#javascriptwasamistake — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #javascriptwasamistake, aggregated by home.social.
-
First it was cool to target just Node, even if TypeScript added some complexity to the bundling process! Worth it!
But then there arrived Deno, which was TypeScript-first and came with a ton of tools built in...but required rewriting to stop using Node-specific data types, to include extensions on every relative import across the entire code base, then finding and referencing deps by HTTPS addresses, and then finding and incorporating another build process for two of the three monorepo packages to maintain the ability to deploy to NPM. Okay, fine I guess...
But wait! Announcing "Web API runtimes" like Bun and Cloudflare workers, and Vercel, requiring test harnesses to try and test future releases in.
It's cool, though: everything still fetches packages from NPM!
Except now there's deno.land/x that can host packages for Deno?! Fine, I'll support that too for the packages that make sense to run server-side.
Except wait there are Deno frameworks with SSR support that can load browser packages from deno.land/x too! But my third, browser-centric package not rewritten for Deno is now suffering from its lack of Deno support!
Maybe rewriting these packages to deploy through JSR.io might eliminate the TypeScript build step, support Node, Deno, and everything else, and make this browser package available in Deno projects?
Well Deno 2.0 did just drop, with workspaces support to boot, so if I'm going to do this, do I also refactor my monorepo to use Deno's workspaces support instead of this working combination of pnpm and Lerna and NX? Except the VS Code Deno extension's "format on save" breaks when you enables workspaces...
...Oh, don't mind me. Just some reminiscing from the OSS maintainer trenches 🥹
#javascript #typescript #deno #node #bun #npm #jsr #javascriptwasamistake
-
First it was cool to target just Node, even if TypeScript added some complexity to the bundling process! Worth it!
But then there arrived Deno, which was TypeScript-first and came with a ton of tools built in...but required rewriting to stop using Node-specific data types, to include extensions on every relative import across the entire code base, then finding and referencing deps by HTTPS addresses, and then finding and incorporating another build process for two of the three monorepo packages to maintain the ability to deploy to NPM. Okay, fine I guess...
But wait! Announcing "Web API runtimes" like Bun and Cloudflare workers, and Vercel, requiring test harnesses to try and test future releases in.
It's cool, though: everything still fetches packages from NPM!
Except now there's deno.land/x that can host packages for Deno?! Fine, I'll support that too for the packages that make sense to run server-side.
Except wait there are Deno frameworks with SSR support that can load browser packages from deno.land/x too! But my third, browser-centric package not rewritten for Deno is now suffering from its lack of Deno support!
Maybe rewriting these packages to deploy through JSR.io might eliminate the TypeScript build step, support Node, Deno, and everything else, and make this browser package available in Deno projects?
Well Deno 2.0 did just drop, with workspaces support to boot, so if I'm going to do this, do I also refactor my monorepo to use Deno's workspaces support instead of this working combination of pnpm and Lerna and NX? Except the VS Code Deno extension's "format on save" breaks when you enables workspaces...
...Oh, don't mind me. Just some reminiscing from the OSS maintainer trenches 🥹
#javascript #typescript #deno #node #bun #npm #jsr #javascriptwasamistake
-
First it was cool to target just Node, even if TypeScript added some complexity to the bundling process! Worth it!
But then there arrived Deno, which was TypeScript-first and came with a ton of tools built in...but required rewriting to stop using Node-specific data types, to include extensions on every relative import across the entire code base, then finding and referencing deps by HTTPS addresses, and then finding and incorporating another build process for two of the three monorepo packages to maintain the ability to deploy to NPM. Okay, fine I guess...
But wait! Announcing "Web API runtimes" like Bun and Cloudflare workers, and Vercel, requiring test harnesses to try and test future releases in.
It's cool, though: everything still fetches packages from NPM!
Except now there's deno.land/x that can host packages for Deno?! Fine, I'll support that too for the packages that make sense to run server-side.
Except wait there are Deno frameworks with SSR support that can load browser packages from deno.land/x too! But my third, browser-centric package not rewritten for Deno is now suffering from its lack of Deno support!
Maybe rewriting these packages to deploy through JSR.io might eliminate the TypeScript build step, support Node, Deno, and everything else, and make this browser package available in Deno projects?
Well Deno 2.0 did just drop, with workspaces support to boot, so if I'm going to do this, do I also refactor my monorepo to use Deno's workspaces support instead of this working combination of pnpm and Lerna and NX? Except the VS Code Deno extension's "format on save" breaks when you enables workspaces...
...Oh, don't mind me. Just some reminiscing from the OSS maintainer trenches 🥹
#javascript #typescript #deno #node #bun #npm #jsr #javascriptwasamistake
-
First it was cool to target just Node, even if TypeScript added some complexity to the bundling process! Worth it!
But then there arrived Deno, which was TypeScript-first and came with a ton of tools built in...but required rewriting to stop using Node-specific data types, to include extensions on every relative import across the entire code base, then finding and referencing deps by HTTPS addresses, and then finding and incorporating another build process for two of the three monorepo packages to maintain the ability to deploy to NPM. Okay, fine I guess...
But wait! Announcing "Web API runtimes" like Bun and Cloudflare workers, and Vercel, requiring test harnesses to try and test future releases in.
It's cool, though: everything still fetches packages from NPM!
Except now there's deno.land/x that can host packages for Deno?! Fine, I'll support that too for the packages that make sense to run server-side.
Except wait there are Deno frameworks with SSR support that can load browser packages from deno.land/x too! But my third, browser-centric package not rewritten for Deno is now suffering from its lack of Deno support!
Maybe rewriting these packages to deploy through JSR.io might eliminate the TypeScript build step, support Node, Deno, and everything else, and make this browser package available in Deno projects?
Well Deno 2.0 did just drop, with workspaces support to boot, so if I'm going to do this, do I also refactor my monorepo to use Deno's workspaces support instead of this working combination of pnpm and Lerna and NX? Except the VS Code Deno extension's "format on save" breaks when you enables workspaces...
...Oh, don't mind me. Just some reminiscing from the OSS maintainer trenches 🥹
#javascript #typescript #deno #node #bun #npm #jsr #javascriptwasamistake
-
First it was cool to target just Node, even if TypeScript added some complexity to the bundling process! Worth it!
But then there arrived Deno, which was TypeScript-first and came with a ton of tools built in...but required rewriting to stop using Node-specific data types, to include extensions on every relative import across the entire code base, then finding and referencing deps by HTTPS addresses, and then finding and incorporating another build process for two of the three monorepo packages to maintain the ability to deploy to NPM. Okay, fine I guess...
But wait! Announcing "Web API runtimes" like Bun and Cloudflare workers, and Vercel, requiring test harnesses to try and test future releases in.
It's cool, though: everything still fetches packages from NPM!
Except now there's deno.land/x that can host packages for Deno?! Fine, I'll support that too for the packages that make sense to run server-side.
Except wait there are Deno frameworks with SSR support that can load browser packages from deno.land/x too! But my third, browser-centric package not rewritten for Deno is now suffering from its lack of Deno support!
Maybe rewriting these packages to deploy through JSR.io might eliminate the TypeScript build step, support Node, Deno, and everything else, and make this browser package available in Deno projects?
Well Deno 2.0 did just drop, with workspaces support to boot, so if I'm going to do this, do I also refactor my monorepo to use Deno's workspaces support instead of this working combination of pnpm and Lerna and NX? Except the VS Code Deno extension's "format on save" breaks when you enables workspaces...
...Oh, don't mind me. Just some reminiscing from the OSS maintainer trenches 🥹
#javascript #typescript #deno #node #bun #npm #jsr #javascriptwasamistake
