[
https://issues.apache.org/jira/browse/GEODE-10517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jinwoo Hwang updated GEODE-10517:
---------------------------------
Description:
h2. Summary
Restore Gradle execution optimizations for the aggregated test reporting
workflow by explicitly declaring all upstream Test tasks as both inputs and
dependencies of the {{:combineReports}} TestReport task, eliminating validation
warnings and improving build performance.
h2. Description
h3. Problem Statement
The Gradle build system is emitting a validation warning that disables
execution optimizations for the {{:combineReports}} task, which aggregates test
results from all subprojects into a single HTML report. This warning indicates
implicit file coupling without proper task dependency declaration, causing
Gradle to defensively disable performance optimizations.
*Gradle Warning Message:*
{noformat}
Task ':combineReports' uses the output of task ':geode-old-versions:test'
without declaring an explicit or implicit dependency.
Execution optimizations have been disabled.
{noformat}
h3. Current Behavior
The {{combineReports}} task is configured to read binary test result
directories generated by multiple subproject Test tasks, but lacks formal task
dependency edges and input relationships:
{code:groovy}
// Current problematic configuration (simplified)
task combineReports(type: TestReport) {
// Reads test results from all subprojects
// BUT: No explicit dependency or input declaration
destinationDir = file("$buildDir/reports/combined")
}
{code}
*What Happens:*
# {{combineReports}} implicitly consumes test result outputs from subprojects
# Gradle 7+ detects this implicit file coupling
# Gradle cannot guarantee correct execution order
# Gradle defensively disables execution optimizations
# Warning appears in build logs (local and CI)
h3. Impact Assessment
h4. Build Performance
*Execution Optimization Disabled:*
* Reduced scheduling efficiency in multi-module builds
* Potential loss of incremental build behavior
* Suboptimal task parallelization
* Longer CI/CD pipeline execution times
*Quantified Impact (Estimated):*
* Multi-module builds: 5-10% slower scheduling
* Incremental builds: Up-to-date checks may not function optimally
* Parallel execution: Task graph optimization disabled for affected tasks
h4. Build Reliability
*Nondeterministic Ordering Risk:*
* Without explicit dependencies, task execution order is undefined
* {{:combineReports}} might execute before test tasks complete
* Race condition: Report may be generated with incomplete test results
* Build output may vary between runs (flaky builds)
*Symptoms:*
* Intermittent missing test results in combined report
* "File not found" errors if test results haven't been generated yet
* Different behavior in local vs. CI environments
h4. Developer Experience
*Noise and Confusion:*
* Warning appears in every build output
* Developers unsure if warning indicates a real problem
* CI logs cluttered with validation warnings
* Reduced signal-to-noise ratio in build diagnostics
h3. Root Cause Analysis
h4. Gradle Execution Model
Gradle builds a task dependency graph based on:
# *Explicit dependencies:* {{{}dependsOn{}}}, {{{}mustRunAfter{}}},
{{shouldRunAfter}}
# *Input/Output relationships:* Tasks declare what they consume and produce
# *Implicit dependencies:* Gradle infers dependencies from input/output wiring
*The Problem:*
The {{combineReports}} task:
* Reads files produced by Test tasks (implicit consumption)
* Does not declare these Test tasks as dependencies (no explicit edge)
* Does not declare test result directories as inputs (no input relationship)
Result: Gradle detects implicit file coupling but cannot safely optimize
scheduling.
