Deep Dive
Shownotes Transcript
Hello and welcome to Python Bytes where we deliver Python news and headlines directly to your earbuds. This is episode 427, recorded April 7th, 2025, and I am Brian Auchin. And I'm Michael Kennedy.
And this episode is brought to you by posit packet, package manager. Um, check their spot out later in the show. Also connect with us if you'd like to on mastodon or blue sky, uh, links are in the show notes. Um, and then also if you'd listen to us, uh,
through a podcast player or something. Occasionally it's fun to stop by and watch us live or at least see past live episodes. You can go to pythonbytes.fm/live. We usually do it Mondays at 10, but sometimes it changes. Anyway, it's fun to have you here and some people hop in and ask questions and it's good.
Also, after the show, when we have everything, all the dust settles, we send out links through our newsletter. So please go to pythonbytes.fm and add your name to the email newsletter, and you'll get the links and some extra content delivered right to your inbox. It's pretty good. Absolutely. Michael, what do you got for us to start with? Let's go on a trip, Brian. Yeah.
Let's go on a trip. Okay. I want to go to take a little trip to GitTown. GitTown 18.3. Let's do it. So GitTown is an interesting project. I believe it's written in Python. I'll double check that in a second. Although it's not totally germane. It's a tool that you can use. And the idea is with the Git CLI, or honestly, most of the Git GUIs, like I'm a big fan of primarily doing Git through
GUI operations using SourceTree, PyCharm, VS Code, whatever I happen to be in, right? I don't go like, well, pretend my editor doesn't know what any of the source control is and go to the terminal and do it and then come back. Like I just prefer it that way, especially, you know, that's why I called out that Git branch explorer, that 2E based one that we talked about last week in Rust. So one of the things you'll find with most of these tools, especially the CLI, but often all of them,
is they're great at the atomic operations of Git and source control, but they don't have workflows built into them. So by that I mean, I want to create a feature branch off of this branch, and as I'm working on the feature branch,
I would like to be able to type a single command and have that branch make sure that it's up to date with it without breaking the work that I'm currently doing. Maybe in the middle of a feature branch I'm working on, I need to create a bug fix branch, do that, come back, sync them. There's like these higher level things that you would like to do, okay? So GitTown is a CLI, I think it's a plugin. I think Git has like plugin type things because the commands are still git space command.
subcommands of Git, right? So it's not a totally separate thing. Anyway, it operates all, it sort of brings all those together. And so if you're, if what I'm about to describe sounds interesting to you, you should really, really watch this seven minute video at git-town.com. And it says, please check it out, right? So just watch this and it talks through the different scenarios, shows you how it works and so on. However, if you jump over to the GitTown website,
GitHub repo, it gives you a much better understanding of all the pieces there. And by written in Python, I mean written in Go. Okay. So Git, it has a couple of different commands. It says I could type Git, like in a Git repo, I could type Git space hack, and that will create a new feature branch that I can explore that will not affect any of the main work. And if I need to throw it away or whatever, it's fine. It's kind of like the exploratory way.
Hack in a positive sense, not hack as in the, I got a hoodie and it's shady in my room sort of thing. Or sync if I'm on top, if I'm on some kind of feature branch, I can update that branch from the main. Cause one of the things you want to do as you're, if you have a long running feature branch is keep periodically pulling from the main branch to
in there so you are less distant from it, right? If for some reason, like when you wanna go back, you want that to be just your changes, not your changes plus all of the conflicts that you've been potentially messing with and you've got another round basically. So this, you would just type git space sync. That's cool. You could type git switch to change branches via UI or once you're happy with your feature branch, you can say git propose and that'll create a merge
It'll create a pull or merge request for a feature branch back on GitHub, things like that. So there's a bunch of cool things like that. Well, I thought it was going to ask it to a dance. Exactly. You know, I mean, I know we haven't talked much and I know this is seventh grade and we're all really shy, but do you go to the dance with me? My dad can drive us. Yeah.
