WEBVTT

00:00:00.100 --> 00:00:01.600
<v Calvin Hendryx-Parker>Hello and welcome to Python Bytes,

00:00:01.790 --> 00:00:04.820
<v Calvin Hendryx-Parker>where we deliver Python news and headlines directly to your earbuds.

00:00:05.140 --> 00:00:07.020
<v Calvin Hendryx-Parker>This is episode 505,

00:00:07.360 --> 00:00:09.480
<v Calvin Hendryx-Parker>recorded on November 17th, 2026.

00:00:10.060 --> 00:00:11.040
<v Calvin Hendryx-Parker>I'm Calvin Hendrix-Parker.

00:00:11.290 --> 00:00:12.240
<v Michael Kennedy>And I'm Michael Kennedy.

00:00:12.660 --> 00:00:13.420
<v Calvin Hendryx-Parker>Happy to have you here.

00:00:13.760 --> 00:00:15.740
<v Calvin Hendryx-Parker>This episode is sponsored by Longlink.

00:00:15.840 --> 00:00:18.280
<v Calvin Hendryx-Parker>There'll be more about that in the break in between.

00:00:18.900 --> 00:00:21.460
<v Calvin Hendryx-Parker>But if you want to connect with Michael and I online,

00:00:21.720 --> 00:00:23.980
<v Calvin Hendryx-Parker>you can find us on Mastodon Bluesky X and LinkedIn.

00:00:24.540 --> 00:00:25.380
<v Calvin Hendryx-Parker>And if you want to find the show,

00:00:25.580 --> 00:00:27.360
<v Calvin Hendryx-Parker>check us out on Mastodon Bluesky X.

00:00:27.740 --> 00:00:32.920
<v Calvin Hendryx-Parker>Join us live if you want to experience this in real time at pythonbytes.fm/live.

00:00:33.110 --> 00:00:33.820
<v Calvin Hendryx-Parker>You can be part of the audience.

00:00:34.440 --> 00:00:39.120
<v Calvin Hendryx-Parker>Usually we're recording Tuesdays at 7 a.m., but another older versions of the videos are there as well.

00:00:39.230 --> 00:00:44.200
<v Calvin Hendryx-Parker>And finally, if you want an artisanal handcrafted digest of each week's show notes in email form,

00:00:44.590 --> 00:00:48.160
<v Calvin Hendryx-Parker>go and join our friends of the show list and we will send it over to you.

00:00:48.260 --> 00:00:49.880
<v Calvin Hendryx-Parker>Michael done a good job of making that happen.

00:00:50.820 --> 00:00:52.260
<v Calvin Hendryx-Parker>Speaking of making things happening,

00:00:52.550 --> 00:00:54.280
<v Calvin Hendryx-Parker>what about job cues?

00:00:54.760 --> 00:00:58.240
<v Michael Kennedy>I have such an interesting project to come up here.

00:00:58.500 --> 00:00:59.460
<v Michael Kennedy>So let's switch over.

00:00:59.620 --> 00:00:59.880
<v Michael Kennedy>Perfect.

00:01:00.360 --> 00:01:01.620
<v Michael Kennedy>Called Tiny Jobs.

00:01:02.120 --> 00:01:06.040
<v Michael Kennedy>So this one comes to us from Pier Giuseppe.

00:01:06.460 --> 00:01:07.920
<v Michael Kennedy>So Pier Giuseppe, this is his project,

00:01:08.120 --> 00:01:08.460
<v Michael Kennedy>send it over.

00:01:08.860 --> 00:01:11.580
<v Michael Kennedy>And I'm going to have to just apologize in advance.

00:01:11.920 --> 00:01:14.120
<v Michael Kennedy>I am on a SQLite mission.

00:01:14.480 --> 00:01:15.460
<v Michael Kennedy>I'm a believer,

00:01:15.700 --> 00:01:15.920
<v Michael Kennedy>Calvin.

00:01:16.200 --> 00:01:19.060
<v Michael Kennedy>And so this is amongst that, but it is.

00:01:19.500 --> 00:01:20.580
<v Calvin Hendryx-Parker>I leaned into it.

00:01:20.660 --> 00:01:22.800
<v Calvin Hendryx-Parker>I've also added a SQLite topic today too.

00:01:23.040 --> 00:01:23.340
<v Michael Kennedy>Incredible.

00:01:23.440 --> 00:01:27.180
<v Michael Kennedy>So people are just going to have to accept like a couple of weeks of SQLite

00:01:27.380 --> 00:01:29.040
<v Michael Kennedy>amazingness because it is,

00:01:29.260 --> 00:01:31.200
<v Michael Kennedy>I honestly think it's really coming into its own.

00:01:31.560 --> 00:01:31.940
<v Michael Kennedy>What do you think?

00:01:32.440 --> 00:01:33.220
<v Calvin Hendryx-Parker>It's always been there.

00:01:33.460 --> 00:01:34.320
<v Calvin Hendryx-Parker>It's an interesting project,

00:01:34.540 --> 00:01:34.620
<v Calvin Hendryx-Parker>right?

00:01:34.920 --> 00:01:35.200
<v Michael Kennedy>It is.

00:01:35.480 --> 00:01:37.080
<v Michael Kennedy>You know, I think it's, I think the wall stuff,

00:01:37.540 --> 00:01:37.660
<v Michael Kennedy>honestly,

00:01:37.900 --> 00:01:38.740
<v Michael Kennedy>the right ahead log

00:01:39.420 --> 00:01:44.000
<v Michael Kennedy>has made the durability sort of much stronger and sort of brought it into like,

00:01:44.100 --> 00:01:45.040
<v Michael Kennedy>you can actually use it now,

00:01:45.280 --> 00:01:45.440
<v Michael Kennedy>you know?

00:01:45.880 --> 00:01:48.760
<v Michael Kennedy>Anyway, like I said, this is only deposit one,

00:01:48.990 --> 00:01:50.280
<v Michael Kennedy>I guess, on that story.

00:01:50.370 --> 00:01:51.320
<v Michael Kennedy>So maybe there'll be more.

00:01:51.920 --> 00:01:52.780
<v Michael Kennedy>I mean, I know there'll be more,

00:01:52.940 --> 00:01:53.740
<v Michael Kennedy>but this is part one.

00:01:53.860 --> 00:01:53.940
<v Michael Kennedy>Anyway,

00:01:54.160 --> 00:01:54.600
<v Michael Kennedy>Tiny Jobs.

00:01:54.730 --> 00:01:58.680
<v Michael Kennedy>So Tiny Jobs is built around fundamentally SQLite,

00:01:58.710 --> 00:02:00.280
<v Michael Kennedy>though it can do Postgres and Redis.

00:02:00.310 --> 00:02:02.440
<v Michael Kennedy>And what it is, is a background job

00:02:02.600 --> 00:02:04.280
<v Michael Kennedy>and cron system for Python.

00:02:04.800 --> 00:02:08.380
<v Michael Kennedy>So imagine I want background jobs like sending email

00:02:09.200 --> 00:02:12.340
<v Michael Kennedy>or rebuilding search indexes or, you know,

00:02:12.460 --> 00:02:15.860
<v Michael Kennedy>updating Parquet files or parsing Parquet files

00:02:16.000 --> 00:02:18.000
<v Michael Kennedy>and moving them into some kind of analytical database.

00:02:18.020 --> 00:02:19.260
<v Calvin Hendryx-Parker>Kick off an agent to do something.

00:02:19.920 --> 00:02:20.120
<v Michael Kennedy>Yeah,

00:02:20.280 --> 00:02:21.920
<v Michael Kennedy>that is definitely the 2026 version,

00:02:22.140 --> 00:02:22.280
<v Michael Kennedy>isn't it?

00:02:22.680 --> 00:02:22.740
<v Calvin Hendryx-Parker>Yeah.

00:02:24.260 --> 00:02:28.520
<v Michael Kennedy>So you can do all of that with maybe no server or maybe a server,

00:02:28.720 --> 00:02:29.500
<v Michael Kennedy>depending on how you want to run it.

00:02:29.770 --> 00:02:35.500
<v Michael Kennedy>This one, it kind of is most likely you would run it as a local separate Docker container

00:02:35.690 --> 00:02:36.500
<v Michael Kennedy>as part of your thing.

00:02:36.550 --> 00:02:40.340
<v Michael Kennedy>So when you say Docker Compose up, like your app starts and this thing starts,

00:02:40.440 --> 00:02:45.200
<v Michael Kennedy>but you can run it in process on things that are async, like FastAPI and so on.

00:02:45.480 --> 00:02:46.500
<v Michael Kennedy>So how does it work?

00:02:46.900 --> 00:02:47.200
<v Michael Kennedy>Well,

00:02:47.550 --> 00:02:52.300
<v Michael Kennedy>you just come down and create a tiny job thing.

00:02:52.570 --> 00:02:53.180
<v Michael Kennedy>You give it,

00:02:54.270 --> 00:02:55.820
<v Michael Kennedy>by default, a SQLite file.

00:02:55.890 --> 00:02:59.180
<v Michael Kennedy>So your connection string is just a file path as it is.

00:02:59.390 --> 00:03:04.560
<v Michael Kennedy>And then you just put app.task or jobs.task, whatever you want to do,

00:03:04.940 --> 00:03:05.480
<v Michael Kennedy>onto functions.

00:03:05.780 --> 00:03:10.400
<v Michael Kennedy>Like the examples are generate PDF or transcode this video.

00:03:10.920 --> 00:03:12.840
<v Michael Kennedy>Given a video ID, you know it's going to go to the database,

00:03:13.030 --> 00:03:14.520
<v Michael Kennedy>pull it up, that sort of thing.

00:03:14.800 --> 00:03:16.440
<v Michael Kennedy>But it has sort of a tenacity,

00:03:16.840 --> 00:03:19.120
<v Michael Kennedy>stamina sort of thing as well,

00:03:19.280 --> 00:03:20.440
<v Michael Kennedy>like five max attempts,

00:03:21.000 --> 00:03:23.320
<v Michael Kennedy>timed out for, you know, 300 seconds,

00:03:23.920 --> 00:03:24.020
<v Michael Kennedy>right?

00:03:24.180 --> 00:03:25.040
<v Michael Kennedy>Because how long is your video?

00:03:25.400 --> 00:03:27.760
<v Michael Kennedy>I could tell you, you might want to up that if you're doing a lot of video work.

00:03:28.380 --> 00:03:29.560
<v Michael Kennedy>I've run those overnight.

00:03:30.640 --> 00:03:32.060
<v Michael Kennedy>But anyway, the point is there, you know?

00:03:32.280 --> 00:03:32.860
<v Calvin Hendryx-Parker>It's really,

00:03:33.320 --> 00:03:33.860
<v Calvin Hendryx-Parker>really elegant.

00:03:34.110 --> 00:03:35.040
<v Calvin Hendryx-Parker>I really like this.

00:03:35.200 --> 00:03:36.320
<v Calvin Hendryx-Parker>I had not seen this before.

00:03:36.610 --> 00:03:39.400
<v Calvin Hendryx-Parker>And now I'm like, how does this not escape my like purview?

00:03:39.820 --> 00:03:40.240
<v Calvin Hendryx-Parker>It's amazing.

00:03:40.240 --> 00:03:40.340
<v Calvin Hendryx-Parker>I know.

00:03:40.600 --> 00:03:42.000
<v Michael Kennedy>It's so good.

00:03:43.040 --> 00:03:48.340
<v Michael Kennedy>we've got you can give it a cron the task instead of just call it you give it a cron one of those

00:03:48.760 --> 00:03:53.700
<v Michael Kennedy>what do you call this cron like signature how about that yeah okay and you give it a time zone

00:03:53.800 --> 00:04:00.040
<v Michael Kennedy>and it just runs like you can say 30 space two place star star star and it just runs on its own

00:04:00.240 --> 00:04:06.220
<v Michael Kennedy>it has guaranteed operation like it'll guarantee that it will handle the jobs even though it's

00:04:06.400 --> 00:04:11.259
<v Michael Kennedy>sqlite does not guarantee they won't run twice if like for some reason you pull the power halfway

00:04:11.280 --> 00:04:11.500
<v Michael Kennedy>through,

00:04:11.720 --> 00:04:13.180
<v Michael Kennedy>it might go, okay, that didn't finish.

00:04:13.340 --> 00:04:13.860
<v Michael Kennedy>I'm going to start over.

00:04:14.240 --> 00:04:15.360
<v Michael Kennedy>If you run it in process,

00:04:16.079 --> 00:04:17.600
<v Michael Kennedy>then it will basically,

00:04:17.900 --> 00:04:20.060
<v Michael Kennedy>if something is running and it's running in process,

00:04:20.299 --> 00:04:21.600
<v Michael Kennedy>it will hold back

00:04:21.760 --> 00:04:24.640
<v Michael Kennedy>the shutdown until it sort of drains its jobs,

00:04:25.060 --> 00:04:26.260
<v Michael Kennedy>which depending on the kind of jobs you got,

00:04:26.300 --> 00:04:28.300
<v Michael Kennedy>it might be awesome. It might be not awesome.

00:04:28.660 --> 00:04:29.400
<v Michael Kennedy>If you're sending email,

00:04:29.760 --> 00:04:30.460
<v Michael Kennedy>probably fine.

00:04:30.720 --> 00:04:31.360
<v Michael Kennedy>If you're

00:04:31.440 --> 00:04:32.080
<v Michael Kennedy>transcoding a video,

00:04:32.440 --> 00:04:33.140
<v Michael Kennedy>probably not fine.

00:04:33.360 --> 00:04:33.840
<v Michael Kennedy>You know what I mean?

00:04:34.220 --> 00:04:35.660
<v Michael Kennedy>So think about what you're doing,

00:04:35.760 --> 00:04:36.560
<v Michael Kennedy>I'm guessing here.

00:04:36.760 --> 00:04:36.840
<v Calvin Hendryx-Parker>Yeah,

00:04:37.040 --> 00:04:37.180
<v Calvin Hendryx-Parker>exactly.

00:04:38.379 --> 00:04:40.840
<v Michael Kennedy>It comes with a Django tasks backend

00:04:41.320 --> 00:04:43.180
<v Michael Kennedy>for the Django, what is that, 6.1 when that came out?

00:04:43.240 --> 00:04:43.300
<v Calvin Hendryx-Parker>Also,

00:04:43.480 --> 00:04:44.440
<v Calvin Hendryx-Parker>again, super elegant,

00:04:44.780 --> 00:04:45.240
<v Calvin Hendryx-Parker>super cool.

00:04:45.460 --> 00:04:46.520
<v Calvin Hendryx-Parker>This is a really, really nice.

00:04:46.680 --> 00:04:48.840
<v Michael Kennedy>Yeah, Django has this tasks thing,

00:04:49.080 --> 00:04:50.200
<v Michael Kennedy>but it doesn't really have an implementation

00:04:50.480 --> 00:04:51.020
<v Michael Kennedy>that comes with it.

00:04:51.120 --> 00:04:53.400
<v Michael Kennedy>And so here's like, I don't want a server.

00:04:53.740 --> 00:04:54.820
<v Michael Kennedy>I just want a file.

