001///////////////////////////////////////////////////////////////////////////////////////////////
002// checkstyle: Checks Java source code and other text files for adherence to a set of rules.
003// Copyright (C) 2001-2025 the original author or authors.
004//
005// This library is free software; you can redistribute it and/or
006// modify it under the terms of the GNU Lesser General Public
007// License as published by the Free Software Foundation; either
008// version 2.1 of the License, or (at your option) any later version.
009//
010// This library is distributed in the hope that it will be useful,
011// but WITHOUT ANY WARRANTY; without even the implied warranty of
012// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
013// Lesser General Public License for more details.
014//
015// You should have received a copy of the GNU Lesser General Public
016// License along with this library; if not, write to the Free Software
017// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
018///////////////////////////////////////////////////////////////////////////////////////////////
019
020package com.puppycrawl.tools.checkstyle.checks.javadoc;
021
022import java.util.Set;
023
024import javax.annotation.Nullable;
025
026import com.puppycrawl.tools.checkstyle.StatelessCheck;
027import com.puppycrawl.tools.checkstyle.api.DetailNode;
028import com.puppycrawl.tools.checkstyle.api.JavadocCommentsTokenTypes;
029import com.puppycrawl.tools.checkstyle.utils.CommonUtil;
030import com.puppycrawl.tools.checkstyle.utils.JavadocUtil;
031
032/**
033 * <div>
034 * Checks the Javadoc paragraph.
035 * </div>
036 *
037 * <p>
038 * Checks that:
039 * </p>
040 * <ul>
041 * <li>There is one blank line between each of two paragraphs.</li>
042 * <li>Each paragraph but the first has &lt;p&gt; immediately
043 * before the first word, with no space after.</li>
044 * <li>The outer most paragraph tags should not precede
045 * <a href="https://www.w3schools.com/html/html_blocks.asp">HTML block-tag</a>.
046 * Nested paragraph tags are allowed to do that. This check only supports following block-tags:
047 * &lt;address&gt;,&lt;blockquote&gt;
048 * ,&lt;div&gt;,&lt;dl&gt;
049 * ,&lt;h1&gt;,&lt;h2&gt;,&lt;h3&gt;,&lt;h4&gt;,&lt;h5&gt;,&lt;h6&gt;,&lt;hr&gt;
050 * ,&lt;ol&gt;,&lt;p&gt;,&lt;pre&gt;
051 * ,&lt;table&gt;,&lt;ul&gt;.
052 * </li>
053 * </ul>
054 *
055 * <p><b>ATTENTION:</b></p>
056 *
057 * <p>This Check ignores HTML comments.</p>
058 *
059 * <p>The Check ignores all the nested paragraph tags,
060 * it will not give any kind of violation if the paragraph tag is nested.</p>
061 *
062 * @since 6.0
063 */
064@StatelessCheck
065public class JavadocParagraphCheck extends AbstractJavadocCheck {
066
067    /**
068     * A key is pointing to the warning message text in "messages.properties"
069     * file.
070     */
071    public static final String MSG_TAG_AFTER = "javadoc.paragraph.tag.after";
072
073    /**
074     * A key is pointing to the warning message text in "messages.properties"
075     * file.
076     */
077    public static final String MSG_LINE_BEFORE = "javadoc.paragraph.line.before";
078
079    /**
080     * A key is pointing to the warning message text in "messages.properties"
081     * file.
082     */
083    public static final String MSG_REDUNDANT_PARAGRAPH = "javadoc.paragraph.redundant.paragraph";
084
085    /**
086     * A key is pointing to the warning message text in "messages.properties"
087     * file.
088     */
089    public static final String MSG_MISPLACED_TAG = "javadoc.paragraph.misplaced.tag";
090
091    /**
092     * A key is pointing to the warning message text in "messages.properties"
093     * file.
094     */
095    public static final String MSG_PRECEDED_BLOCK_TAG = "javadoc.paragraph.preceded.block.tag";
096
097    /**
098     * Constant for the paragraph tag name.
099     */
100    private static final String PARAGRAPH_TAG = "p";
101
102    /**
103     * Set of block tags supported by this check.
104     */
105    private static final Set<String> BLOCK_TAGS =
106            Set.of("address", "blockquote", "div", "dl",
107                   "h1", "h2", "h3", "h4", "h5", "h6", "hr",
108                   "ol", PARAGRAPH_TAG, "pre", "table", "ul");
109
110    /**
111     * Control whether the &lt;p&gt; tag should be placed immediately before the first word.
112     */
113    private boolean allowNewlineParagraph = true;
114
115    /**
116     * Setter to control whether the &lt;p&gt; tag should be placed
117     * immediately before the first word.
118     *
119     * @param value value to set.
