Re: wifi-tui , a new package for FreeBSD 15.1
Date: Mon, 17 Aug 2026 06:13:30 UTC
On Sun, Aug 16, 2026 at 11:03 PM Sulev-Madis Silber < freebsd-current-freebsd-org111@ketas.si.pri.ee> wrote: > so right now i used chatgpt to create myself platform independent multiple > pin gpio interrupt watcher with trailing edge debounce > > i opted to that approach because i couldn't do it myself > > computer did this in 3 seconds. with sources being entire fbsd source tree > with code and manpages, entire mailing list archive and forum content > > iirc it's banned from bugzilla with that anime girl block so it can't use > anything from there > > i mean not even the brightest brain can do that so fast > > google and plain grep -r src and man didn't get me that far quick > > now i have new problem, i can't review it properly because c is hard > > i bet someone can write this straight out of memory or have it sitting on > disk next to electronics project. i'm happy to use that instead > > i probably would rewrite it later myself when i actually understand it > > then i would want to put it somewhere where others can see it. ports. > tree. my blog. whatever > > so that whoever comes and wants to connect some physical switch to freebsd > machine have something ready that (s)he can just piggyback shell or perl on > and be done with it. perhaps later writing their own code > > eh i think i have lpt version of this around. was also hard to create. > that i did with trial and error, unable to imagine what those bits did in io > So I just wanted to point out that the Anubis fronting bugs.freebsd.org is for AI crawlers that are heavily hitting our servers. We banned them because we lack the resources to service them. Code created with LLMs, assuming it meets the project's quality standards and you can explain it in review, is not per-se banned at this time. However, before submitting any large work, it's best to reach out to the relevant people or maintainers to coordinate the effort. LLMs also tend to be terrible at creating the right set of commits (not impossible, just takes a lot of effort). If you'd like a review, it might be better to submit a github pull request and tag me. I'll do my best to assess it if you'd like and think it's ready. Warner > > > On August 17, 2026 6:44:53 AM GMT+03:00, Mario Marietto < > marietto2008@gmail.com> wrote: > >>it's not cruel imo, an application like this needs to be tested > >extensively and have mandatory large set >of unit tests, which themselves > >seems to be vibe coded as well,i am not fully against vibe coding but i > >>realise the amount of care that needs to go in reviewing a vibe coded > >project properly is very high, so >whoever is going to review it needs to > >be aware of that burden > > > > > >Do you have some ideas about how many useful tools these days people > create > >using vibe coding ? A lot of them are good ideas,they help users,they > >deserve more attention and review. Houston we have a problem. With the > >birth of the AI everything has changed and a solution should be found > >to support > >progress (when this brings new possibilities to improve users' lives). I?m > >not against you. You have your good reasons for opting out, but 99% of > >developers do it like you, and that's not good. I myself created a large > >project, a Linux VM using Qemu accelerated by bhyve, months ago, and no > >serious developer has shown interest so far, and I don't think they will. > I > >repeat. Houston, we have a problem. If we want lots of useful, working > code > >to enter FreeBSD and improve it, we need a solution. I don't know what it > >is, though; I'm not so experienced. > > > >On Mon, Aug 17, 2026 at 5:31 AM <polyduekes@proton.me> wrote: > > > >> it's not cruel imo, an application like this needs to be tested > >> extensively and have mandatory large set of unit tests, which themselves > >> seems to be vibe coded as well,i am not fully against vibe coding but i > >> realise the amount of care that needs to go in reviewing a vibe coded > >> project properly is very high, so whoever is going to review it needs > to be > >> aware of that burden > >> > >> > >> -------- Original Message -------- > >> On Monday, 08/17/26 at 08:52 Warner Losh <imp@bsdimp.com> wrote: > >> > >> > >> > >> On Sun, Aug 16, 2026, 8:42 PM Enji Cooper (yaneurabeya) < > >> yaneurabeya@gmail.com> wrote: > >> > >>> > >>> On Aug 16, 2026, at 7:31 PM, developers@atlanexis.com wrote: > >>> > >>> On 2026-08-17 02:55, Enji Cooper (yaneurabeya) wrote: > >>> > >>> On Aug 16, 2026, at 6:46 PM, developers@atlanexis.com wrote: > >>> Good Morning everyone > >>> I'm working on a new package to ease the configuration of WIFI for > >>> FreeBSD, and my current computer doesn't support FreeBSD 15.1 , in the > >>> dmesg the iwlwifi is loading without issue , but after login to the > system > >>> it is not working, have anyone faced this issue previously > >>> > >>> Was this project vibe coded? If so, I have an hourly rate for making > vibe > >>> coded projects work. > >>> Thank you, > >>> -Enji > >>> > >>> > >>> Greetings Enji, > >>> > >>> Yes this was actually coded with Claude Agent > >>> > >>> > >>> Hi, > >>> I was half-joking about the vibe coding hourly rate early, but in > >>> reality, I wasn’t joking at all: I (personally) don’t have the time to > help > >>> someone debug and get a vibe coded application up and running for free. > >>> Maybe someone else does, but I wouldn’t necessarily expect that. > >>> > >> > >> Do you really need to announce that to the world? Seems like a needless, > >> cruel dig. > >> > >> Warner > >> > >> Cheers, > >>> -Enji > >>> > >>> PS Have you seen https://github.com/shazow/wifitui ? Looks like this > >>> project is trying to accomplish what you’re trying to do with golang. > It > >>> already has working Linux and macOS support, so adding FreeBSD support > >>> (hopefully) will achieve whatever you were setting out to do. > >>> > >> > > > >