Say publish = false on the internal shared crates
Thirteen crates in shared/ were publishable-by-default and registered in no
pipeline, so nothing but habit stood between them and crates.io. mnw-assumptions
already said it; the rest now match. tagtree carries its own note because it is
PolyForm Noncommercial, where an accidental publish would offer code on terms
crates.io's audience cannot take.
Author: Max Johnson <me@maxj.phd> · 2026-08-16 22:49 UTC
# Internal shared crate, consumed by path inside this repo and never published.
7
+
# `publish = false` is what makes that a cargo-enforced fact rather than a
8
+
# convention: without it an accidental `cargo publish` is one command away.
9
+
publish = false
6
10
description = "OTA self-update for egui/eframe desktop apps: poll a manifest, verify a minisign signature, download and apply. (Placeholder — implementation in progress.)"
# Internal shared crate, consumed by path inside this repo and never published.
7
+
# `publish = false` is what makes that a cargo-enforced fact rather than a
8
+
# convention: without it an accidental `cargo publish` is one command away.
9
+
publish = false
6
10
description = "The artifact and evidence contract between the system that builds a release and the system that advances it. Names what bytes exist, what source they came from, and what has been proven about them."
# Internal shared crate, consumed by path inside this repo and never published.
7
+
# `publish = false` is what makes that a cargo-enforced fact rather than a
8
+
# convention: without it an accidental `cargo publish` is one command away.
9
+
publish = false
6
10
description = "Shared operator-tool infrastructure: streaming remote-exec, a generic event bus, a disk+broadcast live-log sink, and sqlite helpers. Used by sando-daemon and bento-daemon."