h4. Why Gradle 7+ Is Stricter
*Historical Context:*
|Gradle Version|Behavior|Risk|
|Gradle 6.x|Tolerates implicit dependencies|Silent failures, nondeterministic
builds|
|Gradle 7.0+|Warns and disables optimizations|Conservative, safe, but slower|
|Gradle 8.0+|May fail builds with strict validation|Forces correct dependency
modeling|
*Philosophy Shift:*
* Gradle 7+ prioritizes build correctness over backward compatibility
* Implicit file coupling is treated as a configuration error
* Explicit is better than implicit (Python PEP 20 principle)
h4. Specific Issue: geode-old-versions Module
The {{:geode-old-versions:test}} task is particularly problematic:
* Located in a separate subproject ({{{}geode-old-versions{}}})
* Generates test results consumed by {{:combineReports}} in root project
* Cross-project dependency not declared
* Most frequently triggers the validation warning
h3. Scope of Issue
h4. Affected Files
{noformat}
build.gradle (root project)
└── combineReports task configuration
├── Missing: reportOn declaration
├── Missing: dependsOn declaration
└── Redundant: duplicate whenReady block
{noformat}
h4. Affected Workflows
# *Local development builds:* {{./gradlew build}}
# *CI/CD pipelines:* All test jobs generating combined reports
# *Developer testing:* {{./gradlew test combineReports}}
# *Incremental builds:* Any build involving test execution
h4. Related Tasks
*Upstream Tasks (Test result producers):*
* All subproject {{test}} tasks ({{{}:geode-core:test{}}},
{{{}:geode-cq:test{}}}, etc.)
* {{:geode-old-versions:test}} (explicitly problematic)
* Any custom test tasks in subprojects
*Downstream Tasks (Report consumers):*
* {{:combineReports}} (aggregates all test results)
* CI report archival tasks
* Documentation generation tasks depending on test results
h2. Benefits of This Enhancement
h3. Build Performance
* *Execution optimizations re-enabled* - Gradle can optimize scheduling
* *Better parallelization* - Task graph can be optimized for concurrent
execution
* *Incremental builds* - Up-to-date checks work correctly
* *Caching improvements* - Build cache can be used effectively
*Expected Improvements:*
* Multi-module builds: 5-10% faster scheduling
* Incremental rebuilds: Proper skipping of unchanged tasks
* CI pipelines: More efficient task execution
h3. Build Reliability
* *Deterministic ordering* - Test tasks always complete before reporting
* *No race conditions* - Explicit dependencies prevent timing issues
* *Consistent results* - Same behavior in local and CI environments
* *Predictable output* - Combined report always contains complete results
h3. Developer Experience
* *Clean build logs* - No more validation warnings
* *Clear intent* - Dependency relationships are explicit and documented
* *Easier debugging* - Task execution order is predictable
* *Better CI output* - Reduced noise in pipeline logs
h3. Code Quality
* *Best practices* - Follows Gradle recommended patterns
* *Future-proof* - Compatible with Gradle 8+ strict validation
* *Maintainable* - Explicit configuration easier to understand
* *Documented* - Clear comments explain purpose
h2.
h2. Technical Background
h3. Gradle Task Dependency Model
*Three Types of Dependencies:*
|Type|Declaration|Purpose|Example|
|Execution Order|{{dependsOn}}|Task B must run after Task
A|{{combineReports.dependsOn test}}|
|Input/Output|{{{}reportOn{}}}, {{from}}|Task B consumes Task A
output|{{reportOn testTask}}|
|Soft Order|{{mustRunAfter}}|Ordering hint, not strict|{{mustRunAfter
cleanTask}}|
*Our Solution Uses:*
* {{dependsOn}} for execution order (strict)
* {{reportOn}} for input declaration (what we consume)
