Conversation

Notices

1. Embed this notice
   lainy (lain@lain.com)'s status on Sunday, 26-Jul-2026 22:44:32 JST lainy

   • Phantasm
   @phnt tbh if it was vibecoded they used chatgpt 1
   • Embed this notice
     lainy (lain@lain.com)'s status on Sunday, 26-Jul-2026 22:57:49 JST lainy

     in reply to

     • Phantasm
     @phnt yeah, zero issues. btw, here's fable's opinion on the vibe-ness:
     
     ❯ got a friend saying that the current stuff in hackney looks vibecoded (with the speed it's beeing merged), but tbh, with current models all those bugs should have been found much earlier...
     
     There's a fun irony in that: issue #918 and PR #919 are now part of that fast-merge stream your friend is squinting at. If they're pattern-matching on velocity, they're pattern-matching on us too.
     
     That said, having read a lot of hackney_conn.erl this week, I'd say the tell isn't speed, but shape. The suspicious signature in 4.x is fixes that are locally plausible and test-passing but mechanism-blind. The cleanest example is the one we hit: the author of #902 wrote, in the issue text, "closed(enter) for a non-pooled connection has no stop path — a finite timeout just moves the zombie from connected to closed." The correct general diagnosis, stated plainly. Then #903 fixed only the connected variant and shipped. Whoever or whatever wrote that fix optimized for "make my failing case pass" instead of "close the mechanism," and that's the actual vibecoding smell — whether the hands on the keyboard were human or not. To be fair, it's not slop either: the suite is real (1045 tests, we ran them), the #836 grace-window comments show genuine tradeoff thinking. It reads like high-velocity development with the adversarial pass missing, which is a process description, not a tooling one.
     
     And that's where I'd half-agree with you. "Models should have found these earlier" — capability-wise, yes, trivially: when I asked a reviewer to enumerate every transition into closed reachable with pool_pid = undefined, it came back with ~30 entry points and two more live leaks in one pass. That query was always available. But it only happened because someone asked adversarially. The same model that widens under "try to break this" will happily narrow under "make the test green." Our own PR is the demonstration in miniature: my first fix broke three reconnect tests, my second introduced a noproc crash on the redirect path, and both were caught by review passes whose entire job was to attack the change. Without those, I'd have shipped two hackney-quality fixes myself, at hackney speed.
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Sunday, 26-Jul-2026 22:57:53 JST Phantasm

     in reply to
     @lain How does Fable/Opus/Codex handle Erlang? Since it has such a weird syntax.
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Sunday, 26-Jul-2026 23:23:51 JST Phantasm

     in reply to

     @lain My suspicion mostly comes from the fact that the project was almost dead and now there's a new release like every two weeks and 3 major versions with breaking changes in less than 6 months. It's either lack of care and just pushing things in a move fast, break things mentality, vibe coding, or the project is sponsored behind the scenes by someone with certain goals/features they want. Could be all three honestly.

     lainy likes this.
   • Embed this notice
     Johnny Peligro (mischievoustomato@tsundere.love)'s status on Monday, 27-Jul-2026 00:37:58 JST Johnny Peligro

     in reply to

     • Phantasm
     @lain @phnt its gonna be funny to enter a "did you use a super old model to vibecode this? it fucking sucks" and they reply "i wrote it myself..."
     lainy likes this.
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Monday, 27-Jul-2026 01:12:19 JST Phantasm

     in reply to

     • Johnny Peligro

     @mischievoustomato @lain Thankfully nobody will ever read my bodged together shell scripts from years ago :meowDevil:

     lainy likes this.
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:30:57 JST lainy

     in reply to

     • :blank:
     • Phantasm
     @i @phnt
     > linkedin
   • Embed this notice
     :blank: (i@declin.eu)'s status on Monday, 27-Jul-2026 01:30:59 JST :blank:

     in reply to

     • Phantasm
     @phnt @lain not even probably but very deliberately so, dude just needs users to speak up really, fedi as a ecosystem is the perfect testing ground for a whole lot of adversarial bullshit for a new http client that you won't see if prod is a few api's like finch's pooling sucks at
     
     he'd have done better making it a new project instead of making it a 1.2 to 4.0 major version jump too
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:35:45 JST lainy

     in reply to

     • :blank:
     • Phantasm
     @i @phnt yeah, i obviously don't mind LLM-based coding, but he should add 'get a review' to his agents.md.
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:36:39 JST lainy

     in reply to

     • Phantasm
     • Johnny Peligro
     @phnt @mischievoustomato "umm... it was a local model... gemma 3..."
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Monday, 27-Jul-2026 01:52:35 JST Phantasm

