06:33 <nakilon> so I've got a job

06:35 <nakilon> a typical story: QA dept and their managers once decided that they need web browser automation, but no one was going to hire programmers so they give manual testers with no relevant education the google search query "selenium and ruby" and order: "automate!"

06:39 <nakilon> after 10 years living under a rock they've decided to hire a programmer -- me; and I am supposed to "make it properly" and the success metric is if they now write tests faster; but the problem is that I am making a new testing library but the girls don't even know what is Array#to_h, Object#tap, **kwargs, &block, i.e. they understand nothing

06:40 <nakilon> and so first they took a book about rspec, then a book about ruby, they are reading non-english blogs and a month has passed after they've joined me but they still didn't make no a merge request

06:45 <nakilon> or for example, of course they had no CI and instead they had a PTSD from previous lead who was blaming them in syntax errors in codebase, so I made a jenkins job, etc., and boss asks "he made this thing -- do you need it?", and they tell him: "we don't understand what it is for", etc.

06:45 <nakilon> the question is: how the hell teach them?

10:07 <leah2> phew

10:40 <kjetilho> nakilon: demo time. show the workflow, what happens if you make a typo and how the stuff saves them.

10:59 <leah2> yeah, lots of hand on teaching. they have the benefit of experiencing pain points already, so showing them how it avoids those could help

12:34 <o0x1eef> Those type of tests are more prone to random failures more than anything other I've worked with. So I think it might be hard to make bulletproof. I guess any small step forward is progress

12:38 <leah2> also true -.-

12:38 <leah2> but if you've seen people clicking manually all the time they are happy for every automation :)

12:39 <kjetilho> leah2: don't be so sure...

12:55 <nakilon> they say "good QA engineer is a carefull engineer", advocating with this statement the refusal of any automation or error-avoiding practices

12:55 <nakilon> -- why you do this? -- to avoid this kind of errors -- but you are QA engineer, you should not do the errors!

12:57 <nakilon> i.e. QA dept != automation dept, rather opposite, they are used to do everything manually, etc.

13:57 <o0x1eef> Maybe this is a controversial statement :) They're working against themselves IMO. Same as people who think they can ignore AI. Eventually the world will pass them by.

20:38 <havenwood> FWIW, an install script for macOS that enables MMTk GC and links all the bells and whistles for Ruby 3.4: https://gist.github.com/havenwood/3724538a7a9a3cf90996ec44af0d31f0

20:44 <o0x1eef> Cool. Is it actually an improvement over the traditional GC though ? Or just an experiment ?

20:45 <havenwood> o0x1eef: MMTk say not ready for prod. Works on my machine.

20:45 <havenwood> I think it's meant to just be the first included modular GC but experimental and I dunno how it fares in benchmarks.

20:45 <havenwood> Seems fine so far? ¯\_(ツ)_/¯

20:46 <havenwood> It has various modes it can run in too, at different levels of polish.

20:46 <o0x1eef> It's still cool, and I think that's a nice feature jruby has had for a while (the choice of multiple GCs).