chore(release): 0.0.1-alpha.26 (magic_notifications ^0.2.0)

[anilcancakir]

Cuts alpha.26, carrying the delete-affordance work merged in #121 and the callback answer merged in #122.

Stays on the alpha rail. Nothing here is a public API break in this package: _confirmThenDelete is private, so the only thing that moved for a consumer is the dependency floor, which is what the parenthetical in the title names.

The dependency move is the point of this release. magic_notifications 0.2.0 changed NotificationsListView.onDelete to Future<bool>, and this package's default branch could not compile against it until #122 landed. Publishing this is what carries the fix to an adopter: alpha.25 on pub.dev still pins ^0.1.0 and still returns Future<void>, so anybody upgrading magic_notifications on their own today gets a compile error from this package rather than from their own code.

Six version sites, swept by shape rather than by the old number: pubspec.yaml, CHANGELOG.md, CLAUDE.md, README.md, doc/getting-started/installation.md, and the magicStarterVersion constant behind the command banners.

Verification

• flutter test: 1409 passed
• dart analyze: no issues
• dart format --output=none --set-exit-if-changed .: 0 of 329 changed
• dart pub publish --dry-run: zero warnings

There is no pubspec_overrides.yaml in this worktree, so the resolution above is the one the publish workflow actually sees rather than a friendlier local one. pubspec.lock records:

magic_notifications:
  source: hosted
  url: https://pub.dev
  version: "0.2.0"

which is the release that went out an hour ago, so the 1409 tests ran against the published artefact rather than against a sibling working tree.

Author: Anılcan Çakır anilcan.cakir@gmail.com

[kodizm]

Note

Kodizm (AI-generated). May contain mistakes; verify before acting.

A clean version-only release bump - all six sites agree, the version guard test passes, and the full suite reproduces the 1409 the description claims; two cosmetic notes, nothing blocking.

The diff is six files and no behavioural change: pubspec.yaml, CLAUDE.md, README.md, doc/getting-started/installation.md and magicStarterVersion all move alpha.25 → alpha.26 in lockstep, and CHANGELOG.md converts the [Unreleased] block (the ### Changed from #122, plus ### Added/### Fixed from #121) into a dated [0.0.1-alpha.26] section, leaving [Unreleased] empty. grep for alpha\.2[0-9] finds no site left behind on 25.

Minor

CHANGELOG.md:7 - the entry is dated 2026-09-03, one day in the future (today is 2026-09-02), and it sits directly above [0.0.1-alpha.25] - 2026-09-02. Nothing reads the date - publish.yml triggers on the tag and only runs analyze/format/test - so this is presentation only, but the changelog will read as released a day before it was. (maintainability)

pubspec.yaml:44-46 - the comment block above magic_notifications: ^0.2.0 still narrates 0.1.0 as the floor: "0.1.0 is live on pub.dev, so the floor names it", and "^0.1.0 is >=0.1.0 <0.2.0 ... and admits nothing below the release that carries the API". The constraint below it is ^0.2.0. This landed with #122, not with this diff, so it is out of scope strictly speaking - flagging it only because this is the release that ships that file to pub.dev and the dependency floor is the stated point of the release. (maintainability)

Tests

No new tests, and none are owed - the diff adds no behaviour. The one guard that matters for a version bump already exists and passes: test/cli/starter_artisan_provider_test.dart:58 reads version: out of pubspec.yaml and asserts magicStarterVersion equals it, so the banner constant cannot drift from the published version.

Checks I ran

• flutter test - 1409 passed, 0 failed. Matches the description exactly.
• flutter analyze - No issues found! (ran in 23.5s).
• dart format --output=none --set-exit-if-changed . - Formatted 329 files (0 changed).
• flutter test test/cli/starter_artisan_provider_test.dart - 5/5, including magicStarterVersion matches the version pubspec.yaml declares.
• grep -rn '0\.0\.1-alpha\.2[0-9]' across md/yaml/dart - five sites, all alpha.26; no stale reference outside the changelog history.
• Not verified: the description's pubspec.lock excerpt. pubspec.lock is gitignored (.gitignore:32) and absent from this checkout, so I cannot confirm the 1409 above resolved against magic_notifications 0.2.0 specifically - only that the ^0.2.0 constraint in pubspec.yaml is what the resolver was given. There is no pubspec_overrides.yaml, which I did confirm.
• dart pub publish --dry-run - not run.

[anilcancakir]

Both taken, one applied and one measured away.

The floor comment is a real one and it is fixed. You are right that it landed with #122 rather than with the bump, and right that it matters here anyway: this is the release that ships that file to pub.dev, and an adopter reading the published pubspec would have found the prose arguing against the constraint three lines below it. It now also says why 0.2.0 rather than 0.1.0, which the old text could not: against 0.1.0 the library still compiles, since Future<bool> satisfies a Future<void> parameter, and it is the routes test that does not, because it reads onDelete's return type.

The date is right, and the disagreement is a timezone rather than a mistake. Right now it is 2026-09-02 22:57 UTC and 2026-09-03 01:57 +03. This repository dates by local time, which its own history settles rather than my asserting it:

	value
[0.0.1-alpha.25] heading	2026-09-02
its release commit	2026-09-02 00:42:19 +0300
pub.dev received it	2026-09-01T21:48:14Z

So the previous entry is dated a day AHEAD of the UTC instant it was published at, for exactly the reason this one is. magic_notifications does the same. Leaving it.

Worth recording for the next reader: the two clocks here are three hours apart, so between local midnight and 03:00 they name different days, and "one day in the future" is what UTC sees during that window rather than a defect.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!