Reply desk
Check it, tap Reply, then tap Post in X.
Opened today: 34/10Posted: 46Got a response: 0 (0%) vs 15% baselineModel spend, 24h: $1.31
I work harder than an ugly stripper
a work joke from someone Joe has history with
hope youre getting paid for all that
@nisten · 3h · 215 views · 4 repliestopicadd PSA for those of u vibecoding naked on your laptop with no containers: tell your agents to DELETE Playwright NOW. Even the latest (1.64.1) bundles early Chromium.156 with
CVE-2026-118306
CVE-2026-118331
CVE-2026-118348
CVE-2026-118344
And the Playwright SANDBOX IS OFF BY DEFAULT
container mounts can still expose host files, a separate risk from the browser sandbox
a writable home mount still lets compromised code change host files
id keep host credentials out of the container mounts too
i wouldn't recommend using cloudflare dns as your database (looking at you
@QuinnyPig
), but you should read about how we've scaled it to handle 3x the writes (and all the other benefits we got along the way)
cloudflare just runs at a different scale
The write increase raises a concrete question about DNS update propagation
did propagation time change with the extra write capacity?
was the bottleneck storing changes or pushing them out?
@feross · 4h · 444 views · 1 repliestopicadd On Oct 21, a malicious litellm 1.107.0 hit PyPI and LiteLLM's Docker images. A .pth file ran the stealer on every Python start, no import needed. The GitHub tag was clean, so source review showed nothing. Provenance only helps if something checks it.
Python startup execution can also affect attempts to remove the malicious package
if pip uses that environment, uninstalling can run the stealer again
even cleanup needs a separate python environment if the installed .pth is still there
This repo has 1,455 packages and 27,323 tasks in it.
It just completed CI for my PR in 1m26s.
Work smarter, not harder.
The image shows 500 completed tasks; cache hits would help explain the runtime
how long does that run take with a cold cache?
@valigo · 4h · 2,045 views · 12 repliestopicask I'm gonna crash out. What do you MEAN kotlin can't compile even 1000 lines per second on M5 Pro? What do you MEAN I have to keep kotlinc warm and it takes gigabytes of memory and it's still just under 5k lines a second?
Android people are you ok???
incremental rebuild time would show how much this slows everyday development
is that a full rebuild or an incremental compile?
how long does a one-line change take to rebuild?
@chronark · 3h · 136 views · 0 repliestopicask just ask for performance and there shall be performance
the timing breakdown leaves about 8 seconds outside the subtests
whats taking the other 8 seconds outside the subtests?
is most of the remaining time setup or cleanup?
@_Felipe · 6h · 185 views · 0 repliestopicadd You can checkout the same "branch" on different worktrees. But you have to checkout the commit reference explicitly because two worktrees tracking the same branch would have to mutate in tandem. That would be contradictory to what a worktree is: a disconnected copy.
shared branch refs move, but the other worktree's files and index don't update automatically
the ref moves in both, but the other worktree's files dont
shared refs, separate indexes. committing in one would leave the other out of sync
Integrated the new release process fully into Coolify. The stack of 8 PRs with 48 commits, 3,058 additions and 6,636 deletions is now merged.
Releasing to all of you with v4.6.0.
The update-channel change raises a practical question about existing installations
can existing instances switch update channels without reinstalling?
does the new release process handle rollbacks too?
@isbkch · 8h · 325 views · 4 repliestopicadd two coding agents in one checkout will eat each other's changes.
git worktree add ../app-auth -b auth
git worktree add ../app-billing -b billing
one folder per agent, same repo, separate branches.
the catch: each folder needs its own node_modules, .env and dev port.
worktrees isolate files but agents can still collide through a shared development database
id give each agent its own db too if they're running migrations
separate ports wont isolate database writes if both envs point at the same db