so he does all of this because he dislikes transpiling because supposedly it makes debugging etc harder? does he know about sourcemaps?
he also says he doesnt like type information in his code, so he adda them as comments. the type information is still there, how is this an improvement?
Assembled instructions aren’t even the lowest non-hardware stage in instruction execution. There’s proprietary microcode sitting a level below your typical x86 ISA.
And even then, what if—God forbid—the hardware has errata. A line has to be drawn somewhere between trusting that what you write is logically correct at all stages below it. If someone is unable to trust that the environment they wrote code for works, they better start learning how to create PCBs and writing for FPGAs.
Unless someone is using some language extensions, transpiling from TS to an ECMAScript module using the ESNext target merely drops the type annotations.
If not running the exact same code being developed is an issue, it’s an easy fix.
I do this daily and believe me when I say that I’d trade my kidney for the ability to use TS natively. This looks good on paper but jsdoc notation has lots of flaws and you literally can’t do some things with it. Also, it doesn’t check if the function actually does the thing you described so it needs manual review every time it’s changed.
I haven’t dealt with a larger JS/Node Project in a while, but I like this approach to using TS features in JS.
so he does all of this because he dislikes transpiling because supposedly it makes debugging etc harder? does he know about sourcemaps?
he also says he doesnt like type information in his code, so he adda them as comments. the type information is still there, how is this an improvement?
The issue with transpiling is that the code that’s running in production is not necessarily the one that’s been tested. A source map doesn’t fix that.
I loathe this line of reasoning. It’s like saying “unless you wrote assembly, compiling your code could change what it does.”
Guess what, the CPU reorders/ellides assembly, too! You can’t trust anything!
Haha, what is this, the 90s?
Assembled instructions aren’t even the lowest non-hardware stage in instruction execution. There’s proprietary microcode sitting a level below your typical x86 ISA.
And even then, what if—God forbid—the hardware has errata. A line has to be drawn somewhere between trusting that what you write is logically correct at all stages below it. If someone is unable to trust that the environment they wrote code for works, they better start learning how to create PCBs and writing for FPGAs.
🙈🙉🙊
I know, but I didn’t want to scare the children.
I also chose to pretend it’s just little gnomes moving the bytes around. Less magic.
What are electrons, but a miserable pile of little magic gnomes? But enough talk… have a upvote!
Unless someone is using some language extensions, transpiling from TS to an ECMAScript module using the
ESNexttarget merely drops the type annotations.If not running the exact same code being developed is an issue, it’s an easy fix.
Ugh? Why shouldn’t it be the same code?
Because Browsers can’t run Typescript, they run JavaScript. That’s why the intermediate conversion step isneededd.
But your tests are running on the compiled code too. Nothing can be tested but handwritten assembly, with such approach
That approach was mentioned in the video.
I do this daily and believe me when I say that I’d trade my kidney for the ability to use TS natively. This looks good on paper but jsdoc notation has lots of flaws and you literally can’t do some things with it. Also, it doesn’t check if the function actually does the thing you described so it needs manual review every time it’s changed.