#selenium

OK. There's nothing too bad in there

I’m doing it differently today

But I’m planning on dropping that

Good. That way lay madness

The plan is to run playback on webdriver

though it’ll require all users to have a hub going

cause I can’t spawn driver processes

NP

re: "feeling where you read code and then weep a little?" ... Is that tears of joy or pain?

They had to have that before

cgoldberg-sl: take it any way you want to

I've become visibly angry reading code before. I don't know if I'm just a psychopath or if that's normal

Once we get JPMS in place we'll be able to make standalone binaries. People won't even know that they have java

A binary that includes jre?

Yeah

Won’t that be heavy?

How taxing is jre?

I dunno. I've never run jlink

Seems like go

Puts only the parts in the runtime that you actually use

Exactly

And we're pretty good with not putting crap into selenium

I was surprised that hello world in go is 2mb

Anyway I’ll talk to you if I reach an impass

I’ll try ripping htmlutils out

OK. Here to help

We need to have a chat about link locators

[13selenium] 15AutomatedTester 04force-pushed 06urllib3 from 14f88494f to 1449a58eb: 02https://git.io/fSoAW

13selenium/06urllib3 14da3b0c5 15AutomatedTester: Add urllib3 dependency to install and test systems

13selenium/06urllib3 1449a58eb 15AutomatedTester: Swap out httplib for urllib3

I don’t have a 1-to-1 conversion

Because it’s more like linkGlob than linkText

Makes it hard to move to webdriver spec

One scenario like when entering text in the drodown text box , it displays / populates the values which matches the entered text

clicking on this highlighted value should be done

tomer-sl: I'm not sure I follow

ask simonstewart>

let me explain clearly

there is one drop down box

in which we need to enter text

the resultant values will be displayed based on text entered

we need to select the matched value from drop down

that value will be displayed on dropdown box as selected value

Anil: I'm kind busy with the IDE discussion, shepherding the release, and working on Grid. Probably best to just address the channel

i am able to automate this through selenium , until the value highlighting

They're watching. If someone can help, they will

Sure simon

thanks

Thank you

simonstewart: did you see my PMs? I feel like I have some odd connectivity issues going on.

No PMs on IRC

womp womp

I'll post here then

re: my point about the IDE...

You're right about not running the tests in the IDE. A mis-wording on my part. We don't intend to have people run the tests *in* the IDE (other than playback for one-offs, etc.)

But if the goal is just to export to code for running the tests in parallel, hooking them up to CI, etc., then there's no real need with the command line runner that we've built

But if the goal of exporting to code is different than that (e.g., on-ramp to more robust bits of Selenium or porting IDE tests into an existing Selenium framework) then that makes sense. Just wanted to share my thoughts re: the runner and the benefit it offers in comparison to code export

That's not the reason for exporting the code

The second reason is it: to help act as an on-ramp

Not everyone using IDE intends to continue using it

There's absolutely no way I'd use IDE on a normal project when everything else is written in IntelliJ

So it needs to be more like a Pixar movie. Something for everyone, aye?

Of course not. But why would you use the IDE then?

To quickly bootstrap a series of tests

To file bugs against the existing site

Because I'm not familiar with the webdriver apis, and want to see what people recommend

Because I started with IDE but now want to use a real programming language

one idea for the Selenium IDE: this of the target user as someone who *isn't* writing a whole browser testing framework

perhaps it's a "manual tester" who needs a quick but repeatable script to live for a short-ish period of time

or someone who wants to try out browser automation as a proof of concept

Right, and IDE is great there. No need to export tests

Yeah, that's more what I'm speaking to. The class of user who won't ever export to code. Trying to offer an option for them that's better than "export to code if you want something better"

(food for though, I'll step back now :) )

There's no reason IDE can't be great

^

Right, we just want to make sure we enable both kinds of user

I think it was tourdedave who outlined the idea of a spectrum of capability. Or titus

Thanks for explaining your "why" simonstewart, it's helpful

Might have been titus

simonstewart. Point 4 was me when I joined the current company I work at.

"never code" <------> "codes a little" <------> "proficient developer" <------> "writes damn frameworks" <-----> "will never write a framework again"

`Because I started with IDE but now want to use a real programming language` Actually I was `I started with IDE, and now I want to learn more, what do I do next` - And someone told me. Learn Webdriver and a language.

IDE supports "never code" to "proficient developer"

When I left using IDE I was still at the level of "codes a little"

WebDriver supports "codes a little" to "will never write frameworks again"

Right

So people have made an investment in selenium via IDE

as in IDE you can use jQuery and some JS commands (Or I learnt to)

yeh, now I'm ironically the maintainer of a framework :)

And we'd like them to preserve that investment --- and build upon it --- as they move to webdriver

I still have my old legacy IDE tests somewhere.

But yeh I'm maintaining this POM framework now: https://github.com/natritmeyer/site_prism

What does a great IDE look like to you?

When I started I had some experience, and it did things similar to Excel (Sorry if bad example)

so the way I used macros, was I tested the macro on something rudimentary - Search button e.t.c.

then I copied that code, and edited it 100 times for all the buttons on my site.

However .... The devs used it a diff way. They used the IDE plugin to quickly fill in a form on the "User details page" (We sold holidays) So that allowed them to fill in circa 30 fields in around 1 seconds.

tourdedave: to me?

simonstewart: Yah

luke-sl: interesting

Something that's easy to use

It doesn't need to be Turing complete by a long shot

Sorry for my sudden disappear I'm commuting without network

tomer-sl: no worries

Ideally, I'd be able to do a little abstraction to avoid repetition

Anyway get back to me later on linkText

tomer-sl: np

tomer-sl: what do you need for "linkText"

I didn't understand what you were driving at before

tourdedave: after that, I'd be happy

link in the old IDE is fuzzy.*linkText.*

Rather than an exact match like linkText

simonstewart: Cool, thanks for that

Oh. We have "partial link text" for the imprecise match

Stepping away for a bit.

They can glob

tourdedave: dunno if that makes sense

tourdedave: toodles!

Better than partial

Ahhh...

You can say link=aaaa*

Bam

Well, I guess you'd end up iterating over links, calling "bot.dom.getVisibleText" on each, and using JS's own regex to match against the text

Asking about future conversion to webdriver

Export

I have document.evaluate I can fly

document.evaluate doesn't work for all implementations (looking at IE here)

:)

IDE won't be deployed on IE

tomer-sl: just add a warning that exported regexs will be handled by each language's own regex parser

Sure

Most of the are PCRE, so it should be fine