00:04:54.980 --> 00:04:55.240
<v Michael Kennedy>You know what I mean?

00:04:55.400 --> 00:04:56.240
<v Michael Kennedy>I want a thing to manage.

00:04:56.759 --> 00:04:57.780
<v Calvin Hendryx-Parker>Simplicity is the feature here.

00:04:58.060 --> 00:04:58.260
<v Michael Kennedy>Yeah,

00:04:58.560 --> 00:04:58.800
<v Michael Kennedy>absolutely.

00:04:59.220 --> 00:05:01.360
<v Michael Kennedy>So a lot of cool stuff that you can do here.

00:05:01.580 --> 00:05:02.480
<v Michael Kennedy>And I had,

00:05:02.960 --> 00:05:03.520
<v Michael Kennedy>this is really neat.

00:05:03.680 --> 00:05:05.680
<v Michael Kennedy>This is absolutely on the case.

00:05:05.790 --> 00:05:07.300
<v Michael Kennedy>So I've warned people about SQLite.

00:05:07.300 --> 00:05:09.940
<v Michael Kennedy>They also will have to be warned about free-threaded Python.

00:05:12.060 --> 00:05:12.640
<v Calvin Hendryx-Parker>Lots of news.

00:05:13.180 --> 00:05:18.060
<v Michael Kennedy>Because these are things that are intersecting for me on stuff that I'm working on right now.

00:05:18.380 --> 00:05:18.820
<v Michael Kennedy>Part of my mission.

00:05:18.970 --> 00:05:19.800
<v Michael Kennedy>So anyway.

00:05:20.180 --> 00:05:25.400
<v Calvin Hendryx-Parker>For listeners who don't know, Michael's on a terror to make free-threading available to the world.

00:05:26.220 --> 00:05:26.980
<v Michael Kennedy>That's my mission.

00:05:27.010 --> 00:05:27.120
<v Michael Kennedy>Yeah,

00:05:27.340 --> 00:05:28.640
<v Michael Kennedy>like last week was Gilgard,

00:05:28.800 --> 00:05:28.860
<v Michael Kennedy>right?

00:05:29.060 --> 00:05:29.520
<v Michael Kennedy>G-I-O,

00:05:29.780 --> 00:05:30.100
<v Michael Kennedy>Gilgard.

00:05:30.440 --> 00:05:31.820
<v Michael Kennedy>This week I have something way bigger.

00:05:31.960 --> 00:05:32.120
<v Michael Kennedy>Anyway,

00:05:33.100 --> 00:05:34.140
<v Michael Kennedy>so free-threaded Python.

00:05:34.360 --> 00:05:37.620
<v Michael Kennedy>I talked to Leppie and said, what's the story with free-threading?

00:05:37.820 --> 00:05:38.200
<v Michael Kennedy>I'm on this.

00:05:38.400 --> 00:05:39.320
<v Michael Kennedy>He's like, you know what?

00:05:39.520 --> 00:05:39.900
<v Michael Kennedy>You're right.

00:05:39.940 --> 00:05:40.840
<v Michael Kennedy>Let me make sure it works.

00:05:40.920 --> 00:05:50.340
<v Michael Kennedy>And so the gil stays off on Python 3.14t and dash loops and runs several event loops in one process.

00:05:50.520 --> 00:05:53.800
<v Michael Kennedy>I imagine because I have been juggling this very hard.

00:05:53.820 --> 00:05:54.580
<v Michael Kennedy>It is not easy.

00:05:55.020 --> 00:06:01.740
<v Michael Kennedy>Usually what I'm thinking is happening is each dedicated free-thread gets its own dedicated event loop.

00:06:02.160 --> 00:06:04.300
<v Michael Kennedy>And then because sharing them across threads,

00:06:04.780 --> 00:06:05.220
<v Michael Kennedy>it's kind of,

00:06:05.510 --> 00:06:07.560
<v Michael Kennedy>I don't know, it's like Ghostbusters crossing the wires.

00:06:07.650 --> 00:06:08.640
<v Michael Kennedy>You probably shouldn't do it.

00:06:08.650 --> 00:06:09.440
<v Michael Kennedy>You might be able to.

00:06:10.660 --> 00:06:12.540
<v Calvin Hendryx-Parker>The performance looks pretty awesome,

00:06:12.710 --> 00:06:12.820
<v Calvin Hendryx-Parker>too.

00:06:13.060 --> 00:06:14.580
<v Calvin Hendryx-Parker>You've got some performance numbers here.

00:06:14.980 --> 00:06:15.120
<v Michael Kennedy>Yep.

00:06:15.540 --> 00:06:16.040
<v Michael Kennedy>Where did I put them?

00:06:16.520 --> 00:06:17.200
<v Michael Kennedy>Read them to us here.

00:06:17.420 --> 00:06:17.900
<v Calvin Hendryx-Parker>Oh, here we go.

00:06:18.200 --> 00:06:18.600
<v Calvin Hendryx-Parker>No, I got it.

00:06:18.740 --> 00:06:21.920
<v Michael Kennedy>10 hours of steady load on one SQLite file,

00:06:22.160 --> 00:06:24.220
<v Michael Kennedy>almost 900,000 jobs.

00:06:25.340 --> 00:06:25.520
<v Calvin Hendryx-Parker>Wow.

00:06:25.860 --> 00:06:28.420
<v Michael Kennedy>And this is sort of the durability test.

00:06:28.580 --> 00:06:31.180
<v Michael Kennedy>Like, let me see if I can hammer this and break it.

00:06:31.220 --> 00:06:32.460
<v Michael Kennedy>I have some other interests.

00:06:32.580 --> 00:06:33.340
<v Michael Kennedy>Like, I'm telling you,

00:06:33.460 --> 00:06:34.140
<v Michael Kennedy>SQLite, y'all,

00:06:34.500 --> 00:06:35.160
<v Michael Kennedy>it's out of control.

00:06:35.460 --> 00:06:35.940
<v Michael Kennedy>It's off the hook.

00:06:36.040 --> 00:06:37.000
<v Michael Kennedy>It's something else.

00:06:37.260 --> 00:06:37.460
<v Michael Kennedy>So,

00:06:38.160 --> 00:06:38.260
<v Michael Kennedy>yeah,

00:06:38.640 --> 00:06:41.200
<v Michael Kennedy>900,000 jobs in 10 hours.

00:06:41.620 --> 00:06:41.700
<v Calvin Hendryx-Parker>Yeah.

00:06:41.820 --> 00:06:42.380
<v Michael Kennedy>Nothing lost.

00:06:42.780 --> 00:06:43.240
<v Calvin Hendryx-Parker>That's amazing.

00:06:43.600 --> 00:06:43.900
<v Michael Kennedy>All right.

00:06:44.160 --> 00:06:44.620
<v Calvin Hendryx-Parker>Really impressive.

00:06:45.040 --> 00:06:45.880
<v Calvin Hendryx-Parker>No, that's super cool.

00:06:45.980 --> 00:06:47.900
<v Calvin Hendryx-Parker>So I'm glad you pointed this one out to me.

00:06:48.280 --> 00:06:48.360
<v Calvin Hendryx-Parker>Yeah.

00:06:48.660 --> 00:06:49.760
<v Calvin Hendryx-Parker>Speaking of things being funky,

00:06:50.720 --> 00:06:51.960
<v Calvin Hendryx-Parker>some other cool stuff over here

00:06:52.200 --> 00:06:53.180
<v Calvin Hendryx-Parker>on the SQLite front, too.

00:06:53.460 --> 00:06:53.740
<v Michael Kennedy>All right.

00:06:53.800 --> 00:06:55.300
<v Michael Kennedy>Let's keep the SQLite train going.

00:06:55.440 --> 00:06:56.580
<v Michael Kennedy>This is going to be straight SQLite,

00:06:56.700 --> 00:06:57.440
<v Michael Kennedy>like right down the line.

00:06:57.540 --> 00:06:57.800
<v Michael Kennedy>Let's go.

00:06:58.100 --> 00:06:59.859
<v Calvin Hendryx-Parker>Well, so you're familiar with

00:06:59.880 --> 00:07:03.340
<v Calvin Hendryx-Parker>with func cache or functools cache, correct?

00:07:03.570 --> 00:07:07.860
<v Calvin Hendryx-Parker>So this is Alex Chan's decorator style that stores

00:07:08.220 --> 00:07:09.000
<v Calvin Hendryx-Parker>results in,

00:07:09.310 --> 00:07:09.680
<v Calvin Hendryx-Parker>guess what,

00:07:09.860 --> 00:07:12.320
<v Calvin Hendryx-Parker>SQLite. They survive process exits.

00:07:12.490 --> 00:07:13.140
<v Calvin Hendryx-Parker>So basically,

00:07:13.680 --> 00:07:14.420
<v Calvin Hendryx-Parker>if you've got something

00:07:14.470 --> 00:07:15.420
<v Calvin Hendryx-Parker>where you were using, say,

00:07:15.620 --> 00:07:19.180
<v Calvin Hendryx-Parker>func cache in the past or functools cached in the past, that would

00:07:19.440 --> 00:07:22.480
<v Calvin Hendryx-Parker>disappear when your process stops and your memory goes away.

00:07:22.890 --> 00:07:24.840
<v Calvin Hendryx-Parker>This allows it to survive those

00:07:24.860 --> 00:07:25.180
<v Calvin Hendryx-Parker>restarts.

00:07:25.770 --> 00:07:31.060
<v Calvin Hendryx-Parker>One of the interesting use cases that they used in their example here was for static

00:07:31.150 --> 00:07:31.640
<v Calvin Hendryx-Parker>site generation.

00:07:32.380 --> 00:07:33.140
<v Calvin Hendryx-Parker>For static site generation,

00:07:33.420 --> 00:07:35.780
<v Calvin Hendryx-Parker>a lot of times most things don't change,

00:07:36.220 --> 00:07:36.940
<v Calvin Hendryx-Parker>but you run the

00:07:37.090 --> 00:07:41.860
<v Calvin Hendryx-Parker>process once and then it's done and then now you run it again and it'll go through and redo most of

00:07:41.860 --> 00:07:42.300
<v Calvin Hendryx-Parker>the operation.

00:07:42.610 --> 00:07:48.740
<v Calvin Hendryx-Parker>With this, it can actually pull those resulting values out of SQLite and save you

00:07:48.790 --> 00:07:53.180
<v Calvin Hendryx-Parker>tons and tons of time. So if you've got a large-ish static site,

00:07:53.570 --> 00:07:54.820
<v Calvin Hendryx-Parker>this will save you a ton of time on

00:07:54.840 --> 00:07:59.640
<v Calvin Hendryx-Parker>generation it'll save you for example if you're using ci cd or GitHub actions this will reduce your

00:07:59.960 --> 00:08:04.700
<v Calvin Hendryx-Parker>your GitHub actions run times so obviously saving potentially money or credits or whatever you're

00:08:04.800 --> 00:08:11.960
<v Calvin Hendryx-Parker>using to go out there and build those pieces it works very very similarly to the cache stuff or

00:08:12.140 --> 00:08:18.660
<v Calvin Hendryx-Parker>the functools cache where it's just a you can register the sqlite storage and then you use

00:08:18.600 --> 00:08:21.660
<v Calvin Hendryx-Parker>decorators against the expensive functions,

00:08:21.860 --> 00:08:22.360
<v Calvin Hendryx-Parker>for example,

00:08:22.540 --> 00:08:22.700
<v Calvin Hendryx-Parker>here,

00:08:23.020 --> 00:08:23.960
<v Calvin Hendryx-Parker>and then you can specify

00:08:24.500 --> 00:08:26.660
<v Calvin Hendryx-Parker>how it gets hashed. So under the covers,

00:08:27.100 --> 00:08:27.820
<v Calvin Hendryx-Parker>it is hashing.

00:08:28.240 --> 00:08:28.640
<v Calvin Hendryx-Parker>You can't,

00:08:28.860 --> 00:08:29.880
<v Calvin Hendryx-Parker>for example, you can't hash,

00:08:30.060 --> 00:08:31.680
<v Calvin Hendryx-Parker>use a dictionary as a hash key.

00:08:32.159 --> 00:08:34.659
<v Calvin Hendryx-Parker>You can use it as a value, but you can't store that as a key because

00:08:34.820 --> 00:08:36.120
<v Calvin Hendryx-Parker>the dictionaries themselves are mutable.

00:08:36.500 --> 00:08:39.340
<v Calvin Hendryx-Parker>They figured out how to optimize around turning that

00:08:39.680 --> 00:08:40.820
<v Calvin Hendryx-Parker>dictionary into a string,

00:08:41.060 --> 00:08:44.660
<v Calvin Hendryx-Parker>which is an immutable object in Python, to allow that to be the key for

00:08:44.680 --> 00:08:47.040
<v Calvin Hendryx-Parker>or the pulling back the value out of the cache.

00:08:47.200 --> 00:08:50.860
<v Calvin Hendryx-Parker>That actually led to a cool discovery by them and me now,

00:08:51.220 --> 00:08:53.780
<v Calvin Hendryx-Parker>which is JSON hashed as a string.

00:08:54.280 --> 00:08:55.600
<v Calvin Hendryx-Parker>There's obviously different ways to do it.

00:08:55.650 --> 00:08:57.000
<v Calvin Hendryx-Parker>And then if you want to hash it as the key,

00:08:57.480 --> 00:08:57.640
<v Calvin Hendryx-Parker>there's,

00:08:57.890 --> 00:08:59.880
<v Calvin Hendryx-Parker>for example, you could just use an identity function

00:09:00.260 --> 00:09:01.340
<v Calvin Hendryx-Parker>or an MD5 function,

00:09:01.590 --> 00:09:02.020
<v Calvin Hendryx-Parker>for example.

00:09:02.620 --> 00:09:06.260
<v Calvin Hendryx-Parker>I had never heard of this hashing library called XXH64.

00:09:06.840 --> 00:09:08.620
<v Calvin Hendryx-Parker>I don't know if you've run across that or not either.

00:09:08.630 --> 00:09:09.420
<v Michael Kennedy>No, that's new to me, yeah.

00:09:09.580 --> 00:09:09.660
<v Calvin Hendryx-Parker>Yeah,

00:09:09.810 --> 00:09:10.840
<v Calvin Hendryx-Parker>so converting into JSON,

00:09:11.200 --> 00:09:12.940
<v Calvin Hendryx-Parker>basically we're looking at options like pickle and JSON.

00:09:13.320 --> 00:09:15.360
<v Calvin Hendryx-Parker>Obviously I mentioned JSON as one that got chosen,

00:09:15.940 --> 00:09:19.160
<v Calvin Hendryx-Parker>but if you look here on the performance and time,

00:09:20.200 --> 00:09:22.360
<v Calvin Hendryx-Parker>being able to store that into the database

00:09:22.840 --> 00:09:26.080
<v Calvin Hendryx-Parker>and then compute that hash obviously has a performance impact.

00:09:26.560 --> 00:09:30.480
<v Calvin Hendryx-Parker>The HXXH64 insert and lookups,

00:09:30.550 --> 00:09:32.360
<v Calvin Hendryx-Parker>you can see here are very, very quick,

00:09:32.700 --> 00:09:36.420
<v Calvin Hendryx-Parker>under almost 0.01 seconds in the benchmark

00:09:36.760 --> 00:09:39.720
<v Calvin Hendryx-Parker>compared to like MD5 being 0.1 seconds.

00:09:40.000 --> 00:09:43.040
<v Calvin Hendryx-Parker>So we're looking at like a times 10 speed increase

00:09:43.420 --> 00:09:44.400
<v Calvin Hendryx-Parker>on the MD5 version,

00:09:45.040 --> 00:09:47.940
<v Calvin Hendryx-Parker>which I've made this mistake in the past as well,

00:09:48.080 --> 00:09:50.200
<v Calvin Hendryx-Parker>thinking like, well, MD5 is not necessarily secure,

00:09:50.460 --> 00:09:51.360
<v Calvin Hendryx-Parker>so it must be fast.

00:09:51.580 --> 00:09:52.780
<v Calvin Hendryx-Parker>Not true, not necessarily true.

00:09:53.180 --> 00:09:54.100
<v Calvin Hendryx-Parker>So HH,

00:09:54.500 --> 00:09:55.240
<v Calvin Hendryx-Parker>or good Lord,

00:09:55.500 --> 00:09:56.440
<v Calvin Hendryx-Parker>I'm gonna get that right someday.

00:09:57.839 --> 00:10:00.420
<v Calvin Hendryx-Parker>XXH64 is not cryptographically secure,

00:10:01.000 --> 00:10:02.060
<v Calvin Hendryx-Parker>but that's not the point here.

00:10:02.180 --> 00:10:03.100
<v Calvin Hendryx-Parker>We're using it for identity.

00:10:03.580 --> 00:10:05.320
<v Calvin Hendryx-Parker>And so we were looking for the fastest way

00:10:05.380 --> 00:10:06.500
<v Calvin Hendryx-Parker>to create some kind of a hash.

00:10:06.880 --> 00:10:07.500
<v Calvin Hendryx-Parker>There's an opportunity,

00:10:07.840 --> 00:10:08.519
<v Calvin Hendryx-Parker>potential opportunity

00:10:08.540 --> 00:10:09.240
<v Calvin Hendryx-Parker>for hash collisions,

00:10:09.540 --> 00:10:09.740
<v Calvin Hendryx-Parker>but in this scale-

00:10:09.740 --> 00:10:10.740
<v Michael Kennedy>That is so much faster.

00:10:11.040 --> 00:10:12.100
<v Michael Kennedy>I'm so interested in this.

00:10:12.120 --> 00:10:12.300
<v Calvin Hendryx-Parker>Yeah, yeah,

00:10:12.640 --> 00:10:12.820
<v Calvin Hendryx-Parker>exactly.

00:10:14.060 --> 00:10:15.520
<v Calvin Hendryx-Parker>If nothing else, if you don't even use

00:10:15.640 --> 00:10:18.000
<v Calvin Hendryx-Parker>the SQL-like funky cache piece of this,

00:10:18.520 --> 00:10:20.980
<v Calvin Hendryx-Parker>discovering that there is a faster hashing mechanism

00:10:21.180 --> 00:10:24.700
<v Calvin Hendryx-Parker>when security is not the important part

00:10:24.720 --> 00:10:25.680
<v Calvin Hendryx-Parker>of why you're using it,

00:10:26.140 --> 00:10:28.440
<v Calvin Hendryx-Parker>this is well worth the discovery alone.

00:10:28.620 --> 00:10:28.880
<v Calvin Hendryx-Parker>So yeah.

00:10:29.420 --> 00:10:32.920
<v Michael Kennedy>Let me throw one more thing on hashing

00:10:33.020 --> 00:10:35.940
<v Michael Kennedy>that has caught me out because I was looking at it a lot

00:10:36.060 --> 00:10:37.540
<v Michael Kennedy>with stuff I'm about to talk about.

00:10:37.980 --> 00:10:39.920
<v Michael Kennedy>Python has a built-in hash function,

00:10:40.320 --> 00:10:40.400
<v Michael Kennedy>right?

00:10:40.800 --> 00:10:42.040
<v Michael Kennedy>So why not just use that?

00:10:42.160 --> 00:10:43.260
<v Michael Kennedy>It's the built-in, yeah?

00:10:43.300 --> 00:10:45.520
<v Michael Kennedy>You don't even have to use MD5 or any of those things.

00:10:45.900 --> 00:10:47.340
<v Michael Kennedy>It is not.

00:10:47.720 --> 00:10:50.940
<v Michael Kennedy>If I hash something and then hash it again,

00:10:51.100 --> 00:10:51.840
<v Michael Kennedy>I'll get the same answer.

00:10:52.300 --> 00:10:55.240
<v Michael Kennedy>But if I shut down the process and start it back up again

00:10:55.320 --> 00:10:57.020
<v Michael Kennedy>and hash it again, I get a different answer.

00:10:57.360 --> 00:10:57.680
<v Calvin Hendryx-Parker>Oh, goodness.

00:10:57.680 --> 00:11:00.880
<v Michael Kennedy>So if you save it to a file and you ask,

00:11:01.060 --> 00:11:02.120
<v Michael Kennedy>is it in that file?

00:11:02.240 --> 00:11:03.240
<v Michael Kennedy>It's not in the file anymore.

00:11:03.620 --> 00:11:04.860
<v Michael Kennedy>So stay away.

00:11:05.120 --> 00:11:06.260
<v Michael Kennedy>For this purpose,

00:11:06.440 --> 00:11:08.000
<v Michael Kennedy>for giving an ID,

00:11:08.260 --> 00:11:09.580
<v Michael Kennedy>don't use the build to pick something,

00:11:10.200 --> 00:11:10.940
<v Michael Kennedy>this looks like a good answer.

00:11:12.880 --> 00:11:14.620
<v Calvin Hendryx-Parker>Xxh64 looks like a really good option

00:11:15.139 --> 00:11:17.620
<v Calvin Hendryx-Parker>for reliable hashes that are fast.

00:11:18.620 --> 00:11:18.880
<v Calvin Hendryx-Parker>So yeah,

00:11:19.160 --> 00:11:20.980
<v Calvin Hendryx-Parker>that was super cool to discover that.

00:11:21.260 --> 00:11:22.120
<v Calvin Hendryx-Parker>I think if there's any of that here,

00:11:22.480 --> 00:11:22.820
<v Calvin Hendryx-Parker>cache keys,

00:11:23.200 --> 00:11:24.860
<v Calvin Hendryx-Parker>they're hashed with the JSON bit,

00:11:25.240 --> 00:11:26.320
<v Calvin Hendryx-Parker>they're deliberately lossy,

00:11:26.800 --> 00:11:27.260
<v Calvin Hendryx-Parker>which is fine,

00:11:27.680 --> 00:11:29.080
<v Calvin Hendryx-Parker>because what happens is they go through

00:11:29.420 --> 00:11:32.520
<v Calvin Hendryx-Parker>a couple different iterations on,

00:11:32.900 --> 00:11:34.660
<v Calvin Hendryx-Parker>so I want to jump ahead too quick there,

00:11:35.040 --> 00:11:36.720
<v Calvin Hendryx-Parker>the hashing is, here it is,

00:11:37.160 --> 00:11:40.140
<v Calvin Hendryx-Parker>basically reducing them into a quick hash

00:11:40.420 --> 00:11:41.540
<v Calvin Hendryx-Parker>so that speed was the feature,

00:11:41.940 --> 00:11:44.100
<v Calvin Hendryx-Parker>and then using Pydantic to convert these models

00:11:44.420 --> 00:11:46.280
<v Calvin Hendryx-Parker>to data classes and Python objects,

00:11:46.400 --> 00:11:48.740
<v Calvin Hendryx-Parker>and then refactoring so that each function

00:11:48.840 --> 00:11:50.640
<v Calvin Hendryx-Parker>only takes a minimal set of required data.

00:11:50.900 --> 00:11:52.280
<v Calvin Hendryx-Parker>So this was just an optimization,

00:11:52.540 --> 00:11:52.760
<v Calvin Hendryx-Parker>optimization,

00:11:53.040 --> 00:11:53.260
<v Calvin Hendryx-Parker>optimization,

00:11:53.580 --> 00:11:55.200
<v Calvin Hendryx-Parker>because we're all talking about cache and speed here.

00:11:55.740 --> 00:11:58.200
<v Calvin Hendryx-Parker>The other interesting bit with this specific tool

00:11:58.480 --> 00:12:01.440
<v Calvin Hendryx-Parker>is there's a bit of a dependency tracking here.

00:12:01.720 --> 00:12:02.740
<v Calvin Hendryx-Parker>So if you, for example,

00:12:02.920 --> 00:12:04.960
<v Calvin Hendryx-Parker>or generating a static site generation,

00:12:05.400 --> 00:12:09.260
<v Calvin Hendryx-Parker>you've got a main like index blog page, for example,

00:12:09.660 --> 00:12:10.880
<v Calvin Hendryx-Parker>and you add a new blog post,

00:12:10.890 --> 00:12:13.080
<v Calvin Hendryx-Parker>you would want that main page to get regenerated,

00:12:13.440 --> 00:12:14.540
<v Calvin Hendryx-Parker>but it didn't necessarily change

00:12:14.690 --> 00:12:15.880
<v Calvin Hendryx-Parker>as far as the code's concerned.

00:12:16.190 --> 00:12:18.960
<v Calvin Hendryx-Parker>So you can specify that when these pages change

00:12:18.990 --> 00:12:20.520
<v Calvin Hendryx-Parker>or when this pattern of pages change,

00:12:20.880 --> 00:12:23.000
<v Calvin Hendryx-Parker>regenerate this other potentially expensive page.

00:12:23.280 --> 00:12:25.980
<v Calvin Hendryx-Parker>So it's a little bit like a kind of a graph

00:12:26.490 --> 00:12:27.180
<v Calvin Hendryx-Parker>that you can establish.

00:12:27.350 --> 00:12:29.140
<v Calvin Hendryx-Parker>So you could do something with something like Airflow,

00:12:29.300 --> 00:12:30.180
<v Calvin Hendryx-Parker>but that'd be a little heavy handed.

00:12:31.040 --> 00:12:33.280
<v Calvin Hendryx-Parker>This includes that dependency tracking built in.

00:12:33.310 --> 00:12:35.800
<v Calvin Hendryx-Parker>So you can say registering specific files

00:12:36.340 --> 00:12:37.640
<v Calvin Hendryx-Parker>to track those dependent changes.

00:12:37.980 --> 00:12:40.340
<v Calvin Hendryx-Parker>So if things change in templates, for example,

00:12:40.540 --> 00:12:41.500
<v Calvin Hendryx-Parker>regenerate them.

00:12:42.040 --> 00:12:44.720
<v Calvin Hendryx-Parker>And so it basically shows this kind of reactive aspirations

00:12:44.980 --> 00:12:45.820
<v Calvin Hendryx-Parker>or reactive programming.

00:12:46.440 --> 00:12:48.020
<v Calvin Hendryx-Parker>If you change input article one,

00:12:48.330 --> 00:12:50.720
<v Calvin Hendryx-Parker>then the list of articles two steps away

00:12:50.990 --> 00:12:53.200
<v Calvin Hendryx-Parker>can also change in response to it.

00:12:53.400 --> 00:12:58.500
<v Calvin Hendryx-Parker>So a nice pattern for basically a very minimal DAG situation.

00:12:58.710 --> 00:12:59.940
<v Calvin Hendryx-Parker>So if you didn't want the full blown,

00:13:00.460 --> 00:13:02.780
<v Calvin Hendryx-Parker>Actually, it's kind of worked well with your previous topic

00:13:02.970 --> 00:13:04.160
<v Calvin Hendryx-Parker>about scheduled tasks,

00:13:04.680 --> 00:13:08.020
<v Calvin Hendryx-Parker>combining the previous tiny,

00:13:08.300 --> 00:13:09.380
<v Calvin Hendryx-Parker>tiny, tiny queue,

00:13:09.500 --> 00:13:09.960
<v Calvin Hendryx-Parker>what was that called?

00:13:10.560 --> 00:13:11.340
<v Michael Kennedy>Tiny jobs.

00:13:11.640 --> 00:13:12.000
<v Calvin Hendryx-Parker>Tiny jobs.

00:13:12.140 --> 00:13:14.260
<v Calvin Hendryx-Parker>If you combine tiny jobs with funky cache,

00:13:14.800 --> 00:13:16.260
<v Calvin Hendryx-Parker>you actually get a very lightweight,

00:13:16.840 --> 00:13:18.680
<v Calvin Hendryx-Parker>SQLite based version of Airflow,

00:13:19.090 --> 00:13:19.840
<v Calvin Hendryx-Parker>it seems like to me.

00:13:20.520 --> 00:13:20.880
<v Calvin Hendryx-Parker>I was,

00:13:21.340 --> 00:13:22.860
<v Calvin Hendryx-Parker>again, thought that was a good combo

00:13:22.930 --> 00:13:24.520
<v Calvin Hendryx-Parker>to put together this week's episode.

00:13:24.880 --> 00:13:25.560
<v Michael Kennedy>Yeah, that's really cool.

00:13:25.720 --> 00:13:26.960
<v Michael Kennedy>And I love the names.

00:13:27.690 --> 00:13:28.260
<v Michael Kennedy>I love the names,

00:13:28.560 --> 00:13:29.000
<v Michael Kennedy>they're fun.

00:13:29.420 --> 00:13:29.800
<v Calvin Hendryx-Parker>They're very funky.

00:13:30.190 --> 00:13:30.720
<v Calvin Hendryx-Parker>Very, very funky.

00:13:31.000 --> 00:13:31.900
<v Michael Kennedy>They definitely are funky.

00:13:32.100 --> 00:13:32.280
<v Michael Kennedy>All right.

00:13:32.420 --> 00:13:32.500
<v Michael Kennedy>Now,

00:13:33.090 --> 00:13:36.720
<v Michael Kennedy>before we move on, let me tell you about our sponsor this week.

00:13:37.160 --> 00:13:37.980
<v Michael Kennedy>So Longlink.

00:13:38.240 --> 00:13:40.080
<v Michael Kennedy>Let's talk about Longlink.

00:13:40.480 --> 00:13:45.700
<v Michael Kennedy>Longlink is a really cool platform for building business software.

00:13:46.320 --> 00:13:48.140
<v Michael Kennedy>Saves you a lot of the stuff that you got to do otherwise.

00:13:48.350 --> 00:13:50.340
<v Michael Kennedy>All the plumbing, the foundational stuff.

00:13:50.530 --> 00:13:50.740
<v Michael Kennedy>All right.

00:13:50.880 --> 00:13:53.040
<v Michael Kennedy>So this episode is brought to you by Longlink,

00:13:53.240 --> 00:13:55.660
<v Michael Kennedy>of course, a growing startup out of Switzerland.

00:13:56.200 --> 00:13:59.100
<v Michael Kennedy>How much of your company still runs on spreadsheets?

