Dev to master aug18

.changeset/brave-badgers-arrive.md

@@ -0,0 +1,5 @@
+---
+bump: major
+---
+
+Set Homes Two can now take over from Set Homes v1. /import-homes brings your homes and world blacklist across and lists any v1 config settings that have an equivalent here, the v1 commands and aliases work again, and the plugin tells you at startup when homes are still waiting to be imported.

.changeset/brave-otters-listen.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Tab completion for /import-homes now offers the import sources and then confirm, instead of suggesting player names on every argument. Usage messages also name the command you actually typed, so /uhome-of now says Usage: /uhome-of rather than naming the full command behind the alias.

.changeset/brave-pandas-sing.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+decrease generated jar file size

.changeset/clever-tigers-wander.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+If Set Homes v1 homes are sitting in plugins/SetHomes/homes.yml and you have not imported them yet, the console now says so at startup, naming the file, how many homes are waiting and the command to run. Anyone with sh2.import-homes gets the same reminder in chat when they join. Both stop on their own as soon as any home exists here, so an upgraded server can no longer look empty without explanation.

.changeset/eager-badgers-return.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Added the Set Homes v1 admin commands. /home-of teleports you to another player's home, /delhome-of deletes one, and /uhome-of moves one to where you are standing. /uhome moves one of your own homes. All of them work on players who are offline, as long as they have saved homes, and player and home names now match without regard to case. On a server upgrading from an earlier release, each player has to log in once before the offline lookup can find them.

.changeset/eager-seals-shine.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+/sethome and /home work with no arguments again, creating and finding a home called default, which is the name an unnamed Set Homes v1 home is imported under. /sethome also takes a description straight after the name as it did in v1, so a second word that names an item is read as the icon and the reply tells you which one it chose. Put d in the icon position to keep the whole phrase as the description.

.changeset/gentle-dolphins-wonder.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Tab completion now matches anywhere in a name, not only from the start, so typing net completes to world_nether and ase finds a home called base. Names that start with what you typed are still listed first, so nothing you already do changes.

.changeset/gentle-foxes-config.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+/import-homes sethomes now lists any v1 config.yml settings that have a Set Homes Two equivalent, and the key to set each one under, so a migrating admin does not have to discover a changed teleport delay or a vanished per-group home limit by accident. Nothing is written to config.yml automatically; the report just tells you what to paste in.

.changeset/happy-otters-dance.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Importing from Set Homes v1 no longer loses a home whose name differs only in capitalisation. Set Homes v1 allowed one player to hold both 'base' and 'Base', while home names here ignore case, so the second one is now imported under the next free name such as 'Base2' instead of being silently dropped. The import report and the server log both name it, and the preview now reports the same numbers as the confirm that follows it.

.changeset/kind-herons-shine.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Corrected the spelling in a few of the messages the plugin sends, so they read the same way as the rest of the plugin. The wording of the teleport and import messages is otherwise unchanged. An existing config.yml is not touched, so any message you have already customized stays exactly as you set it.

.changeset/kind-ravens-gather.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+The plugin now sends anonymous usage counts to bStats, so the maintainers can see which commands, menu buttons and settings are used, and whether database errors are happening in the wild. Nothing personal is sent, and the bStats switch in plugins/bStats/config.yml turns it off.

.changeset/lucky-wolves-smile.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Fixed the world blacklist only ever being enforced on the first three worlds of a server. Blacklisting any world beyond those reported success and then did nothing. It is enforced from this release on, so if you have blacklisted more than three worlds, run /blacklist list before you update. Homes in a world that starts being enforced stop being reachable by players who do not hold sh2.bypass-blacklist.

.changeset/plain-lions-report.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+/blacklist add and /blacklist remove no longer report success when the database write actually failed. A failure now says so and the world is left as it was. Permission overrides in config.yml are also matched without regard to the case you write the node in, so a line like SH2.manage-homes now applies rather than being logged as applied and quietly ignored.

