it feels like you’re wrestling with given building blocks, trying to change the way the act, instead of writing the instructions for the program to simply execute

when you need to build a modular system and define map configuration on the fly, web-components make things more complicated.

it’s intended to be used in the way, where you would preconfigure an item (map) in the separate portal gui and pass it to a component as an id prop, but in a power user case this approach isn’t viable.

however, in that particular case of ESRI’s map components library. prior to the switch you had granular control over configurations, could define order of events. now with components you simply wait for them to be mounted, only to then extract and change properties

i’m speaking from the perspective if only one library, which i spent a lot of time building with. that has to be sad, in general i’m all about web standards.

with switch to web components you lose so much control. declarative approach isn’t a silver bullet in all of the cases

Does it mean that nobody gets to keep their watch this year? Or fun run finishers are good to go this time?

what a legend!

John Kelly is the GOAT I'm rooting for right now, no offense Killian

@maplibre.org expects geojson to be EPSG:4326, only to transform it to WebMercator (3857) … 🪦

nginx (?)

at first you learn some new prompt or a technique, try it against your task, get all excited. and come back some time later just to learn that in reality it didn’t do much. and you still have to figure out everything the old way just by yourself

okay, figured it out, it’s an issue with a private browsing mode. not sure if it is a bug or feature. other bowsers’ private modes work fine

and it lines up perfectly with a new drop. gnx might be kendrick's king of limbs. it lines up in my book

`.flow > * + *`! pretty sure i first heard of it from someone on @web-standards.ru, then traced it back to @heydonworks.bsky.social and @bell.bz. been a convert ever since