Another one is you can do things like get ship, which will merge a completed feature branch and then remove it once it's successfully shipped over and so on. So there's like these cool things on top of it, but certainly I recommend watching that video, that seven minute video to see if this is something that's interesting to you. I think I like it. I think it's pretty cool. I could, there's some, there's some cool ones that I always forget about, like get town rename. Yeah.
I mean, renaming is not difficult, but it's nice to have a little minor. Yeah, so...
Absolutely. And if you watch the video, you'll see that there's a lot of, really this thing is mostly built around like, yeah, you could, you're like, well, why can't I just do, you know, get merged from this branch and this and that. And sure, you can do these things. But it has a bunch of little steps to kind of catch the edge cases. Like it'll use stashing for certain operations when it's sure that you don't want any of your changes change and other things. Like,
Each one of these commands usually results in like five or six or seven Git commands. It's neat. It's worth checking out, especially if you like Git via the CLI or you work in a team that wants really structured ways of doing these things. I'm definitely going to check it out. And actually, I used the, I can't remember what we called it last week, the SEG or something like that. But I used it this weekend to look at a Git log from our server.
a project that I hadn't been familiar with to see what was going on. Awesome. Have you tried it on your Windows? Like, I want a Windows terminal follow-up.
What was I using it on? Remember, you were wondering if it would work on Windows. Yeah. Did you do this on your Mac or did you? Using it on my Mac laptop. Okay. Well, the follow-up is still pending then. Okay. And we'll check out GitTown on Windows as well. We'll take a little trip to GitTown. I'm sure it's probably fine. Yeah, that just automates Git, which works everywhere. Okay. Last week, we did, going back to last week, I guess, we did talk...
talk about a couple peps and today I'd like to talk about another pep. This one comes to us from Brett Cannon. Well, I saw him posted on Blue Sky. Pep, five set pipes.
Try that one more time. PEP 751, not 571, 751 has been accepted. So what is this PEP about? It's about lock files. So Brett says this means Python now has a lock file standard that can act as an export target for tools and create some sort of lock file.
And for some tools, the format can act as a primary lock file format as well as instead of some proprietary format. So we already had lock files. A couple, lots of different tools had lock files, but they weren't the same lock file. They were like UV had a lock and I think Poetry had a lock and there's a couple others.
So, uh, so now we have a standardized one and it's actually, I was looking through the standard. It looks pretty great. One of the things I, things that you would expect, you could, uh, do environments and versions and packages and stuff. Um, one of the things I thought was kind of cool. I, uh,
scrolling down to the bottom, was that there's a tool section that is pretty much open-ended, and this allows arbitrary tool configuration. So if other tools like poetry or whatever wanna use, try to use this instead or migrate to it,
but it doesn't kind of cover some of their own configuration, they can override this and put some extra things in here, which I think is pretty cool. There's also packages tool. I'm not sure which, I don't know. When you would use one or the other, but I'll leave it up to some tutorial writer to figure that one out. So this isn't to a, if you wanted to try this out now,
Well, let's look at an example first. The examples is in the video if you want to check out the live version if you're listening. But we're going to link to it as well. It's in the PEP. There's an example. Mostly, this is fairly readable, and I'm pretty on board with how this is set up. It looks pretty good. It's TOML syntax. You've even got things like...
of course, hashes and stuff if you wanna leave the hashes in for making sure it's the right thing. So this is all great. The other bit,
is, well, how can I use this right now? There is, there's some stuff, information about how to teach it, of course. Some discussion around rejected ideas. Oh, I forgot to mention the name. The name of it is pyloc.toml. So, or pyloc something.toml. There's
That's just what they came up with. It doesn't conflict with any other tools right now and it's reasonable, so that's good. If you want to try this right now,
I thought I, I thought we had a implementation. Um, there's a, there's the, there was, okay, I can't find it right now, but there was a, an, an implementation already. Um, that is just an example that it's in Brett's tool, uh, mouse bender. So, um, mouse bender allows you to play with it if you want to play with it right now. So I'm, I, I imagine that some of the other tools will kind of take this up also. And I, I,
I would expect that, like, for instance, UV or something might... I don't know. I'm not involved with UV, but I would expect some of these tools to pick them up pretty quick. But we'll see. Yeah, I'm... Well, first of all, congratulations, Brett Cannon. I know you've been working on this for a long time. And here we are. Super exciting. So I'm sure some conversations at PyCon will be had about this. Yeah. Super neat. So one thing that I like is the format. I saw...