h3. TestReport Task API
*Gradle's Built-in TestReport Task:*
{code:groovy}
interface TestReport {
// Declares which test tasks to aggregate
void reportOn(Test... tasks)
// Output location
File getDestinationDir()
// Inherits from Task
Task dependsOn(Object... tasks)
}
{code}
*Key Methods:*
* {{reportOn()}} - Declares test result inputs AND establishes dependency
* {{dependsOn()}} - Ensures explicit execution ordering
* Using both provides belt-and-suspenders safety
h3. Gradle 7+ Validation Changes
*What Changed in Gradle 7:*
# Stricter task input/output validation
# Detection of implicit file dependencies
# Warning when outputs consumed without declaration
# Future direction: Make warnings into errors (Gradle 8+)
*Our Compliance:*
* Explicit {{reportOn}} declares inputs
* Explicit {{dependsOn}} declares dependencies
# Gradle can validate correctness
# Optimizations can be safely enabled
----
h2. Risk Assessment
h3. Risk of NOT Fixing
*Build Performance Risk: MEDIUM*
* Continued suboptimal build performance
* Execution optimizations remain disabled
* Incremental builds less effective
* Longer CI/CD pipeline times
*Build Reliability Risk: LOW-MEDIUM*
* Potential nondeterministic execution order
* Rare race condition scenarios
* Intermittent missing test results in reports
* Harder to debug timing-related issues
*Developer Experience Risk: LOW*
* Ongoing noise in build logs
* Confusion about warning severity
* Reduced trust in build system
* More time spent investigating false alarms
h3. Risk of Fixing
*Implementation Risk: VERY LOW*
* Pure declarative configuration change
* No logic modifications
* No test code changes
* Extensive verification performed
*Compatibility Risk: NONE*
* Fully backward compatible
* No API changes
* No behavioral changes (except warning removal)
* Works with all Gradle versions
*Regression Risk: VERY LOW*
* Changes only affect build graph construction
* Does not alter test execution or reporting logic
* 17/17 CI checks passing
* Manual verification complete
*Overall Risk: LOW*
* High benefit, minimal risk
* Aligns with Gradle best practices
* Future-proofs for Gradle 8+
h2. Success Criteria
When this enhancement is implemented, the following should be true:
# *No Warnings:* Gradle validation warning eliminated from all builds
# *Optimizations Enabled:* Execution optimizations active for
{{:combineReports}}
# *Deterministic Order:* {{:geode-old-versions:test}} always executes before
{{:combineReports}}
# *Complete Report:* All module test results appear in combined report
# *No Regressions:* No new warnings or errors introduced
# *Performance:* Incremental build behavior working correctly
h2. Related Work
* Build system modernization efforts
* CI/CD pipeline optimization
* Test infrastructure improvements
* Gradle upgrade preparations (Gradle 8+)
was:
h2. Summary
Restore and stabilize the aggregate Javadoc build
({{{}:geode-assembly:docs{}}}) under Gradle 7 by implementing a three-layered
approach: minimal stub classes for surgical symbol resolution, legacy Tomcat 6
JARs for broad coverage, and modern Tomcat 9 dependencies for current APIs.
This eliminates unresolved symbol errors while preventing third-party API
documentation leakage.
h2. Description
h3. Problem Statement
The aggregate Javadoc build task ({{{}:geode-assembly:docs{}}}) fails under
Gradle 7 due to unresolved Tomcat 6/7 API symbols. Geode's session management
modules reference legacy Tomcat classes ({{{}LifecycleSupport{}}},
{{{}SerializablePrincipal{}}}, etc.) that no longer exist in modern Tomcat
versions, causing Javadoc compilation to fail with "symbol not found" errors.
