News for hackers. Stuff that matters.
Re:Unix philosophy and filesystem access makes Claude Code a... by maleldil on Wednesday October 01, 2025 @01:42PM (#45440659)
I have a Just task that runs linters (ruff and pyright, in my case), formatter, tests and pre-commit hooks, and have Claude run it every time it thinks it's done with a change. It's good enough that when the checks pass, it's usually complete.
Re:Unix philosophy and filesystem access makes Claude Code a... by simlevesque on Wednesday October 01, 2025 @02:02PM (#45440937)
A tip for everyone doing this: pipe the linters' stdout to /dev/null to save on tokens.
Re:Unix philosophy and filesystem access makes Claude Code a... by maleldil on Wednesday October 01, 2025 @02:16PM (#45441161)
Why? The agent needs the error messages from the linters to know what to do.
Re:Unix philosophy and filesystem access makes Claude Code a... by jamie_ca on Wednesday October 01, 2025 @02:54PM (#45441723)
If you're running linters for formatting etc, just get the agent to run them on autocorrect and it doesn't need to know the status as urgently.
Re:Unix philosophy and filesystem access makes Claude Code a... by maleldil on Wednesday October 01, 2025 @03:49PM (#45442488)
That's just one part of it. I want the LLM to see type checking errors, failing test outputs, etc.
Re:Unix philosophy and filesystem access makes Claude Code a... by ozim on Wednesday October 01, 2025 @06:20PM (#45444253)
Errors shouldn’t be on stdout ;)
Re:Unix philosophy and filesystem access makes Claude Code a... by dymk on Wednesday October 01, 2025 @09:31PM (#45445533)
“Errors” printed by your linter aren’t errors, they’re reports
Re:Unix philosophy and filesystem access makes Claude Code a... by ozim on Friday October 03, 2025 @10:12AM (#45463244)
So if I want my CICD break on linter errors I have to implement parsing of stdout?

I'd rather have it breakout of the box by going to stderr.

Re:Unix philosophy and filesystem access makes Claude Code a... by joshstrange on Wednesday October 01, 2025 @03:22PM (#45442088)
This is the best way to approach it but if I had a dollar for each time Claude ran “—no-verify” on the git commits it was doing I’d have 10’s of dollars.

Doesn’t matter if you tell it multiple times in CLAUDE.md to not skip checks, it will eventually just skip them so it can commit. It’s infuriating.

I hope that as CC evolves there is a better way to tell/force the model to do things like that (linters, formatters, unit/e2e tests, etc).

Re:Unix philosophy and filesystem access makes Claude Code a... by 35mm on Wednesday October 01, 2025 @04:03PM (#45442692)
I’ve found the same issue and also with Rust sometimes skips tests if it thinks they’re taking too long to compile, and says it’s unnecessary because it knows they’ll pass.
Re:Unix philosophy and filesystem access makes Claude Code a... by pixl97 on Wednesday October 01, 2025 @09:00PM (#45445378)
Even AI understands it's Friday. Just push to to production and go home for the weekend.
Re:Unix philosophy and filesystem access makes Claude Code a... by mediaman on Wednesday October 01, 2025 @09:54PM (#45445658)
We should have a finish hook that, when the AI decides it's run, runs the hook, and gives it to the LLM, and it can decide whether the problem is still there.

Students don't get to choose whether to take the test, so why do we give AI the choice?

Re:Unix philosophy and filesystem access makes Claude Code a... by evertedsphere on Thursday October 02, 2025 @10:26AM (#45450052)
a wrapper script?
Re:Unix philosophy and filesystem access makes Claude Code a... by theshrike79 on Friday October 03, 2025 @06:40AM (#45461320)
(I code mostly in Go)

I have a `task build` command that runs linters, tests and builds the project. All the commands have verbosity tuned down to minimum to not waste context on useless crap.

Claude remembers to do it pretty well. I have it in my global CLAUDE.md sot I guess it has more weight? Dunno.