fix(MoveMailboxModal): make self-move a real no-op and always reset state

[chiliec]

What

Closes #13534

Fixes the dead self-move guard in the move-folder modal.

The bug

onMove() set this.moving = true and then guarded the whole operation with this.mailbox.id !== this.destMailboxId:

• Mailbox::jsonSerialize exposes id as base64_encode($name), while destMailboxId is always the numeric databaseId (or undefined). The two can never be equal, so the "moving a folder into itself is a no-op" check never fired — selecting the current parent for a folder that already sits there issued a rename to its own name, and the resulting IMAP error was only logged (silent no-op for the user).
• The moving = false / $emit('close') cleanup lived inside the finally of that dead branch, so a corrected early exit would have left the dialog's primary button disabled forever (since fix: minor adjustment in moving messages, threads and folders #13431 made the handler's promise the one NcDialogButton awaits).

Fix

• Compare this.mailbox.databaseId against this.destMailboxId.
• Return early (after emitting close) when they match, before setting moving = true.
• The try/finally now covers the actual move, so moving is always reset and the dialog always closes.

Tests

Added src/tests/unit/components/MoveMailboxModal.vue.spec.js covering: same-mailbox no-op (no renameMailbox, moving reset, close emitted), a real move with correct new name, and state reset after a failed move.

Validation

$ npx vitest --run src/tests/unit/components/MoveMailboxModal.vue.spec.js
 Test Files  1 passed (1)
      Tests  3 passed (3)

Genuine RED→GREEN: reverting src/components/MoveMailboxModal.vue to the buggy version makes the "same mailbox is a no-op" test fail (the old code calls renameMailbox when moving into itself); restoring the fix turns it green.

npx eslint on both changed files is clean. I ran the full vitest --run suite too — a handful of unrelated files (TranslationModal, MessageService, store/actions, filePicker, languageDetection) fail non-deterministically with jsdom timeouts on the same count on a clean main checkout without my change, so they are pre-existing and outside this diff.

Happy to adjust.