fix(android): add contentful.java Entry adapter for OptimizedEntry [NT-3807] by fmamud · Pull Request #395 · contentful/optimization

Felipe Mamud (fmamud) · 2026-07-30T10:58:14Z

Summary

Ships a contentful.java → Optimization SDK integration path in optimization-android so integrators stop hand-writing the {sys, fields, metadata} map the resolver requires. Introduces CTEntry, an SDK-owned wrapper around a CDAEntry whose accessors delegate to it, and typed OptimizedEntry(entry: CDAEntry, ...) overloads for Compose and Views that flow the resolved variant back through CTEntry — read with getField<T> / hasField / id instead of as? casts on a raw map. The existing Map<String, Any> public surface is unchanged, so current consumers keep working without edits.

Fixes NT-3807.

What ships

Types

CTEntry (com.contentful.optimization.contentful) — wraps a real CDAEntry; every accessor delegates:
- id → entry.id()
- contentTypeId → entry.contentType()?.id()
- localeCode / createdAt / updatedAt → entry.getAttribute(...)
- getField<T>(name) → entry.getField(name) (walks the real LocalizedResource.Localizer)
- hasField(name) → entry.rawFields().containsKey(name)
- operator get(name): String?
- Three factory methods normalize input to the same wrapped-CDAEntry shape:
  - CTEntry.from(entry: CDAEntry) — thin wrapper (CTEntry(entry)).
  - CTEntry.from(any: Map<String, Any>, fallback = EMPTY) — fabricates a CDAEntry from the resolver Map via reflection (writes package-private fields with a synthetic "_" locale so LocalizedResource.getField resolves).
  - CTEntry.from(json: String, fallback = EMPTY) — parses via JSONTokener, then fabricates.
- Both Map and String forms are fail-soft: on parse or fabrication failure they log a DiagnosticLogger.warning and return the caller-supplied fallback (default empty) rather than throwing.
- Serializers: toFoundation(): Map<String, Any> (delegates to the SDK's internal toOptimizedEntryMap(entry) adapter — centralizes the metadata block the resolver requires), toJSON(): String.
ResolvedOptimizedEntry.entry changed from Map<String, Any> to CTEntry. Both the Map and CDAEntry code paths wrap their output the same way; the underlying map is reachable via .entry.toFoundation() as an escape hatch.

Compose / Views surface

OptimizedEntry(entry: CDAEntry, ..., content: (CTEntry) -> Unit) — new Compose overload.
OptimizedEntryView.setEntry(entry: CDAEntry, ...) — new Views overload.
Existing OptimizedEntry(entry: Map<String, Any>, ...) and setEntry(entry: Map) / setContentRenderer((Map) -> View) signatures untouched — the base composable internally holds a CTEntry but hands the callback a Map via .toFoundation() for backwards compatibility.

Adapter behavior

metadata: {tags, concepts} always emitted, populated from CDAMetadata when present.
Nested resolved links (CDAEntry/CDAAsset) walked recursively; back-edges emit an unresolved Link stub via ancestor-set tracking so a real cycle doesn't recurse forever, and diamonds still fully expand on both branches.
Sys optional attrs (createdAt, updatedAt, revision, locale, space, environment) surfaced when the CDA response populated them, omitted when absent — matches raw sys optionality.
Assets carry description, fileName, contentType, details.{size, image.{width, height}}, and url (not just title + file.url).
CDARichNode re-serialized to the raw {nodeType, data, content} JSON shape. Embedded resource nodes (CDARichEmbeddedBlock / CDARichEmbeddedInline) matched before the generic CDARichHyperLink case, so an embedded entry no longer emits data.uri and loses its target; plain URI hyperlinks emit data.uri.
Date field values ISO-8601 formatted (previously fell through to Date.toString(), unusable JSON).
Fail-soft resolver: resolveOptimizedEntry wraps the baseline in CTEntry.from(baseline) once; the resolver-output parse passes the baseline CTEntry as fallback, so an unparseable bridge result decays to baseline content instead of a blank render.

