Search
112 results for “develwithoutacause”
-
Ok, think I figured this out. A few things going on here.
1. I wasn't rebuilding dependencies with #Turbo correctly.
2. Local #PDS instances are limited to `.test` and `.example` by default.
3. Browsers create TIDs from `@atproto/common-web` (not the more widely-documented `@atproto/common`).I still got some weird "invalid handle character" errors last night and a couple times today, but after resolving the `serviceHandleDomains` that seems to have gone away, so I'm a bit confused what happened with that one.
-
@develwithoutacause @eliajf #Jest is definitely the industry standard, and has a lot of backing behind it - great interactive CLI and works great with tools like #testing-library and #jsdom - would definitely recommend checking that option out!
-
I'm also just now realizing the game doesn't search your inputs for the solution as a subpath, meaning you need to wait a couple seconds to reset and try again.
That might explain why my inputs seem less than consistent.
-
I've been playing #Tunic and really enjoying it overall, but in the late game, these Holy Cross puzzles are really getting on my nerves.
I'm like 5 puzzles in a row now where I solved it conceptually but just didn't read the path the same way the game intended and end up looking up each one.
-
I've been playing #Tunic and really enjoying it overall, but in the late game, these Holy Cross puzzles are really getting on my nerves.
I'm like 5 puzzles in a row now where I solved it conceptually but just didn't read the path the same way the game intended and end up looking up each one.
-
RE: https://social.treehouse.systems/@jnkrtech/116519264248279742
Did you know? #JavaScript recently became #Baseline Widely Available.
You can now use it without needing progressive enhancement!
-
#Idea: Ban an #AI agent from writing a file with arbitrary text.
Instead, force it to use a minimal set of tools for the specific task it is given. This can potentially create a more deterministic migration / transformation of source code while still using AI as a decision maker.
Ultimately the agent could always execute `echo "..." > file.txt`, and it's hard to prevent that while still allowing it the flexibility it wants.
I'm just wondering if constraining the agent in the right way would potentially increase consistency while leveraging it as a decision engine.
-
#rules_prerender tries to go a step further with partial prerendering at build time (reducing rendering work on the server) and driving *bundling* by what gets rendered, not just how resources are loaded on the client. We can eagerly load all required JS and CSS up front.
-
Cool to see #MDN exploring a unique #frontend architecture which better meets their needs.
https://developer.mozilla.org/en-US/blog/mdn-front-end-deep-dive/
A lot of the core ideas here are stuff I've been trying to formalize in #HydroActive and #rules_prerender. Islands of reactivity on a mostly static page and driving resource loading by what gets rendered are both key features most FE systems lack.
-
-
Exploring an #AI workflow for automatically fixing flaky tests, and having reasonable success with it.
https://github.com/angular/angular/pull/68026
#Bazel is actually a particularly nice fit with features like `--runs_per_test`. I expect any Bazel #web workspace could probably use this prompt mostly as-is.
-
RE: https://hostux.social/@defcon201/116324484796527679
Quick #story time: I was on a team which accidentally shipped sourcemaps to production!
The production server would serve sourcemaps for local dev, and we built sourcemaps in prod to deobfuscate errors but they weren't provided to the server in the prod environment.
You might see the hazard there. An innocuous change to our #Bazel build structure accidentally pulled in sourcemaps into the server, which it then helpfully served.
This was an app in the #Google monorepo too, so we were revealing all our source code as well as all our dependencies!
-
Free #startup #idea: A #GitHub bot which daily/weekly runs all the tests in your repo, looks for flakey targets, and then creates #AI generated fixes to improve stability.
#Nrwl #Nx self-healing CI is the closest existing example I can think of, but maybe there's something else out there I'm not aware of.
-
Free #startup #idea: A #GitHub bot which daily/weekly runs all the tests in your repo, looks for flakey targets, and then creates #AI generated fixes to improve stability.
#Nrwl #Nx self-healing CI is the closest existing example I can think of, but maybe there's something else out there I'm not aware of.
-
Free #startup #idea: A #GitHub bot which daily/weekly runs all the tests in your repo, looks for flakey targets, and then creates #AI generated fixes to improve stability.
#Nrwl #Nx self-healing CI is the closest existing example I can think of, but maybe there's something else out there I'm not aware of.
-
Free #startup #idea: A #GitHub bot which daily/weekly runs all the tests in your repo, looks for flakey targets, and then creates #AI generated fixes to improve stability.
#Nrwl #Nx self-healing CI is the closest existing example I can think of, but maybe there's something else out there I'm not aware of.
-
Free #startup #idea: A #GitHub bot which daily/weekly runs all the tests in your repo, looks for flakey targets, and then creates #AI generated fixes to improve stability.
#Nrwl #Nx self-healing CI is the closest existing example I can think of, but maybe there's something else out there I'm not aware of.
-
I remain convinced that the #Git CLI is a terrible UX and an especially bad fit for AI agents. Evidence:
```
GIT_SEQUENCE_EDITOR="python3 -c \"import sys; p=sys.argv[1]; c=open(p).read(); open(p,'w').write(c.replace('pick 29507ade4d6', 'edit 29507ade4d6'))\"" \
git rebase -i 29507ade4d6^
```I only asked it to take some local changes, break them up, and amend them to some existing individual commits and it is thrashing *wildly* but I'm scared to abort or it'll leave me in a potentially worse state.
I'm been playing with #Jiujitsu on the side, I wonder if that might work better?
-
-
I was just asking #Gemini about #BlueSky video hosting limitations and if I could work around them by self-hosting a #PDS, and it pointed me at "Bluesky's video.bsky.app transcoding sidecar" which is apparently... a Rick Roll?
Did I just get Rick Rolled by my #AI?
What's even more baffling is that it's not an arbitrary link. Someone at BlueSky had to make video.bsky.app redirect to that specific video. Someone got paid to do this.
-
Whoo! Took down Consort (actually hitless!) and wrapped my RL1 All Remembrances run.
Had a lot of fun with this and I'm glad I did it, but I think now it's time to play something I haven't beaten multiple times already.
-
Whoo! Took down Consort (actually hitless!) and wrapped my RL1 All Remembrances run.
Had a lot of fun with this and I'm glad I did it, but I think now it's time to play something I haven't beaten multiple times already.
-
Whoo! Took down Consort (actually hitless!) and wrapped my RL1 All Remembrances run.
Had a lot of fun with this and I'm glad I did it, but I think now it's time to play something I haven't beaten multiple times already.