120     * @since 6.9
121     */
122    public void setAllowNewlineParagraph(boolean value) {
123        allowNewlineParagraph = value;
124    }
125
126    @Override
127    public int[] getDefaultJavadocTokens() {
128        return new int[] {
129            JavadocCommentsTokenTypes.NEWLINE,
130            JavadocCommentsTokenTypes.HTML_ELEMENT,
131        };
132    }
133
134    @Override
135    public int[] getRequiredJavadocTokens() {
136        return getAcceptableJavadocTokens();
137    }
138
139    @Override
140    public void visitJavadocToken(DetailNode ast) {
141        if (ast.getType() == JavadocCommentsTokenTypes.NEWLINE && isEmptyLine(ast)) {
142            checkEmptyLine(ast);
143        }
144        else if (JavadocUtil.isTag(ast, PARAGRAPH_TAG)) {
145            checkParagraphTag(ast);
146        }
147    }
148
149    /**
150     * Determines whether or not the next line after empty line has paragraph tag in the beginning.
151     *
152     * @param newline NEWLINE node.
153     */
154    private void checkEmptyLine(DetailNode newline) {
155        final DetailNode nearestToken = getNearestNode(newline);
156        if (nearestToken != null && nearestToken.getType() == JavadocCommentsTokenTypes.TEXT
157                && !CommonUtil.isBlank(nearestToken.getText())) {
158            log(newline.getLineNumber(), newline.getColumnNumber(), MSG_TAG_AFTER);
159        }
160    }
161
162    /**
163     * Determines whether or not the line with paragraph tag has previous empty line.
164     *
165     * @param tag html tag.
166     */
167    private void checkParagraphTag(DetailNode tag) {
168        if (!isNestedParagraph(tag)) {
169            final DetailNode newLine = getNearestEmptyLine(tag);
170            if (isFirstParagraph(tag)) {
171                log(tag.getLineNumber(), tag.getColumnNumber(), MSG_REDUNDANT_PARAGRAPH);
172            }
173            else if (newLine == null || tag.getLineNumber() - newLine.getLineNumber() != 1) {
174                log(tag.getLineNumber(), tag.getColumnNumber(), MSG_LINE_BEFORE);
175            }
176
177            final String blockTagName = findFollowedBlockTagName(tag);
178            if (blockTagName != null) {
179                log(tag.getLineNumber(), tag.getColumnNumber(),
180                        MSG_PRECEDED_BLOCK_TAG, blockTagName);
181            }
182
183            if (!allowNewlineParagraph && isImmediatelyFollowedByNewLine(tag)) {
184                log(tag.getLineNumber(), tag.getColumnNumber(), MSG_MISPLACED_TAG);
185            }
186            if (isImmediatelyFollowedByText(tag)) {
187                log(tag.getLineNumber(), tag.getColumnNumber(), MSG_MISPLACED_TAG);
188            }
189        }
190    }
191
192    /**
193     * Determines whether the paragraph tag is nested.
194     *
195     * @param tag html tag.
196     * @return true, if the paragraph tag is nested.
197     */
198    private static boolean isNestedParagraph(DetailNode tag) {
199        boolean nested = false;
200        DetailNode parent = tag.getParent();
201
202        while (parent != null) {
203            if (parent.getType() == JavadocCommentsTokenTypes.HTML_ELEMENT) {
204                nested = true;
205                break;
206            }
207            parent = parent.getParent();
208        }
209
210        return nested;
211    }
212
213    /**
214     * Determines whether or not the paragraph tag is followed by block tag.
215     *
216     * @param tag html tag.
217     * @return block tag if the paragraph tag is followed by block tag or null if not found.
218     */
219    @Nullable
220    private static String findFollowedBlockTagName(DetailNode tag) {
221        final DetailNode htmlElement = findFirstHtmlElementAfter(tag);
222        String blockTagName = null;
223
224        if (htmlElement != null) {
225            blockTagName = getHtmlElementName(htmlElement);
226        }
227
228        return blockTagName;
229    }
230
231    /**
232     * Finds and returns first html element after the tag.
233     *
234     * @param tag html tag.
235     * @return first html element after the paragraph tag or null if not found.
236     */
237    @Nullable
238    private static DetailNode findFirstHtmlElementAfter(DetailNode tag) {
239        DetailNode htmlElement = getNextSibling(tag);
240
241        while (htmlElement != null
242                && htmlElement.getType() != JavadocCommentsTokenTypes.HTML_ELEMENT) {
243            if (htmlElement.getType() == JavadocCommentsTokenTypes.HTML_CONTENT) {
244                htmlElement = htmlElement.getFirstChild();
245            }
246            else if (htmlElement.getType() == JavadocCommentsTokenTypes.TEXT
247                    && !CommonUtil.isBlank(htmlElement.getText())) {
248                htmlElement = null;
249                break;
250            }
251            else {
252                htmlElement = htmlElement.getNextSibling();
253            }
254        }
255        if (htmlElement != null
256                && JavadocUtil.findFirstToken(htmlElement,
257                        JavadocCommentsTokenTypes.HTML_TAG_END) == null) {
258            htmlElement = null;
259        }
260
261        return htmlElement;
262    }
263
264    /**
265     * Finds and returns first block-level html element name.
