Repository navigation
release 0.0.9908 - #148
release 0.0.9908#148
Conversation
avocet-bot
left a comment
There was a problem hiding this comment.
Review — #148 "release 0.0.9908"
Reviewed head SHA: d99107e0178fc191df6543320501bc2845566939
Verdict: Approve — no blocking issues.
What this PR is
deploy-cli is the deno deploy command-line tool. Its version is declared in a single source of truth, deno.json (field version), and version.ts imports that JSON and re-exports it as VERSION, so the CLI reports the same number at runtime. Third-party dependencies are declared as JSR/npm specifiers in deno.json imports and pinned to exact resolved versions + integrity hashes in deno.lock.
This is a routine release PR: it cuts a new version and rolls in work that already landed on main via separate PRs. The diff touches only two files:
deno.json:version0.0.9907→0.0.9908;@deno/framework-detectspecifier^0.3.3→^0.3.4.deno.lock: the three framework-detect lines — the top-level specifier map (~0.3.3→~0.3.4), the package entry (@deno/framework-detect@0.3.3→@0.3.4, integrity6276…ddc8→a8ab…7649, transitive deps unchanged), and the workspace-member dep list (~0.3.3→~0.3.4).
The PR body credits already-merged #146 (deno deploy sends the app directory's deno.json on every deploy) and #147 (--build-timeout accepts a unit suffix like 10m/600s). Those changes are not in this diff — they merged earlier — so they are out of scope here; this diff is purely the version string plus the single JSR dependency bump (denoland/deployng#3774: workspace/app directories such as ./packages/a/ are now detected).
Verification performed
Reviewed at the live head SHA d99107e (matches headRefOid), so no checkout drift. The pr-review-toolkit:code-reviewer agent plus independent checks on a fresh clone of the PR head confirmed:
-
Specifier ↔ lock consistency — consistent, no stale pin.
deno.jsonreads^0.3.4;deno.lockresolves0.3.4in all three locations. The lock keys read~0.3.4whiledeno.jsonreads^0.3.4; this is Deno's canonical normalization, not drift — for a0.x.yversion^0.3.4and~0.3.4denote the identical range (>=0.3.4 <0.4.0), and the same^(json)/~(lock) pairing existed for0.3.3before this PR.grepoverdeno.lock/deno.jsonon the PR head finds no residual0.3.3/~0.3.3/^0.3.3reference. -
Version bumped consistently — single source of truth.
deno.json:7is the only occurrence of the release version anywhere in the repo (grepfor9907/9908returns only that line).version.tsimportsdeno.jsondynamically rather than hard-coding, andrs_lib/Cargo.tomlcarries an independent0.0.0unrelated to the CLI release version. -
Integrity hash — verified against JSR.
sha256ofjsr.io/@deno/framework-detect/0.3.4_meta.jsonequalsa8abfde986605990b2729ba778421d65209acdd122174d79fe2e54b487c97649, an exact match for the lock. 0.3.4 is the currentlateston JSR. The 0.3.4 dependency graph (jsr:@std/jsonc,jsr:@std/path@^1.1.2) is identical to 0.3.3's, so the unchangeddependenciesarray in the lock is correct and both transitive packages are already resolved — no new/missing lock entry. -
No correctness/regression/security risk from this diff. A patch-level bump of one JSR dependency plus the version string. PR CI (
jsr,deno) is green.
Findings
None. This is a correct, low-risk release PR. The only thing that superficially looks like a discrepancy — ^ in deno.json vs ~ in deno.lock — is expected Deno canonicalization, not lockfile drift.
Review skill evidence
Invoked the pr-review-toolkit:code-reviewer agent (via the Task/Agent tool) to perform the dependency/lockfile/version verification, supplemented by direct gh/git/grep checks on a fresh clone of the PR head and a JSR _meta.json integrity-hash check.
Bumps the version in
deno.jsonto0.0.9908. Includes:deno deploysends the app directory's deno.json on everydeploy, so its
deploysection applies on every deploy rather thanonly at
create--build-timeoutaccepts a unit suffix (10m,600s)@deno/framework-detect0.3.3 -> 0.3.4 (denoland/deployng#3774):workspace and app directories like
./packages/a/are found