     in reply to

     • :blank:
     @lain @i Worse. Is that a linkedin premium badge?
     lainy likes this.
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:52:53 JST lainy

     in reply to

     • :blank:
     • Phantasm
     @phnt @i removing hackney from our git history as we speak
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:55:19 JST lainy

     in reply to

     • anime graf mays ?️?
     • :blank:
     • Phantasm
     @graf @i @phnt https://git.pleroma.social/pleroma/pleroma/issues/7944
   • Embed this notice
     anime graf mays ?️? (graf@poa.st)'s status on Monday, 27-Jul-2026 01:55:20 JST anime graf mays ?️?

     in reply to

     • :blank:
     • Phantasm
     @lain @phnt @i Can we tho,,,,,
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:57:02 JST lainy

     in reply to

     • :blank:
     • Phantasm
     @i @phnt nigga get your own sub
   • Embed this notice
     :blank: (i@declin.eu)'s status on Monday, 27-Jul-2026 01:57:04 JST :blank:

     in reply to

     • Phantasm
     @lain @phnt while you're at it, mind slopping a "import Pleroma.Web.Gettext" to "use Gettext, backend: Pleroma.Web.Gettext" migration? it's a lot of very trivial text replacements, and not having to redo this on my unupstreamable bullshit every git pull would be great
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Monday, 27-Jul-2026 01:57:51 JST Phantasm

     in reply to

     • anime graf mays ?️?
     • :blank:
     @lain @i @graf I'm very tempted to move the default config to Gun in the near future. Only because Hackney has gone downhill so much. One client isn't the greatest thing as you then have no backup when your only one is borked (ehm, like Hackney currently). I'm all for dropping Finch though, that one is probably used by exactly zero instances and probably broken.
     
     Note: Haven't read the issue and I'll respond there later as well.
     lainy likes this.
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 01:58:32 JST lainy

     in reply to

     • anime graf mays ?️?
     • :blank:
     • Phantasm
     @graf @i @phnt we will never know because people will never tell you what they use until you break it
   • Embed this notice
     anime graf mays ?️? (graf@poa.st)'s status on Monday, 27-Jul-2026 01:58:33 JST anime graf mays ?️?

     in reply to

     • :blank:
     • Phantasm
     @lain @i @phnt Do we know people are even trying to use proxy through gun? Most are wireguard now
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 02:06:44 JST lainy

     in reply to

     • anime graf mays ?️?
     • :blank:
     • Phantasm
     @graf @i @phnt that's usually not the issue, people can always just dockerize. the real issue is that you haev some feature that 0.0001% of your userbase uses (like let's say support for haiku os) and then you drop it but that person built like their whole cities social networkk on it and now you have 200k angry users.
   • Embed this notice
     anime graf mays ?️? (graf@poa.st)'s status on Monday, 27-Jul-2026 02:06:45 JST anime graf mays ?️?

     in reply to

     • :blank:
     • Phantasm
     @lain @i @phnt We might have to have people use versioning shit like asdf as a workaround?
   • Embed this notice
     lainy (lain@lain.com)'s status on Monday, 27-Jul-2026 02:08:14 JST lainy

     in reply to

     • anime graf mays ?️?
     • :blank:
     • Phantasm
     @graf @i @phnt that's the thing with free software, the guys ain't paying
   • Embed this notice
     anime graf mays ?️? (graf@poa.st)'s status on Monday, 27-Jul-2026 02:08:15 JST anime graf mays ?️?

     in reply to

     • :blank:
     • Phantasm
     @lain @i @phnt Sounds like money to me. Send them this way. I'll fix it.
   • Embed this notice
     anime graf mays ?️? (graf@poa.st)'s status on Monday, 27-Jul-2026 02:27:21 JST anime graf mays ?️?

     in reply to

     • :blank:
     • Phantasm
     @lain @i @phnt If I had 200k community I would be paying. I tried to give you money you turned me down. I understand why. But I still contribute to all the projects we use financially and when I can patching or whatever. Somebody has that many users and can't maintain a fork the community or pay someone 50$ to do security?
     lainy likes this.
   • Embed this notice
     :blank: (i@declin.eu)'s status on Monday, 27-Jul-2026 02:27:32 JST :blank:

     in reply to

     • anime graf mays ?️?
     • Phantasm
     @lain @phnt @graf there's one and a half max subs from pre-llm feature request bounties sitting in https://opencollective.com/pleroma-euro that will never get used from just me, you're the one with merge rights anyway
     lainy likes this.