/SeleniumHQ/selenium/issues/2856 [C-server]

also… don't you format stackstraces better? ;)

stack trace?

the response was a stacktrace

with a NullPointerException in it

this? https://gist.github.com/titusfortner/fc0bec840a...

yes

Sorry :)

the error implies the response didn't have "width" in the value

right

it’s a GET

annnnnnnd...

which value return value?

the response object from geckodriver

sorry, still caffeinating

ah, well, it works when I go directly to geckodriver

you are leaving out crucial pieces to the puzzle :)

do you have the output of the standalone server during the error?

It’s one of my specialties… I typically do a $DEBUG = true and copy/paste what I get back. Is there a log somewhere that is more helpful?

maybe a geckodriver log?

by default the standalone doesn't log to a file

let me see if I can pass that option into remote server

Do I need to start the Remote server with a switch in order to get the --log for geckodriver?

looks like no… you can't :)

hmm, maybe we should allow that ;)

I’ve not done much configuration of remote driver in general

also… looking at #2855 … java appears to be doing the best it can… it will throw a TimeoutException with the message "script timeout" … can you show the actual error response object the server returns?

03titusfortner open Wrong Exception for Script Timeout - https://github.com/SeleniumHQ/selenium/issues/2855 [C-server]

It was working before? :)

because you were relying on status codes ;)

not strings :)

and simon changed everything...

it should be returning with the status code too, but need to test it further…

Shouldn’t I be relying on status codes?

probably should start relying on strings now

at least you'll start needing to do that come 4.0

I guess since there are a bunch of different strings that return the same status codes

got it

sigh

not saying there isn't a bug to fix in java ;) but it's for the legacy status code bridge

spec still says it needs to throw that error code, though :)

or are the codes different?

what? there are no "code" aka numbers

oh different

can you show what the json response object is?

ok, I see. JWP returns a status code 28

*should return

yep

and they are all 408 errors in W3c

ugh

ok

#thankssimon :)

but i think standalone should still be replying to remotes in JWP

so… there's the bug ;)

Also, seeing the same issue for Element position as I am Window position

position?

bugged was logged for size? ;)

oh

right

#moarcoffee

Ok, so “string” response for #2855 is still ‘timeout’ instead of ‘script timeout'

03titusfortner open Wrong Exception for Script Timeout - https://github.com/SeleniumHQ/selenium/issues/2855 [C-server]

So even were I to change how it’s parsing errors, it’s still giving back the wrong one

ok… i'll look more

thanks. I’d say I wish I understood the java code to be of more help, but...

wow, no, I filed this bug back in August: #2554 and forgot about it, it should be guarded out, what did I do to unguard it… hrrm

03titusfortner open Script Timeouts - https://github.com/SeleniumHQ/selenium/issues/2554 [C-server]

i think i have a fix though ;)

Yay

yuck… worst is forgetting about your coffee… taking a sip and having it be so cold!

lol

I indeed hate that

spit it right back out into the cup… ;)

hmm, didn't fix it yet.

pretty sure i know where the issue is… just need a debugger attached

pretty sure #2856 & #2857 are related

03titusfortner open Window Size returning error - https://github.com/SeleniumHQ/selenium/issues/2856 [C-server]

03titusfortner open Element Position & Remote Server - https://github.com/SeleniumHQ/selenium/issues/2857 [C-server]

Is Element Size no longer a thing in the w3c?

correct, size and position combined: https://w3c.github.io/webdriver/webdriver-spec....

oh, right which is why they both fail

I implemented that correctly

whew

titusfortner-sl: the thing is, the server still assumes clients will speak JWP instead of W3C :) getting size should work, right?

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

13selenium/06master 149453cc7 15Luke Inman-Semerau: need a new ScriptTimeoutException in order to pass appropriate error codes to clients....

two bugs, one push ;)