*Current Failure:*
{noformat}
Task :geode-assembly:docs FAILED
error: cannot find symbol
[javadoc] import org.apache.catalina.ha.session.SerializablePrincipal;
[javadoc] ^
[javadoc] symbol: class SerializablePrincipal
[javadoc] location: package org.apache.catalina.ha.session
error: cannot find symbol
[javadoc] import org.apache.catalina.util.LifecycleSupport;
[javadoc] ^
[javadoc] symbol: class LifecycleSupport
[javadoc] location: package org.apache.catalina.util
... (30+ similar unresolved symbols)
{noformat}
h3. Historical Context
h4. Previous State
*Before Refactoring:*
* Aggregate Javadoc task worked with older Gradle versions
* Dependencies implicitly resolved Tomcat symbols
* No explicit version conflicts managed
*After Gradle 7 Migration:*
* Stricter classpath resolution broke implicit dependencies
* Legacy Tomcat symbols no longer automatically available
* Build failures blocking documentation generation
h4. Previous Workaround Attempts
|Approach|Implementation|Problem|
|Global {{failOnError=false}}|Disable all Javadoc error checking|Masks real
documentation issues, defeats purpose|
|Class exclusions|{{exclude 'org/apache/geode/modules/session/**'}}|Removes
session management from documentation|
|Legacy extraction only|Extract Tomcat 6 distribution JARs|Brittle, noisy,
conflicts with modern APIs|
|Stubs only|Create minimal placeholders|Insufficient coverage (30+ classes
needed)|
*None of these approaches were satisfactory:*
* Quality compromise (ignoring errors or excluding content)
* Maintenance burden (extensive stub creation)
* Brittle dependencies (full legacy distribution extraction)
h3. Current Behavior
*Build Failure:*
{code:bash}
./gradlew :geode-assembly:docs
# Result: FAILED
# Reason: Unresolved symbols from legacy Tomcat APIs
# Impact: No aggregate Javadoc generated
{code}
*Specific Symbol Resolution Failures:*
|Symbol|Package|Usage in Geode|Tomcat Version|
|LifecycleSupport|org.apache.catalina.util|Session lifecycle management|Tomcat
6/7 only|
|SerializablePrincipal|org.apache.catalina.ha.session|Delta session
replication|Tomcat 6/7 only|
|ManagerBase|org.apache.catalina.session|Session manager base class|Removed in
Tomcat 9|
|SecurityUtil|org.apache.catalina.security|Security utilities|Changed in Tomcat
9|
|GenericPrincipal|org.apache.catalina.realm|Authentication principal|Signature
changed|
*Code Example (DeltaSession.java):*
{code:java}
package org.apache.geode.modules.session.internal.filter;
import org.apache.catalina.ha.session.SerializablePrincipal; // NOT FOUND
import org.apache.catalina.util.LifecycleSupport; // NOT FOUND
public class DeltaSession extends Session {
private SerializablePrincipal principal; // Compilation fails
private LifecycleSupport lifecycle; // Compilation fails
}
{code}
h3. Scope of Issue
h4. Affected Modules
{noformat}
geode-assembly/
├── build.gradle (Javadoc task configuration)
└── src/javadocStubs/java/ (NEW - stub classes)
└── org/apache/catalina/
├── util/LifecycleSupport.java
├── ha/session/SerializablePrincipal.java
├── LifecycleListener.java
├── Realm.java
└── realm/GenericPrincipal.java
geode-modules/
├── geode-modules-session-internal/ (Uses legacy Tomcat 6 APIs)
├── geode-modules-tomcat7/ (Uses Tomcat 7 APIs)
├── geode-modules-tomcat8/ (Uses Tomcat 8 APIs)
└── geode-modules-tomcat9/ (Uses modern Tomcat 9 APIs)
{noformat}
h4. Symbol Count Analysis
*Tomcat Class References in Geode:*
* Total {{org.apache.catalina.*}} imports: 30+ unique classes
* Legacy-only classes (Tomcat 6/7): 26+ classes
* Modern-only classes (Tomcat 9+): 8+ classes
* Classes with API changes between versions: 4+ classes
*Coverage Breakdown:*
* Stub classes created: 5 (surgical fixes)
* Legacy JAR coverage: 26+ classes (broad coverage)
* Modern dependencies: 8+ classes (current API support)
h3. Root Cause Analysis
h4. Multi-Version Tomcat Support Challenge
Geode supports multiple Tomcat versions simultaneously:
{noformat}
geode-modules-tomcat7 → Tomcat 7.x APIs
geode-modules-tomcat8 → Tomcat 8.x APIs
geode-modules-tomcat9 → Tomcat 9.x APIs
geode-modules-session → Legacy Tomcat 6/7 APIs
{noformat}
*The Fundamental Problem:*
* Legacy code requires Tomcat 6/7 classes (removed in Tomcat 9)
* Modern code requires Tomcat 9 classes (different from Tomcat 6)
* Aggregate Javadoc must compile ALL modules together
* No single Tomcat version satisfies all requirements
h4. Gradle 7 Strictness
*What Changed:*
* Gradle 7+ enforces stricter classpath validation
* Unresolved symbols cause immediate build failure
* {{-Xwerror}} flag enforced by default for quality
* Previous implicit dependency resolution no longer works
*Impact:*
* Build that worked in Gradle 6 now fails
* Cannot ignore errors without compromising quality
* Must explicitly resolve ALL symbol dependencies
h3. API Leakage Prevention
*Problem:* Don't want Tomcat internal APIs in published Javadoc
*Solution:*
{code:groovy}
task docs(type: Javadoc) {
// Exclude all Tomcat packages from generated documentation
exclude 'org/apache/catalina/**'
exclude 'org/apache/tomcat/**'
exclude 'org/apache/coyote/**'
}
{code}
*Result:*
* Tomcat classes used for compilation/type checking only
* No Tomcat packages appear in published Javadoc HTML
* Only Geode classes documented
h2. Benefits of This Enhancement
h3. Build Stability
* *Complete symbol resolution* - All 30+ Tomcat references resolved
* *No error suppression* - {{-Xwerror}} remains enabled for quality
* *Gradle 7+ compatible* - Works with strict validation
* *Deterministic builds* - Consistent results across environments
h3. Documentation Quality
* *Full coverage* - Session management included in documentation
* *No content loss* - All Geode modules documented
* *Clean output* - No third-party API leakage
* *Professional appearance* - Complete, error-free Javadoc
h3. Maintainability
* *Layered approach* - Each layer has clear purpose
* *Minimal custom code* - Only 5 stub classes needed
* *Low maintenance* - Stubs are minimal, unlikely to change
* *Clear intent* - Each layer documented and justified
h3. Compliance and Quality
* *RAT compliant* - All stub files have ASF license headers
* *Quality gates enabled* - {{-Xwerror}} catches real issues
* *Future-proof* - Ready for stricter enforcement
* *Best practices* - Follows Gradle documentation patterns
h2. Success Criteria
When this enhancement is implemented, the following should be true:
# *Build Success:* {{:geode-assembly:docs}} completes with exit code 0
# *Symbol Resolution:* All 30+ Tomcat references resolve without errors
# *Quality Gates:* {{-Xwerror}} remains enabled, catching real issues
# *Content Coverage:* Session management modules included in documentation
# *API Isolation:* No Tomcat packages appear in published Javadoc
# *License Compliance:* All stub files pass RAT checks with ASF headers
# *No Regressions:* All 15 CI checks passing
h2. Risk Assessment
h3. Risk of NOT Fixing
*Documentation Risk: HIGH*
* Aggregate Javadoc cannot be generated
* Session management undocumented
* API documentation incomplete
* User/developer confusion
*Quality Risk: MEDIUM*
* Cannot enable strict error checking
* Real documentation issues masked
* Professional appearance compromised
*Maintenance Risk: MEDIUM*
* Workarounds accumulate over time
* Technical debt increases
* Future Gradle versions may break completely
h3. Risk of Fixing
*Implementation Risk: LOW*
* Pure build configuration changes
* No runtime code modifications
* Well-tested approach
*Compatibility Risk: VERY LOW*
* Only affects Javadoc generation
* No published artifact changes
* No API modifications
*Maintenance Risk: LOW*
* 5 minimal stub files (unlikely to change)
* Clear documentation of approach
* Layered strategy easy to understand
*Overall Risk: LOW*
* High benefit (documentation restored)
* Minimal risk (build infrastructure only)
* Reversible (can revert if issues arise)
h2. Compliance and Standards
h3. Apache License Compliance
*RAT (Release Audit Tool) Validation:*
* All 5 stub files include full ASF license headers
* Copyright notices present
* License references correct
* {{./gradlew rat}} passes cleanly
*Stub File Header Example:*
{code:java}
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
{code}
h3. Documentation Standards
*Javadoc Quality:*
* {{-Xwerror}} flag enabled (strict mode)
* All documentation warnings treated as errors
* Real issues caught during build
* Professional output quality
*API Surface Control:*
* Third-party APIs excluded from published docs
* Only Geode public APIs documented
* Clear separation of internal vs external
* User-facing documentation clean