Dependency

com.contentful.java:java-sdk:10.6.0 added as compileOnly (with okhttp-jvm excluded to avoid the okhttp5 duplicate-class conflict against okhttp-android). Consumers that only pass entry Maps aren't forced onto it.
The Map / json CTEntry.from(...) paths write into contentful.java's package-private fields (attrs, defaultLocale, rawFields, fields, contentType, metadata) at construction. Pinned to 10.6.0; a future rename fails at runtime with a clear NoSuchFieldException.

Docs

documentation/internal/sdk-knowledge/native/android.md — reconciled to the new surface (added CTEntry to state/result types, dedicated CTEntry section describing the CDAEntry-backed storage + reflection fabrication, updated resolveOptimizedEntry fail-soft bullet, noted the CDAEntry overload on OptimizedEntry / setEntry).
Compose and Views integration guides — code samples updated for CTEntry (.entry.toFoundation() where the old sample assigned .entry into a Map variable), added CDAEntry-overload examples, prose reflects the CDAEntry-backed shape.
documentation/concepts/android-sdk-runtime-and-interaction-mechanics.md — states that .entry is a CTEntry wrapping a CDAEntry.

Tests

73 unit tests pass. New coverage:
- CTEntryTest — three factory methods, fail-soft with caller-supplied fallback, JSON round-trip, accessor semantics (including the JVM generic-erasure gotcha), plus three Option-A-specific tests: from(entry) reads through real CDAEntry accessors, the fabricated CDAEntry from a Map reads through the real Localizer (regression guard), and malformed sys.contentType degrades gracefully.
- CDAEntryAdapterTest — whole-tree identity checks (bare entry sys/contentType shape; metadata always present when CDAMetadata is null; tag/concept link references; nested link expansion; back-edge cycle stub; diamond expansion; three-node cycle; self-reference; five-level linear chain; unresolved-link stub pass-through; asset happy path + no-description/no-image edge cases; asset with no file; Date field ISO-8601; URI hyperlink data.uri vs embedded-resource data.target).
./gradlew assembleRelease — clean.
Reference implementation APKs (:compose:assembleDebug + :views:assembleDebug) — both build cleanly against the new SDK, confirming the public API stays backwards-compatible.
Maestro E2E on Pixel_8_API_34: Compose app 34/34 flows passed in 6m 9s (identify/reset, view/click tracking, preview-panel scenarios 1–7, live-updates locking, profile screens, refresh). Views app 30/34 passed before manual emulator shutdown aborted the final 4; the ADB stack trace right after confirms those failures are shutdown-artifact, not SDK regressions.

Notes

Side note: iOS covers the same surface in fix(swift): add contentful.swift Entry adapter for OptimizedEntry [NT-3808] #393 — Contentful.Entry → resolver-map mapping, a CTEntry type on the resolver output, and a typed OptimizedEntry(entry:) initializer. This PR takes the analogous shape on Android; iOS chose an SDK-owned Codable envelope internally because Contentful.Entry.init(from:) requires a LocalizationContext only a live decode carries, while Android's CDAEntry can be fabricated from a Map via reflection into package-private fields — so Android delegates to a real CDAEntry directly rather than a Codable envelope.
Related follow-up: NT-3856 tracks contentful.java unwrapping / typed EntryMapping compatibility, mirroring iOS's NT-3854 for EntryDecodable.

🤖 Generated with Claude Code

Addresses Daviti's PR feedback (#395 review comment). Mirrors iOS PR 393's identity-check style: rather than digging into the resulting Map field-by-field with `as?` casts, each test now names the entire expected `{sys, fields, metadata}` map and asserts against it with `assertEquals`. Kotlin's `Map.equals` is entry-set equal (key-order-insensitive) and `List.equals` is positional — same semantics as iOS's `JSONValue` tree comparison in `CTEntryTests.swift`. The upside: an unexpected extra key or shape drift anywhere in the tree now fails the test; before, only the handful of keys a test happened to name were checked. Adds a small set of expected-shape helpers (`sys(id, contentTypeId)`, `linkStub(id, linkType)`, `emptyMetadata`) so each test's literal reads as the shape it intends, not as boilerplate. 70/70 unit tests pass; release build clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

