refactor: extract KeyboardLayoutHandler from GeneralKeyboardIME (Part 15) - #426 by prince-0408 · Pull Request #692 · scribe-org/Scribe-Android

prince-0408 · 2026-08-31T20:52:21Z

Description

This PR is Part 15 in modularizing GeneralKeyboardIME for #426.

It extracts keyboard layout XML resolution (getCurrentKeyboardLayoutXML, getPrimarySymbolKeyboardLayoutXML, getKeyboardLayoutForState), container width calculation (getKeyboardWidth), input-type keyboard layout matching (getKeyboardLayoutXMLForInputType, shouldUseNumericKeyboard), and keyboard view re-creation (recreateKeyboard) out of GeneralKeyboardIME.kt into a standalone helper class KeyboardLayoutHandler.

Detailed Changes Table

File / Component	Changes Applied	Detailed Impact
`KeyboardLayoutHandler.kt`	Created standalone helper encapsulating layout XML resolution (`getCurrentKeyboardLayoutXML`, `getPrimarySymbolKeyboardLayoutXML`, `getKeyboardLayoutForState`), container width calculation (`getKeyboardWidth`), input-type layout matching (`getKeyboardLayoutXMLForInputType`, `shouldUseNumericKeyboard`), and view re-creation (`recreateKeyboard`).	Centralizes all keyboard XML layout resolution and width calculation logic in a dedicated helper class.
`GeneralKeyboardIME.kt`	Instantiated `layoutHandler` delegate and delegated layout methods (`getCurrentKeyboardLayoutXML`, `getKeyboardWidth`, `recreateKeyboard`, `getKeyboardLayoutForState`). Delegated companion input type layout helpers (`getKeyboardLayoutXMLForInputType`, `shouldUseNumericKeyboard`). Removed dead extension `coerceInSafe`.	Fulfills the final component extraction step of #426 by decoupling layout resolution and container width calculations from `GeneralKeyboardIME.kt` while preserving 100% backward compatibility.
`KeyboardLayoutHandlerTest.kt`	Added unit tests covering layout XML resolution across letter mode, symbol mode, symbol-shift mode, state-based conjugation layouts, and input type matching.	Ensures unit test coverage for layout handler logic.
`CHANGELOG.md`	Added entry under `### ♻️ Code Refactoring` detailing the extraction of `KeyboardLayoutHandler` from `GeneralKeyboardIME`.	Satisfies `ci_changelog_check` workflow requirement for pull requests targeting `main`.

Key Benefits

Decoupled Layout Resolution & Width Calculations: Removes layout XML selection, conjugation grid resolution, and container width math from GeneralKeyboardIME.kt.
Improved Testability & Clean Architecture: Layout selection and input-type matching can now be tested independently without instantiating full Android service components.

Related Issue

Refactors part of #426

andrewtavis · 2026-08-31T21:14:34Z

Great that this is open, @prince-0408! Would this be the last PR to close the issue?

prince-0408 · 2026-09-01T10:48:54Z

Great that this is open, @prince-0408! Would this be the last PR to close the issue?

Thanks @andrewtavis! Not the last one -- Part 16 (introducing a KeyboardIMEContext interface contract as suggested by @Roniscend in #689) is already raised. That should be the final PR to close #426.

Roniscend · 2026-09-01T13:13:05Z

-                } else {
-                    when (dataSize) {
-                        DATA_SIZE_2 -> R.xml.conjugate_view_2x1
-                        DATA_CONSTANT_3 -> R.xml.conjugate_view_1x3


Naming nit , DATA_CONSTANT_3 should be DATA_SIZE_3 to match DATA_SIZE_2 wherever data_constant is mentioned

Updated DATA_CONSTANT_3 to DATA_SIZE_3 for consistency.

Roniscend · 2026-09-01T13:15:25Z

+import android.text.InputType.TYPE_MASK_CLASS
+import be.scri.R
+import be.scri.models.ScribeState
+import be.scri.services.GeneralKeyboardIME


Circular import: KeyboardLayoutHandler (helpers) and GeneralKeyboardIME (services) import each other, the handler isn't truly decoupled, it just holds a full ime reference.

Good catch. In this PR (Part 15), KeyboardLayoutHandler followed the same structure as previous helper extractions. In the dedicated follow-up #693 (Part 16), we introduce the KeyboardIMEContext interface contract across all handlers, removing the GeneralKeyboardIME reference from KeyboardLayoutHandler and breaking the circular import across the codebase.

Roniscend · 2026-09-01T13:19:02Z

-                TYPE_CLASS_NUMBER, TYPE_CLASS_DATETIME, TYPE_CLASS_PHONE -> true
-                else -> false
-            }
+        internal fun shouldUseNumericKeyboard(inputType: Int): Boolean = KeyboardLayoutHandler.shouldUseNumericKeyboard(inputType)


Wrappers still hang around shouldUseNumericKeyboard and getKeyboardLayoutXMLForInputType in the companion object just forward to KeyboardLayoutHandler now ,could remove them and call KeyboardLayoutHandler directly if nothing else still uses these.

