🐐 The Greatest Episode of All Time 🐐

This show is supported by you. Stick around till the break to hear more about that. Welcome to Radio Cup A Go for no. No. Just do the normal one. Okay. This is Cup A Go for 08/23/2025. Keep up to date with the important happenings in the Go community in about fifteen minutes per week. I'm Shay Nehmad.

And I'm Jonathan Hall. And I'm so confused.

That, like, morning DJ radio thing?

Was that morning, or was that, like,

Monday night? Coming to you live. Yeah. Hauling the oats. It's right in the morning, giving you all the harsh truths about rust. Go. Blah blah blah.

I think I think maybe you missed a calling, man.

Man, I wish I could work at the radio. I actually did, real radio work. Have you ever done real radio work?

I don't think so. No.

I've done. I've had four shows. Let's see if anybody can dig that up. It was, like, when I was 14 for a school project.

I did

four shows on radio Jerusalem against violence. It was very, very nice.

Nice.

How are you doing? What's that clicking sound I'm hearing?

My fingers are cramping and and and confused and hitting all the wrong keys. I I got this new Kinesis Freestyle Pro keyboard today. Congrats. Yeah.

Can you, like, pull it up to the can you pull it up to the camera? I'll take a screenshot. Oh my god. Very cool.

So I used it for a couple hours, and then I went back to my regular keyboard for the livestream I was doing because I didn't wanna be fumbling over my keys on on live TV or whatever. But I'm back to it again.

In media res, this person doesn't know how to type. It was a rough adjustment. I have the ZSA Moonlander. It was a rough adjustment. They sent me an email. It was crazy. It's called, Typing Yoga. I think it's, like, free. I don't think it's just for ZSA customers, which is actually a meditation online guided meditation session

Oh, wow.

For, like, ten minutes while typing. It's like, feel your space key. How does it make you feel? How relax your shoulders, blah blah blah, whatever.

My goodness.

And unironically, it was pretty good at, like, easing me into my new keyboard.

So I I did consider the Moonlander, and and I may consider that yet still in the future. But I debated, should I go, like, all the way to the extreme or should I sort of do baby steps? And some of the advice I got online was to go to a split keyboard for sure rather than one of the, like, molded ones, like the Microsoft Ergo keyboard. And if it doesn't make the pain go away, then then consider the the full upgrade to the ortholinear style and everything.

I will say, the one nice thing about the ZSA Moonlander is it's very portable. It has this nice carrying case, so it could help me, with it coming to all these meetups that we're about to talk about. That sounds handy. And also, unironically, pulling it up at a meetup or a hackathon, at least in San Francisco, is basically the equivalent of showing up with, like, a hot date to the party. All the engineers are like, oh my god, I love your keyboard.

Yeah. So we have a couple, to talk about. The first one is sort of an unofficial one. Well, it's official, unofficial. We've already talked about GopherCon twenty twenty five coming up in New York City. But some of the folks on the channel, some of the listeners to CupiGo are planning to get together in GopherCon in the NYC. So if you are going to be there, check out the thread on our Slack channel, Cup A Go. I'll put a link directly to that thread in the show notes so you can find it easily.

If you do get up to that after party, take a selfie, send it to us on

the Gopher Slack. Looks like there'll be at least three folks, I think, have said they're gonna be there. So, yeah, send us a photo of your of your Cup O Go after party at GopherCon.

Not not this year, but maybe next year. I'll plan my schedule better, and I'll go. When is the GopherCon? Just for if people haven't listened to the previous episodes. This is like the GopherCon. Right? That's the big one.

Yeah. The GopherCon is coming up starting August 26, so that's, right around the corner here next Tuesday.

Wait. August 26? No way. Should be pretty cool.

It will already be over by the time you listen to our next episode. Our next episode, not this one. Not this one. Unless you listen to this one late, which maybe you do, and welcome. Thanks for listening even if it's late.