.changeset/quiet-herons-arrive.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Set Homes now publishes to the original Set Homes project page, so servers running the older Set Homes are offered this as an update on the page they already watch. The download is named SetHomes rather than SetHomesTwo. The plugin folder and your existing config.yml are unaffected.

.changeset/quiet-otters-jump.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+/set-max-homes now has the alias /setmax, matching Set Homes v1. Its usage message also names whichever form you typed, rather than always naming /set-max-homes.

.changeset/quiet-otters-names.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Importing from Set Homes v1 or EssentialsX now resolves each player's name from the server's own player cache, so admin commands like /get-player-homes and /home-of work on a migrated player immediately instead of requiring them to log in first. A player the server has never seen still imports fine; they are picked up on their next join, same as before.

.changeset/tidy-tigers-wonder.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+The Set Homes v1 import report no longer applies the colour codes found in v1 messages, so a coloured v1 message no longer turns the rest of the advice line red or unreadable. Codes are shown as an ampersand instead, and the note says where to copy the exact original from. Also, the "you have no homes yet" notice now uses the same wording and is shown as plain information rather than red from /homes, /list-homes and the homes menu alike.

.changeset/tidy-wolves-travel.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Permission defaults can now be changed from config.yml with no permissions plugin installed. Uncomment the permissions block and set any sh2 node to true, false, op or not-op. Two bundles, sh2.player and sh2.admin, move a whole role at once, and three new bypass permissions were added for admins: sh2.bypass-max-homes, sh2.bypass-blacklist and sh2.bypass-teleport-delay. Operators hold those three by default, the same way Set Homes v1 granted homes.config_bypass through homes.*, so an operator is not held to the teleport delay, the world blacklist or the home limit. Set any of them to false in the permissions block if you would rather they were.

.changeset/warm-badgers-wonder.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+Set Homes now refuses to start when a Set Homes v1 jar is still in your plugins folder, and the console says which file to move and what to run next. Both plugins provide /sethome, /home and /delhome and v1 wins those names, so running the two side by side used to split your players' homes between them with nothing in the log to show for it.

.changeset/wobbly-hedgehogs-import.md

@@ -0,0 +1,5 @@
+---
+bump: patch
+---
+
+/import-homes sethomes now brings across the Set Homes v1 world blacklist along with the homes, so a server that had worlds blocked in v1 does not silently lose that protection on migration. Re-running the import never blacklists a world twice.

.github/workflows/changeset.yml

@@ -3,7 +3,10 @@ name: Changeset
 # Its own workflow rather than a job in Tests: a job skipped by an `if:` still
 # reports a check run, so the push event would publish a second, skipped
 # "changeset" result for the same commit.
-on: pull_request
+on:
+  pull_request:
+    # labeled and unlabeled so applying the downmerge label re-runs the gate.
+    types: [opened, synchronize, reopened, labeled, unlabeled]
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
@@ -24,6 +27,8 @@ jobs:
         env:
           GH_TOKEN: ${{ github.token }}
           PR: ${{ github.event.pull_request.number }}
+          PR_BASE: ${{ github.event.pull_request.base.ref }}
+          PR_LABELS: ${{ join(github.event.pull_request.labels.*.name, ',') }}
         run: |
           set -euo pipefail
           # Written to a file rather than piped, so a failing gh call cannot be

