package com.google.googlejavaformat.java;

import static com.google.common.base.Preconditions.checkArgument;

import static com.google.common.base.Preconditions.checkElementIndex;

import static java.util.Arrays.stream;

import com.google.common.collect.ImmutableList;

import com.sun.tools.javac.parser.Tokens.TokenKind;

import com.sun.tools.javac.parser.UnicodeReader;

import com.sun.tools.javac.util.Context;

import java.lang.reflect.Method;

import java.util.ArrayList;

import java.util.Collections;

import java.util.Comparator;

import java.util.HashSet;

import java.util.List;

import java.util.Objects;

import java.util.Set;

import org.checkerframework.checker.nullness.qual.Nullable;

/** A wrapper around javac's lexer. */

class JavacTokens {

final class JavacTokens {

/** The lexer eats terminal comments, so feed it one we don't care about. */

// TODO(b/33103797): fix javac and remove the work-around

@@ -51,6 +54,8 @@ static class RawTok {

private final int endPos;

RawTok(String stringVal, TokenKind kind, int pos, int endPos) {

checkElementIndex(pos, endPos, "pos");

checkArgument(pos < endPos, "expected pos (%s) < endPos (%s)", pos, endPos);

this.stringVal = stringVal;

this.kind = kind;

this.pos = pos;

@@ -136,13 +141,30 @@ public static ImmutableList<RawTok> getTokens(

int last = 0;

for (Token t : javacTokens) {

if (t.comments != null) {

// javac accumulates comments in reverse order

for (Comment c : Lists.reverse(t.comments)) {

if (last < c.getSourcePos(0)) {

tokens.add(new RawTok(null, null, last, c.getSourcePos(0)));

int pos = c.getSourcePos(0);

int length;

if (pos == -1) {

// We've found a comment whose position hasn't been recorded. Deduce its position as the

// first `/` character after the end of the previous token.

//

// javac creates a new JavaTokenizer to process string template arguments, so

// CommentSavingTokenizer doesn't get a chance to preprocess those comments and save

// their text and positions.

//

// TODO: consider always using this approach once the minimum supported JDK is 16 and

// we can assume BasicComment#getRawCharacters is always available.

pos = source.indexOf('/', last);

length = CommentSavingTokenizer.commentLength(c);

} else {

length = c.getText().length();

}

tokens.add(

new RawTok(null, null, c.getSourcePos(0), c.getSourcePos(0) + c.getText().length()));

last = c.getSourcePos(0) + c.getText().length();

if (last < pos) {

tokens.add(new RawTok(null, null, last, pos));

}

tokens.add(new RawTok(null, null, pos, pos + length));

last = pos + length;

}

}

if (stopTokens.contains(t.kind)) {

@@ -181,6 +203,32 @@ public static ImmutableList<RawTok> getTokens(

/** A {@link JavaTokenizer} that saves comments. */

static class CommentSavingTokenizer extends JavaTokenizer {

private static final Method GET_RAW_CHARACTERS_METHOD = getRawCharactersMethod();

private static @Nullable Method getRawCharactersMethod() {

try {

// This is a method in PositionTrackingReader, but that class is not public.

return BasicComment.class.getMethod("getRawCharacters");

} catch (NoSuchMethodException e) {

return null;

}

}

static int commentLength(Comment comment) {

if (comment instanceof BasicComment && GET_RAW_CHARACTERS_METHOD != null) {

// If we've seen a BasicComment instead of a CommentWithTextAndPosition, getText() will

// be null, so we deduce the length using getRawCharacters. See also the comment at the

// usage of this method in getTokens.

try {

return ((char[]) GET_RAW_CHARACTERS_METHOD.invoke(((BasicComment) comment))).length;

} catch (ReflectiveOperationException e) {

throw new LinkageError(e.getMessage(), e);

}

}

return comment.getText().length();

}

CommentSavingTokenizer(ScannerFactory fac, char[] buffer, int length) {

super(fac, buffer, length);

}

@@ -288,4 +336,6 @@ protected AccessibleReader(ScannerFactory fac, char[] buffer, int length) {

super(fac, buffer, length);

}

}

private JavacTokens() {}

}