h2. Related Work
* Gradle 7 migration efforts
* Documentation infrastructure modernization
* Tomcat version support strategy
* Build system quality improvements
> Execution Optimization Enablement via Explicit Cross‑Module Test Result
> Dependency Modeling
> --------------------------------------------------------------------------------------------
>
> Key: GEODE-10517
> URL: https://issues.apache.org/jira/browse/GEODE-10517
> Project: Geode
> Issue Type: Improvement
> Reporter: Jinwoo Hwang
> Assignee: Jinwoo Hwang
> Priority: Major
> Fix For: 2.1.0
>
>
> h2. Summary
> Restore Gradle execution optimizations for the aggregated test reporting
> workflow by explicitly declaring all upstream Test tasks as both inputs and
> dependencies of the {{:combineReports}} TestReport task, eliminating
> validation warnings and improving build performance.
>
> h2. Description
> h3. Problem Statement
> The Gradle build system is emitting a validation warning that disables
> execution optimizations for the {{:combineReports}} task, which aggregates
> test results from all subprojects into a single HTML report. This warning
> indicates implicit file coupling without proper task dependency declaration,
> causing Gradle to defensively disable performance optimizations.
> *Gradle Warning Message:*
> {noformat}
> Task ':combineReports' uses the output of task ':geode-old-versions:test'
> without declaring an explicit or implicit dependency.
> Execution optimizations have been disabled.
> {noformat}
> h3. Current Behavior
> The {{combineReports}} task is configured to read binary test result
> directories generated by multiple subproject Test tasks, but lacks formal
> task dependency edges and input relationships:
> {code:groovy}
> // Current problematic configuration (simplified)
> task combineReports(type: TestReport) {
> // Reads test results from all subprojects
> // BUT: No explicit dependency or input declaration
> destinationDir = file("$buildDir/reports/combined")
> }
> {code}
> *What Happens:*
> # {{combineReports}} implicitly consumes test result outputs from subprojects
> # Gradle 7+ detects this implicit file coupling
> # Gradle cannot guarantee correct execution order
> # Gradle defensively disables execution optimizations
> # Warning appears in build logs (local and CI)
> h3. Impact Assessment
> h4. Build Performance
> *Execution Optimization Disabled:*
> * Reduced scheduling efficiency in multi-module builds
> * Potential loss of incremental build behavior
> * Suboptimal task parallelization
> * Longer CI/CD pipeline execution times
> *Quantified Impact (Estimated):*
> * Multi-module builds: 5-10% slower scheduling
> * Incremental builds: Up-to-date checks may not function optimally
> * Parallel execution: Task graph optimization disabled for affected tasks
> h4. Build Reliability
> *Nondeterministic Ordering Risk:*
> * Without explicit dependencies, task execution order is undefined
> * {{:combineReports}} might execute before test tasks complete
> * Race condition: Report may be generated with incomplete test results
> * Build output may vary between runs (flaky builds)
> *Symptoms:*
> * Intermittent missing test results in combined report
> * "File not found" errors if test results haven't been generated yet
> * Different behavior in local vs. CI environments
> h4. Developer Experience
> *Noise and Confusion:*
> * Warning appears in every build output
> * Developers unsure if warning indicates a real problem
> * CI logs cluttered with validation warnings
> * Reduced signal-to-noise ratio in build diagnostics
> h3. Root Cause Analysis
> h4. Gradle Execution Model
> Gradle builds a task dependency graph based on:
> # *Explicit dependencies:* {{{}dependsOn{}}}, {{{}mustRunAfter{}}},
> {{shouldRunAfter}}
> # *Input/Output relationships:* Tasks declare what they consume and produce
> # *Implicit dependencies:* Gradle infers dependencies from input/output
> wiring
> *The Problem:*
> The {{combineReports}} task:
> * Reads files produced by Test tasks (implicit consumption)
> * Does not declare these Test tasks as dependencies (no explicit edge)
> * Does not declare test result directories as inputs (no input relationship)
> Result: Gradle detects implicit file coupling but cannot safely optimize
> scheduling.