.github/workflows/issue-status.yml

@@ -0,0 +1,114 @@
+name: Issue status
+
+# Moves issues across the project board as their work progresses. GitHub does
+# not link an issue to a pull request based on dev, so the closing keywords are
+# parsed out of the body here rather than read back from the API.
+on:
+  push:
+    branches: ['issue-*']
+  pull_request:
+    types: [opened, reopened, ready_for_review, edited, closed]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: false
+
+permissions: {}
+
+env:
+  PROJECT_NUMBER: '1'
+  GH_TOKEN: ${{ secrets.RELEASE_TOKEN }}
+
+jobs:
+  in-progress:
+    # A branch deletion after its pull request merges can also fire a push
+    # event, with no branch left to check out.
+    if: github.event_name == 'push' && github.event.deleted == false
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out
+        uses: actions/checkout@v4
+
+      - name: Move the issue along
+        env:
+          BRANCH: ${{ github.ref_name }}
+        run: |
+          set -uo pipefail
+          source scripts/issue-status.sh
+
+          issue="$(printf '%s' "$BRANCH" | sed -nE 's/^issue-([0-9]+)-.*/\1/p')"
+          if [ -z "$issue" ]; then
+            echo "Branch $BRANCH does not name an issue - nothing to do."
+            exit 0
+          fi
+
+          # Pushing to a branch that is already under review, by merging the
+          # base branch in for example, must not report it as in progress.
+          # synchronize is not a trigger type, so this job is the only thing
+          # that runs on such a push.
+          if [ "$(gh pr list --head "$BRANCH" --state open --json number --jq 'length')" = "0" ]; then
+            state="none"
+          else
+            state="open-pr"
+          fi
+
+          mapfile -t plan < <(push_plan "$state")
+          bash scripts/set-issue-status.sh "$issue" "${plan[0]}" "${plan[@]:1}"
+
+  pull-request:
+    # Fork runs get no secrets, so GH_TOKEN would be empty and every external
+    # contribution would show a red check it has no way to fix.
+    if: github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == github.repository
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out
+        uses: actions/checkout@v4
+
+      - name: Move the referenced issues
+        env:
+          # Through the environment, never interpolated into the script: a body
+          # is attacker controllable text.
+          BODY: ${{ github.event.pull_request.body }}
+          MERGED: ${{ github.event.pull_request.merged }}
+          BASE: ${{ github.event.pull_request.base.ref }}
+          ACTION: ${{ github.event.action }}
+        run: |
+          set -uo pipefail
+          source scripts/issue-status.sh
+
+          if [ "$ACTION" = "closed" ]; then
+            if [ "$MERGED" != "true" ] || [ "$BASE" != "dev" ]; then
+              echo "Closed without merging into dev - nothing to do."
+              exit 0
+            fi
+            status="Ready for release"
+            # Every state except Done: a merge into dev always beats an
+            # earlier state, but a down-merge from master can repeat a
+            # Closes reference for an issue that already shipped, and that
+            # must not drag it backwards out of Done.
+            allowed=(Todo unset "In progress" "In review" "Ready for release")
+          else
+            status="In review"
+            # Guards an edit on an already merged pull request from pulling
+            # the issue back out of Ready for release or Done.
+            allowed=(Todo unset "In progress")
+          fi
+
+          refs="$(printf '%s' "$BODY" | closing_refs)"
+          if [ -z "$refs" ]; then
+            echo "No closing references in the body - nothing to do."
+            exit 0
+          fi
+
+          failures=0
+          while IFS= read -r issue; do
+            [ -n "$issue" ] || continue
+            if ! bash scripts/set-issue-status.sh "$issue" "$status" "${allowed[@]}"; then
+              printf 'FAILED: could not set #%s to %s\n' "$issue" "$status" >&2
+              failures=$((failures + 1))
+            fi
+          done <<< "$refs"
+
+          if [ "$failures" -gt 0 ]; then
+            exit 1
+          fi

.github/workflows/release.yml

@@ -16,6 +16,8 @@ concurrency:
 
 permissions:
   contents: write
+  issues: write
+  pull-requests: read
 
 jobs:
   release:
@@ -96,6 +98,17 @@ jobs:
           echo "current=$CURRENT"
           echo "next=$NEXT"
 
+      - name: Dry run - list the issues this release would close
+        if: steps.plan.outputs.release == 'true' && inputs.dry_run
+        env:
+          GH_TOKEN: ${{ github.token }}
+          VERSION: ${{ steps.plan.outputs.next }}
+          DRY_RUN: '1'
+        run: |
+          set -uo pipefail
+          PREV_TAG="$(git describe --tags --abbrev=0 2>/dev/null || true)" \
+            bash scripts/close-released-issues.sh
+
       - name: Dry run - resolve BukkitDev game versions
         if: steps.plan.outputs.release == 'true' && inputs.dry_run
         env:
@@ -116,6 +129,17 @@ jobs:
                 end
             '
 
