Our engineers are fucking morons. And this guy was the dumbest of the bunch. If you think Netflix hires top tier talent, you don't know Netflix.
The trick is to use my massive brain to root cause several significant outages, discover that most of them originate in code written by the same employee, and notice that said employee liked to write things like
// @ts-nocheck
// obligatory disabling of typescript: static typing is hard, so why bother with it?
async function upsertWidget() {
try {
// await api.doSomeOtherThings({ ... })
// 20 line block of commented-out useless code
// pretend this went on much longer
let result = await api.createWidget({ a, b, c })
if (!result.ok) {
result = await api.createWidget({ a, b }) // retries for days! except with different args, how fun
if (!result.ok) {
result = await api.updateWidget({ a, b, c }) // oh wait, this time we're updating
}
}
// notice that api.updateWidget() can fail silently
// also, the three function calls can each return different data, I sure am glad we disabled typescript
return result
} catch (error) {
return error // I sure do love this pattern of returning errors and then not checking whether the result was an error or the intended object
}
}
function doSomething() {
const widget = await upsertWidget()
}
I am so tired of these people. I am not a 10x rockstar engineer and not without flaws, but they are just so awful and draining, and they never seem to get caught in time to stop them ruining perfectly fine companies. Every try>catch>return is like an icy cat hand from the grave reaching up to knock my coffee off my desk.
So again, maybe they're a bad employee but it seems like nothing's done to even try and minimize the risks they present.
At that point it's not one person being obnoxious and never approving their team members diffs and more of a team effort to do so.
But at minimum if you have a culture of trying to improve your codebase you'll inevitably set up tests, ci/cd with checks, etc. before any code can be deployed. Which should really take any weight of responsibility out of any one member of the team. Whether trying to put out code or reject said code.
Which, yes, does raise interesting questions about how someone who can't be trusted to handle errors in an API call ended up in a senior enough position to do that.