… [NT-3807] Ships a `CDAEntry -> OptimizedEntry` conversion path so integrators stop hand-writing the `{sys, fields, metadata}` mapping: `CDAEntry.toOptimizedEntryMap()`, new `OptimizedEntry(entry: CDAEntry, ...)` Compose overload and `OptimizedEntryView.setEntry(entry: CDAEntry)` / `setResolvedContentRenderer(...)` View overloads, and typed accessors on the existing `ResolvedOptimizedEntry` (`id`, `contentTypeId`, `getField<T>`, `getEntry`, `getEntries`, `getAsset`) plus a `ResolvedAsset` peer. Metadata is always emitted (`{tags, concepts}`) so the resolver's entry guard cannot silently fall back to baseline. Asset mapping surfaces `description`, `fileName`, `contentType`, and `details.{size, image.{width, height}}` instead of only `title` / `file.url`. Rich Text URI hyperlinks are emitted with `data.uri` and split from embedded resource nodes (`data.target`). `Date` field values are ISO-8601 formatted instead of falling through to `Date.toString()`. `contentful.java:java-sdk:10.6.0` is added as `compileOnly` (with `okhttp-jvm` excluded to avoid the okhttp5 duplicate-class conflict with `okhttp-android`), so consumers that only pass entry Maps are not forced to depend on it. Existing `OptimizedEntry(entry: Map<String, Any>, ...)` and `setContentRenderer((Map<String, Any>) -> View)` signatures are unchanged. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Removed prose explaining what code already says. Kept only the non-obvious "why" notes: cycle-guard rationale on entryToMap, embedded-node ordering in richNodeToMap, okhttp-jvm exclusion, JVM lambda-erasure explainer on setResolvedContentRenderer, and the reflection fixture note. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Mirrors the scenarios OptimizationEntryMappingTests.swift covers: - Sys optional attrs: createdAt, updatedAt, revision, locale, space, environment. Emitted when the CDA response populates them (via CDAResource.getAttribute) and omitted when absent — matches the optionality of the raw sys block. - Link resolution: diamond expansion, 3-node cycle, self-referencing entry, five-level linear chain, and a raw unresolved link stub inside a field. - Asset edge cases: no description → key omitted; non-image file → details.image omitted; asset with no file → sys and title preserved with an empty file record. - Rich Text: plain URI hyperlink emits data.uri and content (proving the branch ordering fix — embedded resource cases match before the generic hyperlink case, so a plain hyperlink no longer emits data.target). New RawEntryMapPassThroughTest confirms the "raw entries" path: a hand-built Map<String, Any> flows through the base OptimizedEntry(Map) overload and reads through ResolvedOptimizedEntry's typed accessors — no contentful.java dependency needed at runtime for that integration. 64/64 tests passing. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