How's go 01/28? Tell us if you're listening to this one and a half years in the future. You know what? I have I have heard of listeners that actually war work through the backlog, which is not how I expected people to consume the show.

I I agree. Yeah. But I guess

if you give up some people a an API, they will find all ways to use it, even if you didn't plan on them. It's like an RSS feed you should subscribe to and listen every week. Oh, oh, I can download any episode whenever I want? Alright. I'll listen to it like years after.

Yeah. So this new conference, let me just read off the website the part I can. Let's go conf. Let's go conf twelve three hundred lightning talks. There's a bunch of

other stuff here that I can't read. Maybe you can? So not really, but I'll try. There's a listener called Pavel Druzhin who's hosting this Go conference in Moscow in Russia. It's relevant to Russians. I'm trying to learn Russian, so please forgive me, but I'll try. Pavel vanash Nanashemslakkanal, Raskas Almenia, Konferniz Pagot. About Go, you got it? I'm gonna stop here. I'm gonna sorry, guys.

Yes. Have.

In true yeah. In true Russian fashion, there's a Telegram channel about the about the, like, event, if you wanna join. I actually joined it, and I see a screenshot from our channel, which is pretty funny. The it's, like, about the one twenty to one twenty one x tools breakage, so it's making the rounds. Yeah. But, yeah, the the conference looks pretty pretty serious. It's in Moscow, September 12. The lineup of, speakers and the topics looks pretty cool.

There's one talk here I really wanna see, except that I want it to be in English. Reducing Sego in the video platform, the path from FFmpeg to PureGo. I've been using working on a project that uses SEGO to call FFmpeg, and a PureGo version would be amazing.

It's it it definitely looks like there's a lot of talks that seem interesting. The program just seems great overall. If you're in the Moscow area around that time, like, be sure to drop by. I don't know if it's relevant to American listeners to get on a plane to Russia right now. I I definitely can't.

That's

alright. Let's talk about releases.

I wanna try to jog your memory a little bit, Shadi. Do you remember back in March, we had Grant Nelson on the show?

Yes, of course.

Yeah. We talked you remember what we talked about?

Yeah. It was like a Gopher JS, Workiva. Wait. I am honestly, I'm opening the transcript, but Of course.

Yeah. So we we had go Grant Nelson on who was doing some work on Go for JS, trying to get it updated to support generics.

Go for JS is the compiler from Go to JavaScript.

JavaScript. Right. Right. Go to WebAssembly is built into the Go compiler, but Go to JavaScript is a separate project. And I'm excited to announce that just this last week, his work finally came to a point of completion and Go for JS 1.19 beta two was released with support for generics.

I even see you approved some of the pull requests. Like, I'm seeing you approving upgrading Node. Js from 12 to 18.

That that was a big one. Oh

my god. Yeah. So who is this relevant for?

Also, before I answer that, also, since our last conversation with Grant, Grant has been made one of the official. Now three of us are official maintainers of the Go for JS project. Who's it for? It's it targets a lot of the same people who might wanna use WebAssembly, but it lets you integrate with JavaScript libraries in ways that WebAssembly doesn't. So if you need your Go code and JavaScript code to talk to each other, Go for JS is probably what you want to be using.

Well, well done, Grant. This seems impossible basically. Yes. It honestly seems like looking at the release notes like an impossible list of features.

He has done a great job and I've been doing my best to review his pull requests. Many of them I like, I don't see anything wrong, but that doesn't mean I understand everything that's happening. For sure. So big thanks to Grant for making the making that happen.

An update on previous news, do you remember the stricter JSON tag checking thing, the the security foot guns? I'm jogging your memory now.

Mhmm. Mhmm.

Blog post version of your knowledge in a sense.

Oh, okay.

It's from the trail of bits blog, and, Vesco Franco, I hope I'm saying that correctly, wrote a whole write up about JSON, JSON v two, XML, YAML, about, like, surprising behaviors. So following that discussion, we actually told you about the proposal being opened a few, episodes ago, and now it's closed. It's been accepted. And even someone two hours ago suggested that they could take it up if no one is working on it yet. Yeah.