00:13:59.600 --> 00:14:00.400
<v Michael Kennedy>I don't know what the answer is.

00:14:00.720 --> 00:14:01.920
<v Michael Kennedy>Probably it's too much.

00:14:02.320 --> 00:14:03.560
<v Calvin Hendryx-Parker>Probably too much, even in my own.

00:14:03.960 --> 00:14:05.860
<v Michael Kennedy>In all of ours, it's probably too much.

00:14:06.300 --> 00:14:07.120
<v Michael Kennedy>So for most businesses,

00:14:07.440 --> 00:14:07.840
<v Michael Kennedy>it's a lot.

00:14:07.900 --> 00:14:08.960
<v Michael Kennedy>Approvals live in email threads,

00:14:09.660 --> 00:14:11.100
<v Michael Kennedy>tracking lives in shared Excel files.

00:14:11.560 --> 00:14:13.040
<v Michael Kennedy>Off-the-shelf software rarely fits,

00:14:13.200 --> 00:14:14.500
<v Michael Kennedy>so you've got to work around it.

00:14:14.580 --> 00:14:16.460
<v Michael Kennedy>So AI has actually changed that quite a bit.

00:14:16.470 --> 00:14:17.720
<v Michael Kennedy>And I'm seeing that with other companies

00:14:17.810 --> 00:14:18.340
<v Michael Kennedy>that are working with it.

00:14:18.340 --> 00:14:19.800
<v Michael Kennedy>They're like, oh, here's our new internal app.

00:14:19.860 --> 00:14:21.340
<v Michael Kennedy>I'm like, oh, that looks familiar.

00:14:22.180 --> 00:14:24.500
<v Michael Kennedy>Or I could see the origin of where that might come from.

00:14:24.700 --> 00:14:27.560
<v Michael Kennedy>So custom business software is now cheap to build,

00:14:27.720 --> 00:14:31.380
<v Michael Kennedy>but keeping it simple and affordable to maintain is not as easy.

00:14:31.840 --> 00:14:33.700
<v Michael Kennedy>So that's the problem that Longlink is solving.

00:14:34.120 --> 00:14:36.820
<v Michael Kennedy>You write your business apps in plain Python with familiar libraries

00:14:36.990 --> 00:14:38.260
<v Michael Kennedy>and standard development practices.

00:14:38.820 --> 00:14:40.740
<v Michael Kennedy>Longlink provides the platform underneath.

00:14:41.080 --> 00:14:41.540
<v Michael Kennedy>Authentication,

00:14:41.920 --> 00:14:42.460
<v Michael Kennedy>permissions,

00:14:42.900 --> 00:14:43.260
<v Michael Kennedy>deployment,

00:14:43.540 --> 00:14:44.020
<v Michael Kennedy>storage,

00:14:44.200 --> 00:14:44.420
<v Michael Kennedy>routing,

00:14:44.820 --> 00:14:45.680
<v Michael Kennedy>logging, all handled.

00:14:46.000 --> 00:14:47.080
<v Michael Kennedy>You focus on the business logic.

00:14:47.280 --> 00:14:47.920
<v Michael Kennedy>Purchase approvals,

00:14:48.120 --> 00:14:49.720
<v Michael Kennedy>vendor onboarding, work orders,

00:14:50.080 --> 00:14:51.320
<v Michael Kennedy>and complaint tracking

00:14:51.530 --> 00:14:53.580
<v Michael Kennedy>to start one process and then build the next

00:14:53.590 --> 00:14:56.900
<v Michael Kennedy>and sort of work back through these apps that need modernization.

00:14:57.240 --> 00:14:58.400
<v Michael Kennedy>So get started in one command,

00:14:58.680 --> 00:15:00.560
<v Michael Kennedy>uvx long link in it.

00:15:00.800 --> 00:15:02.540
<v Michael Kennedy>Platform has a free tier, no credit card required.

00:15:02.940 --> 00:15:04.560
<v Michael Kennedy>The project is open source and in beta,

00:15:04.740 --> 00:15:06.460
<v Michael Kennedy>and the team would love you to try it out and share your feedback.

00:15:06.650 --> 00:15:07.240
<v Michael Kennedy>And while you're there,

00:15:07.510 --> 00:15:09.580
<v Michael Kennedy>they would also really appreciate a GitHub star.

00:15:09.820 --> 00:15:12.520
<v Michael Kennedy>So check it out at pythonbytes.fm/long link.

00:15:13.080 --> 00:15:13.380
<v Michael Kennedy>The link,

00:15:13.570 --> 00:15:15.580
<v Michael Kennedy>no joke, is in your podcast player show notes.

00:15:15.860 --> 00:15:16.320
<v Calvin Hendryx-Parker>Is it long?

00:15:17.500 --> 00:15:18.940
<v Michael Kennedy>Actually, ironically or surprisingly,

00:15:19.230 --> 00:15:20.540
<v Michael Kennedy>it is not long.

00:15:20.710 --> 00:15:21.180
<v Michael Kennedy>How about that?

00:15:21.300 --> 00:15:21.540
<v Calvin Hendryx-Parker>That's great.

00:15:21.720 --> 00:15:22.200
<v Calvin Hendryx-Parker>Great to hear.

00:15:22.480 --> 00:15:22.580
<v Michael Kennedy>Yeah.

00:15:22.940 --> 00:15:25.320
<v Michael Kennedy>All right. Let's go back over to you.

00:15:25.480 --> 00:15:25.760
<v Michael Kennedy>What's next?

00:15:26.000 --> 00:15:26.000
<v Calvin Hendryx-Parker>Yeah,

00:15:26.620 --> 00:15:29.080
<v Calvin Hendryx-Parker>so I can't resist a great set of tools

00:15:29.600 --> 00:15:30.880
<v Calvin Hendryx-Parker>and this one is no exception.

00:15:31.699 --> 00:15:33.380
<v Calvin Hendryx-Parker>I love local based tools,

00:15:33.580 --> 00:15:35.120
<v Calvin Hendryx-Parker>I love terminal based tools

00:15:35.140 --> 00:15:36.240
<v Calvin Hendryx-Parker>and this one is both.

00:15:36.500 --> 00:15:38.760
<v Calvin Hendryx-Parker>And so this is a fast keyboard driven

00:15:39.380 --> 00:15:41.700
<v Calvin Hendryx-Parker>TUI API client and webhook inspector

00:15:41.880 --> 00:15:43.600
<v Calvin Hendryx-Parker>written in Python called Ghostman.

00:15:43.920 --> 00:15:45.080
<v Calvin Hendryx-Parker>So if you check it out,

00:15:45.740 --> 00:15:47.460
<v Calvin Hendryx-Parker>the GitHub repository for it,

00:15:47.740 --> 00:15:49.320
<v Calvin Hendryx-Parker>it is no cloud,

00:15:49.600 --> 00:15:50.080
<v Calvin Hendryx-Parker>no electron,

00:15:50.720 --> 00:15:51.700
<v Calvin Hendryx-Parker>100% local first.

00:15:52.040 --> 00:15:54.060
<v Calvin Hendryx-Parker>It hits all the marks for me.

00:15:54.640 --> 00:15:55.780
<v Calvin Hendryx-Parker>It's also MIT licensed.

00:15:56.200 --> 00:15:56.780
<v Calvin Hendryx-Parker>It just,

00:15:57.340 --> 00:15:59.000
<v Calvin Hendryx-Parker>again, looks like a really cool tool.

00:15:59.300 --> 00:16:00.560
<v Calvin Hendryx-Parker>I've started playing with it a little bit

00:16:00.650 --> 00:16:01.700
<v Calvin Hendryx-Parker>in preparation for the show.

00:16:02.540 --> 00:16:05.840
<v Calvin Hendryx-Parker>They mentioned installing it via pip from PyPI,

00:16:06.180 --> 00:16:08.320
<v Calvin Hendryx-Parker>but I would recommend just uvx Ghostman.

00:16:08.660 --> 00:16:10.700
<v Calvin Hendryx-Parker>If you just type uvx Ghostman into your terminal,

00:16:11.220 --> 00:16:13.680
<v Calvin Hendryx-Parker>you will get a beautiful little looking terminal thing.

00:16:13.840 --> 00:16:18.020
<v Calvin Hendryx-Parker>So for those of you who've used Postman or Bruno

00:16:18.520 --> 00:16:22.899
<v Calvin Hendryx-Parker>or some of the other web GUI versions of API client

00:16:22.920 --> 00:16:26.720
<v Calvin Hendryx-Parker>testing out and exploring various REST APIs,

00:16:27.280 --> 00:16:29.600
<v Calvin Hendryx-Parker>this is an option you can use just right there in your

00:16:29.700 --> 00:16:29.880
<v Calvin Hendryx-Parker>terminal.

00:16:30.730 --> 00:16:33.860
<v Calvin Hendryx-Parker>It just launched, first release was like Halloween,

00:16:34.460 --> 00:16:35.900
<v Calvin Hendryx-Parker>and then this most recent release came

00:16:36.040 --> 00:16:40.860
<v Calvin Hendryx-Parker>out November 3rd at 1.0.1.8, built on Textual, HTTPX,

00:16:41.120 --> 00:16:41.520
<v Calvin Hendryx-Parker>and Rich,

00:16:41.720 --> 00:16:43.400
<v Calvin Hendryx-Parker>so it's pure Python under the hood,

00:16:43.650 --> 00:16:44.280
<v Calvin Hendryx-Parker>which also makes,

00:16:44.540 --> 00:16:48.759
<v Calvin Hendryx-Parker>you know, warms my heart when I hear that there are these pure Python tools still

00:16:48.780 --> 00:16:49.040
<v Calvin Hendryx-Parker>out there.

00:16:49.480 --> 00:16:52.200
<v Calvin Hendryx-Parker>You can do things like, you know, obviously the standard get post put,

00:16:52.440 --> 00:16:52.540
<v Calvin Hendryx-Parker>delete,

00:16:52.780 --> 00:16:53.620
<v Calvin Hendryx-Parker>and patch requests.

00:16:54.120 --> 00:16:57.140
<v Calvin Hendryx-Parker>It'll show DNS and time to first byte. So you get a little bit of performance

00:16:57.360 --> 00:16:57.980
<v Calvin Hendryx-Parker>stuff going on in there.

00:16:58.440 --> 00:17:02.300
<v Calvin Hendryx-Parker>And it will also color your JSON responses in response.

00:17:02.780 --> 00:17:03.500
<v Calvin Hendryx-Parker>It's got a built

00:17:03.560 --> 00:17:03.660
<v Calvin Hendryx-Parker>in,

00:17:04.100 --> 00:17:06.839
<v Calvin Hendryx-Parker>the neat thing, it's got a built in webhook inspector.

00:17:07.020 --> 00:17:08.720
<v Calvin Hendryx-Parker>So if you use the keyboard, there's a

00:17:08.720 --> 00:17:11.339
<v Calvin Hendryx-Parker>couple of keyboard shortcuts for it, which is also kind of nice,

00:17:11.500 --> 00:17:12.319
<v Calvin Hendryx-Parker>big fan of keyboard shortcuts.

00:17:12.980 --> 00:17:15.600
<v Calvin Hendryx-Parker>Control W will start the webhook listener on port 9090.

00:17:15.880 --> 00:17:17.859
<v Calvin Hendryx-Parker>You can actually point your application to

00:17:17.880 --> 00:17:22.660
<v Calvin Hendryx-Parker>it if it's posting webhooks and inspect what comes out the other side without having to resort to

00:17:22.839 --> 00:17:27.360
<v Calvin Hendryx-Parker>like netcat and some other stuff i i used to do that quite a bit i would just like open up a netcat

00:17:27.740 --> 00:17:32.420
<v Calvin Hendryx-Parker>watch the text just flow in but here you can get it actually like syntax highlighted which was nice

00:17:32.740 --> 00:17:36.980
<v Michael Kennedy>uh the webhooks are the worst i mean they're really valuable but they're so hard to debug

00:17:37.360 --> 00:17:42.619
<v Calvin Hendryx-Parker>hard to debug exactly hard to debug you can build collections just like you could in postmans and

00:17:42.640 --> 00:17:48.580
<v Calvin Hendryx-Parker>you can put them it looks for them under a.ghostman folder inside the current directory or you can

00:17:48.580 --> 00:17:52.740
<v Calvin Hendryx-Parker>have a global one at your home directory with a.ghostman version of it there.

00:17:53.100 --> 00:17:53.500
<v Calvin Hendryx-Parker>Again fully

00:17:53.640 --> 00:17:56.260
<v Calvin Hendryx-Parker>keyboard driven it's very very lean.

00:17:56.680 --> 00:18:00.320
<v Calvin Hendryx-Parker>The author claims 30 megabytes of RAM and a cold start of

00:18:00.460 --> 00:18:01.700
<v Calvin Hendryx-Parker>about 150 milliseconds.

00:18:02.100 --> 00:18:04.600
<v Calvin Hendryx-Parker>So we all love a fast tool like that.

00:18:04.780 --> 00:18:06.000
<v Calvin Hendryx-Parker>Actually I've got a demo for you. I'll

00:18:06.120 --> 00:18:07.360
<v Calvin Hendryx-Parker>show you a quick little demo.

00:18:07.870 --> 00:18:10.720
<v Calvin Hendryx-Parker>I can put up my screen again here share screen.

00:18:11.080 --> 00:18:11.839
<v Calvin Hendryx-Parker>I'll go to my

00:18:11.860 --> 00:18:12.000
<v Calvin Hendryx-Parker>terminal.

00:18:12.480 --> 00:18:12.920
<v Calvin Hendryx-Parker>Here we go.

00:18:13.290 --> 00:18:16.640
<v Calvin Hendryx-Parker>So here is going to add it back to the stage. There we go.

00:18:16.960 --> 00:18:17.780
<v Calvin Hendryx-Parker>This is what it

00:18:17.780 --> 00:18:18.480
<v Calvin Hendryx-Parker>looks like when it launches.

00:18:18.780 --> 00:18:22.560
<v Calvin Hendryx-Parker>It's got a couple of examples built into here. So you can just like

00:18:22.900 --> 00:18:23.460
<v Calvin Hendryx-Parker>load that up,

00:18:23.800 --> 00:18:26.340
<v Calvin Hendryx-Parker>choose what you want. And I'm using the mouse with it,

00:18:26.520 --> 00:18:26.860
<v Calvin Hendryx-Parker>oddly enough.

00:18:27.120 --> 00:18:27.720
<v Calvin Hendryx-Parker>It's a full rich,

00:18:28.220 --> 00:18:28.340
<v Calvin Hendryx-Parker>you know,

00:18:28.600 --> 00:18:29.360
<v Calvin Hendryx-Parker>2E interface,

00:18:30.020 --> 00:18:30.640
<v Calvin Hendryx-Parker>which is kind of nice,

00:18:30.840 --> 00:18:32.020
<v Calvin Hendryx-Parker>but you can create posts,

00:18:32.450 --> 00:18:32.560
<v Calvin Hendryx-Parker>public,

00:18:33.180 --> 00:18:33.600
<v Calvin Hendryx-Parker>you know, it's got the

00:18:33.880 --> 00:18:35.360
<v Calvin Hendryx-Parker>GitHub example here for Octocat.

