วีดีโอ

Luckily Alison has re-upped our contract, so we know she wants the functionality we have delivered.

Thank you. As I hadn’t rehearsed this session, my surprise at the failing test was genuine. To be honest though, I doubt whether the time spent writing that test was actually worthwhile. The consequences of leaving a class open are hardly severe, and it’s telling that the name of the test got out of sync because it was an identifier that wasn’t updated in a refactoring. So maybe the takeaway should be - you can use tests to maintain code structure invariants, and you can set up tests to let you know when an event has occurred.

Thank you. I think some of my later work is better, or at least better (video) edited, so please look around.

I think this is the same question as @dom asked? This would be a better solution, marred only by my not knowing how to make it work! At this point having something that does the job allows me to get on with actual programming, but, as you say, with a bit of an unsatisfactory feeling.

@@eatthepi if the flywayMigrate task output returns if new migrations have been applied, yes you could. But if we are looking at two tasks that need to always run together the declarative approach stops making sense and it's better/easier to just use "doLast{}" on flywayMigrate.

That’s it, hold my feet to the fire! I have plans to look at click to edit, but one thing at a time. The fabulous Ivan wrote plugins.jetbrains.com/plugin/9424-okey-doke-support

Good question, thank you. I did wonder while I was filming. We would still have to make the jooq task depend on migrate in order to have the migrate happen, but jooq would run, with its delays, whether or not the migrate actually did anything. I suppose it is possible to run one task only if another actually ran? But this works, and seems the least complicated, if not the best expression of the actual task dependencies.

@@PairingWithDuncan Yes, apparently this can be achieved using outputs.upToDateWhen { false } on the flywayMigrate task (I haven't personally tested it).

I think I know what you mean by the attitude towards complexity, but like democracy, Gradle is the worst solution except for all the others

@@PairingWithDuncan This is true, sadly. Here's hoping Amper can rescue us from the madness. I'm in the middle of some KMP fun and games myself, and I find I spend more time fiddling with Gradle than I do actually developing features.

There's also Declarative Gradle that's being worked on, which might make the builds for most projects (that don't need extra logic) more simple to understand, eventually?

:-) I figure that at least I’m aiming at a relatively stationary part of Gradle, so stand a chance of hitting.

Oh wow, I’ve been using Gradle for, what, 10 years now, and no one has ever told me about that. Must investigate thank you.

I had heard that there was editor support, but haven’t seen it in IntelliJ. There is a warning (at least in the build.gradle) if there is a CVE against a library, but I didn’t see support for other version updates.

Do you need to talk about it ;-? I do feel that Kotlin is going through a bit of a lull at the moment. We have to put up with constant tooling change for not perceived progress on the language front. The changes, Gradle aside, are rarely breaking though, so I tolerate them. Java is also evolving, but I don’t think we can say either won, and I still prefer Kotlin to Java or Scala for team productivity and fun.

Good question thank you. They do! Gradle-versions-plugin has something complicated, but I’m told that refreshVersions can be configured with refreshVersions { enableBuildSrcLibs() rejectVersionIf { candidate.stabilityLevel.isLessStableThan(current.stabilityLevel) } }

Hilarious intro 😂

@TimSchraepen thank you. It’s the only bit I script, so it’s good to know my efforts aren’t wasted

I think this sums up my view. Neither is good enough, Kotlin is a bit better.

In my opinion is a lot better, the problem with groovy gradle files is that people at some point tend to write scripts that works by accident, they dont declare plugins but declare extensions for them and the plugin is applied in different file etc. This leads to IDE being completly useless, hovewer in kotlin due to type safety it forces you to write better code, and IDE is actually helpful.

At the risk of flogging a dead elephant, this is on my list for next week!

All the more content to thrill you with

You jest, but Jetbrains are working on a new build system...

No, not even. Gradle is working on a new declarative DSL. As for Amper, so far, there's been conflicting information on whether it's a layer on top of gradle or not. If it is, then I fear it will add too many new failure points to be worthwhile.

@oharaandrew314 Isn’t Amper both? The problem is that we want a JFDI build, where we can also customise everything when we have specialist requirements

I suppose my thought was that if I had other builds to convert, the extension would save a bunch of manual work.

To be honest, because when I tried that in a run-through (I prefer the term to rehearsal ;-) it just dug me in to a deeper hole.

Thanks for the feedback. I’m slowly learning where I expect it to be faster than doing things by hand. At the moment these bulk transformations seem to be a good bet.

Thank you for the info. I’ve been thinking of introducing a version catalog, although yet another file format!

I did wonder about the name, thank you. It’s surprisingly hard to record and RTFS at the same time.

Thank you. I think I have accumulated a bit of compiler warning debt - the Playwright deprecation in particular, but that should be fixable when I next see it. I have to confess that I don't tend to check library changelogs unless something goes wrong. Which is a bit of a risk, but with good integration tests hasn't yet bitten me (that I remember ;-).

That's good to know thanks. I may try to see if they can add some style!

this is really useful, specially if you're gonna use htmx all over the place

For attributes that have a predefined set of possible values one can create an enum containing the posibilities, to take advantage of typesafety enum class HxSwap { /** * Replace the inner html of the target element. * */ innerHTML, /** * Replace the entire target element with the response * */ outerHTML, .... and so on } /** * The hx-swap attribute allows you to specify how the response will be swapped in relative to the [hxTarget] of an * AJAX request. If you do not specify the option, the default is htmx.config.defaultSwapStyle (innerHTML). * * @see hxSwapOob * */ internal var HTMLTag.hxSwap: HxSwap get() = HxSwap.valueOf(attributes.getValue("hx-swap")) set(value) { attributes["hx-swap"] = value.name }

I also did this but I never implemented the get()’s. Do you happen to know when they’re used?

@@TimSchraepen actually we never use the getter, atleast i havent found a situation where it's useful to read the value of an attribute, it's actually better to define extension functions, that way we can do it in a more semántic way hxSwap(HxSwap.outerHTML, "from:body") // or whatever fun HTMLTag.hxSwap(value: HxSwap, vararg modifier: String = emptyArray()) { attributes["hx-swap"] = value.name + modifier.joinToString(prefix = " ", separator = " ") }

That's a good tip thank you.

Glad it was helpful! The shortcuts are shown with plugins.jetbrains.com/plugin/7345-presentation-assistant-for-2023-2, although I note that the IDE now has that built in (in a more in-your-face way). I get lots of questions about the test progress bar. It was written by the inimitable @dmitrykandalov. To use it install his Liveplugin (plugins.jetbrains.com/plugin/7282-liveplugin) and then this gist gist.github.com/dmcg/1f56ac398ef033c6b62c82824a15894b

Thank you for your kind words. As you say, contexts are solving multi-layer propagation, and so require intervention at each layer. I’m thinking of reverting most of this refactor in order to redo it when context parameters are finally available (skipping a compiler version).

Maybe it’s a bit clickbait, but yes, it hasn’t been communicated well! And context(_:IO) is just a shame!

@@PairingWithDuncan I don't even think your video is clickbait. There's been somewhat of a "panic" (for lack of a better word) in the context-using community, and it's unfortunate because the reality isn't that bad at all. `context(_: IO)` is annoying, but that's a minor change that can be easily added in the future. I think right now they don't want to add `context(IO)` because there's an idea to have different semantics for that (which would basically match the current CR semantics) but that idea hasn't been explored fully.

For me the infuriating thing about AI is that it can and does give useful results, but, at present, not reliably enough to be economical in my day to day work.

I really appreciate you taking the time to comment. The last few videos have had more downvotes than usual, so I’m wondering what people want to see, so thanks for letting me know.

Thank you. I hadn’t realised that I would be publishing this on my birthday, so it’s a nice bonus to share it with viewers.

It’s my happy place! I’m thinking of organising a Gilded Rose Con next year.

🤔 I think that it has legs, especially as I know two people who know how to write IntelliJ plugins! Have you seen this th-cam.com/video/HsAdTP8T5Ao/w-d-xo.html

@@PairingWithDuncan oh that’s interesting! Thanks for sharing!

Well you seem to be way better at getting TH-cam subscribers than I am! My advice would be to build contacts in the industry by meeting other developers, and build a portfolio of work that you can show off on GitHub

I’m pleased we managed to produce something interesting enough!

Erm?

Very knowledgeable and generous with his time

This is an interesting point. PhD knowledge may be good when we want an answer to a factual question, or a summary on a topic. No one has a PhD in my codebase though, so here we want good inference, not knowledge. Plus ‘clever’ people sometimes make bad developers. Was it Joel Spolsky who wanted to hire “smart, and gets things done”?

My experience has been that for tasks that I know how to do, AI is a bit frustrating. And I know how to do the majority of tasks in my projects. But others are reporting significant productivity gains, and frustration is often a function of lack of experience, so I’m trying to work through to find the sweet spot.

Thank you! More interesting chat about chat to come.

I appreciate the feedback. I’m trying to feature more actual pairing on the channel, but in the meantime there are quite a few with Dmitry to enjoy

I'll put details in the shownotes too

I did warn you in the introduction!

@@PairingWithDuncan You did.

It's certainly nice to have options, and I've learned a lot by using it.

That’s really useful to know, thank you

So many people I respect report having such good results that I think I must be doing something wrong. Maybe my coding style, choice of language and libraries is just not conducive to AI assistance, or maybe I’m just not patient enough. I hope Stephan will help me work out which.