That's a stretch. Nice thinking on your feet feet there. On your foot?

Yeah. The ones I haven't shot, the ones I haven't shot yet,

for sure.

Alright. So that's the proposal we updated on. Actually, all through from the blog post to the proposal to the, acceptance, now one of us one of y'all need to go implement it. But there's a new proposal we you wanted to tell me about, something with errors.

Yeah. There's a couple of new ones. Let's start with the errors one. So the the proposal so new is a relative term here. When I say new, I mean it's been recently added to the the weekly discussion about features. The the proposal itself was created back in 2022 on March 25. So the proposal's been there a little while, but the proposal is to add a new variant of the errors. As function. I don't know how often you use that function. Do use it very often?

Yeah. Like when you need to find, like translate errors in catch thing, right?

Kind of, yeah. So like you have an error and you want to know, does

it You want to deduce if this error is of type x, because then you handle it differently.

And then you can, like, look at maybe the fields of that particular error type or whatever. Right. Normally, I

talk with APIs, right, I I check if it's a HP error and then, oh, if it's a four two nine, I'll just wait for Exactly. A while and retry If it's a 400, I'll raise it back up somewhere.

Exactly. So yeah, it's sort of like a recursive type assertion on the error interface, right? So the proposal was to make a new variant of that that's generics aware. So errors. As takes two arguments, original error and then a pointer to the error target that it would assert into.

If functionally doesn't propose, anything over, what already exists and it looks the same, like you have to give the customer and the actual instance anyway, What's the benefit other than the fact that it's, like, it justifies a a generic solve? But you see you see, we did need it.

I think I think the annoying thing about the current approach is you have to create this error value that you may or may not use, which I don't I don't think from a performance or operational standpoint it's any different. But like from a code cleanliness standpoint, you have to create this extra thing and then maybe you use it, maybe you don't. With the new version, you can at least put that in an if block. So you could you have your error assignment inside of the I don't remember what they call it, but like the preamble in your if statement and then the value is is the new variable value is constrained to that scope. So it's it's a tiny bit cleaner from a code readability standpoint.

There is one thing about the error as which looks weird. You almost never pass to the standard library, like, sort of input output argument types. You know what I mean? Yeah. Like, you pass the error, which is makes sense.

I'm gonna I'm gonna remember something that I think is an argument for the proposal, and that is that I sometimes get confused about what type of argument to pass as the second argument to errors dot as. It's supposed to be a pointer to an error value, but when your error value is a struct, like sometimes you I don't know if it needs to be a pointer to the struct or a pointer to a pointer to the struct, and that kind of depends on how your receivers are set up. This doesn't have that ambiguity. So that that's a in my opinion, a bonus that makes it worth voting for. Cool.

I like it, though. I like that we're trying to improve the language. Yeah. Generally. Like, oh, we have the this new tooling. There is a way to do it. Maybe we can do it better. And now with the Modernizer, you know what I mean? Like, since you have the Modernizer,

it's It'll fix.

It's slightly yeah. It's slightly less evil to me to suggest improvements that might make the language better, but are not necessarily whereas because in a couple of years, everybody's gonna run that thing on their code base, then as long as you don't break the old one, that's fine. It's not that bad to have two ways to do things. Maybe five ways is not great. Like, I would be super happy if we had a new time format in Go. Yes. I'm bringing it up again.

Speaking of GoFix, have you ever run GoFix and noticed that all your references to XNet context disappeared?

No. No. I've never used the XNet context. I just used the context.

You've never used it.

I might have, but my first usage of Go was in an air gapped environment. Uh-huh. So none of the code, like, made it with me. You know what I mean? I

see. I see.

And all my open source, work, like not necessarily open source, but things that happened on the Internet are, like, since 2018 ish, 2019. So I think 2019, like, it was just a context package. Right?