Removed the forwarding wrappers from GeneralKeyboardIME companion object and updated all call sites and unit tests to invoke KeyboardLayoutHandler directly.

andrewtavis · 2026-09-05T10:15:09Z

@prince-0408: Would you be able to rebase and resolve conflicts? Anything else needed after that, or would you say that we're ready for a final review? :)

Roniscend · 2026-09-23T16:38:35Z

+ * Encapsulates keyboard XML layout resolution, symbol layout mapping,
+ * keyboard width calculations, state-based layout XML selection, and view re-creation.
+ */
+class KeyboardLayoutHandler(


Couples to the concrete GeneralKeyboardIME instead of KeyboardIMEContext like every sibling,needs handler,getKeyboardLayoutXML(), defaultConjugateLayoutXML, enterKeyType, isFloatingMode, and isNumericKeyboardActive as a property added to the interface first.

Roniscend · 2026-09-23T16:40:23Z

+
+    @Test
+    fun testGetCurrentKeyboardLayoutXML_letterMode() {
+        every { mockIme.keyboardMode } returns 0


Stubs getKeyboardLayoutXML() to return R.xml.keys_symbols, so the assertion passes even if the mode-check branch were inverted

Roniscend · 2026-09-23T16:48:44Z

After these changes , this pr will also be good to merge also then we can ask Andrew for a final review

andrewtavis · 2026-09-25T17:48:08Z

Looks like we're good to move to this one post resolving conflicts, @prince-0408 :) Once we're done here, maybe you and @Roniscend can support @growabeard with the upcoming PR and then we can finalize the issue! 😊

… 15) - scribe-org#426

andrewtavis · 2026-09-27T08:07:14Z

Thanks for the rebase, @prince-0408! @Roniscend, if you have a moment for an initial review, that'd be very helplful :) I'll hopefully be done with the Scribe-Data work later today.

Roniscend · 2026-09-27T12:00:46Z

Thanks for the rebase, @prince-0408! @Roniscend, if you have a moment for an initial review, that'd be very helplful :) I'll hopefully be done with the Scribe-Data work later today.

Sure, Also I have reviewed this pr twice and after these changes we are good to merge this pr, requesting a final review from your side
Thanks

andrewtavis · 2026-09-27T15:47:45Z

Hi @prince-0408, @Roniscend, @angrezichatterbox: We have an issue from something getting into main that is now breaking a proper keyboard load. This is likely from the refactor issues, which I seem to have not done a good enough job doing final reviews on. I assume that the issue will be easy to fix, but let's fix it as a part of this PR or send something along quickly in a different PR. See the image below, which is from a Pixel 9 Android 36 emulator:

andrewtavis requested review from Roniscend and andrewtavis August 31, 2026 21:14

Roniscend reviewed Sep 1, 2026

View reviewed changes

prince-0408 added the refactor Refactor code to improve quality label Sep 1, 2026

prince-0408 force-pushed the refactor/extract-keyboard-layout-handler-426 branch 2 times, most recently from 250fe75 to 2ab5b6a Compare September 5, 2026 17:56

andrewtavis added the no-changelog No changelog entry is needed for this pull request label Sep 13, 2026

This was referenced Sep 13, 2026

refactor: extract SuggestionUIHandler from GeneralKeyboardIME (Part 11) (#426) #685

Merged

Code improvements to make Scribe ready for future features #426

Open

Roniscend suggested changes Sep 23, 2026

View reviewed changes

prince-0408 added 2 commits September 27, 2026 03:26

refactor: extract KeyboardLayoutHandler from GeneralKeyboardIME (Part…

35da63d

… 15) - scribe-org#426

refactor: address review comments on KeyboardLayoutHandler

e0d0c14

prince-0408 force-pushed the refactor/extract-keyboard-layout-handler-426 branch from 625680c to e0d0c14 Compare September 26, 2026 21:57

andrewtavis requested a review from Roniscend September 27, 2026 08:06

andrewtavis mentioned this pull request Sep 27, 2026

Keyboard doesn't load properly on API 36 on a Pixel 9 emulator #707

Open

2 tasks

Conversation

prince-0408 commented Aug 31, 2026

Description

Detailed Changes Table

Key Benefits

Related Issue

Uh oh!

andrewtavis commented Aug 31, 2026

Uh oh!

prince-0408 commented Sep 1, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

andrewtavis commented Sep 5, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Roniscend commented Sep 23, 2026

Uh oh!

andrewtavis commented Sep 25, 2026

Uh oh!

andrewtavis commented Sep 27, 2026

Uh oh!

Roniscend commented Sep 27, 2026

Uh oh!

andrewtavis commented Sep 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants