text
stringlengths
0
2.35k
[1531.38 --> 1532.70] I don't want to be forgetting anyone.
[1532.84 --> 1533.74] So if I forgot you.
[1533.80 --> 1535.50] No, I think we should spend time on this.
[1536.66 --> 1537.58] Don't forget anyone.
[1537.94 --> 1539.48] We should not spend time on this.
[1539.66 --> 1540.86] Mind you, you're just telling me a number.
[1540.86 --> 1545.54] Even if you forgot a Michael, they don't know which one they've, they don't know that they've
[1545.54 --> 1546.06] been forgotten.
[1546.28 --> 1547.12] It's just, there you go.
[1547.20 --> 1547.82] I think you're safe.
[1548.30 --> 1551.88] So everyone can assume that I included them in the list of Michaels.
[1552.10 --> 1552.42] Yeah.
[1552.56 --> 1554.22] So calm down, Michaels.
[1554.34 --> 1554.94] You were counted.
[1555.36 --> 1557.76] All Michaels have been accounted for.
[1559.50 --> 1559.90] Excellent.
[1559.90 --> 1564.38] I was thinking before we go on to the next topic, if anybody wants to read about this,
[1564.56 --> 1567.08] the issue number is 44167.
[1567.70 --> 1571.50] And at the end of the issue, which is very short, there's a link to the full proposal
[1571.50 --> 1573.28] design, which is very long.
[1573.54 --> 1576.34] And you can read that carefully and get the full picture.
[1576.72 --> 1576.88] Yeah.
[1577.00 --> 1578.54] It looks very well written.
[1578.88 --> 1582.62] And we'll post the link to all of these in the show notes.
[1582.68 --> 1585.58] So you'll be able to go and actually look at the original issues.
[1585.58 --> 1590.96] And honestly, like notice that some of these issues aren't created by members of the Go
[1590.96 --> 1596.24] team or even popular contributors like Daniel who've contributed massively.
[1596.86 --> 1600.26] Sometimes these come from just people in the community that have a problem that they want
[1600.26 --> 1601.54] to solve or something they care about.
[1601.86 --> 1606.20] So we do get stuck in basically, because you never know, you might get some improvements
[1606.20 --> 1607.94] made and that'd be great for everyone.
[1615.58 --> 1624.16] We are going to send three, two, one.
[1624.62 --> 1629.90] I'm Karhara Zhu, host of Ship It, a show with weekly episodes about getting your best ideas
[1629.90 --> 1631.88] into the world and seeing what happens.
[1632.22 --> 1637.28] We talk about code, ops, infrastructure, and the people that make it happen like charity
[1637.28 --> 1638.36] majors from Honeycomb.
[1638.70 --> 1641.64] We act like great engineers make great teams.
[1641.84 --> 1643.28] And it's exactly the opposite.
[1643.28 --> 1646.92] In fact, it is great teams that make great engineers.
[1647.58 --> 1650.82] And Dave Farley, one of the founders of Continuous Delivery.
[1651.18 --> 1653.98] Start off assuming that we're wrong rather than assuming that we're right.
[1654.24 --> 1655.22] Test our ideas.
[1655.36 --> 1656.86] Try and falsify our ideas.
[1657.00 --> 1658.98] Those are better ways of doing work.
[1659.04 --> 1661.28] And it doesn't really matter what work it is that you're doing.
[1661.42 --> 1663.10] That stuff just works better.
[1663.60 --> 1669.54] We even experiment on our own open source podcasting platform so that you can see how we implement
[1669.54 --> 1672.60] specific tools and services within changelog.com.
[1672.90 --> 1674.68] What works and what fails.
[1675.08 --> 1678.92] It's like there's a brand new hammer and we grab hold of it and everyone gathers around.
[1679.02 --> 1682.82] We put our hand out and we strike it right on our thumb.
[1683.06 --> 1685.90] And then everybody knows that hammer really hurts.
[1686.06 --> 1688.54] When you strike it on your thumb, I'm glad those guys did it.
[1688.60 --> 1689.42] I've learned something.
[1689.58 --> 1690.24] Instead, yeah.
[1690.24 --> 1694.98] I think that's a very interesting perspective, but I don't see that way.
[1695.16 --> 1695.38] Okay.
[1695.50 --> 1698.60] It's an amazing analogy, but I'm not sure that applies here.
[1698.96 --> 1701.24] Listen to an episode that seems interesting or helpful.
[1701.40 --> 1703.04] And if you like it, subscribe today.
[1703.16 --> 1704.30] We'd love to have you with us.
[1708.98 --> 1711.06] Does anyone have the M1 chip?
[1711.52 --> 1712.60] Apple's M1.
[1713.04 --> 1715.30] I have it on my personal laptop.
[1715.68 --> 1716.30] Yeah, that counts.
[1716.80 --> 1717.62] It's fast, isn't it?
[1717.82 --> 1718.78] Oh yeah, it's great.
[1718.78 --> 1721.64] I've been surprised with how fast it is.
[1721.96 --> 1722.28] Me too.
[1722.48 --> 1726.02] I got a new MacBook Pro recently and it's phenomenal.
[1726.24 --> 1726.64] Absolutely.
[1727.32 --> 1730.96] But Go had support for the M1 chip for quite a while, didn't it?
[1731.24 --> 1732.38] What does that look like?
[1732.44 --> 1734.08] How do we support another chip?
[1734.38 --> 1738.26] Could someone just briefly, and I do mean briefly, like we don't have to get into the
[1738.26 --> 1740.24] weeds of it, but what do we have to do?
[1740.30 --> 1745.38] Is it literally, we have to add some kind of mapping file for all the instructions so
[1745.38 --> 1748.18] that a compiler knows what to compile them into?
[1748.18 --> 1750.18] And it's different if it's a different chip?
[1750.56 --> 1752.86] Because there's also the Rosetta 2 stuff.
[1753.00 --> 1758.70] So that even if a binary on these new architectures hasn't been built for that architecture, this
[1758.70 --> 1759.76] is translation layer.
[1760.14 --> 1764.70] And to be honest, they're still lightning fast, like as far as I can see when I run programs
[1764.70 --> 1765.24] like that.
[1765.74 --> 1767.08] But there are some improvements coming.
[1767.14 --> 1767.58] Is that right?
[1767.58 --> 1773.42] I do seem to recall that when the M1 first came out, Go did already support ARM64.
[1773.74 --> 1776.16] So the 64 version of the ARM architecture.
[1776.50 --> 1782.58] But binary's build for Go targeting the architecture didn't work out of the box for one reason,
[1782.66 --> 1787.16] because there wasn't a Darwin slash ARM64 port yet.
[1787.16 --> 1791.92] So Go did support Mac, and it supported ARM64, but not together yet.
[1792.04 --> 1797.26] So they needed to add some glue code to essentially make those two work together.
[1797.60 --> 1802.20] And I think the other major work they had to do was the whole thing about signing binaries,
[1802.68 --> 1805.94] because I think the M1 was the first machine that required all binaries to be signed.
[1806.14 --> 1810.00] So they had to teach the linker how to sign binaries locally, something like that.
[1810.00 --> 1810.72] Yeah, yeah.
[1811.04 --> 1811.90] Oh, that's very cool.
[1812.14 --> 1813.68] Well, I just noticed it started working.
[1814.12 --> 1820.00] There's also a lot of work that needs to be done when we're signing binaries for,
[1820.28 --> 1824.42] when we're making releases, when Apple makes changes to their operating system,