So, yeah, for those of you who don't remember or who weren't using Go eighteen million years ago when this was relevant, xnetcontext was the original context package that existed there for, I don't know, a while. And old code, know, code that predates 2016 will sometimes still reference Xnet context. That was sort of their experimental version of it, you know, and then they upgraded to the standard library. Xnet context still exists. The new proposal now is to deprecate that package.

What's the implication of making it deprecated? I think the main thing

is that it would cause GoMod to tell you that you're referencing a deprecated package and encourage you to upgrade.

And then if you don't care about it, you can just move on with your life.

Can still ignore it. Yeah.

So what's the the the benefit is Go team has to maintain less code, and they can focus on really important things. And the downside is, like, there's this one guy with a tinfoil hat unwilling to upgrade beyond Go 1.7 Well who's gonna get a little a little warning when he does Go mod. I mean, come on. What's the question?

Yeah. So the original proposal was also included to pass possibly tag and delete the package. I don't think they're actually seriously taking that into consideration, at least not right now. There's no I I see no harm in leaving the package there, but deprecating it would encourage people to upgrade.

And and Google the is hurting for, like, server storage space.

No. And the new context package has a bunch of new functionality that could be beneficial that maybe people aren't aware of if they're still using this for some weird reason. So, yeah, I don't think it's a big deal.

I'm willing to delete some videos from my Google Photos account if they need space back. And I assume, like, one ten second video that I have of, you know, my daughter accidentally picking up the phone and filming the floor is is can it can have, like, a million of these packages in their service for it.

If you are Alright. Personally attached to that XNet context package, go speak your mind right now because it's about to be accepted. So

If anybody opposes this, speak now or forever hold your shaw.

That's right.

As mentioned at the top of the show, this show is supported by you. The main way to support the show is directly financially via Patreon. We don't make money off of this, but this does like cost time and money for us to produce. So if you wanna support us, the best way to do that is to join our Patreon. We wanna re welcome jose underscore d underscore s for rejoining the Patreon.

But if you do cancel, please re add yourself.

Yeah. That's no. I mean, if you cancel and you can't pay for it, it's fine. This is all just for fun. Like, we do appreciate your support.

have found the greatest library of all time. It is truly the GOAT.

Oh my

god. It's is called drum roll, please. GOAT. GOAT is the Go Ascii tool. What GOAT can do for you? It converts Ascii art into SVGs.

Honestly, it's crazy. Like, look at the examples.

This is pretty amazing. I'm imagining it. I'm curious how sophisticated this is. Can I convert a mermaid chart into ASCII and then into an SVG?

It sounds like a perfect, job for an LLM, but I've tried it and it sucks. Yeah. Like, it can't do visual thinking.

Oh, you're right. Yeah. That was a terrible idea.

But the the sort of diagrams you can do here are really advanced, way more advanced than Mermaid.

Yeah. Definitely.

You can have these all these, like, super complicated arrows, overlapping line decoration. It's, like, really unbelievable.

I mean, at some point, I I get I'm thinking, like, why would I wanna spend so much time doing that much ASCII art? Like, it would be easier to pull this up in an actual GUI tool in some cases, but I love it. It's a great tool.

I think it's just fun. I think it's fun. I don't know if it's good, but it's fun. Anyway. And again, it's called the Goat, which is great.

It is. Yeah. Check that one out. What do you have for us?

Fun with slices.

I saw this Fun with

You know these sorts of blog posts where someone is like, I'm new to a thing. I'm just gonna play around with it a little bit to learn it. That's how I like to learn things. I admit that recently I've been I've fallen into the trap of having LLMs teach me about things. I'll just, like, ask an AI, how does x work?

Sweet. I think that's it for our show. Thanks for listening.

Yeah. See you next week, and hopefully in San Francisco if you want to join my meetup.

Program exited. Goodbye.