nathants

45 ms·

631.
by nathants 4y ago
using: xinput disable $id. code here: https://gist.github.com/nathants/e6b98ef1bb996492195a0cfa15b...
632.
Show HN: Adhoc remote execution in AWS Lambda(github.com)
2 points by nathants 4y ago|0 comments
633.
by nathants 4y ago
yes, but i have used this in onsite work as well. i grep for a few critical apps, and avoid locking if any are open. screen recording, video call, etc. lock starts as soon as they exit.
634.
by nathants 4y ago
physically moving around is the trick. standing or sitting perfectly still for long periods is terrible for mind and body. i have a python program on my laptop that is typically running 35 minute sessions. then it locks my computer, requiri
635.
by nathants 4y ago
can confirm kagi is very good.
636.
by nathants 4y ago
the last good ideas in js land were react and google’s closure compiler. reagent[1], a react wrapper, has had a stable and tiny api for a decade. easy, fun, and effective sdlc[2] on mature technology is easier now than ever. there are no mo
637.
Show HN: Fullstack web should be easy(github.com)
5 points by nathants 4y ago|0 comments
638.
by nathants 4y ago
whether a site is js heavy, spa, or txt served by nginx doesn’t matter. it doesn’t matter at all. what matters is how the site feels to use on a mobile browser over cellular. there are only two options: - good - annoying
639.
by nathants 4y ago
github exposes an api to render markdown the same way it renders readmes. i use that to render markdown to html with a tiny scaffold: https://github.com/nathants/render
640.
Show HN: Easy and fun browser testing from Python with pyqt6-webengine(github.com)
1 points by nathants 4y ago|0 comments
641.
by nathants 4y ago
everything is fine. people need to chill. if you don’t have a good sdlc[1], build one. if you’re in an environment where that is impossible, change environments. there’s never been a better time than right now to build software of all sorts
642.
by nathants 4y ago
everything is fine. i’ve gone full spa[1]. i don’t think i could go back. my starter project is here[2]. backend is go on lambda. frontend is clojurescript with react and mui, compiled down to a single html file with js inlined, and pre-gzi
643.
by nathants 4y ago
i’m not. i do my own thing[1] with the go sdk. 1. https://github.com/nathants/libaws
644.
by nathants 4y ago
vpc and ec2 are the big ones. they both support tagging atomically during create though, so it’s fine? for ec2 i’m not sure i’d want unique names per instance. typically i prefer groups of instances with the same name.
645.
by nathants 4y ago
this is the ideal approach for most scenarios. fewer knobs, simpler tooling.
646.
by nathants 4y ago
aws is absolutely the worst featured and least stable cloud provider. except for all the others. sometimes, being least bad is a great accomplishment.
647.
by nathants 4y ago
one still must test on multiple machines before being confident it doesn’t just work on one’s machine. cloud is no different. we must test.
648.
by nathants 4y ago
something that can help reduce the need for state is aggressively partitioning systems into separate aws accounts. then you can KNOW that no other random infrastructure should exist in an account. terraform definitely does help coordinate t
649.
by nathants 4y ago
using names to uniquely identify infrastructure is very convenient and how most of aws already works. why not take advantage of that instead mapping name to uuid via state? in what scenario are non-stable infrastructure names useful?
650.
by nathants 4y ago
deletion is more complicated than creation. what if it’s stateful infrastructure like a db, and contains important state? is it already backed up? would it be expensive to restore? what else aren’t we thinking of. even in a stateful model,
651.
by nathants 4y ago
amen! as another commenter here rightly put: state or idempotency, pick one. hello? how is this even a question? i have a hard time believing that idempotent infrastructure mutations are not the right move almost always. shameless plug[1],
652.
by nathants 4y ago
the basics are: - atom is a compare and swap data structure, its lets you mutate state - in your atom you hold a hashmap, deeply nested, that contains ALL state describing your app - your react components reference that state with: (:some-h
653.
by nathants 4y ago
tbh how can one NOT use react? it’s insanely simple. is this not the way? https://reagent-project.github.io/ api basically unchanged since 2013 btw.
654.
by nathants 4y ago
cdk is for great good, and i’m glad it’s putting the spring back in your step! a rising tide raises all ships. i can’t wait until i’m embarrassed by how slow and annoying my aws sdlc is compared to sota. then i’ll have to do better. embarra
655.
by nathants 4y ago
to be fair, deeply understanding the aws primitives, their apis, and their relationships, is a lot more work than slapping together some cf json. while i enjoy it, and value the experience of it, i realize not everyone feels the same or has
656.
by nathants 4y ago
extremely well said. bravo. i don’t enjoy cdk/cf, and want to explore further, but i’m glad cf exists. if i were to use one of the big 4 iac, it would be cf. it’s not that much worse than the others, and it’s documented and used in the
657.
by nathants 4y ago
no startup should start on aws unless they have years of expertise with it. unless they can tape over all the knobs they don’t need with little notes that say: peel off when your on the moon. this is what i’ve done here. i believe there sho
658.
by nathants 4y ago
aws absolutely should be easy, and needn’t sacrifice any capability to do so. i can choose to use none of apigateway but wiring up a domain to a lambda. i haven’t removed any of its other functionality, i’ve just chosen not to use it. it wi
659.
by nathants 4y ago
it’s certainly possible this was the original intent, or just an accident and they leaned into it. we’ll never know. either way the essential core of primitives is down there, just wrapped in multiple levels of nutshell. the sre excellence
660.
by nathants 4y ago
i disagree. i believe that aws is hard because it provides the feature set that is the union of the needs of all customers, and that the customers are massive company inc. inside of that bloated feature set, there exists an essential core o

More ›