00:18:35.730 --> 00:18:36.520
<v Calvin Hendryx-Parker>You can get the response.

00:18:36.860 --> 00:18:38.720
<v Calvin Hendryx-Parker>Scrolling works as you would expect it,

00:18:38.780 --> 00:18:41.740
<v Calvin Hendryx-Parker>or you can go fully full blown keyboard shortcuts.

00:18:42.180 --> 00:18:43.600
<v Calvin Hendryx-Parker>If I hit the webhook UI,

00:18:43.850 --> 00:18:46.580
<v Calvin Hendryx-Parker>you can see right there webhook listener started on 9090.

00:18:46.720 --> 00:18:48.120
<v Calvin Hendryx-Parker>So now if I was supposed to post something into it,

00:18:48.290 --> 00:18:51.000
<v Calvin Hendryx-Parker>it would show up over here in the in the UI.

00:18:51.440 --> 00:18:51.580
<v Calvin Hendryx-Parker>Again,

00:18:51.880 --> 00:18:53.340
<v Calvin Hendryx-Parker>really, really cool local first.

00:18:53.630 --> 00:18:55.660
<v Calvin Hendryx-Parker>You can see there's the collections piece there.

00:18:55.660 --> 00:18:56.640
<v Calvin Hendryx-Parker>You can create a new collection,

00:18:57.260 --> 00:18:59.480
<v Calvin Hendryx-Parker>start in adding in payloads and webhooks and headers.

00:18:59.780 --> 00:19:01.280
<v Calvin Hendryx-Parker>And it's basically Postman.

00:19:01.720 --> 00:19:03.560
<v Calvin Hendryx-Parker>But for those of us who love the terminal,

00:19:03.870 --> 00:19:04.780
<v Calvin Hendryx-Parker>I found a new tool.

00:19:04.930 --> 00:19:05.400
<v Calvin Hendryx-Parker>I love it.

00:19:05.590 --> 00:19:06.520
<v Michael Kennedy>This is very Calvin.

00:19:06.570 --> 00:19:06.980
<v Michael Kennedy>I love it.

00:19:07.900 --> 00:19:15.180
<v Calvin Hendryx-Parker>it's so true yeah it looks textual it is yeah it's textual and httpx which again just beautiful

00:19:15.530 --> 00:19:21.780
<v Calvin Hendryx-Parker>uh and this is running inside of my 3d accelerated terminal kitty which i am a huge fan of because

00:19:21.880 --> 00:19:27.420
<v Calvin Hendryx-Parker>everyone needs a 3d accelerated terminal why would you not doing at least 240 frames a second on your

00:19:27.520 --> 00:19:33.539
<v Calvin Hendryx-Parker>terminal you're doing a terminal i mean speed is the feature michael speed is the feature but on

00:19:33.560 --> 00:19:39.120
<v Michael Kennedy>you know that is a joke sort of but at the same time if you get if you get flickering and stuff

00:19:39.240 --> 00:19:43.800
<v Michael Kennedy>in your terminal then that's also not fun that's really that's annoying it's no good no good

00:19:44.080 --> 00:19:49.580
<v Michael Kennedy>it's no good all right something that i think is good i want to i want to talk about the next thing

00:19:49.980 --> 00:19:55.540
<v Michael Kennedy>this is the big thing for for me this week so i've been a big fan of disk cache and disk cache i even

00:19:55.640 --> 00:20:03.520
<v Michael Kennedy>did a podcast episode on talk python with vincent warmerdom and we both said should we cover this

00:20:03.540 --> 00:20:08.520
<v Michael Kennedy>It literally is possibly one of the best caching options bar none out there.

00:20:08.520 --> 00:20:09.140
<v Michael Kennedy>It is so,

00:20:09.380 --> 00:20:09.680
<v Michael Kennedy>so good.

00:20:10.120 --> 00:20:13.220
<v Michael Kennedy>But the person who maintains it just ghosted the world, which is fine.

00:20:13.240 --> 00:20:15.940
<v Michael Kennedy>I'm not putting them on blast or nothing for that.

00:20:16.120 --> 00:20:17.720
<v Michael Kennedy>But it leaves a big gap,

00:20:18.080 --> 00:20:18.200
<v Michael Kennedy>right?

00:20:18.720 --> 00:20:20.060
<v Michael Kennedy>And had a bunch of problems,

00:20:20.440 --> 00:20:21.140
<v Michael Kennedy>a bunch of GitHub issues,

00:20:21.320 --> 00:20:22.680
<v Michael Kennedy>a bunch of unanswered pull requests.

00:20:22.940 --> 00:20:24.600
<v Michael Kennedy>And I thought, well, this is part of my mission,

00:20:24.780 --> 00:20:24.880
<v Michael Kennedy>right?

00:20:24.940 --> 00:20:26.260
<v Michael Kennedy>Like I'm trying to make free.

00:20:27.020 --> 00:20:30.400
<v Michael Kennedy>My mission is I want a foundation I can build web apps in

00:20:30.800 --> 00:20:32.740
<v Michael Kennedy>that is safe and free-threaded Python.

00:20:32.980 --> 00:20:34.840
<v Michael Kennedy>And while you're at it, let's just see if we can make it fast.

00:20:34.910 --> 00:20:35.000
<v Michael Kennedy>Yeah.

00:20:35.400 --> 00:20:40.200
<v Michael Kennedy>So my big announcement is I rebuilt disk cache and something called disk cache oxide.

00:20:40.600 --> 00:20:42.260
<v Michael Kennedy>Calvin, can you guess what that might be written in?

00:20:43.340 --> 00:20:43.700
<v Calvin Hendryx-Parker>Let me guess.

00:20:44.500 --> 00:20:46.620
<v Calvin Hendryx-Parker>I'm sensing a trend here in your projects.

00:20:47.190 --> 00:20:47.320
<v Calvin Hendryx-Parker>Yes.

00:20:47.780 --> 00:20:48.260
<v Calvin Hendryx-Parker>Is it Rust?

00:20:48.640 --> 00:20:50.080
<v Michael Kennedy>It is a Rust core.

00:20:50.460 --> 00:20:50.900
<v Calvin Hendryx-Parker>Love that.

00:20:51.380 --> 00:20:52.100
<v Michael Kennedy>With Python.

00:20:52.720 --> 00:20:53.880
<v Michael Kennedy>It's a Python only cache.

00:20:53.980 --> 00:20:55.160
<v Michael Kennedy>This is not a Rust library.

00:20:55.600 --> 00:20:57.340
<v Michael Kennedy>So introducing disk cache oxide,

00:20:57.840 --> 00:20:58.460
<v Michael Kennedy>Python cache,

00:20:58.860 --> 00:20:59.240
<v Michael Kennedy>Rust engine,

00:20:59.760 --> 00:21:00.440
<v Michael Kennedy>SQLite storage.

00:21:00.780 --> 00:21:03.000
<v Michael Kennedy>And this thing is an absolute monster.

00:21:03.250 --> 00:21:04.880
<v Michael Kennedy>I mean, it's out of control.

00:21:05.490 --> 00:21:08.780
<v Michael Kennedy>So this is powering Python by Spatafim right now

00:21:09.140 --> 00:21:10.540
<v Michael Kennedy>and the courses and talk Python.

00:21:10.840 --> 00:21:13.320
<v Michael Kennedy>And it's actually been running all the web apps

00:21:13.500 --> 00:21:14.580
<v Michael Kennedy>for maybe a week,

00:21:15.460 --> 00:21:16.480
<v Michael Kennedy>week and a half, something like that.

00:21:16.720 --> 00:21:18.480
<v Michael Kennedy>I didn't want to announce it and have some kind of issue.

00:21:18.680 --> 00:21:20.940
<v Michael Kennedy>I'm like, let's just live with it for a while,

00:21:21.360 --> 00:21:21.960
<v Michael Kennedy>see how it's going.

00:21:22.110 --> 00:21:23.320
<v Michael Kennedy>And it's been flawless.

00:21:23.820 --> 00:21:27.160
<v Michael Kennedy>So let me just scroll down here and pull up some numbers.

00:21:27.300 --> 00:21:31.680
<v Michael Kennedy>on a single SQLite file with one thread,

00:21:32.200 --> 00:21:33.560
<v Michael Kennedy>just straight gil python,

00:21:33.860 --> 00:21:35.020
<v Michael Kennedy>nothing fancy yet.

00:21:35.280 --> 00:21:39.080
<v Michael Kennedy>It does in a 99 to one read-write ratio.

00:21:39.780 --> 00:21:40.940
<v Michael Kennedy>So 99 reads,

00:21:41.190 --> 00:21:41.540
<v Michael Kennedy>one write.

00:21:41.570 --> 00:21:42.720
<v Michael Kennedy>That's pretty cash standard.

00:21:42.960 --> 00:21:44.960
<v Michael Kennedy>There's also a graph for a thousand reads and one write.

00:21:44.990 --> 00:21:49.980
<v Michael Kennedy>It does 670,000 reads and writes a second.

00:21:51.060 --> 00:21:51.560
<v Michael Kennedy>A second,

00:21:51.780 --> 00:21:52.080
<v Michael Kennedy>okay?

00:21:52.660 --> 00:21:54.700
<v Calvin Hendryx-Parker>Michael, isn't it just good enough?

00:21:55.940 --> 00:21:56.680
<v Calvin Hendryx-Parker>Do we need this?

00:21:57.420 --> 00:22:02.820
<v Michael Kennedy>yes we do because the previous one was also very impressive like i i do not take away my my um

00:22:03.280 --> 00:22:08.560
<v Michael Kennedy>positive comments about regular disc cache yeah but if and here remember the free thread mission

00:22:08.610 --> 00:22:14.580
<v Michael Kennedy>this is my mission what happens if i take that same awesome disc cache not mine the original

00:22:15.000 --> 00:22:20.300
<v Michael Kennedy>and i use it in a multi-threaded situation it goes from 400 000 requests a second to 19 000

00:22:20.760 --> 00:22:25.860
<v Michael Kennedy>woof it's 20 times slower if i try to use it in a free threaded way yeah broken as that

00:22:25.960 --> 00:22:26.340
<v Calvin Hendryx-Parker>It's no good.

00:22:26.420 --> 00:22:28.800
<v Michael Kennedy>And it's just, it was built for a different time, right?

00:22:29.240 --> 00:22:30.860
<v Michael Kennedy>And I would be totally content if,

00:22:31.460 --> 00:22:32.520
<v Michael Kennedy>I think the guy's name is Greg.

00:22:32.740 --> 00:22:34.360
<v Michael Kennedy>Greg behind it would like go, yeah,

00:22:34.520 --> 00:22:35.960
<v Michael Kennedy>we're totally going to ramp this thing up.

00:22:36.100 --> 00:22:37.320
<v Michael Kennedy>But he got hired at OpenAI.

00:22:37.740 --> 00:22:38.960
<v Michael Kennedy>And then like the next day,

00:22:39.120 --> 00:22:40.620
<v Michael Kennedy>like responses on GetUp.

00:22:40.860 --> 00:22:41.640
<v Michael Kennedy>Like there was just,

00:22:42.020 --> 00:22:42.920
<v Michael Kennedy>maybe it's a corporate thing.

00:22:43.040 --> 00:22:43.180
<v Michael Kennedy>I don't know.

00:22:43.240 --> 00:22:44.800
<v Michael Kennedy>It's just like not allowed to even reply,

00:22:44.980 --> 00:22:45.480
<v Michael Kennedy>but done.

00:22:46.000 --> 00:22:49.140
<v Michael Kennedy>So I went through and I said, I want to take a similar API,

00:22:49.440 --> 00:22:50.900
<v Michael Kennedy>not identical, but almost the

00:22:51.040 --> 00:22:51.180
<v Michael Kennedy>same.

00:22:51.460 --> 00:22:53.760
<v Michael Kennedy>and I want to build this free-threaded,

00:22:54.290 --> 00:22:54.940
<v Michael Kennedy>multi-threaded,

00:22:55.210 --> 00:22:56.920
<v Michael Kennedy>Rust-based version of this.

00:22:57.180 --> 00:22:57.260
<v Michael Kennedy>Yeah,

00:22:57.470 --> 00:22:58.160
<v Michael Kennedy>and that's what we got.

00:22:58.380 --> 00:22:59.120
<v Michael Kennedy>It's super,

00:22:59.560 --> 00:22:59.880
<v Michael Kennedy>super neat.

00:22:59.970 --> 00:23:01.640
<v Michael Kennedy>And it does a bunch of things

00:23:01.840 --> 00:23:04.300
<v Michael Kennedy>that the regular disk cache does it.

00:23:04.400 --> 00:23:06.740
<v Michael Kennedy>And I think this is really one of them,

00:23:06.900 --> 00:23:08.180
<v Michael Kennedy>going to be one of the really best

00:23:08.520 --> 00:23:09.420
<v Michael Kennedy>caching libraries for Python.

00:23:09.760 --> 00:23:10.940
<v Calvin Hendryx-Parker>So here's how it works.

00:23:11.100 --> 00:23:13.400
<v Calvin Hendryx-Parker>So you're ready to deploy onto Py0 now?

00:23:13.680 --> 00:23:14.060
<v Michael Kennedy>Yes,

00:23:14.290 --> 00:23:14.400
<v Michael Kennedy>yeah,

00:23:14.600 --> 00:23:15.120
<v Michael Kennedy>100%.

00:23:15.120 --> 00:23:17.540
<v Michael Kennedy>So you give it just SQLite file

00:23:17.770 --> 00:23:19.660
<v Michael Kennedy>and it can be durable or not.

00:23:19.720 --> 00:23:24.860
<v Michael Kennedy>So the serialization for disk cache and a lot of these caches, I think I saw it in the other one, is pickle.

00:23:25.380 --> 00:23:26.440
<v Michael Kennedy>That's just wrong to me.

00:23:26.740 --> 00:23:28.480
<v Michael Kennedy>Like it's a security violation.

00:23:28.670 --> 00:23:30.960
<v Michael Kennedy>You hear this a lot with hugging face and those types of things.

00:23:31.200 --> 00:23:31.280
<v Michael Kennedy>Like,

00:23:31.640 --> 00:23:32.620
<v Michael Kennedy>hey, if we use pickle,

00:23:32.840 --> 00:23:35.080
<v Michael Kennedy>somebody could insert a malicious script.

00:23:35.250 --> 00:23:38.680
<v Michael Kennedy>And then when you just load the file, it runs this Python and it's bad.

00:23:39.000 --> 00:23:40.020
<v Michael Kennedy>Plus it's just not durable.

00:23:40.160 --> 00:23:44.900
<v Michael Kennedy>Like if I change the name of a field and I rerun it, like I can't parse it anymore.

00:23:45.000 --> 00:23:45.140
<v Michael Kennedy>Well,

00:23:45.380 --> 00:23:45.740
<v Michael Kennedy>that's broken.

00:23:45.960 --> 00:23:47.440
<v Michael Kennedy>You know, like I can't control it.