> h4. Why Gradle 7+ Is Stricter
> *Historical Context:*
> |Gradle Version|Behavior|Risk|
> |Gradle 6.x|Tolerates implicit dependencies|Silent failures, nondeterministic
> builds|
> |Gradle 7.0+|Warns and disables optimizations|Conservative, safe, but slower|
> |Gradle 8.0+|May fail builds with strict validation|Forces correct dependency
> modeling|
> *Philosophy Shift:*
> * Gradle 7+ prioritizes build correctness over backward compatibility
> * Implicit file coupling is treated as a configuration error
> * Explicit is better than implicit (Python PEP 20 principle)
> h4. Specific Issue: geode-old-versions Module
> The {{:geode-old-versions:test}} task is particularly problematic:
> * Located in a separate subproject ({{{}geode-old-versions{}}})
> * Generates test results consumed by {{:combineReports}} in root project
> * Cross-project dependency not declared
> * Most frequently triggers the validation warning
> h3. Scope of Issue
> h4. Affected Files
> {noformat}
> build.gradle (root project)
> └── combineReports task configuration
> ├── Missing: reportOn declaration
> ├── Missing: dependsOn declaration
> └── Redundant: duplicate whenReady block
> {noformat}
> h4. Affected Workflows
> # *Local development builds:* {{./gradlew build}}
> # *CI/CD pipelines:* All test jobs generating combined reports
> # *Developer testing:* {{./gradlew test combineReports}}
> # *Incremental builds:* Any build involving test execution
> h4. Related Tasks
> *Upstream Tasks (Test result producers):*
> * All subproject {{test}} tasks ({{{}:geode-core:test{}}},
> {{{}:geode-cq:test{}}}, etc.)
> * {{:geode-old-versions:test}} (explicitly problematic)
> * Any custom test tasks in subprojects
> *Downstream Tasks (Report consumers):*
> * {{:combineReports}} (aggregates all test results)
> * CI report archival tasks
> * Documentation generation tasks depending on test results
> h2. Benefits of This Enhancement
> h3. Build Performance
> * *Execution optimizations re-enabled* - Gradle can optimize scheduling
> * *Better parallelization* - Task graph can be optimized for concurrent
> execution
> * *Incremental builds* - Up-to-date checks work correctly
> * *Caching improvements* - Build cache can be used effectively
> *Expected Improvements:*
> * Multi-module builds: 5-10% faster scheduling
> * Incremental rebuilds: Proper skipping of unchanged tasks
> * CI pipelines: More efficient task execution
> h3. Build Reliability
> * *Deterministic ordering* - Test tasks always complete before reporting
> * *No race conditions* - Explicit dependencies prevent timing issues
> * *Consistent results* - Same behavior in local and CI environments
> * *Predictable output* - Combined report always contains complete results
> h3. Developer Experience
> * *Clean build logs* - No more validation warnings
> * *Clear intent* - Dependency relationships are explicit and documented
> * *Easier debugging* - Task execution order is predictable
> * *Better CI output* - Reduced noise in pipeline logs
> h3. Code Quality
> * *Best practices* - Follows Gradle recommended patterns
> * *Future-proof* - Compatible with Gradle 8+ strict validation
> * *Maintainable* - Explicit configuration easier to understand
> * *Documented* - Clear comments explain purpose
> h2.
