Enforce file size on Java inputs

[romani]

From #11156 (review)

All suppressions can be resolved in separate PRs for each Check individually.

Rationale: big input files a hard to review, hard to debug.

All existing inputs are moved to suppression at 3b523bd

checkstyle/config/checkstyle_resources_suppressions.xml

Line 174 in b37ec9c

<!-- Until https://github.com/checkstyle/checkstyle/issues/11163 -->

Now we need to update all that inputs to follow limit 120 and remove related suppressions. Each Check inputs should be in separate Pull Request

Example of update: 8673001

Test method names and input file name should be similar and describe what we are testing. Example from link above:

• Test method name is testJavadocVariableOnInnerClassFields
• Input file name is InputJavadocVariableOnInnerClassFields.java

List of checks for which we need to enforce file size on inputs -

• AbbreviationAsWordInName
• AnnotationLocation
• Antlr4AstRegression
• Antlr4AstRegressionUncommon
• AnonInnerLength
• AvoidEscapedUnicodeCharacters
• CatchParameterName
• ClassFanOutComplexity
• CommentsIndentation
• ConstantName
• CorrectSummaryJavaDocCheck
• CovariantEquals
• CyclomaticComplexity
• DefaultComesLast
• DeclarationOrder
• DesignForExtension
• EmptyBlock
• EmptyCatchBlock
• EmptyLineSeparator
• EqualsAvoidNull
• EqualsHashCode
• FallThrough
• FileLength
• FileTabCharacter
• FinalClass
• FinalLocalVariable
• FinalParameters
• HiddenField
• IllegalInstantiation
• IllegalTokenText
• Indentation
• IndentationCorrectIfAndParameter
• InnerAssignment
• Java14InstanceofWithPatternMatching
• Java14Records
• Java14SwitchExpression
• Java14TextBlocks
• JavaDocTagContinuationIndentation
• JavaNCSS
• JavaParserFullJavaIdentifierSupport
• JavadocMethod
• JavadocMetadataScraper
• JavadocPackage
• JavadocTagContinuationIndentation
• JavadocType
• JavadocVariable
• LeftCurly
• LineLength
• LocalFinalVariableName
• LocalVariableName
• MainComplexityOverflow
• MemberName
• MethodCount
• MethodLength
• MethodName
• MissingJavadocMethod
• MissingJavadocType
• ModifierOrder
• MultipleStringLiterals
• NPathComplexity
• NeedBraces
• NoEnumTrailingComma
• NoFinalizer
• NoWhitespaceAfter
• NoWhitespaceBefore
• OneStatementPerLine
• OuterTypeNumber
• PackageName
• ParenPad
• ParameterName
• ParameterNumber
• RecordComponentNumber
• RedundantModifier
• Regexp
• RegexpMultiline
• RegexpOnFilename
• RegexpSingleline
• RegexpSinglelineJava
• RequireThis
• RightCurly
• Rule332NoLineWrap
• Rule3SourceFile
• Rule412NonEmptyBlocks
• Rule42BlockIndentation
• Rule44ColumnLimit
• Rule451WhereToBreak
• Rule452IndentContinuationLines
• Rule461VerticalWhitespace
• Rule462HorizontalWhitespace
• Rule4641Indentation
• Rule487Modifiers
• Rule525NonConstantFieldNames
• Rule64Finalizers
• StaticVariableName
• SummaryJavadoc
• SuppressionCommentFilter
• TodoComment
• TypecastParenPad
• UpperEll
• VariableDeclarationUsageDistance
• VisibilityModifier
• WhitespaceAfter
• WhitespaceAround
• XpathMapperStringConcat

[Vyom-Yadav]

@romani Here is some data-

Total Num Of Lines: 226357
Total Num Of Files: 1976
Median of number of lines: 44.0
The average number of lines: 114.55313765182186

This includes Input files from both ../resources and ../resources-noncompilable.
How I generated the data: https://gist.github.com/Vyom-Yadav/76cedd6a5286496e25cf762b31ac2e10

Edit-
Not to forget we have a file with 65553 lines https://github.com/checkstyle/checkstyle/blob/master/src/test/resources/com/puppycrawl/tools/checkstyle/checks/avoidescapedunicodecharacters/InputAvoidEscapedUnicodeCharactersAllEscapedUnicodeCharacters.java and many more files which are used for testing stack overflow errors, etc.
I will send a fix once we decide the bottleneck on the number of lines an Input file can have.

[romani]

I am ok to make max limit as 120 lines, I glanced through files, all big files are NOT for good reason.

@pbludov , @rnveach , @nmancus1 , @strkkk , please confirm that you are ok with this.
there definetely might be exceptions, and we can keep them as big files.

[rnveach]

I would prefer a round number like 200.

Configuration javadoc, package, imports, class definitions, and the necessary blank lines add up.

[romani]

We do not use imports that much, header comment is not that big. Please show me input file that is good to be above 120 lines. There might few only.

[pbludov]

Long files list: https://gist.github.com/pbludov/fe3786436e177939e2041e3596626296
355 files > 120 lines
200 files > 200 lines

We can start with a limit of 200 lines, and then reduce the limit to 120.

[romani]

count of files is not that I worry about. This task will be perfect for introduction to project, so it will be long lasting and done eventually - ok.

Having two iterations: to 200 and then to 120 will be a lot of efforts. Contributor will spend same time to split files from xxxxx to 200 as to xxxx to 120.

[nrmancuso]

I am ok to make max limit as 120 lines...

I agree that this is a good limit under normal circumstances. Limiting file lengths will help to make violations more clear, and to reduce/ eliminate extension of existing test inputs.

In PR for this issue, we should decide on big files to keep with comment explaining why for each suppression, then have contributors simply select suppression with no comment to work on.

[romani]

First PR should do suppression "until #11163". If we decide to keep big input, we should make special comment.

[strkkk]

I am ok with 120

[Vyom-Yadav]

Here is some more data to help you decide whether it will be 120 or 200.

Total Num Of Lines: 54466
The average number of lines: 27.521980798383023

This number is composed of the following-

/*
SomeCheck
max = (default)200


*/

package com......

public class InputSomeCheck {

    void method() {
    }

}

Total num of lines before top-level class declaration - 10 (includes the class declaration)
Total num of empty lines (for readability of code) - 2 (After class declaration)
Total = 10 + 2 = 12
Here is the data for every input file (.java) - https://gist.github.com/Vyom-Yadav/44af5b511fd2d7b89eacacb62fed7ea5

[romani]

We come to majority agreement to make it 120.

We can always keep in permanent suppression cases that need more for good reason or we can make limit higher later on.
For now let's try 120.

@Vyom-Yadav , can you send PR that will do validation and placing all files above limit to temporary suppression ?

[nrmancuso]

@MANISH-K-07 @Vyom-Yadav now that we have opened an issue on input file/test method naming conventions, please mark your comments as outdated to clean this issue up.