#selenium

hi

G'day, folks!

zavelevsky and corevo (or, indeed, anyone who's familiar with yarn and webpack): I could with a hand getting the IDE build using atoms built from a snapshot

barancev: yt?

Firefox Nightly and High Sierra aren't best buddies yet

I have a branch I’ve been testing on

trying to load atoms via google closure compiler

but I’m having trouble with the closure dependency tree

closure uses regex to replace goog requires with the correct object

which doesn’t happen when you compile each file in it’s own context

tomer-sl: eh?

I'd be interested to see how you hooked the compiler into the webpack config

I can then look at the flags we use in selenium to generate the files

ok

It’s not staged yet, I’ll link it in a bit

OK

Thank you

:note p0deje We need to update to a version of jruby that supports java 9. Any idea if there is one?

ok!

!m selbot2

You're doing good work, selbot2!

simonstewart-sl https://github.com/jruby/jruby/issues/4111

apparently it will be supported in 9.2

That'll be good to see

Any idea when that'll be :)

no

but

I’ll meet JRuby author tomorrow

so I’m going to ask :)

[13selenium] 15shs96c pushed 2 new commits to 06master: 02https://git.io/vdOeA

13selenium/06master 14893b377 15Valery Yatsynovich: Move building of locators to How enum...

13selenium/06master 148126e10 15Simon Stewart: Add missing copyright header and hook small test into suite

[13selenium] 15shs96c closed pull request #4543: Move building of locators to How enum (06master...06master) 02https://git.io/v5kg5

Good idea

I’m in Budapest for https://www.euruko2017.org and he’s a speaker there

Ah! _THE_ Charles Nutter

I also see that there are some workarounds for the latest JRuby which we might just use

we’re not extensive users of JRuby anyway

Did we ever update to a more modern snapshot?

no

Hahaha.

there were some problems

We're in for a bumpy ride

wait-wait

Let's wait for 9.2, and then go all in on the madness

simonstewart: here

I'm using selenium to test my webservice, that uses google accounts for logging in to authorize

I have a google test account that I use, and the test suite is run on a cloud service like AWS. google detects the account as a bot and hits me with a captcha

I'm starting to put some thought into Java 9 and how we structure our code

Anyone know a workaround for that? like being able to set the account as a test account for my service, or something? I haven't found anything so far.

Jonas__: grab the cookies and set them

Hmm, that's interesting. I'll try that!

barancev: I'm trying to decide whether it's worth using the maven packages as good points for java 9 module boundaries

simonstewart: I'd say "yes"

That's my gut instinct

What do folks want to hear about in the State of the Union at SeConf?

I’d say about PhantomJS

I think many people use it

They're welcome to, but it's going to break

Yeah

So it would be great to mention that :)

simonstewart-sl https://github.com/SeleniumHQ/selenium-ide/blob...

Ah. Yeah

based on the deprecated status of PhantomJS in the bindings, we are also thinking about dropping the docker image it is the least pulled anyway

That looks less perplexing than I thought it would

I

diemol-sl: I guess people can grab the last released version

exactly, we won't remove it, but we don't want to release it anymore

SGTM

tomer-sl: I guess if we add a `test` that the path is in `src/atoms` we can ensure that gets built one way, right?

Yes

Good :)

But if you look at the compiled assets

Closure compiler doesn't resolve the dependencies

It won't without the right flags

Advanced?

simonstewart: I have a suggestion to use trunk-based policy for selenium-ide repository too

Module not found: Error: Can't resolve 'closure-compiler-loader' in '/Users/shs/src/selenium/selenium-ide/src'

I tried looking into the BUCK config

tomer-sl: nah. That's just the compilation level

barancev: I want IDE back in the main tree ASAP

barancev: what does trunk-based policy mean?

me too

Run yarn

especially because I disliked peru

Wait

barancev: it's a stop gap

Yes you should run yarn

tomer-sl: ok. I issed the "do a pull" step :)

simonstewart-sl how come you are so omnipresent?

Eh?

p0deje-sl: prefer rebasing to merging

Able to work and chat

Ah!

got it

I'm awful at both when I do them at the same time :)

p0deje-sl: https://trunkbaseddevelopment.com

Paul Hammant wrote it. He's a star

And one of the original selenium developers

so we have linear history

I always rebase

A linear history that makes sense

But I never work in master

I assume you have to work in master to do trunk-based

yes, you develop in branches, but then you rebase to master

like with SVN

p0deje-sl: I have short-lived local branches, which I constantly rebase on master

When I'm done, I sqash the commits, do another rebase, and merge back the single change to be the new HEAD of master

same for me

03shs96c 31 minutes ago - Add missing copyright header and hook small test into suite | https://github.com/SeleniumHQ/selenium/commit/8...

except for massive changes

like I did recently with Ruby bindings tests

Massive changes, I try and land a bit at a time

With each commit improving the state of master

So if I get hit by a bus, at least the world is a little nicer

It's the Campsite Rule

"Leave the campsite in the same condition or better than when you arrived"