00:23:47.580 --> 00:23:48.060
<v Michael Kennedy>Nothing like that.

00:23:48.260 --> 00:23:50.480
<v Michael Kennedy>So this by default uses message pack,

00:23:50.840 --> 00:23:53.340
<v Michael Kennedy>which is a super tight binary format for storing things.

00:23:53.960 --> 00:23:57.120
<v Michael Kennedy>And also has JSON as another built-in option.

00:23:57.260 --> 00:23:58.400
<v Michael Kennedy>And if you really want, you can say pickle,

00:23:58.640 --> 00:23:59.700
<v Michael Kennedy>but try not to, you know?

00:23:59.920 --> 00:24:00.560
<v Calvin Hendryx-Parker>Now why risk it?

00:24:00.760 --> 00:24:01.080
<v Michael Kennedy>Yeah, exactly.

00:24:01.440 --> 00:24:02.300
<v Calvin Hendryx-Parker>Those things can execute.

00:24:03.000 --> 00:24:04.280
<v Michael Kennedy>Yeah, so you can come up with this cache.

00:24:04.280 --> 00:24:04.640
<v Michael Kennedy>You can say,

00:24:04.830 --> 00:24:07.180
<v Michael Kennedy>if something goes wrong with the GC serialization,

00:24:07.340 --> 00:24:09.700
<v Michael Kennedy>just throw that away and just say it's a miss,

00:24:09.990 --> 00:24:10.080
<v Michael Kennedy>right?

00:24:10.260 --> 00:24:11.620
<v Michael Kennedy>So instead of trying to deal with a bunch of errors,

00:24:11.860 --> 00:24:14.380
<v Michael Kennedy>you're just like, look, if I can't deserialize it, it's a cache.

00:24:14.460 --> 00:24:16.840
<v Michael Kennedy>What else am I going to do besides just say it's not there?

00:24:17.760 --> 00:24:20.160
<v Michael Kennedy>Or you can let it throw exceptions and so on.

00:24:20.260 --> 00:24:26.300
<v Michael Kennedy>And then you just cache get item or bracket item and then set as expirations.

00:24:26.520 --> 00:24:26.660
<v Michael Kennedy>Now,

00:24:26.880 --> 00:24:28.740
<v Michael Kennedy>a lot of these SQLite local caches,

00:24:29.380 --> 00:24:32.800
<v Michael Kennedy>it's up to you to kind of manage the files

00:24:33.040 --> 00:24:37.900
<v Michael Kennedy>and make sure at some point in your app you call expire to get rid of the other ones

00:24:37.960 --> 00:24:39.860
<v Michael Kennedy>or you can set a limit of how many it can have.

00:24:39.860 --> 00:24:43.100
<v Michael Kennedy>You can say delete and delete those if it's too big and call them and so on.

00:24:43.420 --> 00:24:49.560
<v Michael Kennedy>This has a dedicated worker maintenance thread in Rust that's periodically going around and just cleaning up.

00:24:49.720 --> 00:24:50.820
<v Michael Kennedy>So it's a little more tidy.

00:24:51.160 --> 00:24:56.440
<v Michael Kennedy>You can even use it for job scheduling and cross-process coordination.

00:24:56.840 --> 00:24:58.740
<v Michael Kennedy>So you can say, like, I want to claim a job.

00:24:58.880 --> 00:25:00.740
<v Michael Kennedy>I'll add this job, give it my ID,

00:25:01.060 --> 00:25:01.420
<v Michael Kennedy>and expire.

00:25:01.480 --> 00:25:04.560
<v Michael Kennedy>And either it's already in there, which means it's running,

00:25:04.760 --> 00:25:07.740
<v Michael Kennedy>or you've added it, which means, hey, you're starting this job.

00:25:07.860 --> 00:25:12.900
<v Michael Kennedy>And the fact that you expire the previous one is you just keep trying to add this job.

00:25:12.980 --> 00:25:13.740
<v Michael Kennedy>It does git added.

00:25:13.820 --> 00:25:15.040
<v Michael Kennedy>That means the previous one expired.

00:25:15.120 --> 00:25:15.900
<v Michael Kennedy>So it's time to do,

00:25:16.060 --> 00:25:17.420
<v Michael Kennedy>I don't know, refresh feeds or whatever.

00:25:17.840 --> 00:25:21.140
<v Michael Kennedy>And because SQLite is cross-atomic and cross-process,

00:25:21.880 --> 00:25:24.460
<v Michael Kennedy>this basically is multiprocess coordination.

00:25:25.240 --> 00:25:26.000
<v Michael Kennedy>So if I have, say,

00:25:26.580 --> 00:25:29.140
<v Michael Kennedy>Granian or GUnicorn or whatever with four workers,

00:25:29.440 --> 00:25:32.740
<v Michael Kennedy>they can use this to have one of the workers periodically run one of the jobs,

00:25:33.120 --> 00:25:33.660
<v Michael Kennedy>which is pretty neat.

00:25:33.960 --> 00:25:37.120
<v Michael Kennedy>There's a lot of documentation that I've put together.

00:25:37.240 --> 00:25:40.840
<v Michael Kennedy>And by a lot, I don't think you would, until you start reading,

00:25:41.040 --> 00:25:43.640
<v Michael Kennedy>you're like, why am I not making progress in the scroll bar?

00:25:44.060 --> 00:25:45.920
<v Michael Kennedy>There's, by way of comparison,

00:25:46.610 --> 00:25:49.980
<v Michael Kennedy>The Great Gap Spee is 45,000 words of the book,

00:25:50.240 --> 00:25:50.520
<v Michael Kennedy>the novel.

00:25:50.980 --> 00:25:53.680
<v Michael Kennedy>This is 70,000 words of like, hey, I'm new to caching.

00:25:53.790 --> 00:25:55.460
<v Michael Kennedy>What do I do on this situation?

00:25:55.710 --> 00:25:56.140
<v Michael Kennedy>How do I,

00:25:56.520 --> 00:25:56.780
<v Michael Kennedy>you know,

00:25:57.010 --> 00:25:58.320
<v Michael Kennedy>how do I work with serialization?

00:25:58.500 --> 00:26:00.760
<v Michael Kennedy>How do I work with like groups of named caches?

00:26:01.080 --> 00:26:06.340
<v Michael Kennedy>Or there's like recipes for different common things you do with like coordinated workers,

00:26:06.580 --> 00:26:06.640
<v Michael Kennedy>right?

00:26:06.700 --> 00:26:07.480
<v Michael Kennedy>that thing I just talked about,

00:26:07.680 --> 00:26:10.680
<v Michael Kennedy>or using it for rate limiting, or using it for work cues,

00:26:11.080 --> 00:26:13.980
<v Michael Kennedy>and so on. So this is a really fun project.

00:26:14.380 --> 00:26:17.340
<v Michael Kennedy>Let me see if I can find the free threading.

00:26:18.260 --> 00:26:20.580
<v Michael Kennedy>Let's see, I think I was trying just under performance,

00:26:20.840 --> 00:26:22.480
<v Michael Kennedy>because this is also gets interesting.

00:26:22.880 --> 00:26:25.600
<v Michael Kennedy>So if I want to do a get 0.9 microseconds,

00:26:26.140 --> 00:26:29.340
<v Michael Kennedy>that is less than one one thousandths of a millisecond.

00:26:29.740 --> 00:26:34.380
<v Michael Kennedy>I want to do a set, it's only 0.018 milliseconds.

00:26:34.880 --> 00:26:35.800
<v Michael Kennedy>But if you come down here,

00:26:35.980 --> 00:26:37.120
<v Michael Kennedy>and you look at free threading,

00:26:38.350 --> 00:26:39.140
<v Michael Kennedy>for four threads,

00:26:39.540 --> 00:26:42.280
<v Michael Kennedy>it does 1.3 million operations,

00:26:42.660 --> 00:26:43.500
<v Michael Kennedy>reads and writes a second.

00:26:43.880 --> 00:26:44.600
<v Calvin Hendryx-Parker>That's pretty incredible.

00:26:44.940 --> 00:26:45.160
<v Michael Kennedy>Yeah.

00:26:45.330 --> 00:26:46.100
<v Michael Kennedy>And this is SQLite.

00:26:46.100 --> 00:26:46.340
<v Michael Kennedy>Here we go.

00:26:46.840 --> 00:26:47.680
<v Calvin Hendryx-Parker>We're modern now.

00:26:47.860 --> 00:26:48.600
<v Michael Kennedy>And I'll tell you what,

00:26:48.760 --> 00:26:49.480
<v Michael Kennedy>this is not,

00:26:49.530 --> 00:26:49.660
<v Michael Kennedy>oh,

00:26:49.940 --> 00:26:50.140
<v Michael Kennedy>yeah,

00:26:50.320 --> 00:26:50.740
<v Michael Kennedy>it's in Rust,

00:26:50.840 --> 00:26:51.480
<v Michael Kennedy>so of course it's fast.

00:26:51.540 --> 00:26:52.960
<v Michael Kennedy>No, this is Python code

00:26:53.030 --> 00:26:56.020
<v Michael Kennedy>doing 1.3 million reads and writes a second.

00:26:56.280 --> 00:26:57.620
<v Michael Kennedy>It is so,

00:26:58.000 --> 00:26:58.440
<v Michael Kennedy>so cool.

00:26:59.600 --> 00:27:00.380
<v Michael Kennedy>What else can I say?

00:27:00.520 --> 00:27:02.000
<v Michael Kennedy>I guess it's super tested.

00:27:02.070 --> 00:27:02.440
<v Michael Kennedy>Like I said,

00:27:02.500 --> 00:27:04.560
<v Michael Kennedy>it's been running all the websites for a while,

00:27:04.680 --> 00:27:12.520
<v Michael Kennedy>But also I've been having, I had to reboot my mini, which is why I was messing with my camera at the beginning of my mini pro because I've had it do so much work.

00:27:12.640 --> 00:27:12.840
<v Michael Kennedy>It was,

00:27:13.380 --> 00:27:14.320
<v Michael Kennedy>it'll run like, Hey,

00:27:14.660 --> 00:27:16.600
<v Michael Kennedy>I'm going to do a big change to the way,

00:27:16.760 --> 00:27:19.420
<v Michael Kennedy>like the work stealing cue or whatever sort of works.

00:27:19.480 --> 00:27:20.700
<v Michael Kennedy>And so let's,

00:27:20.900 --> 00:27:21.560
<v Michael Kennedy>let's make that change.

00:27:21.600 --> 00:27:25.680
<v Michael Kennedy>But I also want to test that there's like no race conditions and deadlocks and leaks and so on.

00:27:25.940 --> 00:27:27.400
<v Michael Kennedy>So I'll run it for like 24 hours,

00:27:28.120 --> 00:27:33.160
<v Michael Kennedy>17 workers doing 150,000 requests a second each for a day.

00:27:34.100 --> 00:27:40.420
<v Michael Kennedy>And it'll have this chaos monkey in there, like killing off the processes periodically and like seeing what's going on.

00:27:40.450 --> 00:27:41.660
<v Michael Kennedy>And it's been taken.

00:27:41.660 --> 00:27:44.540
<v Michael Kennedy>I think I did about 1.5 trillion cash operations.

00:27:44.800 --> 00:27:45.720
<v Michael Kennedy>Probably destroyed my drive.

00:27:45.720 --> 00:27:46.760
<v Calvin Hendryx-Parker>That is a big number.

00:27:48.460 --> 00:27:51.260
<v Michael Kennedy>It's 1,500 billion operations.

00:27:51.700 --> 00:27:54.040
<v Calvin Hendryx-Parker>Yeah, I think people recognize how big those numbers are.

00:27:54.040 --> 00:27:55.180
<v Calvin Hendryx-Parker>We throw them around these days.

00:27:55.320 --> 00:27:56.280
<v Calvin Hendryx-Parker>That's a very big number.

00:27:56.540 --> 00:28:03.340
<v Michael Kennedy>One final thing, super interesting that this has that might be not obvious is this fan out cash.

00:28:03.780 --> 00:28:04.420
<v Michael Kennedy>So one of the,

00:28:04.880 --> 00:28:08.740
<v Michael Kennedy>you know, you've like, oh, I got SQLite, I have this problem, like where there's like this write contention or whatever.

00:28:09.020 --> 00:28:09.140
<v Calvin Hendryx-Parker>Yeah.

00:28:09.290 --> 00:28:12.720
<v Michael Kennedy>So what you can do is you can say, I want eight SQLite files,

00:28:13.130 --> 00:28:14.340
<v Michael Kennedy>and it's going to automatically,

00:28:14.680 --> 00:28:18.300
<v Michael Kennedy>given the key, it's going to use the hash and like a way to figure out where that goes.

00:28:18.300 --> 00:28:19.180
<v Calvin Hendryx-Parker>So you can shard.

00:28:19.560 --> 00:28:25.080
<v Michael Kennedy>Yeah, so you can shard into eight or more SQLite things using multi,

00:28:25.320 --> 00:28:26.380
<v Michael Kennedy>when you're doing multi-threading,

00:28:26.790 --> 00:28:32.520
<v Michael Kennedy>then each independent thread could potentially choose a different shard, in which case those are independent SQLite operations.

00:28:33.060 --> 00:28:33.880
<v Michael Kennedy>So super,

00:28:34.320 --> 00:28:34.720
<v Michael Kennedy>super cool.

00:28:35.060 --> 00:28:36.040
<v Michael Kennedy>Now I tried to run this.

00:28:36.500 --> 00:28:38.500
<v Michael Kennedy>Yeah, I tried to run this with the old one

00:28:38.840 --> 00:28:41.620
<v Michael Kennedy>and it just crashed because it like leaked handles

00:28:41.860 --> 00:28:42.800
<v Michael Kennedy>or couldn't deal with it.

00:28:42.800 --> 00:28:42.940
<v Calvin Hendryx-Parker>It's like, nope.

00:28:43.960 --> 00:28:44.800
<v Michael Kennedy>It was fine on one,

00:28:44.920 --> 00:28:46.040
<v Michael Kennedy>but as soon as you started to do threading,

00:28:46.080 --> 00:28:47.120
<v Michael Kennedy>it's like, nope, there's too many crash.

00:28:47.380 --> 00:28:49.100
<v Michael Kennedy>I'm like, yeah, no, this one doesn't crash.

00:28:49.320 --> 00:28:50.200
<v Michael Kennedy>It totally takes it.

00:28:50.340 --> 00:28:52.080
<v Michael Kennedy>So there's a lot of interesting things.

00:28:52.150 --> 00:28:53.580
<v Michael Kennedy>I don't think, oh, do I have a graph here?

00:28:53.640 --> 00:28:53.840
<v Michael Kennedy>Yeah,

00:28:54.260 --> 00:28:54.380
<v Michael Kennedy>again,

00:28:54.840 --> 00:28:57.180
<v Michael Kennedy>there's a lot of ops you can do here.

00:28:57.260 --> 00:28:57.900
<v Michael Kennedy>This is very fun.

00:28:58.120 --> 00:28:59.960
<v Michael Kennedy>So I'm really, really proud of this library.

