#selenium

titusfortner-sl: ESR - 1 was to allow folks who want to use an old Firefox version to use an old Firefox version

titusfortner-sl: probably

I don't see myself editing the code for the XPI much, but we might get some changes transitively if we modify the atoms

I'm not sure right terminology here. The "charter" expires soon. To move from "draft" to "released" means having everything in the spec nailed down, and two implementations that support it completely?

[13selenium] 15shs96c pushed 1 new commit to 06master: 02https://git.io/vy3ui

13selenium/06master 14f6fde71 15Simon Stewart: Bump the version number to 3.3...

titusfortner-sl: we need to get to Candiate Review (CR)

Which is the stage we're at

Once we have two independent implementations, and we're happy with the test suite, we can move to REC

The charter expires on the 31st March, but we can keep moving to REC assuming we have a realistic end date for the CR period

We're in CR, or CR is when we're happy with everything in the draft and don't plan to make updates?

Judging from the inputs from the browser vendors, it seems as if the two implementations will be geckodriver and the IEDriver

We're in CR right now

!m jimevans

You're doing good work, jimevans!

We're still waiting to hear from Google about whether chromedriver could be one of the implementations

So we will or won't need an extension based on current progress?

We do need an extension

AutomatedTester and I are responsible for making that happen

But we can't get the extension until we know how long we need

And for _that_ we need to get the members of the WG to agree on a date

well, if there's a w3c beta binary or toggle for chromedriver, it would be great to start testing with it

so the "by christmas" won't fly for them

lol

lucast: nope :)

:P

titusfortner-sl: there is. send a capability of "w3c" with a value of `true`

I believe "end of Q2" is what Mozilla are shooting for

whoa

I think jimevans is aiming for "ASAP"

But he has work that gets in the way

I'm planning on getting the selenium server acting as a proper intermediary node ASAP too

i'll put a stake in the ground and shoot for 31 march. i need to get some work approval to spend some time on it.

I think the spec is now pretty solid for endpoint node implementors

i think its super cool that you are both invested enough to want to spend work time doing it, and that your work can allow it

titusfortner-sl: let me point out that w3c capability should be sent as part of the "chromeOptions" dictionary.

@jim

So does that mean Se 4 release before end of June, or we can start tidying things up for a release in July?

:)

@jimevans I now remember @lucast trying this out and complaining that everything was failing

it wasn't pretty, no.

se4 will ship once the spec hits REC + whatever time we need to sort things out

so, Christmas

there were a fair number of gaps when i tried it about a month ago, but it wasn't disastrous.

i didn't do a ton of analysis on what worked and what didn't, tbh. i was just looking to see if there was an implementation at least underway.

still, exciting stuff...

As for the 3.3, we're waiting for geckodriver update & fixing session creation before releasing?

I need more coffee

AutomatedTester: geckodriver doesn't do the full new session payload yet, right? ("firstMatch"/"alwaysMatch")

s/fixing session creation/adding value key wrapper on session creation

titusfortner-sl: yeah. In the java-land barancev has a patch he wants to land to the FirefoxDriver

I want that in too

I have one PR that is approved that I'm going to merge. I'm probably not going to get to the other one I'm working on if we're talking about tomorrow or Thursday release. But I'm sure 3.4 won't be far off :-D

lets make sure to get to double digits so we can have the chromedriver issue of people downloading .9

We do seem to have picked up pace again

[13selenium] 15shs96c pushed 1 new commit to 06master: 02https://git.io/vy32n

13selenium/06master 14efdec2c 15Simon Stewart: Update the java changelog

Huh. Fired up Slack on my laptop and it seems to have forgotten every single channel I was in

Any way it can just pick up everything from Slack on my phone

?

i believe you need to add each channel for each app

they're all technically different "accounts"

That's... frustrating

just to add to your reasons of loving slack :)

hi titusfortner-sl @celsovjf from twitter here! :slightly_smiling_face:

Hey fernandes-sl welcome

Ooh, @simonstewart on Slack?

thank you :slightly_smiling_face:

I'm looking. I don't like it

lol

Meh

Going back to IRC

Doesn't eat memory on my laptop

nom nom nom

At the point I keep 60 some odd chrome tabs open it kind of swamps my laptop memory issues :)

Have you looked at the Great Suspender for Chrome?

It's great

+1

No, that's going to keep my laptop from hanging while writing to page file after I open two more Saucelabs tabs? :)

jimevans: no and when it does we will support the old way too

Furrfu. I need to make the selenium server speak the w3c protocol to clients that use it

I'd been putting this off

need to keep the sleeves down to wipe the tears

anyway... supper time... bbiab

hahaha, I'm seeing this failing test on the latest Firefox Nightly (was working a few weeks ago) and I'm all ready to complain about it...

and when I troubleshoot it...

i see it is regarding the change *I made to the spec properly implemented

titusfortner-sl: hahaha

If I fix it, I don't think the backward compatibility will work, but I think we're ok at this point still telling people to use Nightly with geckodriver?

sounds like 52 might be mostly complete

Why Nightly?

titusfortner-sl: what's the problem you're trying to solve?

well, let me verify the fix with Prod

I definitely need more coffee

that would be an easier problem to solve

titusfortner-sl: people should really be able to use any version of geckodriver and any version of firefox, and any version of selenium :)

But that's hard to manage sanely

With Java, I'm aiming for geckodriver and "geckodriver - 2"

And anything else I'll ignore

simonstewart: i won't sign off on that ("any version of components in any combination") as a valid goal of my portions of the project.

"Ideally"

That's why we have the supported browser versions

ok, fix didn't work anyway

Here is the problem: https://gist.github.com/titusfortner/fc58040afb...

works on prod FF, doesn't work on nightly

and this is only via Standalone Server

(in master)

Well, I have work to do on the selenium server

I though doing `page load` -> `pageLoad` would fix it, but it does not

titusfortner-sl: the whole format changed for the timeouts payload.

yes, but the change doesn't work either :)

you should be sending `{ "pageLoad": <value> }` no "type" property at all.

Alright. So the server has four cases to deal with. w3c -> w3c, w3c -> json, json -> w3c, json -> json

Pass though is easy

@jimevans ahh, I see that now

let me update to that and see what happens

Hi, I'm unable to set a firefox profile of my choosing in C# using the profileManager.GetProfile() method, is this currently an unresolved issue?

OMG. I can delete so much code.....

Using firefox 51.0.1 & Webdriver 3.2.0

genesis_: what's going wrong?

@jimevans this is giving me an unknown error with FF Nightly / Standalone as well: `http://127.0.0.1:4444/wd/hub/session/a67759b8-b549-483c-8575-a3bad92d17c9/timeouts | {"pageLoad":2000}`

titusfortner-sl: i'm just pointing out what the proper payload *should* be. does the remote end (standalone or geckodriver) respond properly to that? ¯\_(ツ)_/¯

@jimevans Am I doing this inherently wrong? http://pastebin.com/r9M4GgmA firefox loads, but a rust_moz temp profile and not the one I specify.