> h2. Technical Background
> h3. Gradle Task Dependency Model
> *Three Types of Dependencies:*
> |Type|Declaration|Purpose|Example|
> |Execution Order|{{dependsOn}}|Task B must run after Task
> A|{{combineReports.dependsOn test}}|
> |Input/Output|{{{}reportOn{}}}, {{from}}|Task B consumes Task A
> output|{{reportOn testTask}}|
> |Soft Order|{{mustRunAfter}}|Ordering hint, not strict|{{mustRunAfter
> cleanTask}}|
> *Our Solution Uses:*
> * {{dependsOn}} for execution order (strict)
> * {{reportOn}} for input declaration (what we consume)
> h3. TestReport Task API
> *Gradle's Built-in TestReport Task:*
> {code:groovy}
> interface TestReport {
> // Declares which test tasks to aggregate
> void reportOn(Test... tasks)
>
> // Output location
> File getDestinationDir()
>
> // Inherits from Task
> Task dependsOn(Object... tasks)
> }
> {code}
> *Key Methods:*
> * {{reportOn()}} - Declares test result inputs AND establishes dependency
> * {{dependsOn()}} - Ensures explicit execution ordering
> * Using both provides belt-and-suspenders safety
> h3. Gradle 7+ Validation Changes
> *What Changed in Gradle 7:*
> # Stricter task input/output validation
> # Detection of implicit file dependencies
> # Warning when outputs consumed without declaration
> # Future direction: Make warnings into errors (Gradle 8+)
> *Our Compliance:*
> * Explicit {{reportOn}} declares inputs
> * Explicit {{dependsOn}} declares dependencies
> # Gradle can validate correctness
> # Optimizations can be safely enabled
> ----
> h2. Risk Assessment
> h3. Risk of NOT Fixing
> *Build Performance Risk: MEDIUM*
> * Continued suboptimal build performance
> * Execution optimizations remain disabled
> * Incremental builds less effective
> * Longer CI/CD pipeline times
> *Build Reliability Risk: LOW-MEDIUM*
> * Potential nondeterministic execution order
> * Rare race condition scenarios
> * Intermittent missing test results in reports
> * Harder to debug timing-related issues
> *Developer Experience Risk: LOW*
> * Ongoing noise in build logs
> * Confusion about warning severity
> * Reduced trust in build system
> * More time spent investigating false alarms
> h3. Risk of Fixing
> *Implementation Risk: VERY LOW*
> * Pure declarative configuration change
> * No logic modifications
> * No test code changes
> * Extensive verification performed
> *Compatibility Risk: NONE*
> * Fully backward compatible
> * No API changes
> * No behavioral changes (except warning removal)
> * Works with all Gradle versions
> *Regression Risk: VERY LOW*
> * Changes only affect build graph construction
> * Does not alter test execution or reporting logic
> * 17/17 CI checks passing
> * Manual verification complete
> *Overall Risk: LOW*
> * High benefit, minimal risk
> * Aligns with Gradle best practices
> * Future-proofs for Gradle 8+
> h2. Success Criteria
> When this enhancement is implemented, the following should be true:
> # *No Warnings:* Gradle validation warning eliminated from all builds
> # *Optimizations Enabled:* Execution optimizations active for
> {{:combineReports}}
> # *Deterministic Order:* {{:geode-old-versions:test}} always executes before
> {{:combineReports}}
> # *Complete Report:* All module test results appear in combined report
> # *No Regressions:* No new warnings or errors introduced
> # *Performance:* Incremental build behavior working correctly
> h2. Related Work
> * Build system modernization efforts
> * CI/CD pipeline optimization
> * Test infrastructure improvements
> * Gradle upgrade preparations (Gradle 8+)
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