like the hash is on the same line as the actual requirement pinned version there. So I've tried to do hashes in my requirements.txt files and the hashes so overwhelm the version of the thing that you actually have in the diff that it's just like, well, just it's a bunch of hash requirements
variations when I look at the diff and I got a hunt for the actual new version, like, well, what one of these changed because the version changed and what is the new version and so on. So the way this looks, it looks like it's going to be, well, it might differ also on the end for the hashes, but the first thing that appears on the line is the actual version number wheel and so on. So I think that's going to be really nice to look at. So what are
I might put hashes back in. Yeah, I think it's kind of cool that there's an upload time. So if you don't want to deal with hashes, but you do want to grab the upload time as a way to check for things, that's cool. Yeah, you just kind of know how old is that as well. Yeah. Right. I mean, for things that are not using Calver or they're abusing Calver or whatever, you know. Is it five minutes old or is it five months old? That might matter to you. Yeah. Yeah, yeah. Very cool. So awesome. I'm here for it. Do.
Do you know what else is awesome? What is that, Brian? I think Posit's pretty awesome. They are awesome. Very much. So I'd like to take a moment to thank our sponsor. This episode of Python Bytes is brought to you by the folks at Posit. Posit has made a huge investment in the Python community lately. Known originally for RStudio, they've been building out a suite of tools and services for Team Python. Over the past few years, we've learned...
some pretty scary terms like typesquatting, supply chain attack, obfuscated code, and more. These all orbit around the idea that you install Python packages, when you install Python packages, you're effectively running arbitrary code off the internet, on your dev machine, and usually even your production servers. That thought alone makes me shudder. This doesn't even touch on the reproducibility issues surrounding external packages, but
there are tools to help. Posit Package Manager can solve both problems for you. Think of Posit Package Manager as your personal package concierge. You use it to build your own package repository within your firewall that keep your project safe. You can upload your own internal packages to share or import packages directly from PyPI. Your team members can install from these repos in the normal ways using tools like
pip, poetry, and UV. Posit Package Manager will help you manage updates, ensuring you're using the latest, most secure version of your packages. But Posit Package Manager also takes point-in-time snapshots of your repos, which you can use to rerun your code reproducibly...
rerun your code reproducibility in the future. Posit Package Manager reports on packages with known CVEs and other vulnerabilities so you can keep ahead of threats. And if you need the highest level of security, you can even run Posit Package Manager in an air-gapped environment.
If you work on a data science team where security matters, you owe it to you and your org to check out Posit Package Manager. Visit pythonbytes.fm slash ppm today and get a three-month free trial and see if it's a good fit. That's pythonbytes.fm slash ppm. The link is in your podcast player's show notes. Thank you, Posit, for supporting Python Bytes. Indeed. Thanks, Posit. It's a cool project.
having a little bit more control over your packages is always a good thing. Well, Brian, I want to carry on with the theme of Git and talk about three more Git projects, Git topics, I believe at least. Who? So we know where, like this is sort of the thing they teach you in like early writing classes, the
All the W's, right? Who, what, where, when, why? How are we covering these things? We've gone from where to who, and now we're talking about git-who. So I've always enjoyed the git blame, or basically more broadly, like source control blame. Like, all right, who did this?
Do you see that line of code? Who do I talk to about this? Either they are not very good or we need to talk to them, or there's something way more complicated going on here and I need to understand it better. But who do I talk to about this line of code? You know what I mean? And the way that works is you can run that on a file and it shows you line by line who made the last change to that particular line, right? Editors support that.
super easy these days, like in PyCharm, it'll actually put an overlay of who wrote the function or whatever last. If you click on that, it'll sort of slide out the get blame. But that's a line by line type of thing. What if instead I wanted to know as...
an entire project. Like if I go to the repo and I type get who dash L, it will say who has made how many commits, how many changes, how many lines, when did they last do that as a summary of the project instead of a summary of the line. Cool, right? So that's super easy to do. And
You can go into subdirectories and sort of explore those and go, who's created changes on this particular subdirectory or whatever. So I think this is really neat. It's sort of a natural extension of Git Blame. Okay, so you can give it a particular file even and just say who modified this file? Yeah, I would think it would say how many contributions there are to that particular file.
I don't know if you can give it a file, but if there was a single file in a directory, you sure could. - That's pretty cool. - Probably you can give it a file as well. Yeah, it just sort of shows you who has been active in this repo or this subdirectory of a repo lately and how much.
Yeah, the blame part is one of the reasons why I'm careful if I try to do just a prettying up of code. Do you want to break it? Well, no, I don't want to be blamed for it. So if you're just lining up something to make it look nicer, be careful because you're now the owner of that line of code. Yeah, that's an interesting angle. I wonder if you should almost create a formatter account and just...
let the formatter do the formatting. Well, actually that's one of the, one of the lovely things about putting, putting like a separating, separating formatting stuff. So I was like, that's one of the things I try to teach people is to if cleaning up is great, but do it in a separate commit. Don't do it in the same commit as a bunch of actual code changes. So, sure. But it doesn't fix the blame part. No, it doesn't really fix the blame part, but
Another thing that the, that get town thing does is it'll, it uses squashing pretty aggressively. So when you do all your feature branches and you make 20 changes, you can squash that into a single commit as part of the publication PR flow. And so then at least when people go look at the blame, it's like all part of it.
one commit that has to do with this feature rather than rough formatting adjusted adjusted space you know like randomly like if you get it too small your comments are like really really not historically relevant you know like this is the formatting statement like great and why did when were we doing this you know what i mean yeah actually i kind of forget that not everybody does that so i always do this squat do a squash workflow you're a natural squasher okay natural squasher yeah
Well, mostly because it's way easier to roll back stuff. So if there's a, if something that like really causes a downstream problem, you kind of want to roll back all of that work together. Yeah. Yeah. That's a good point. Well,
So next up is watcha. Watcha. Okay. Watch, G-H-A, but there's no capitalization. So watchga. This comes to us from Ned Matchelder, as in not he sent it to us, but he created it. And this is just another get thing real quick. I think people find it relevant. So what you can do is you can install this, and then it will allow you to watch in real time the status of your GBA.
GitHub Actions. So imagine there's a bunch of changes that have come in. You're running talks for different versions of Python on your thing. There's some pending other requests that are queued up and you
And you're waiting for them to run. And you just want to know, like, well, when are these done? How long are they taking? So you can just point this at a GitHub repo that has a GitHub action. This is awesome. It's pretty cool, right? Yeah. Yes. I often have like a get tab open to watch this stuff. But having it right in the command line. Nice. Yeah. And it's like a live updating, very 2E feel, right? It's a very advanced sort of UI. It's not just a stream of text, but it's...
the UI is updating and changing with progress bars for each action and so on. Yeah, cool. So yeah, well done on that, Ned. That's super cool. All right, that's it for my main items. Over to you, Brian. Okay, so are we on? No, for your last one. You got one more, don't you? Oh, I got one more? Yeah, I do. I think you do. Yeah, so next up is I'd like to talk about UV a little bit. Yes.
So this is, this is, this is Dave, this Dave J. So Dave Johnson article share, it's called share Python scripts like a pro UV and pep 723 for easy to easy deployment. And we've talked about a lot of these pieces in different, different episodes, but I really, I really like the, I really like this write up. It's a sort of kind of a,
If you want to write Python scripts with dependencies, that was an issue before and now it's not, but all the different parts of how to do this. So it starts out with talking about what UV is and Uv and PEP 723. And that's the PEP that allowed you to put some stuff at the top of a file to include your dependencies and stuff.
So, um, to set the stage, we don't have to dig into what this is, but he gives a, a command line script called word lookup.
that has some things like arg parse that's built in, but it also uses HTTPX. So that's an extra dependency that's outside of the script. And this is if you're doing a web scraping script, you're gonna often have like some extra dependency or something. And so then that's the script, teaches you how to install UV, but you probably already have it installed. And then...
