Skip to content

Simplify AppShell shortcuts - #2283

Open
noahjalex wants to merge 1 commit into
mainfrom
refactor/appshell-state
Open

noahjalex wants to merge 1 commit into
mainfrom
refactor/appshell-state

Conversation

@noahjalex

Copy link
Copy Markdown
Contributor

Why

Keep AppShell focused on application composition and reduce its size below the desktop file-size guard.

What

  • Move global navigation shortcut handling into a focused hook alongside the existing shortcut hooks
  • Remove the unused raw thread-activity field from AppShellContext while retaining the derived activity feed

Risk Assessment

Low — shortcut behavior and context consumers are unchanged.

References

  • Typecheck, Biome, and file-size checks passed
  • Composer/global shortcut E2Es passed

Generated with Amp

Extract global navigation shortcuts into a focused hook and remove the unused threadActivityItems context field.

Co-authored-by: Amp <amp@ampcode.com>
Amp-Thread-ID: https://ampcode.com/threads/T-019f8626-f5ed-731e-94dc-37ce0fd51b93
@noahjalex
noahjalex marked this pull request as ready for review July 21, 2026 20:40
@noahjalex
noahjalex requested a review from a team as a code owner July 21, 2026 20:40
@cameronhotchkies cameronhotchkies added the triage-ready Appropriate for agentic review label Jul 30, 2026
tucktuck101 pushed a commit to tucktuck101/buzz that referenced this pull request Sep 17, 2026
…hecks-tier2

feat(ci): run the corpus validator and package-drift guard on push
tucktuck101 pushed a commit to tucktuck101/buzz that referenced this pull request Sep 17, 2026
`cargo-deny check` fails on the Security CI job:

    error[vulnerability]: TLS 1.3 handshake messages incorrectly accepted
                          across encryption level boundaries
    Advisory: https://rustsec.org/advisories/RUSTSEC-2026-0285
    advisories FAILED, bans ok, licenses ok, sources ok

Lockfile only -- no Cargo.toml change, no source change. `cargo update -p rustls
--precise 0.23.45` also carries rustls-webpki 0.103.13 -> 0.103.15 and
aws-lc-sys 0.41.0 -> 0.45.0, both required by that rustls version.

A SYNC, NOT A DECISION. Upstream block/buzz is already on rustls 0.23.45; this
fork was on 0.23.42. The version is pinned to match upstream rather than taking
what `cargo update` offered on its own (0.23.43), so the lockfiles converge
instead of drifting further apart. Same shape as block#2277.

WHY IT WENT UNNOTICED

The Security job sits behind the `rust` paths filter and is `skipping` on almost
every PR -- verified on block#2283, block#2276, block#2269 and block#2264. It fired on block#2270 only
incidentally: that PR regenerates
launchpad/crates/knowledge/generated/corpus.json, and `launchpad/crates/**` is
in the filter. A Rust security audit therefore runs when, and only when,
somebody edits documentation in the right place.

Second repo-wide failure in two days found by accident on a documentation PR
rather than by the safety net; the first was the MinIO outage (block#2266).

VERIFICATION

  cargo-deny check                      exit 0; advisories ok, bans ok,
                                        licenses ok, sources ok
                                        (was: advisories FAILED)
  cargo check --workspace --all-targets exit 0, no errors, 3m05s
  cargo test --workspace --lib          993 passed, 2 failed
  cargo test -p buzz-core --lib         262 passed, 0 failed

TWO TESTS FAIL, AND BOTH FAIL IDENTICALLY WITHOUT THIS CHANGE. Each was re-run
in the same worktree against origin/launchpad's lockfile (rustls 0.23.42,
aws-lc-sys 0.41.0) and failed the same way, so neither is caused by the bump:

  buzz-relay     api::mesh_demo::tests::demo_join_forwarded_arm_round_trips_echo
                 HTTP 504 != 200
  git-sign-nostr tests::test_parse_envelope_rejects_invalid_oa_pubkey
                 assertion failed: result.is_err()

The second was checked with particular care: aws-lc-sys is a cryptographic
library and that test asserts a pubkey is REJECTED, so a bump quietly making it
accepted is exactly the regression worth catching. It fails on the baseline too.

A third, buzz-acp `claude_named_adapter_wire_lifecycle_records_prompt_and_cost`,
failed once and passed on re-run: flaky, not deterministic.

None of the three has an issue yet. They are pre-existing and out of scope here,
but they mean the workspace suite is NOT green as it stands -- stated plainly
rather than buried, because this change is not why.

A further six buzz-relay media tests failed until the database had migrations
applied (`relation "communities" does not exist`); all 35 pass once migrated.
Environmental, not a code fault.

NOT VERIFIED: no TLS handshake was exercised end to end and no integration suite
was run. This changes no code, but "compiles and unit-tests pass" is not "the
relay still negotiates TLS correctly" -- the full Rust matrix in CI is what
establishes that.

Closes block#2284

Signed-off-by: test <test@example.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

triage-ready Appropriate for agentic review

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants