[dcalhoun]

What?

Limit auto-focus to new, empty posts.

Why?

Aligns more closely with the WP Admin experience.

How?

• Add an internal isNewEmptyPost check for iOS and Android
• Check for no server ID, empty title, and empty content

Testing Instructions

Test the demo app on iOS and Android with a site added through the demo app's login, rather than the built-in wp-env site, which has no Browse button on Android. Use a device or a simulator/emulator with the software keyboard enabled.

1. In the site's WP Admin, add a new post, type only a title, and click Save draft.
2. In the demo app, select the site and tap Start to open a new post. Verify that the keyboard opens with the caret in the title.
3. Close the editor, tap Browse, and open the post from step 1. Verify that the keyboard doesn't open. On trunk, it does.
4. To open a new post with only a title, as a reopened local draft looks to the editor, temporarily add .setTitle("Trip notes") to applyDemoAppDefaults in ios/Demo-iOS/Sources/Views/SitePreparationView.swift. On Android, change .setTitle("") to .setTitle("Trip notes") in SitePreparationViewModel.kt. Tap Start, and verify that the keyboard doesn't open. On trunk, it does.
5. To hide the title, temporarily add .setShouldHideTitle(true) to applyDemoAppDefaults on iOS. On Android, change .setHideTitle(false) to .setHideTitle(true). Tap Start, and verify that the keyboard opens with the caret in the content. This PR shouldn't change this case.

Revert the demo app edits afterward.

Accessibility Testing Instructions

N/A, no navigation changes.

Screenshots or screencast

N/A, no visual changes.

---

AI-generated details

Native autofocus (#191) was meant for new posts but checked only for empty content. It now uses the same rule as the web editor's new posts (#794): no server ID (missing or ≤ 0), and a blank title and content.

• Adds an internal isNewEmptyPost to EditorConfiguration on iOS and Android.
• iOS focus(force:) and Android's autofocus on load use it. focus(force: true) still overrides it.
• On Android, isNewEmptyPost and GBKitGlobal share an internal serverPostId (null for a new post), so autofocus and the web editor can't disagree about whether a post is new. Only the bridge's Post.id still encodes a new post as -1.

Behavior changes

• Still focuses: a new post, and a new title-less editor such as comments.
• Stops focusing: an existing server post with empty content, and a local draft with only a title.
• Public API: a host's focus() call without force on iOS no longer focuses an existing post with empty content.
• Android bridge: a negative post ID converted to UInt now reaches the web editor as -1 rather than its own negative value. Both mean a new post.

The native calls stay because core's title focus only moves DOM focus. iOS's becomeFirstResponder() and Android's requestFocus() and showSoftInput raise the keyboard, and GBK's editor.focus() covers editors with a hidden title, which core never focuses.

Testing

• Unit tests for isNewEmptyPost on both platforms, and for the Android -1 mapping in GBKitGlobalTest. The iOS package tests and Android library unit tests pass, and both linters pass.
• Not yet tested on a device.

🤖 Generated with Claude Code

https://claude.ai/code/session_01YNDBX7RNXn3NsfDuV1PkNV

[wpmobilebot]

XCFramework Build

This PR's XCFramework is available for testing. Add the following to your Package.swift:

.package(url: "https://github.com/wordpress-mobile/GutenbergKit", branch: "pr-build/823")

_{Built from ac09603}

[nbradbury]

Claude had one medium finding, but it's not something that really needs to be addressed, imo.

Severity	Merge	Location	Issue	Impact
Medium	Could fix	GBKitGlobal.kt:122	The web editor now gets -1 for every negative post ID on Android, while iOS still passes the ID through unchanged (configuration.postID ?? -1 in GBKitGlobal.swift).	A host passing (-5L).toUInt() used to send post.id = -5 and now sends -1, but iOS still sends -5; that changes what goes over the wire in a commit labelled a refactor and leaves the two platforms disagreeing.

[dcalhoun]

@nbradbury I don't believe it is all that valuable to address here. I do not believe any post passes negative IDs aside from -1 and doing so wouldn't change the outcome. We could potentially clean up the divergence later.

[nbradbury]

All good