00:29:00.160 --> 00:29:03.020
<v Michael Kennedy>I think it's going to be super useful

00:29:03.040 --> 00:29:03.400
<v Michael Kennedy>for people.

00:29:03.920 --> 00:29:04.240
<v Michael Kennedy>Check it out.

00:29:04.680 --> 00:29:05.760
<v Michael Kennedy>I'd love for you to try it.

00:29:05.920 --> 00:29:08.240
<v Calvin Hendryx-Parker>Is this full drop-in for disk cache ready,

00:29:08.520 --> 00:29:10.040
<v Calvin Hendryx-Parker>or do people need to make some modifications?

00:29:11.020 --> 00:29:12.080
<v Michael Kennedy>No, it's full drop-in.

00:29:12.240 --> 00:29:13.580
<v Michael Kennedy>One thing that it doesn't do

00:29:13.580 --> 00:29:17.160
<v Michael Kennedy>is it doesn't take your existing cache files.

00:29:17.340 --> 00:29:19.020
<v Michael Kennedy>So if you've got a cache that was hard to compute,

00:29:19.520 --> 00:29:24.960
<v Michael Kennedy>it's because I disagree with the philosophy of how to store the files. I don't want them as pickles,

00:29:25.160 --> 00:29:29.820
<v Michael Kennedy>and there's other things as well. It actually creates a separate folder path that it goes down.

00:29:30.380 --> 00:29:32.720
<v Michael Kennedy>So in that sense, you'd have to rebuild the cache.

00:29:33.390 --> 00:29:36.040
<v Michael Kennedy>But other than that, it's like 99% identical.

00:29:36.500 --> 00:29:39.240
<v Michael Kennedy>The one thing that's not there yet is the Django cache,

00:29:39.480 --> 00:29:40.120
<v Michael Kennedy>which is coming,

00:29:40.360 --> 00:29:40.500
<v Michael Kennedy>folks.

00:29:40.550 --> 00:29:41.980
<v Michael Kennedy>But give me a moment.

00:29:42.180 --> 00:29:43.000
<v Michael Kennedy>It's a lot of work.

00:29:43.010 --> 00:29:43.880
<v Michael Kennedy>It's taking me some time.

00:29:44.660 --> 00:29:46.880
<v Michael Kennedy>But I'm working on the Django cache and some other ideas.

00:29:47.320 --> 00:29:50.500
<v Michael Kennedy>I'm thinking that there's really interesting things, like large files,

00:29:50.960 --> 00:29:52.600
<v Michael Kennedy>kind of like almost an S3 type of thing,

00:29:53.010 --> 00:29:54.340
<v Michael Kennedy>which you don't necessarily want to put in the database,

00:29:54.570 --> 00:29:56.940
<v Michael Kennedy>but kind of like an aside storage of files.

00:29:57.160 --> 00:29:58.360
<v Michael Kennedy>I'm going to put them in the cache,

00:29:58.440 --> 00:29:59.660
<v Michael Kennedy>and when the thing expires in the cache,

00:29:59.770 --> 00:30:01.640
<v Michael Kennedy>it manages the files coming and going.

00:30:01.850 --> 00:30:03.220
<v Michael Kennedy>And I think those kinds of things

00:30:03.270 --> 00:30:04.400
<v Michael Kennedy>would be really fun to do as well.

00:30:04.460 --> 00:30:04.520
<v Calvin Hendryx-Parker>Yeah,

00:30:04.750 --> 00:30:06.000
<v Calvin Hendryx-Parker>and I see this like Mac and Linux.

00:30:07.100 --> 00:30:08.000
<v Calvin Hendryx-Parker>What's the holdup on Windows?

00:30:08.460 --> 00:30:08.980
<v Michael Kennedy>No, it's on Windows.

00:30:09.400 --> 00:30:09.980
<v Michael Kennedy>It is on Windows.

00:30:10.480 --> 00:30:10.580
<v Michael Kennedy>Yeah,

00:30:10.780 --> 00:30:12.520
<v Michael Kennedy>if my docs say that it's not on Windows,

00:30:12.720 --> 00:30:13.120
<v Michael Kennedy>it is.

00:30:13.760 --> 00:30:14.320
<v Michael Kennedy>It just takes,

00:30:14.580 --> 00:30:16.760
<v Michael Kennedy>these wheels take a long time to build

00:30:16.860 --> 00:30:20.680
<v Michael Kennedy>because the wheels are profile-guided optimized wheels.

00:30:21.200 --> 00:30:22.260
<v Michael Kennedy>So what I have is I have a whole profiling,

00:30:22.880 --> 00:30:24.780
<v Michael Kennedy>like a series of stuff that it runs through

00:30:24.850 --> 00:30:27.180
<v Michael Kennedy>and it'll first build them bare

00:30:27.520 --> 00:30:31.580
<v Michael Kennedy>and then it builds them with prof to make sure the test fast then it will build them with profiling

00:30:31.900 --> 00:30:36.380
<v Michael Kennedy>on take the results of the profiling and then rebuild them optimized with like that information

00:30:36.940 --> 00:30:41.540
<v Michael Kennedy>yeah and just doing that on windows was harder it's not super hard it's just more the automation

00:30:41.750 --> 00:30:47.860
<v Michael Kennedy>of it um so no i it's if you pip uv pip install or pip install it you'll get the windows wheels now

00:30:48.060 --> 00:30:52.700
<v Calvin Hendryx-Parker>as well it's very cool awesome michael like wow thanks like i can't wait to see what next week

00:30:52.720 --> 00:30:57.760
<v Calvin Hendryx-Parker>brings uh what what oxide project are we going to get next week i just you know i mean this graph is

00:30:57.820 --> 00:31:01.760
<v Michael Kennedy>sort of this first graph that's right on the home page it just tells like why why do we need

00:31:01.860 --> 00:31:06.520
<v Michael Kennedy>this like because the python space is not it's just not ready for three free threading and people

00:31:06.540 --> 00:31:12.980
<v Michael Kennedy>have not put an effort into it and this project is abandoned effectively right much much like

00:31:13.539 --> 00:31:17.600
<v Michael Kennedy>chameleon is not abandoned but it's just like it's no one is putting the time and energy go like we

00:31:17.760 --> 00:31:22.680
<v Michael Kennedy>really need to push this forward and the adoption of free threaded python is not going to work if

00:31:22.700 --> 00:31:24.700
<v Michael Kennedy>is the graph everyone gets. It's just like from

00:31:25.000 --> 00:31:26.640
<v Michael Kennedy>high numbers to just collapse if you

00:31:26.760 --> 00:31:28.660
<v Michael Kennedy>try to use it. Like, well, why would I do this? This is like way

00:31:28.820 --> 00:31:30.100
<v Michael Kennedy>more compute and it doesn't work.

00:31:30.280 --> 00:31:30.620
<v Michael Kennedy>Well, it

00:31:30.740 --> 00:31:31.040
<v Michael Kennedy>can work.

00:31:31.300 --> 00:31:32.700
<v Michael Kennedy>It just doesn't work unless you

00:31:33.420 --> 00:31:34.760
<v Michael Kennedy>specifically put some focus on

00:31:34.760 --> 00:31:35.160
<v Michael Kennedy>it, you know? Yeah.

00:31:35.360 --> 00:31:36.560
<v Calvin Hendryx-Parker>You're doing a good public service

00:31:36.720 --> 00:31:36.840
<v Calvin Hendryx-Parker>here.

00:31:37.020 --> 00:31:37.920
<v Michael Kennedy>Thank you. Thank you.

00:31:38.080 --> 00:31:38.600
<v Calvin Hendryx-Parker>All right.

00:31:39.100 --> 00:31:40.660
<v Calvin Hendryx-Parker>Well, let's head

00:31:40.740 --> 00:31:41.920
<v Calvin Hendryx-Parker>over into extras and see.

00:31:42.160 --> 00:31:42.760
<v Calvin Hendryx-Parker>I've got

00:31:42.760 --> 00:31:44.340
<v Calvin Hendryx-Parker>a couple extras for us.

00:31:44.980 --> 00:31:46.640
<v Calvin Hendryx-Parker>And one of them is the Django

00:31:46.940 --> 00:31:48.680
<v Calvin Hendryx-Parker>Software Foundation board

00:31:49.040 --> 00:31:50.820
<v Calvin Hendryx-Parker>nominations are open until December 2nd.

00:31:51.320 --> 00:31:52.660
<v Calvin Hendryx-Parker>There are five

00:31:52.680 --> 00:31:55.500
<v Calvin Hendryx-Parker>of the seven board positions open for election this year.

00:31:56.040 --> 00:31:58.780
<v Calvin Hendryx-Parker>Go sign up if you are interested in participating

00:31:58.860 --> 00:32:01.120
<v Calvin Hendryx-Parker>in the governance of an amazing open source project.

00:32:02.240 --> 00:32:02.660
<v Calvin Hendryx-Parker>Kaplan Moss,

00:32:02.820 --> 00:32:03.560
<v Calvin Hendryx-Parker>still involved in the project,

00:32:03.780 --> 00:32:04.940
<v Calvin Hendryx-Parker>especially from a governance standpoint,

00:32:05.520 --> 00:32:07.060
<v Calvin Hendryx-Parker>has been helping out getting this all going.

00:32:07.440 --> 00:32:09.780
<v Calvin Hendryx-Parker>And we wanna thank those who have spent the time

00:32:10.080 --> 00:32:11.360
<v Calvin Hendryx-Parker>and effort on the last round.

00:32:11.600 --> 00:32:13.540
<v Calvin Hendryx-Parker>So Jeff Triplett, Abigail Avogato,

00:32:14.260 --> 00:32:16.460
<v Calvin Hendryx-Parker>Paolo Melchua and Tom Carrick,

00:32:16.560 --> 00:32:17.180
<v Calvin Hendryx-Parker>the time, dedication,

00:32:17.560 --> 00:32:18.280
<v Calvin Hendryx-Parker>and efforts they've spent.

00:32:18.700 --> 00:32:21.760
<v Calvin Hendryx-Parker>So go join the Django Software Foundation board

00:32:21.780 --> 00:32:25.260
<v Calvin Hendryx-Parker>and participate in what happens next for that.

00:32:25.780 --> 00:32:28.180
<v Calvin Hendryx-Parker>The one last other extra I wanted to put in here

00:32:28.320 --> 00:32:29.760
<v Calvin Hendryx-Parker>is for those of you who don't know,

00:32:30.940 --> 00:32:34.940
<v Calvin Hendryx-Parker>one of the more prolific maintainers of Python has passed,

00:32:35.380 --> 00:32:35.840
<v Calvin Hendryx-Parker>Tim Peters,

00:32:36.180 --> 00:32:36.980
<v Calvin Hendryx-Parker>passed in July.

00:32:37.500 --> 00:32:38.500
<v Calvin Hendryx-Parker>And I wanted to remind people,

00:32:38.880 --> 00:32:41.220
<v Calvin Hendryx-Parker>this is one of my favorite Easter eggs

00:32:41.440 --> 00:32:42.140
<v Calvin Hendryx-Parker>in the whole language,

00:32:42.640 --> 00:32:45.700
<v Calvin Hendryx-Parker>is the fact the Zen of Python is inside of,

00:32:46.060 --> 00:32:47.520
<v Calvin Hendryx-Parker>and that was authored by Tim.

00:32:47.660 --> 00:32:50.140
<v Calvin Hendryx-Parker>And if you just go import this from a REPL

00:32:50.160 --> 00:32:52.280
<v Calvin Hendryx-Parker>or just do Python 3 minus C,

00:32:52.440 --> 00:32:53.080
<v Calvin Hendryx-Parker>import this,

00:32:53.540 --> 00:32:55.080
<v Calvin Hendryx-Parker>you will be able to read one of the things

00:32:55.110 --> 00:32:56.340
<v Calvin Hendryx-Parker>I refer to constantly,

00:32:56.830 --> 00:32:57.760
<v Calvin Hendryx-Parker>which is the Xenopython.

00:32:57.980 --> 00:33:01.960
<v Calvin Hendryx-Parker>So we're obviously sad to see that light extinguish.

00:33:02.360 --> 00:33:06.020
<v Calvin Hendryx-Parker>And he was a great contributor to the language itself too.

00:33:06.780 --> 00:33:07.500
<v Calvin Hendryx-Parker>But Xenopython,

00:33:07.660 --> 00:33:09.260
<v Calvin Hendryx-Parker>what a great legacy in my mind.

00:33:09.540 --> 00:33:10.940
<v Calvin Hendryx-Parker>And everyone should go do that.

00:33:11.080 --> 00:33:13.240
<v Calvin Hendryx-Parker>I tell people to print it out, frame it,

00:33:13.350 --> 00:33:14.620
<v Calvin Hendryx-Parker>and put it on their wall in the bathroom

00:33:14.670 --> 00:33:15.600
<v Calvin Hendryx-Parker>so they can read it every day

00:33:15.720 --> 00:33:16.300
<v Calvin Hendryx-Parker>when they're brushing their teeth.

00:33:17.200 --> 00:33:17.900
<v Calvin Hendryx-Parker>Words to live by.

00:33:18.140 --> 00:33:18.320
<v Michael Kennedy>Yeah,

00:33:18.830 --> 00:33:19.980
<v Michael Kennedy>the Xenopython is amazing.

00:33:20.080 --> 00:33:25.460
<v Michael Kennedy>and it's had a huge effect but honestly that's just a part right like it's just a small part i mean

00:33:25.820 --> 00:33:30.540
<v Calvin Hendryx-Parker>it's one i cherish because i when i was when i would teach people python that's where i'd start

00:33:30.900 --> 00:33:34.580
<v Calvin Hendryx-Parker>with that because there's so many great things there he's obviously contributed way bigger things

00:33:34.820 --> 00:33:39.780
<v Michael Kennedy>to the language itself well he will be missed thanks for for bringing that up on the show yeah

00:33:40.160 --> 00:33:44.640
<v Michael Kennedy>all right michael you got some extras i have some extras i will try to be quick because i have

00:33:44.660 --> 00:33:46.120
<v Michael Kennedy>Extras start page,

00:33:46.600 --> 00:33:49.680
<v Michael Kennedy>the search engine has always been nice, but you know,

00:33:49.840 --> 00:33:50.740
<v Michael Kennedy>I actually wrote an

00:33:50.810 --> 00:33:52.760
<v Michael Kennedy>article and complained about how like the whole front,

00:33:53.060 --> 00:33:55.280
<v Michael Kennedy>like everything above the fold is now ads,

00:33:55.380 --> 00:33:56.040
<v Michael Kennedy>which I'm like, this is,

00:33:56.440 --> 00:33:58.740
<v Michael Kennedy>but you could also put your money where the mouth is and say like,

00:33:58.820 --> 00:34:00.940
<v Michael Kennedy>I'd rather support the world than have ads.

00:34:01.160 --> 00:34:04.660
<v Michael Kennedy>And so they opened up a paid $4 a month for their

00:34:04.810 --> 00:34:06.000
<v Michael Kennedy>search tools with zero ads.

00:34:06.480 --> 00:34:08.820
<v Michael Kennedy>And you know what, to me, I thought that's, it's pretty neat. So

00:34:09.280 --> 00:34:11.340
<v Michael Kennedy>I don't know. I just thought I'd give that a quick shout out for them.