How do I add that thing? You don't have to memorize the format of how to put that up at the top because if you say UV add script and then you give it your script name and then what your dependency is, it automatically fills that out for you. It sticks that script block at the top of your code, which is awesome. And I think I knew that but forgot about it, so I'm glad it's in there to just add that little script tag.
And then, uh, talks about why, how this is different and similar to, to, um, what the pipe project at Tomo and the relation to that, which I kind of like that extra teaching. It isn't necessary that you know that for scripts, but it's good for people to have that, uh, knowledge about the whole ecosystem. Um, talks about running it with, uh, with UV, UV run, um,
And then jumps down and saying things like you can do UV run, but you can also just add that UV run as a shebang for Linux and Mac users. You can add that to the top and it gives the specific environment SUV run script that you have to stick that at the top of that. That is it.
Super interesting, actually. And then you don't have to say Python or UV run. You can just say the script name if it's an executable and it'll just work, which is awesome. And I didn't get into trying this with Windows, but apparently you can kind of do some of this around how to get this invocable on Windows as well. I haven't tried this, but you can use Py and other things. So there's some help for Windows.
respects shebangs on windows, which is something that adds. Okay, cool. And then, um, uh, he recommends creating like a, like a work where, uh, an extra script with a command line, a CMD script to launch your, your script with. So how to, how to do this, how to complete the loop with windows. So really great write up, um, for people, uh, writing utility scripts in Python. So,
That's actually really nice. Yeah. Well done. Yeah. So that's all for our main topics. We both have extras. Do you want to jump into your extras first or do you want me to do ours? I'll do mine real quick. Sure. I got a few. I'm going to try to go quick through them. One, I talked about PySQL Scribe, which if you're writing raw SQL for your code, you
but you want to have some kind of tooling support to help you in Python, this PySQL scribe lets you express through code the structure of your tables and their relationships. And then you can ask it like, hey, what would the query of this be? And then it'll generate it based on what database...
scheme or dialect like SQLite versus Microsoft SQL Server query syntax you wanted to generate and so on. And I proposed when we talked about it that, hey, would it be cool if there was an ability just pointed at a bunch of create scripts that defines your 100 tables and it just knows that? Well, Daniel, who runs a project, said that is a good idea and just released a version that you can just point it at a directory of create scripts.
and it'll use them to understand your entire data structure. Good job, Daniel. So a little bonus there. We talked about, this is a long running one. We talked about formatting Markdown, which with MD format, which I've started to adopt. I will talk about that next, but Hugo,
points out like, Hey, I use prettier, which has been around for a long time. It now has a markdown format. So that's really cool from 2017. People can check that out. And I'll link to that in the show notes. If you want to use that, um,
I, however, am continuing on with MD format and I ran it against the entire Markdown structure of my upcoming TalkPython in production book. And it's like I said, all Markdown. And one of the things I discovered, I don't know if you knew it, Brian, but I definitely did not. This is a really interesting linting philosophy. When you run it against Markdown, it will render the Markdown to HTML, then render
pretty it, then, you know, clean it up and then it'll run it through the markdown parser again. And if the HTML differs, it will roll back that change. Oh, that's cool. Yeah. So one of my chapters, apparently I put a whole bunch of plugins and goes, one of the plugins is like not great because it changed the format. And I'm like, Oh, which one is it? But I like them all.
I don't know. Maybe I need to debug it right now. I'm just not formatting one chapter. And you know what chapter is? 13. Obviously, obviously it's 13. Anyway, what I found, the command that I think is really, really good is to use, I wrote it down. It's in the show notes, but I must find it. It is to get, say, MD format dash dash number. So it renumbers those things, which is great.
And then this one, I thought I would like it to wrap it, but all of my editors actually preserve the wrapping format so that it can kind of help you see the structure a little better, even though Markdown Render doesn't. And so that looked really crappy. So if you say dash dash wrap no, it'll actually explicitly unwrap if you've accidentally got some in there. So I like number and wrap no, not just...
not change it, but actually unravel the wrapping, which I thought was really cool. And as part of that, we did talk about UV, sorry, we talked about pipx inject. And last time you pointed out correctly that uv tool install dash dash with is the equivalent. But I wanted to point out that when you, and I put it in the show notes, if you want to use with and uv tool start like inject, you have to say with as many times as there are extras. So uv tool install MD format with MD format GFS.
GFM, dash dash width, MD format front matter, dash dash width, and so on. So if you want to get that inject feel, you've got to like width separately every single parameter, every single extra or whatever.
So anyway, that's a thing. Another one you can see right now that if you're watching the live stream or see the pictures that I'm using the Zen browser, which I really come to love. It's built on top of Firefox, which is great. And I really want this like nice clean format. We just have the tabs as icons on the side and everything's just at the top and it gives you more space, right? Well, Firefox, if you are a Firefox purist has now come up with this
tabs on the left side of things, which is a brand new feature that they just released last week. I think it was, but if you grab that little width of the tab and you slide it over to their icon size, it looks exactly like what I have on the screen right now, Brian, where it has just icons on the left for the tabs and, and a address bar across the top. The new Firefox is pretty good. There's still stuff missing that Zen brings. That's really cool. But anyway, I wanted to point people that that's an option now. Nice. Yeah. Finally, last thing for,
on my end, is a follow-up on UV from James Fallon. Falcon, sorry, James Falcon. Thanks for sending this in. He says, hey, I'm also a UV enthusiast, and I was holding out to using UV because there's one case it hasn't solved. However, you guys convinced me to work through it and get it figured out. So he maintains a library supported across multiple Python versions and needs to deal with bugs and specific ones. So like, it doesn't work this way in 3.8, but it works fine in 3.10. So the idea was to have...
Originally, I was using PyENV because you can create virtual environments with those different ones and then activate them by name. But the way UV worked originally, it didn't. So Woody said, hey, look, I'm creating just a location for virtual environments on my profile, maybe a subfolder for the project. And then I use now use AXE.
UV to create version-specific virtual environments, you know, 3.11, 3.12, 3.13, and so on for that project. And then he includes a batch file that you can just say, work on this virtual environment project name, and it'll automatically find it and activate for you and so on. So kind of gives us an example of like how you might sort of PyEV, ENV your way around different versions using UV. That's pretty cool. Yeah, so there's a little batch script that...
people can check out if they want to use it. Nice. Okay. In the show notes. Yep. That's it for my extras. Well, so for my extras today, so it's April, right? And one of the things we do in April in on April 1st, sometimes there's April fool's pranks that happen. And one of the things, so there's, there's, there's good April fool's pranks and then there's bad April fool's pranks. So I'm going to cover both.
So first off, I think a good one, which was a video put out by, well, it was forwarded by the PSF, but it says Cult Repo. I'm not sure. I looked and I'm like, that is a really weird thing to kind of do an official joke from the PSF. That Cult Repo group, they make a bunch of documentaries about, like highly produced documentaries about cults.
Programming languages and their origins or programming frameworks. Like they have some on Vue.js and they have some on Angular. That's where it's from. Okay, so the video is breaking Guido van Rossum returns as Python's BDFL. And it's a two-minute video, or almost two minutes. Hilarious. I loved it. It shows him typing of the...
the saying that the everybody, what is the standards people? Oh, the the board's no longer. - Steering council. - Steering council is no longer necessary. Couple of things I wanted to point out that I loved about this video.
uh, Brett Cannon, he's, um, he's in it and he's, uh, listed as famous Python quotationist, um, is his title. And then, uh, then Guido goes ahead and says one of the things that he said, he Guido says, as I always say, uh, come for the language, uh, stay for the community. And that's not a Guido quote. That's a Brett quote. So, um,
the famous quote Python quotation is gets crossed out after, after Guida says that, which is pretty funny.