- Compose OptimizedEntry(CDAEntry, ...) content callback now receives a CDAEntry (the winning variant, or the baseline as fallback), not a ResolvedOptimizedEntry. The variant CDAEntry is already reachable through the baseline's nt_experiences → nt_variants link graph — no decoder/parser needed. Adds a public findVariantEntry(baseline, id) helper. - Drop toOptimizedEntryMap as an extension on CDAEntry in favor of a standalone toOptimizedEntryMap(entry: CDAEntry). - Remove setResolvedContentRenderer from OptimizedEntryView (unused). - Revert ResolvedOptimizedEntry's typed accessors and ResolvedAsset now that no SDK caller needs them. - Drop the corresponding accessor / raw-map pass-through test files; add 4 findVariantEntry tests to CDAEntryAdapterTest. 58/58 tests passing. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Adds a CTEntry class that wraps the resolver's `{sys, fields, metadata}` map, matching the design of iOS's CTEntry in PR 393. Three constructors normalize the input to the same internal shape: from(CDAEntry) via the existing toOptimizedEntryMap; from(any: Map) with an org.json round-trip for JSON-safety normalization; from(json: String) via JSONTokener. Both Map- and JSON-based constructors accept a `fallback: CTEntry` (defaulting to empty), log a warning through DiagnosticLogger on a parse failure, and return the fallback instead of throwing. ResolvedOptimizedEntry.entry changes from Map<String, Any> to CTEntry. OptimizationClient.resolveOptimizedEntry wraps the baseline once and passes it as the fallback for the resolver-output parse — an unparseable bridge result now decays to the mapped baseline (baseline content renders) rather than an empty entry. Compose OptimizedEntry(CDAEntry, ...) now hands the content callback a CTEntry built from the resolver's output map (no more findVariantEntry graph walk; the CTEntry represents the winning variant regardless of whether it's reachable via nt_experiences → nt_variants). The base OptimizedEntry(Map, ...) overload is untouched — its Map-based callback still fires via `result.entry.toFoundation()`. Same story for OptimizedEntryView: internal state is CTEntry, public setContentRenderer signature stays (Map) -> View. CDAEntry.toOptimizedEntryMap is now internal (only CTEntry.from(CDAEntry) needs it). findVariantEntry is removed along with its 4 tests. New CTEntryTest (13 tests) covers accessors, three constructors, fallback behavior, and JSON round-trip. Validation: - unit tests: 70/70 pass - release build clean - both reference impl APKs assemble cleanly - Maestro E2E (Pixel_8_API_34): Compose 34/34 pass, Views 30+/34 pass before manual emulator shutdown aborted the remaining flows Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Removed section headers, restatements of what code says, and rationale prose that's redundant with clear method/test names. Kept only the "why" notes a future reader can't recover from the code alone: - ancestor cycle rationale on entryToMap - embedded-node-before-hyperlink ordering in richNodeToMap - CTEntry.from fallback docstring (one line each) - okhttp-jvm exclusion rationale - JVM generic-erasure gotcha in the getField test - reflection-fixture note explaining why we set private fields 70/70 tests still pass. -101 lines. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Updates documentation/internal/sdk-knowledge/native/android.md to reflect: - CTEntry added to the state/result types row with source pointer. - OptimizedEntry / OptimizedEntryView table rows note the CDAEntry overload and its CTEntry callback (Compose) or Map renderer (Views). - resolveOptimizedEntry fail-soft bullet: baseline is wrapped via CTEntry.from(baseline), the resolver-output parse passes the baseline CTEntry as fallback, and ResolvedOptimizedEntry.entry is a CTEntry. - selectedOptimizations semantics bullet mentions the setEntry(CDAEntry) overload alongside the Map one. - New CTEntry section documents the three factory methods, fail-soft behavior with DiagnosticLogger warnings, serializers, and mirrored CDAEntry accessors. - New OptimizedEntry(CDAEntry) / setEntry(CDAEntry) section documents the overload routing and the compileOnly contentful.java dep with the okhttp-jvm exclusion. pnpm knowledge:check passes; every new source pointer resolves. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Guides and the Android runtime concept doc referenced ResolvedOptimizedEntry.entry as a Map<String, Any>; it's now a CTEntry. Fixes the load-bearing prose and one broken assignment in the Compose guide's DirectResolution pattern, adds CDAEntry-overload notes to both guides, and clarifies the CTEntry shape in the concept doc. Compose guide: fix `resolvedEntry = ....entry` to `.entry.toFoundation()`; add a CDAEntry overload example alongside the base Map example; update the ResolvedOptimizedEntry description to name CTEntry and its accessors. Views guide: name CTEntry in the ResolvedOptimizedEntry description, mention the typed `setEntry(CDAEntry)` overload, unwrap `.entry` via `.toFoundation()` in the direct-resolution snippet. Concept doc: state that `.entry` is a CTEntry and reference its accessors + `toFoundation()` escape hatch. pnpm knowledge:check passes; format:check clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Addresses Daviti's PR feedback (#395 review comment). Mirrors iOS PR 393's identity-check style: rather than digging into the resulting Map field-by-field with `as?` casts, each test now names the entire expected `{sys, fields, metadata}` map and asserts against it with `assertEquals`. Kotlin's `Map.equals` is entry-set equal (key-order-insensitive) and `List.equals` is positional — same semantics as iOS's `JSONValue` tree comparison in `CTEntryTests.swift`. The upside: an unexpected extra key or shape drift anywhere in the tree now fails the test; before, only the handful of keys a test happened to name were checked. Adds a small set of expected-shape helpers (`sys(id, contentTypeId)`, `linkStub(id, linkType)`, `emptyMetadata`) so each test's literal reads as the shape it intends, not as boilerplate. 70/70 unit tests pass; release build clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

iOS PR 393's `CTEntry` holds a Codable `CDA.Entry` struct (sys/fields/metadata as typed properties) built once from `Contentful.Entry` at construction, not the raw dictionary — its docstring says "shares the resolved shape with Entry, not the type." This aligns Android on the same shape: replaces the internal `Map<String, Any>` with a typed `Envelope` data class (`sys: Sys?`, `fields: Map<String, Any>`, `metadata: Metadata?`), reconstructed from the resolver Map at `from(Map/JSON/CDAEntry)` and back to a Map via `toFoundation()` for consumers and the bridge. Public API is unchanged. `id`, `contentTypeId`, `localeCode`, `createdAt`, `updatedAt`, `getField<T>`, `hasField`, `operator get(name)`, `toFoundation`, `toJSON`, and the three `from(...)` factories all read/write the same shape they did before — internal accessors now walk typed properties instead of `as?` casts on nested Maps. Sys carries its explicit typed fields (id/type/contentTypeId/createdAt/ updatedAt/revision/locale) plus an `extras` bag so a resolver Map with e.g. `space`/`environment` link refs round-trips through the envelope unchanged. 70/70 unit tests pass; release build clean; both reference APKs (:compose:assembleDebug, :views:assembleDebug) still build clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Full delegation to contentful.java's CDAEntry: CTEntry stores a private val entry: CDAEntry and every accessor calls through (id → entry.id(); contentTypeId → entry.contentType()?.id(); localeCode/createdAt/updatedAt → entry.getAttribute(...); getField<T> → entry.getField(name); hasField → entry.rawFields().containsKey). from(entry: CDAEntry) is a thin wrapper; from(any: Map) and from(json) fabricate a CDAEntry via reflection so downstream reads always go through real contentful.java accessors. Public API unchanged. toFoundation()/toJSON() still use the SDK adapter so the resolver Map shape is preserved for the JS bridge. Tradeoff: production code now reflects into contentful.java's package-private fields (attrs, defaultLocale, rawFields, fields, contentType, metadata, tags, concepts) on the Map/JSON paths. Pinned to contentful.java 10.6.0; a future rename fails at runtime with a clear NoSuchFieldException. New Option A-specific tests: - from(entry) reads through the real CDAEntry accessors - fabricated CDAEntry from a Map reads through the real CDAEntry accessors (regression guard for the Localizer shape) - fabricated entry with a bad contentType Map falls back to null 73/73 unit tests pass; release build clean; both reference APKs (:compose:assembleDebug, :views:assembleDebug) still build clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Follow-up to `refactor(android): back CTEntry with a real CDAEntry (Option A)` — CTEntry is no longer an envelope wrapping the resolver Map; it wraps a real CDAEntry and delegates every accessor. Docs still called it "the SDK-owned typed view over the entry map." - android.md KB — the CTEntry section now describes the wrapper shape, the three factory methods (from(CDAEntry) trivial; from(Map)/from(json) fabricate a CDAEntry via reflection into package-private fields), and the pinned 10.6.0 dependency contract. - Compose + Views integration guides — reword the "typed view over the entry map" to "wrapper around a CDAEntry whose accessors delegate to it". No code sample changes. - Runtime concept doc — same reword. pnpm knowledge:check passes; all source pointers still resolve. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Drops all reflection into contentful.java's package-private fields. CTEntry now holds a nested `CTEntry.Entry` data class (SDK-owned, with typed Sys/Metadata/Link sub-classes) that Gson serializes natively. Three factories, all reflection-free: - from(entry: CDAEntry): walks the live CDAEntry via the existing toOptimizedEntryMap adapter and Gson-decodes the result into Entry. - from(any: Map): rejects self-referential input up-front via a hand-rolled ancestor-set cycle check (Gson's serializer would blow its stack otherwise), then round-trips through Gson. - from(json: String): decodes directly with gson.fromJson. EMPTY is a plain Entry(sys=null, fields=emptyMap(), metadata=null). fabricateEntry, fabricateContentType, fabricateMetadata, localizeFields, setPrivateField, FABRICATED_LOCALE and the org.json helper cluster all gone from production. Renames toFoundation() to toMap() to fit Kotlin idioms; toJSON pairs with toMap the way toList pairs with toSet. Four SDK call sites and five doc mentions updated to match. ISO-8601 formatting: switched Locale.US to Locale.ROOT in both CTEntry.kt and CDAEntryAdapter.kt, and wrapped CDAEntryAdapter's formatter in ThreadLocal (fixes a latent race where multiple coroutine dispatchers could hit the shared SimpleDateFormat). New JSON-identity tests exercise the full String → CTEntry → toJSON round-trip and compare parsed trees structurally (order-insensitive on objects, positional on arrays, numeric equality across Int/Double so Gson's default number widening doesn't false-positive). Validation: - 75/75 unit tests pass - SDK assembleRelease clean - Reference APKs (:compose:assembleDebug, :views:assembleDebug) clean - pnpm knowledge:check passes (1295 source pointers resolve) - Prettier clean on the four touched docs Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Prior docs described the CTEntry-wrapping-CDAEntry-with-reflection design. After the envelope refactor, CTEntry is backed by an SDK-owned `CTEntry.Entry` data class that Gson serializes natively — no reflection, no wrapping of a live CDAEntry. - android.md KB section rewritten: dropped the "fabricates CDAEntry via reflection" claim, described the SDK-owned Entry data class shape and the Gson round-trip, kept the `getField<T>` returns-Double-on-JSON-boundary note that iOS has too. Source pointer chain updated. - Compose + Views integration guides: reword "SDK-owned wrapper around a CDAEntry" to "SDK-owned view over the resolved entry" and list the full accessor surface (`id`, `getField<T>`, `hasField`, `contentTypeId`, `createdAt`, `updatedAt`, `localeCode`, `toMap()`). - Runtime concept doc: same reword. pnpm knowledge:check passes; Prettier clean. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

…y tests as JSON identity CTEntry now owns the full JSON round-trip (`from(CDAEntry)` / `from(Map)` / `from(json)` → `toMap()` / `toJSON()`), so the standalone `CDAEntryAdapter` and its test file are dead. Rewrite `CTEntryTest` around `assertJsonIdentity`: parse wire JSON, wrap in a `CTEntry`, serialize back, and compare parsed trees — with zero reflection, in production or tests. The `Entry.from(CDAEntry)` walk stays covered by the reference-implementation E2E, which is the only place a real `CDAEntry` graph exists. Docs: correct the SDK-knowledge entry to say `fields: Map<String, JSONValue>` and note the `JSONValueTypeAdapter` + reflection-free stance. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

CTEntry.kt initializes a top-level `Gson` instance and reads `com.google.gson.*` from a Kotlin file-scope val, so the class-init runs the first time anything touches CTEntry. contentful.java (which transitively brings Gson) is declared `compileOnly` so consumers who never touch CDAEntry aren't forced onto it — but that means Gson was absent from the app's runtime classpath, and the compose/views reference apps (both Map-only consumers) crashed with `java.lang.NoClassDefFoundError: com/google/gson/GsonBuilder` at CTEntry's <clinit> during first render, taking down every E2E flow that got past `fab-visible`. Add Gson as a direct `implementation` dep on the SDK so the runtime classpath has it regardless of whether the consumer imports contentful.java. contentful.java stays compileOnly. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

David Nalchevanidze (nalchevanidze) reviewed Jul 30, 2026

View reviewed changes

Comment thread ...ContentfulOptimization/src/main/kotlin/com/contentful/optimization/compose/OptimizedEntry.kt Outdated

David Nalchevanidze (nalchevanidze) reviewed Jul 30, 2026

View reviewed changes

Comment thread ...entfulOptimization/src/main/kotlin/com/contentful/optimization/contentful/CDAEntryAdapter.kt Outdated

David Nalchevanidze (nalchevanidze) reviewed Jul 30, 2026

View reviewed changes

Comment thread ...ntfulOptimization/src/main/kotlin/com/contentful/optimization/core/ResolvedOptimizedEntry.kt Outdated

David Nalchevanidze (nalchevanidze) reviewed Jul 30, 2026

View reviewed changes

Comment thread ...ntentfulOptimization/src/main/kotlin/com/contentful/optimization/views/OptimizedEntryView.kt Outdated

David Nalchevanidze (nalchevanidze) reviewed Jul 31, 2026

View reviewed changes

Comment thread ...entfulOptimization/src/main/kotlin/com/contentful/optimization/contentful/CDAEntryAdapter.kt Outdated

David Nalchevanidze (nalchevanidze) reviewed Jul 31, 2026

View reviewed changes

Comment thread ...ContentfulOptimization/src/main/kotlin/com/contentful/optimization/compose/OptimizedEntry.kt Outdated

Felipe Mamud (fmamud) requested a review from David Nalchevanidze (nalchevanidze) July 31, 2026 12:37

Felipe Mamud (fmamud) marked this pull request as ready for review July 31, 2026 14:01

Felipe Mamud (fmamud) requested review from Lotfi Anwar L Arif (Lotfi-Arif), Tim Beyer (TimBeyer) and Charles Hudson (phobetron) as code owners July 31, 2026 14:01

David Nalchevanidze (nalchevanidze) reviewed Aug 3, 2026

View reviewed changes

Comment thread ...ContentfulOptimization/src/test/kotlin/com/contentful/optimization/contentful/CTEntryTest.kt

David Nalchevanidze (nalchevanidze) reviewed Aug 3, 2026

View reviewed changes

Comment thread ...oid/ContentfulOptimization/src/main/kotlin/com/contentful/optimization/contentful/CTEntry.kt Outdated

Felipe Mamud (fmamud) force-pushed the NT-3807-contentful-client-compat branch from cb2e053 to 2454d28 Compare August 3, 2026 08:06

Felipe Mamud (fmamud) requested a review from David Nalchevanidze (nalchevanidze) August 3, 2026 09:17

Felipe Mamud (fmamud) requested a review from Elton Muteswa (elton-muteswa-cf) as a code owner August 3, 2026 12:14

Felipe Mamud (fmamud) force-pushed the NT-3807-contentful-client-compat branch from 4afcb3f to f35b8b9 Compare August 3, 2026 12:15

Felipe Mamud (fmamud) changed the title ~~fix(android): add contentful.java CDAEntry adapter for OptimizedEntry [NT-3807]~~ fix(android): add contentful.java Entry adapter for OptimizedEntry [NT-3807] Aug 3, 2026

David Nalchevanidze (nalchevanidze) previously approved these changes Aug 3, 2026

View reviewed changes

Felipe Mamud (fmamud) dismissed David Nalchevanidze (nalchevanidze)’s stale review via edff3ee August 3, 2026 14:09

Charles Hudson (phobetron) approved these changes Aug 3, 2026

View reviewed changes

David Nalchevanidze (nalchevanidze) approved these changes Aug 3, 2026

View reviewed changes

Felipe Mamud (fmamud) and others added 5 commits August 3, 2026 16:29

Felipe Mamud (fmamud) and others added 11 commits August 3, 2026 16:29

Felipe Mamud (fmamud) force-pushed the NT-3807-contentful-client-compat branch from d90fe0f to 76f6645 Compare August 3, 2026 14:29

Felipe Mamud (fmamud) merged commit 525c55a into main Aug 3, 2026
67 of 68 checks passed

Felipe Mamud (fmamud) deleted the NT-3807-contentful-client-compat branch August 3, 2026 14:50

Charles Hudson (phobetron) mentioned this pull request Aug 3, 2026

chore(repo): release packages for main #397

Open

Felipe Mamud (fmamud) mentioned this pull request Aug 3, 2026

fix(android): declare okhttp-android runtime dep and add CTEntry.imageURL [NT-3807] #406

Merged

6 tasks

Navigation Menu

Uh oh!

fix(android): add contentful.java Entry adapter for OptimizedEntry [NT-3807] - #395

fix(android): add contentful.java Entry adapter for OptimizedEntry [NT-3807]#395
Felipe Mamud (fmamud) merged 16 commits into
mainfrom
NT-3807-contentful-client-compat

Felipe Mamud (fmamud) commented Jul 30, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

Felipe Mamud (fmamud) commented Jul 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

What ships

Tests

Notes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Felipe Mamud (fmamud) commented Jul 30, 2026 •

edited

Loading