URGENT ADFA-3604: Fix R8 shrink/optimize crashes live on stage (#1609)

ADFA-3604: Fix R8 shrink/optimize bugs found in on-device testing

On-device verification of the shrunk release build (requested in review)
surfaced five distinct R8 bugs, each fixed here:

1. R8's optimize pass retargeted a call to CoGo's own isTestMode() to the
   unrelated, device-missing dalvik.system.VMRuntime.isTestMode, crashing
   every release build on launch. Disabled optimization (-dontoptimize);
   shrinking is what actually cuts dex size, optimization was the unsafe
   part given the R8/Kotlin 2.3.0 metadata mismatch already on record in
   this file (see the existing StopWatch workaround).

2. Protobuf-lite resolves generated message fields by name via reflection;
   shrinking removed a "field with no direct bytecode reference"
   (SyncMetaModels$SyncMeta.projectModelInfo_), crashing project sync.

3-5. lsp/kotlin's own service registrations (kt-lsp.xml and
   AnalysisApiServiceProviders.kt's ::class-literal registrar, which
   PicoContainer instantiates reflectively) and Caffeine's runtime cache
   implementation selection all hit the same class of bug: a reflective
   lookup R8 can't trace from static analysis. Each fix revealed another
   instance elsewhere in the same dependency graph, so rather than keep
   discovering them one on-device crash at a time, keep the whole
   lsp/kotlin runtime dependency graph (org.jetbrains.kotlin, Caffeine,
   kotlin-reflect, kotlin-script, coroutines internals, streamex, Trove)
   whole.

Verified end to end on a physical device: clean Kotlin analysis session
init, and a live diagnostic ("Expecting member declaration") confirming
the analysis engine correctly parses and resolves code post-shrink.

Net result: release dex settles at ~85 MB (down from the 119.4 MB
baseline), short of the initially-measured 28.8 MB because that number
came from a build that silently corrupted the Kotlin LSP -- this is the
verified-correct number.

Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>

Author: davidschachterADFA

app/proguard-rules.pro

@@ -4,6 +4,16 @@
 -dontnote **
 -dontobfuscate
 
+# ADFA-3604: R8's optimize pass (inlining/method merging) can silently
+# retarget a call to the wrong method when it fails to parse Kotlin 2.3.0
+# metadata (see "R8: An error occurred when parsing kotlin metadata" in the
+# build log) -- observed corrupting calls to utils.TestModeUtilsKt.isTestMode
+# into calls to the unrelated, device-missing dalvik.system.VMRuntime.isTestMode,
+# crashing every release build on launch. Shrinking (dead-code removal) is
+# what actually cuts dex size; the optimize passes are the risky part given
+# this R8/Kotlin version mismatch, so disable only optimization.
+-dontoptimize
+
 -keep class javax.** { *; }
 -keep class jdkx.** { *; }
 
@@ -19,6 +29,42 @@
 # Builder model implementations
 -keep class com.itsaky.androidide.builder.model.** { *; }
 
+# lsp/kotlin registers its own IntelliJ project/application services -- some
+# by class name in lsp/kotlin/src/main/resources/META-INF/kt-lsp/kt-lsp.xml,
+# the rest via ::class literals in
+# lsp/kotlin/.../registrar/AnalysisApiServiceProviders.kt, which PicoContainer
+# then instantiates reflectively via each class's no-arg constructor. A
+# ::class literal doesn't count as an actual `new` call to R8, so it kept
+# stripping "unused" no-arg constructors one class at a time as each was
+# discovered on-device (ClassNotFoundException on DirectInheritorsProvider,
+# then a PicoInitializationException on ModuleDependentsProvider's missing
+# constructor). Keep the whole package rather than list every implementation
+# class in AnalysisApiServiceProviders.kt individually.
+-keep class com.itsaky.androidide.lsp.kotlin.compiler.services.** { *; }
+
+# Kotlin Analysis API (bundled in subprojects/kotlin-analysis-api, used by the
+# Kotlin LSP). subprojects/kotlin-analysis-api/consumer-rules.pro keeps every
+# class this jar's own IntelliJ plugin XML descriptors and ServiceLoader
+# entries reference by name, but on-device testing kept surfacing distinct
+# reflection paths that narrow list didn't cover -- not just inside this jar,
+# but in other lsp/kotlin runtime dependencies too (Caffeine picks a cache
+# implementation from dozens of codegenned variant classes at runtime; a
+# protobuf-lite message field is resolved by name string; lsp/kotlin's own
+# kt-lsp.xml above; and a NullPointerException deep in IntelliJ's own
+# JavaCoreApplicationEnvironment bootstrap, verified absent on an unshrunk
+# debug build). Each fix was quick but the next gap kept appearing elsewhere
+# in the same dependency graph, so rather than keep discovering them one
+# on-device crash at a time, keep every runtime dependency lsp/kotlin pulls in
+# whole. This gives up the dex-size reduction for this whole dependency graph;
+# see ADFA-3604 for the size trade-off.
+-keep class org.jetbrains.kotlin.** { *; }
+-keep class com.github.benmanes.caffeine.** { *; }
+-keep class kotlin.reflect.** { *; }
+-keep class kotlin.script.** { *; }
+-keep class kotlinx.coroutines.internal.** { *; }
+-keep class one.util.streamex.** { *; }
+-keep class gnu.trove.** { *; }
+
 # Eclipse
 -keep class org.eclipse.** { *; }
 
@@ -88,6 +134,14 @@
 
 -keep class com.itsaky.androidide.treesitter.** { *; }
 
+# Protobuf lite: the generated runtime resolves message fields by name via
+# reflection (the info string baked into newMessageInfo()), so shrinking a
+# "field with no direct bytecode reference" breaks it at runtime with a
+# NoSuchFieldException (observed on SyncMetaModels$SyncMeta.projectModelInfo_).
+-keepclassmembers class * extends com.google.protobuf.GeneratedMessageLite {
+    <fields>;
+}
+
 # Retrofit 2
 -dontwarn retrofit2.**
 -keep class retrofit2.** { *; }