And then a little later, who was this? Barry Warsaw walks in and says, Guido, are you ready to release Python 2.8? Which is never going to happen, right? So that's pretty good. Guido says indentation is enlightenment. And he says, or Barry gives him credit for saying that
Um, the being Pythonic means becoming one with white space, um, which is good. Anyway, the, um, the, and then it kind of wraps it up with, um, uh, Guido's announcing that he's going to implement a new keyword. Maybe it's kind of like if, but way more Pythonic. So maybe something, and it gives an example of maybe print Python, the documentary coming this summer. And,
And I kind of hope that one is real. I hope we get a documentary coming. Well, it's already the summer, or it's almost the summer. If we get a documentary. Well, we're going to get it soon, then that'll be great. And if this cult repo company makes it, I think it'll be pretty good. Yeah.
when I first saw this, I was like, I saw a tweet come on. I'm like, Hey, whatever. Funny, a funny little joke. And I just thought it was a tweet. Then Nick wall sent it over and like, wait, you should check this out. It's funny. And I'm like, watch them. Oh, this is, there's way more here than I realized. This is, this is well done. This is good. Okay. So that's a, that's a good way to do a April fool's prank. I've never done one. I've never done like a tech prank. So I decided to try my hand at that this year.
It did not go that well. So I'm, and you know how, you know, this, the, have you ever heard, like, if you have to explain a joke and it's not funny, this is me explaining a joke. So it's not funny. So there's PyTest repeat. So I've been doing on testing code. I've been doing the series on PyTest plugins.
And PyTestRepeat is one that I use all the time. And I noticed that it hadn't been being tested on, it only was tested on 312, it's not tested on 313 yet or 314. And okay, so 314 is 3.14. Py starts with 3.14, but Py is an irrational number.
And it doesn't repeat. So the joke I was trying to make was I put out an episode on testing code on April 1st, Python 3.4 won't repeat with PyTest repeat. Um,
And I guess it was, and I even tried to make it, I thought, obvious to say, unfortunately, it doesn't seem to work on 3.14, even though there's no rational, like no irrational. Yeah, that was nice. I listened to that episode. It was well done. Anyway, I got some feedback of people confused and wondering why I was testing 3.14 anyway, since it was an alpha and...
Anyway, I had to explain like to five different people that it's actually fine. So the TLDR is three, PyTest repeat works fine with Python 3.14. So don't worry about it.
Anyway, how not to do an April Fool's prank. So I guess that's all of our extra items, right? It sure is. We'll see how much I got to explain this joke, huh? No, I think this one is fine. This one is good. So this comes to us from the Programming Humor Twitter feed, and it says, what's the best thing you found in code? And they are almost always about...
comments in this thread. So here's a... This is actually Python. So this is a comment that somebody found in the code and they posted it.
But if you look through and you look at the comments below, there's a glorious set of 132 comments of similar. They're not all great. So here's what it says at the top of this thread. It says, as a comment, you may think that this function is obsolete and it doesn't seem to do anything and you would be correct. But when we remove this function, for some reason, the whole program crashes and we can't figure out why. So here it will stay.
So I entitled the joke, so here it will stay. And let me see if I can find a few. This is my favorite one, though. This probably should have been the top line one, but this came to us as a response from the original. Somebody named Jeppe pointed out and said, not mine, but I find this kind of funny. And this is a C-based type of language because it's slash slash for the comment.
Dear programmer, when I wrote this code, only God and I knew how it worked. Now only God knows it. Therefore, if you're trying to optimize this routine and it fails, most surely it will, please increase the counter as a warning to the next person. And then there's a variable below that says total hours wasted here, 254. So good.
That's good. And on and on it goes. If you don't want the rest of your, if you want to lose maybe a half hour, you can go right down here and there's a bunch of good stuff in there. So there it is.
That's good. Yeah, I think that's a good one. Yeah. I've only a couple of times have actually at least I have at least one memory of some code that didn't look like it did anything except for obviously it was getting run and the timing. I'm sure it was just timing. Like if you took it out, it changed the timing of the rest of the program and then it didn't work. So yeah, I've actually run into it. You would uncover some kind of concurrency bug that it was defensive against. Yeah.
Yeah. Yeah. C codes are like that. So yeah. It's hard to see the errors. That's true. Hard to see the errors. That's good. It wasn't. No, no, it wasn't. We better quit. We better just call it an episode. Okay. All right. Thanks a lot. Bye. Bye.