Zsh Mailing List Archive
Messages sorted by:
Reverse Date,
Date,
Thread,
Author
Re: [PATCH 0/1] compaudit: add ZSH_COMPLETION_USERS parameter
- X-seq: zsh-workers 54884
- From: dana <dana@xxxxxxx>
- To: "Petros Amoiridis" <petros@xxxxxxxxxxxx>
- Cc: zsh-workers@xxxxxxx
- Subject: Re: [PATCH 0/1] compaudit: add ZSH_COMPLETION_USERS parameter
- Date: Tue, 30 Jun 2026 10:29:44 -0500
- Archived-at: <https://zsh.org/workers/54884>
- Feedback-id: i9be146f9:Fastmail
- In-reply-to: <20260630132020.75264-1-petros@workbrew.com>
- List-id: <zsh-workers.zsh.org>
- References: <20260630132020.75264-1-petros@workbrew.com>
On Tue 30 Jun 2026, at 08:20, Petros Amoiridis wrote:
> Trust stays opt-in.
was this written by an llm?
i have a patch to compinit/compaudit that i've been sitting on which
adds a -t option ('trust' i guess) to specify user ids or names to
append to _i_owners (you don't need to pre-resolve anything, the u
glob qualifier supports names directly)
this was mostly intended to address users complaining about compaudit
errors when they run zsh via sudo. so you could do this in zshrc (if it
fits your threat model, obv there's still some risk here):
if [[ -n $SUDO_USER$DOAS_USER ]]; then
compinit -Dt $SUDO_USER:$DOAS_USER
else
compinit
fi
but it would work more generally too
i never submitted it, though, because oliver thinks compaudit is
pointless in the first place. he seems to feel pretty strongly about
that, maybe we could form a consensus on it
dana
Messages sorted by:
Reverse Date,
Date,
Thread,
Author