00:34:11.780 --> 00:34:11.960
<v Michael Kennedy>Also,

00:34:12.560 --> 00:34:13.200
<v Michael Kennedy>have you noticed my

00:34:13.220 --> 00:34:16.679
<v Calvin Hendryx-Parker>browser is a little different. I don't know if you've... I haven't, but I'm very curious. I saw

00:34:16.860 --> 00:34:21.100
<v Calvin Hendryx-Parker>this listed in the extra. I peeked at this one and I'm very curious. Look how tight and like how much

00:34:21.379 --> 00:34:22.419
<v Michael Kennedy>space I have on my screen,

00:34:22.600 --> 00:34:24.560
<v Michael Kennedy>on my little like streaming screen.

00:34:24.740 --> 00:34:24.980
<v Calvin Hendryx-Parker>True.

00:34:25.560 --> 00:34:26.659
<v Calvin Hendryx-Parker>And I've gone through a

00:34:26.659 --> 00:34:30.679
<v Calvin Hendryx-Parker>bunch of different browsers recently and someone has just informed, which is usually my children,

00:34:30.820 --> 00:34:32.940
<v Calvin Hendryx-Parker>that my most recent browser choice is still problematic.

00:34:33.899 --> 00:34:35.700
<v Michael Kennedy>You know what? I mean, yes,

00:34:35.919 --> 00:34:36.440
<v Michael Kennedy>but like

00:34:36.780 --> 00:34:38.399
<v Michael Kennedy>so is Firefox and some of its origins.

00:34:38.580 --> 00:34:39.520
<v Michael Kennedy>So like... Right.

00:34:40.360 --> 00:34:42.599
<v Michael Kennedy>There was somebody that wrote a parser that

00:34:42.580 --> 00:34:43.020
<v Michael Kennedy>was a jerk.

00:34:43.240 --> 00:34:43.659
<v Michael Kennedy>And so,

00:34:43.960 --> 00:34:44.260
<v Michael Kennedy>I mean,

00:34:44.780 --> 00:34:44.879
<v Michael Kennedy>yes,

00:34:45.040 --> 00:34:45.620
<v Michael Kennedy>but at the same time,

00:34:45.780 --> 00:34:45.860
<v Calvin Hendryx-Parker>right.

00:34:47.500 --> 00:34:49.020
<v Michael Kennedy>But I'm sorry, using helium,

00:34:49.540 --> 00:34:52.120
<v Michael Kennedy>which I've talked about before. And it's actually Python plus Chrome,

00:34:52.340 --> 00:34:53.460
<v Michael Kennedy>which is an interesting angle.

00:34:53.679 --> 00:34:57.520
<v Michael Kennedy>Like all of those sort of interactions is like, I believe Python and stuff, at least it was when

00:34:57.520 --> 00:34:57.820
<v Michael Kennedy>we talked.

00:34:58.080 --> 00:34:59.100
<v Michael Kennedy>So this is really neat.

00:34:59.260 --> 00:35:00.820
<v Michael Kennedy>It's privacy first, like brave.

00:35:01.100 --> 00:35:01.200
<v Michael Kennedy>Um,

00:35:01.880 --> 00:35:03.440
<v Michael Kennedy>and it has vertical tabs

00:35:03.480 --> 00:35:06.880
<v Michael Kennedy>on the left, or you can have this super condensed view. And this is like, at first, when I saw this,

00:35:06.880 --> 00:35:09.160
<v Michael Kennedy>I like, Oh, this would be great for streaming, but it's so hard to navigate.

00:35:09.540 --> 00:35:10.380
<v Michael Kennedy>I can just hit like

00:35:10.400 --> 00:35:15.420
<v Michael Kennedy>a button and then just i get a little drop down which is like vertical tabs but it's not you can

00:35:15.420 --> 00:35:19.520
<v Michael Kennedy>do auto hide there's even less than this calvin a hotkey to hide and close the vertical tabs on

00:35:19.520 --> 00:35:24.280
<v Calvin Hendryx-Parker>the side oh okay i'm sold yes i know i know that was your whole thing i am moving over today now

00:35:24.440 --> 00:35:28.760
<v Michael Kennedy>it's it's still a pretty new project so you know take that with a grain of salt and like all the

00:35:28.880 --> 00:35:34.300
<v Calvin Hendryx-Parker>things that you should consider but it is people who know me know i'm an early adopter i know exactly

00:35:34.320 --> 00:35:34.420
<v Michael Kennedy>Well,

00:35:35.000 --> 00:35:35.680
<v Michael Kennedy>welcome to the show.

00:35:35.720 --> 00:35:36.800
<v Michael Kennedy>You fit perfectly in, right?

00:35:37.220 --> 00:35:37.840
<v Michael Kennedy>That's what this is about.

00:35:38.120 --> 00:35:40.600
<v Michael Kennedy>So, but it is based on Chrome

00:35:41.080 --> 00:35:43.940
<v Michael Kennedy>and they promise to release Chrome updates frequently

00:35:44.260 --> 00:35:45.220
<v Michael Kennedy>and if there's any kind of issue.

00:35:45.320 --> 00:35:47.040
<v Michael Kennedy>So keep that in mind, but I'm enjoying it.

00:35:47.460 --> 00:35:47.680
<v Michael Kennedy>All right,

00:35:48.060 --> 00:35:49.240
<v Michael Kennedy>Tallyman, my little app

00:35:50.380 --> 00:35:54.260
<v Michael Kennedy>that lets you get graphs of like how much,

00:35:55.740 --> 00:35:57.800
<v Michael Kennedy>what type of code you're using in your project

00:35:58.100 --> 00:36:00.840
<v Michael Kennedy>and generate little shareable graphics

00:36:01.580 --> 00:36:02.860
<v Michael Kennedy>that you can put images and so on.

00:36:03.000 --> 00:36:04.020
<v Michael Kennedy>Got a major release.

00:36:04.580 --> 00:36:06.240
<v Michael Kennedy>So I went through it and did a whole bunch of changes.

00:36:06.420 --> 00:36:08.240
<v Michael Kennedy>So if you're using that, give it an update.

00:36:09.200 --> 00:36:11.760
<v Michael Kennedy>I want to just call out in a not positive way,

00:36:12.240 --> 00:36:12.760
<v Michael Kennedy>bunny.net,

00:36:13.020 --> 00:36:15.900
<v Michael Kennedy>which I love bunny.net for their CDN and their DNS and stuff.

00:36:16.300 --> 00:36:18.100
<v Michael Kennedy>But I had the crappiest experience with them.

00:36:18.320 --> 00:36:20.120
<v Michael Kennedy>So they have fonts.bunny.net,

00:36:20.400 --> 00:36:21.200
<v Michael Kennedy>which is an alternative,

00:36:21.380 --> 00:36:23.820
<v Michael Kennedy>an API compatible alternative to Google Fonts.

00:36:24.060 --> 00:36:24.540
<v Michael Kennedy>That's cool.

00:36:24.720 --> 00:36:25.680
<v Michael Kennedy>Privacy focused, cool.

00:36:26.200 --> 00:36:26.940
<v Michael Kennedy>You know what's not cool?

00:36:27.280 --> 00:36:29.540
<v Michael Kennedy>Is if you build your website with it four years ago

00:36:29.840 --> 00:36:31.620
<v Michael Kennedy>and you set a CSP policy

00:36:31.620 --> 00:36:33.900
<v Michael Kennedy>so people can't run crap on your website.

00:36:34.220 --> 00:36:34.460
<v Michael Kennedy>Like, hey,

00:36:34.620 --> 00:36:34.920
<v Michael Kennedy>security,

00:36:35.080 --> 00:36:35.700
<v Michael Kennedy>that is a thing,

00:36:35.920 --> 00:36:36.000
<v Michael Kennedy>right?

00:36:36.060 --> 00:36:36.720
<v Calvin Hendryx-Parker>It is definitely a thing.

00:36:37.060 --> 00:36:39.340
<v Michael Kennedy>Well, they just decided we're going to not run that anymore.

00:36:39.520 --> 00:36:41.780
<v Michael Kennedy>And so we're just serving everything out of JS Deliver,

00:36:41.980 --> 00:36:44.800
<v Michael Kennedy>which is no hurt on JS, no shade on JS Deliver.

00:36:45.320 --> 00:36:46.000
<v Michael Kennedy>But I was at the,

00:36:46.260 --> 00:36:46.720
<v Michael Kennedy>go ahead.

00:36:47.020 --> 00:36:49.120
<v Calvin Hendryx-Parker>Yeah, but no one loves a brown bag release.

00:36:49.680 --> 00:36:49.880
<v Michael Kennedy>Yeah.

00:36:50.220 --> 00:36:53.240
<v Michael Kennedy>So I was at the Apple store with my daughter helping us

00:36:53.680 --> 00:36:54.340
<v Michael Kennedy>get something set up.

00:36:54.680 --> 00:36:56.900
<v Michael Kennedy>And the iPhone Duo was at it.

00:36:56.920 --> 00:36:57.220
<v Michael Kennedy>Oh, cool.

00:36:57.320 --> 00:36:58.620
<v Michael Kennedy>Let me see what the website looks on that.

00:36:59.140 --> 00:37:00.180
<v Michael Kennedy>And I load it.

00:37:00.180 --> 00:37:01.060
<v Michael Kennedy>I'm like, that looks amazing.

00:37:01.240 --> 00:37:03.640
<v Michael Kennedy>I click one thing, it says too many requests. I'm like, what?

00:37:04.000 --> 00:37:06.060
<v Michael Kennedy>Because the CSP policy didn't allow

00:37:06.170 --> 00:37:11.320
<v Michael Kennedy>any of the fonts or any of the other things that needed. And so every page load was sending 1,600

00:37:11.640 --> 00:37:14.340
<v Michael Kennedy>CSP violations to my server. I'm like, oh my God,

00:37:14.540 --> 00:37:14.640
<v Michael Kennedy>dude.

00:37:15.240 --> 00:37:16.860
<v Michael Kennedy>And it stopped serving all the fonts. I'm

00:37:16.920 --> 00:37:18.340
<v Michael Kennedy>like, that is not a cool thing to do.

00:37:18.570 --> 00:37:19.400
<v Michael Kennedy>So if you,

00:37:19.880 --> 00:37:20.360
<v Michael Kennedy>long story short,

00:37:20.600 --> 00:37:22.280
<v Michael Kennedy>the extra is if you use bunny.net

00:37:22.480 --> 00:37:23.740
<v Michael Kennedy>fonts and you have a CSP policy,

00:37:23.950 --> 00:37:24.780
<v Michael Kennedy>you need to review this.

00:37:25.120 --> 00:37:25.380
<v Michael Kennedy>All right.

00:37:25.800 --> 00:37:26.560
<v Michael Kennedy>Really quick. Last thing.

00:37:26.880 --> 00:37:27.119
<v Michael Kennedy>Did I

00:37:27.060 --> 00:37:27.480
<v Michael Kennedy>say I'm on a kick?

00:37:28.000 --> 00:37:33.920
<v Michael Kennedy>I'm doing a members only Talk Python tech meetup. So please join the members,

00:37:34.160 --> 00:37:35.700
<v Michael Kennedy>come see the tech meetup or watch it after.

00:37:36.240 --> 00:37:38.300
<v Michael Kennedy>SQLite might actually be all you need.

00:37:38.460 --> 00:37:39.120
<v Michael Kennedy>So this,

00:37:39.580 --> 00:37:42.160
<v Michael Kennedy>this is not even on any of the SQLite stuff I've talked about.

00:37:42.560 --> 00:37:43.800
<v Michael Kennedy>This is a totally different and even

00:37:44.160 --> 00:37:45.820
<v Michael Kennedy>crazier SQLite story.

00:37:46.120 --> 00:37:48.360
<v Michael Kennedy>And it's going to be amazing December 2nd.

00:37:48.520 --> 00:37:49.660
<v Michael Kennedy>So it's in the show notes.

00:37:50.000 --> 00:37:50.400
<v Michael Kennedy>Hopefully

00:37:50.460 --> 00:37:52.460
<v Calvin Hendryx-Parker>you can be there. I thought attention was all I needed.

00:37:52.700 --> 00:37:53.180
<v Michael Kennedy>It is,

00:37:53.680 --> 00:37:54.800
<v Michael Kennedy>but then attention helps you get

00:37:54.780 --> 00:38:01.100
<v Michael Kennedy>sequel light perfect um you know hat tip to the um transformers original paper yeah that's not

00:38:01.190 --> 00:38:06.020
<v Calvin Hendryx-Parker>funny but you know what is what is funny two jokes some jokes you got two jokes yes i got two jokes

00:38:06.360 --> 00:38:09.820
<v Michael Kennedy>i've not peaked good i couldn't even decide i couldn't even decide something they're both going

00:38:09.920 --> 00:38:15.080
<v Michael Kennedy>in because what happened this week maybe last week this week i think it was last week all the things

00:38:15.340 --> 00:38:21.160
<v Michael Kennedy>happened last week has anyone tried to like i don't know check their outlook mail i'm kind of scared

00:38:21.180 --> 00:38:27.180
<v Calvin Hendryx-Parker>about the next friday the 13th this last friday the 13th was pretty bad yes i know i'm i am not

00:38:27.180 --> 00:38:32.140
<v Michael Kennedy>a fan of friday 13th let's don't go down that path so remember clippy yes yeah the little thing says

00:38:33.000 --> 00:38:38.120
<v Michael Kennedy>um the joke says decommissioned just means the one machine in the building nobody will

00:38:38.400 --> 00:38:44.580
<v Michael Kennedy>ever dare touch it's clippy pops up and says would you like to try powering up the decommissioned

00:38:44.740 --> 00:38:50.299
<v Michael Kennedy>exchange 2010 server that you never bothered to remove from the rack so that you can get your mail

00:38:50.320 --> 00:38:57.060
<v Calvin Hendryx-Parker>back this bit me this weekend i powered i powered off a piece of hardware i had in my little home

00:38:57.220 --> 00:39:00.520
<v Calvin Hendryx-Parker>rack that i'd forgotten about for like the last year that i thought i'd taken out of service

00:39:01.060 --> 00:39:05.520
<v Calvin Hendryx-Parker>sure enough stuff stopped working on the network things were talking to it yep it's they're in

00:39:05.520 --> 00:39:12.260
<v Michael Kennedy>there you don't know very real the other one also outlook based it has a picture of like earth rise

00:39:12.600 --> 00:39:17.760
<v Michael Kennedy>from the moon right yeah and it says the prophecy has been fulfilled i have two versions of outlook

00:39:18.140 --> 00:39:25.360
<v Calvin Hendryx-Parker>neither of them is working it has been foretold so are we all it's all behind us now done yeah

00:39:25.370 --> 00:39:29.160
<v Michael Kennedy>we're good to go it's all fine we've got what's done what's done is done it's done i think this

00:39:29.360 --> 00:39:32.940
<v Calvin Hendryx-Parker>episode's done what do you think i think so too thank you michael to you all next week