Codestin Search App

UncleOwen · 2025-09-21T19:55:08Z

Describe the PR

This PR shrinks the reported locations for some rules. There is more to do on this front, but I'd like to get this part merged.

Related issues

Related: [java] Single Line Warnings #730, [java] Review reported locations of rules #3769.

Ready?

Added unit tests for fixed bug/feature
Passing all unit tests
Complete build ./mvnw clean verify passes (checked automatically by github actions)

pmd-actions-helper · 2025-09-21T20:11:49Z

Documentation Preview

Compared to main:
This changeset changes 45390 violations,
introduces 684 new violations, 0 new errors and 0 new configuration errors,
removes 846 violations, 0 errors and 0 configuration errors.

Regression Tester Report

(comment created at 2026-01-22 18:35:15+00:00 for d15e3b0)

adangel

Thanks for the PR!

Before merging this, I'd like to clarify, that we really report on useful lines (e.g. report on the class identifier instead of on the class keyword).

After this PR, I'd suggest the following approach:

create a list of all rules
Cross out the rules, for which you created a test in this PR already
select one single rule
create a test case "Verify location of violation" with the expected report location. Keep in mind, that the end-line seems to where github actions would show the annotation (see #3769)
fix this rule
repeat with 2 until no rules are left
extend documentation "how to write/test a rule" to describe, that not only a positive and negative test case is required, but also a verification of the report location test.
Because now, we fixed "some rules" but we don't know which exactly, and when we have actually done #3769. This makes it hard to monitor progress on this front.

Additionally, we might need to think about extending the regression tester. I think, we currently only compare the beginLine, but maybe we should compare the complete report location for each violation. But maybe this is not needed, if we have for every rule such a test case. Maybe that's sufficient.

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/ASTCompilationUnit.java

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/ASTSwitchStatement.java

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/rule/design/ExcessiveImportsRule.java

pmd-java/src/main/resources/category/java/codestyle.xml

.../net/sourceforge/pmd/lang/java/rule/bestpractices/xml/AbstractClassWithoutAbstractMethod.xml

.../test/resources/net/sourceforge/pmd/lang/java/rule/bestpractices/xml/NonExhaustiveSwitch.xml

pmd-java/src/test/resources/net/sourceforge/pmd/lang/java/rule/codestyle/xml/NoPackage.xml

...src/test/resources/net/sourceforge/pmd/lang/java/rule/codestyle/xml/TooManyStaticImports.xml

pmd-java/src/test/resources/net/sourceforge/pmd/lang/java/rule/design/xml/TooManyMethods.xml

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/rule/design/ExcessiveImportsRule.java

oowekyala · 2025-11-27T11:43:23Z

.../src/main/java/net/sourceforge/pmd/lang/java/rule/internal/AbstractJavaCounterCheckRule.java

            int threshold = getProperty(reportLevel);
            if (metric >= threshold) {
-                asCtx(data).addViolation(node, metric, threshold);
+                asCtx(data).addViolationWithPosition(t, t.getAstInfo(), getReportLocation(t), getMessage(), metric, threshold);


This is a slight conflict with #6260 because this overload of addViolationWithPosition is removed in that PR. If that PR is merged before this one, the correct incantation would be

Suggested change

asCtx(data).addViolationWithPosition(t, t.getAstInfo(), getReportLocation(t), getMessage(), metric, threshold);

asCtx(data).at(t.atLocation(getReportLocation(t))).warnWithArgs(metric, threshold);

adangel · 2026-01-07T13:53:45Z

Note: I'm going to rebase this a couple of times...

adangel · 2026-01-07T15:22:44Z

Status update (see also #730): The two rules AbstractClassWithoutAbstractMethod and NonExhaustiveSwitch of category "Best Practices" are done now.

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/AbstractTypeDeclaration.java

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/ASTCompilationUnit.java

adangel · 2026-01-16T12:57:48Z

When looking at the regression report, the following additional rules have been changed:

AtLeastOneConstructor
ClassWithOnlyPrivateConstructorShouldBeFinal
CommentDefaultAccessModifier
CommentRequired
ShortClassName
TestClassWithoutTestCases
UnnecessaryModifier
UseUtilityClass

Once the new report is available, maybe less rules are affected by this change. But if there are more rules affected (because e.g. we changed all rules that report on ClassDeclaration), we should already handle them here in this PR.

I'm also inclined to actually implement pmd/pmd-regression-tester#148 before merging this change, just to be sure, the new report location makes sense and we don't inadvertently change other rules as well.

So, this is maybe not ready even for 7.21.0.

…e location of the violation

… keyword. Define a report location for ASTCompilationUnit

… import.

…ynchronized keyword.

…eyword.

…class

adangel · 2026-01-22T18:17:07Z

I've update it now and added test cases for other affected rules (AbstractClassWithoutAnyMethod, AvoidDollarSigns, ClassNamingConventions, CyclomaticComplexity, DataClass, ExhaustiveSwitchHasDefault, GodClass, ImplicitFunctionalInterface, JUnit5TestShouldBePackagePrivate, MissingSerialVersionUID, MissingStaticMethodInNonInstantiatableClass, SwitchDensity, TooFewBranchesForSwitch, NcssCount, SingleMethodSingleton, LocalHomeNamingConvention, LocalInterfaceSessionNamingConvention). They all report on the identifier of class/interfaces now or on the switch keyword.

I fixed CouplingBetweenObjects to correctly recognize nested classes. Hopefully the results make more sense now (lets wait for regression tester result).

Two rules are affected, but I didn't include them, as they would not be finished completely:

ControlStatementBraces -> reported on synchronized statement in braceless if. Now, only reports on the synchronized keyword rather than on the whole block. But in general, we still might report multiple lines, if the first statement of a braceless if takes multiple lines. This rule should be tackled in a different PR (along with rethinking the decision to report on the first statement of a braceless if and not on the braceless if statement directly)
FinalizeDoesNotCallSuperFinalize -> when reported on synchronized statement, now reporting only on synchronized keyword.

In total, we have now updated 13% of our Java rules (37 of 289 rules).

Next steps:

Finish Fix #148: Add support for comparing entire rule violation location pmd-regression-tester#149 and release a new regression tester
Update this PR
Merge this PR

adangel requested changes Oct 24, 2025

View reviewed changes

adangel added this to the 7.19.0 milestone Oct 27, 2025

adangel self-assigned this Nov 27, 2025

adangel modified the milestones: 7.19.0, 7.20.0 Nov 27, 2025

oowekyala reviewed Nov 27, 2025

View reviewed changes

adangel modified the milestones: 7.20.0, 7.21.0 Dec 28, 2025

adangel mentioned this pull request Jan 7, 2026

[java] Single Line Warnings #730

Open

adangel marked this pull request as draft January 7, 2026 13:42

adangel force-pushed the issue-730-ExcessiveImportsRule-location-of-violation branch from bf4235c to 0e585ff Compare January 7, 2026 15:18

oowekyala requested changes Jan 8, 2026

View reviewed changes

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/AbstractTypeDeclaration.java Outdated Show resolved Hide resolved

pmd-java/src/main/java/net/sourceforge/pmd/lang/java/ast/ASTCompilationUnit.java Outdated Show resolved Hide resolved

adangel force-pushed the issue-730-ExcessiveImportsRule-location-of-violation branch from 56b87d5 to 78f0bb7 Compare January 16, 2026 12:52

UncleOwen and others added 13 commits January 16, 2026 17:52

ExcessiveImportsRule: Don't violate the whole file, just the imports

2d3c02a

ExcessiveParameterList/ExcessivePublicCount: Add tests that verify th…

a905fc0

…e location of the violation

Add more tests that verify the location of violations

f38cc15

TooManyMethods: Don't violate the whole class, just the class keyword.

ffa496d

CouplingBetweenObjects: Don't violate the whole class, just the class…

93c94d9

… keyword. Define a report location for ASTCompilationUnit

TooManyStaticImports: Don't violate the whole class, just the last(!)…

f83e99f

… import.

TooManyFields: Don't violate the whole class, just ClassDeclaration.

542e1b3

AvoidSynchronizedStatement: Don't violate the whole block, just the s…

48640b1

…ynchronized keyword.

NonExhaustiveSwitch: Don't violate the whole block, just the switch k…

49d4536

…eyword.

AbstractClassWithoutAbstractMethod: Report on identifier

f290ea0

Add test for classdeclaration reportlocation

2b47d57

TooManyMethods: Report on identifier

032cd9e

NoPackage: Report on identifier of class/type

d7307bf

[doc] Update release notes

b26811f

adangel force-pushed the issue-730-ExcessiveImportsRule-location-of-violation branch from 78f0bb7 to b26811f Compare January 16, 2026 16:55

adangel added 23 commits January 17, 2026 11:54

Test pmd/pmd-regression-tester#149

911442c

Test pmd/pmd-regression-tester#149 - Part 2

d7af301

AbstractClassWithoutAnyMethod: Report on identifier of class

087834b

AvoidDollarSigns: report on identifiers

09ba813

ClassNamingConventions: report on identifier of types

9298378

CouplingBetweenObjects: fix reporting issue for nested classes

c9e742b

CyclomaticComplexity: Report on identifier of class/method

ceb1744

DataClass: Report on identifier of class

cbb2b2b

ControlStatementBraces: Verify synchronized statement

bfb966f

EmptyControlStatement: Verify switch/synchronized statements

e6115c3

ExhaustiveSwitchHasDefault: report on switch keyword

4ef9ad3

GodClass: Report on identifier of class

0af317e

ImplicitFunctionalInterface: Report on identifier of interface

3430038

JUnit5TestShouldBePackagePrivate: Report on identifier of class/method

e367a72

LocalHomeNamingConvention: Report on identifier of interface

ae0f22d

LocalInterfaceSessionNamingConvention: Report on identifier of interface

c189780

MissingSerialVersionUID: Report on identifier of class

d3bf988

MissingStaticMethodInNonInstantiatableClass: Report on identifier of …

403bf99

…class

NcssCount: Report on identifier of class/method

49a834b

SingleMethodSingleton: Report on identifier of class

b47427f

SwitchDensity: Report on switch keyword

c27f71c

TooFewBranchesForSwitch: Report on switch keyword

7c54d73

[doc] Update release notes

d15e3b0

adangel modified the milestones: 7.21.0, 7.22.0 Jan 29, 2026

adangel modified the milestones: 7.22.0, 7.23.0 Feb 26, 2026

	asCtx(data).addViolationWithPosition(t, t.getAstInfo(), getReportLocation(t), getMessage(), metric, threshold);
	asCtx(data).at(t.atLocation(getReportLocation(t))).warnWithArgs(metric, threshold);

Uh oh!

Conversation

UncleOwen commented Sep 21, 2025

Describe the PR

Related issues

Ready?

Uh oh!

pmd-actions-helper bot commented Sep 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

adangel left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

oowekyala Nov 27, 2025

Choose a reason for hiding this comment

Uh oh!

adangel commented Jan 7, 2026

Uh oh!

adangel commented Jan 7, 2026

Uh oh!

Uh oh!

Uh oh!

adangel commented Jan 16, 2026

Uh oh!

adangel commented Jan 22, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

pmd-actions-helper bot commented Sep 21, 2025 •

edited

Loading