+      # Must run before Commit and tag creates v$NEXT, or the range below is
+      # empty and nothing closes.
+      - name: Record the previous release tag
+        id: prev
+        if: steps.plan.outputs.release == 'true'
+        run: |
+          set -uo pipefail
+          tag="$(git describe --tags --abbrev=0 2>/dev/null || true)"
+          echo "tag=$tag" >> "$GITHUB_OUTPUT"
+          echo "Previous release tag: ${tag:-none}"
+
       - name: Commit and tag
         if: steps.plan.outputs.release == 'true' && !inputs.dry_run
         env:
@@ -126,15 +150,15 @@ jobs:
           git add pom.xml README.md .changeset
           git commit -m "chore(release): $NEXT"
           git tag -a "v$NEXT" -m "Release $NEXT"
-          git push origin HEAD:master --follow-tags
+          git push --atomic origin HEAD:master --follow-tags
           git ls-remote --exit-code --tags origin "refs/tags/v$NEXT" > /dev/null \
             || { echo "Tag v$NEXT did not reach the remote"; exit 1; }
 
       - name: Rename the jar to the release convention
         if: steps.plan.outputs.release == 'true' && !inputs.dry_run
         env:
           NEXT: ${{ steps.plan.outputs.next }}
-        run: cp "target/SetHomesTwo-$NEXT.jar" "SetHomesTwo.V$NEXT.jar"
+        run: cp "target/SetHomesTwo-$NEXT.jar" "SetHomes.V$NEXT.jar"
 
       - name: Publish to BukkitDev
         if: steps.plan.outputs.release == 'true' && !inputs.dry_run
@@ -150,6 +174,18 @@ jobs:
           VERSION: ${{ steps.plan.outputs.next }}
         run: |
           NOTES=$(bash scripts/release.sh notes --readme README.md --version "$VERSION")
-          gh release create "v$VERSION" "SetHomesTwo.V$VERSION.jar" \
-            --title "SetHomesTwo V$VERSION" \
+          gh release create "v$VERSION" "SetHomes.V$VERSION.jar" \
+            --title "Set Homes V$VERSION" \
             --notes "$NOTES"
+
+      # An issue closes when the commit that fixed it is contained in the commit
+      # being released. That is what keeps work sitting on dev open when an
+      # immediate fix ships from master, and closes an immediate fix that never
+      # passed through dev.
+      - name: Close the issues this release contains
+        if: steps.plan.outputs.release == 'true' && !inputs.dry_run
+        env:
+          GH_TOKEN: ${{ github.token }}
+          PREV_TAG: ${{ steps.prev.outputs.tag }}
+          VERSION: ${{ steps.plan.outputs.next }}
+        run: bash scripts/close-released-issues.sh

.github/workflows/tests.yml

@@ -28,6 +28,20 @@ jobs:
       - name: Build and test
         run: mvn -B verify
 
+      - name: Prove the shaded libraries were relocated
+        run: |
+          jar="$(ls target/SetHomesTwo-*.jar | head -n 1)"
+          listing="$(unzip -l "$jar")"
+          if echo "$listing" | grep -q ' org/bstats/'; then
+            echo "org.bstats was not relocated"; exit 1
+          fi
+          if ! echo "$listing" | grep -q ' com/samleighton/sethomestwo/lib/bstats/'; then
+            echo "relocated bstats classes are missing"; exit 1
+          fi
+          if echo "$listing" | grep -q ' net/wesjd/'; then
+            echo "anvilgui was not relocated"; exit 1
+          fi
+
       - name: Run the release script tests
         run: bash scripts/test-release.sh
 
@@ -37,6 +51,9 @@ jobs:
       - name: Run the BukkitDev publish tests
         run: bash scripts/test-publish-bukkitdev.sh
 
+      - name: Run the issue status parsing tests
+        run: bash scripts/test-issue-status.sh
+
       # Rehearses the release on every pull request. The changelog heading went
       # missing in a docs change, and nothing noticed until a release ran.
       - name: Check a release could be applied