266     *
267     * @param htmlElement block-level html tag.
268     * @return block-level html element name or null if not found.
269     */
270    @Nullable
271    private static String getHtmlElementName(DetailNode htmlElement) {
272        final DetailNode htmlTagStart = htmlElement.getFirstChild();
273        final DetailNode htmlTagName =
274                JavadocUtil.findFirstToken(htmlTagStart, JavadocCommentsTokenTypes.TAG_NAME);
275        String blockTagName = null;
276        if (BLOCK_TAGS.contains(htmlTagName.getText())) {
277            blockTagName = htmlTagName.getText();
278        }
279
280        return blockTagName;
281    }
282
283    /**
284     * Returns nearest node.
285     *
286     * @param node DetailNode node.
287     * @return nearest node.
288     */
289    private static DetailNode getNearestNode(DetailNode node) {
290        DetailNode currentNode = node;
291        while (currentNode != null
292                && (currentNode.getType() == JavadocCommentsTokenTypes.LEADING_ASTERISK
293                    || currentNode.getType() == JavadocCommentsTokenTypes.NEWLINE)) {
294            currentNode = currentNode.getNextSibling();
295        }
296        if (currentNode != null
297                && currentNode.getType() == JavadocCommentsTokenTypes.HTML_CONTENT) {
298            currentNode = currentNode.getFirstChild();
299        }
300        return currentNode;
301    }
302
303    /**
304     * Determines whether or not the line is empty line.
305     *
306     * @param newLine NEWLINE node.
307     * @return true, if line is empty line.
308     */
309    private static boolean isEmptyLine(DetailNode newLine) {
310        boolean result = false;
311        DetailNode previousSibling = newLine.getPreviousSibling();
312        if (previousSibling != null && (previousSibling.getParent().getType()
313                == JavadocCommentsTokenTypes.JAVADOC_CONTENT
314                || insideNonTightHtml(previousSibling))) {
315            if (previousSibling.getType() == JavadocCommentsTokenTypes.TEXT
316                    && CommonUtil.isBlank(previousSibling.getText())) {
317                previousSibling = previousSibling.getPreviousSibling();
318            }
319            result = previousSibling != null
320                    && previousSibling.getType() == JavadocCommentsTokenTypes.LEADING_ASTERISK;
321        }
322        return result;
323    }
324
325    /**
326     * Checks whether the given node is inside a non-tight HTML element.
327     *
328     * @param previousSibling the node to check
329     * @return true if inside non-tight HTML, false otherwise
330     */
331    private static boolean insideNonTightHtml(DetailNode previousSibling) {
332        final DetailNode parent = previousSibling.getParent();
333        DetailNode htmlElement = parent;
334        if (parent.getType() == JavadocCommentsTokenTypes.HTML_CONTENT) {
335            htmlElement = parent.getParent();
336        }
337        return htmlElement.getType() == JavadocCommentsTokenTypes.HTML_ELEMENT
338                && JavadocUtil.findFirstToken(htmlElement,
339                    JavadocCommentsTokenTypes.HTML_TAG_END) == null;
340    }
341
342    /**
343     * Determines whether or not the line with paragraph tag is first line in javadoc.
344     *
345     * @param paragraphTag paragraph tag.
346     * @return true, if line with paragraph tag is first line in javadoc.
347     */
348    private static boolean isFirstParagraph(DetailNode paragraphTag) {
349        boolean result = true;
350        DetailNode previousNode = paragraphTag.getPreviousSibling();
351        while (previousNode != null) {
352            if (previousNode.getType() == JavadocCommentsTokenTypes.TEXT
353                    && !CommonUtil.isBlank(previousNode.getText())
354                || previousNode.getType() != JavadocCommentsTokenTypes.LEADING_ASTERISK
355                    && previousNode.getType() != JavadocCommentsTokenTypes.NEWLINE
356                    && previousNode.getType() != JavadocCommentsTokenTypes.TEXT) {
357                result = false;
358                break;
359            }
360            previousNode = previousNode.getPreviousSibling();
361        }
362        return result;
363    }
364
365    /**
366     * Finds and returns nearest empty line in javadoc.
367     *
368     * @param node DetailNode node.
369     * @return Some nearest empty line in javadoc.
370     */
371    private static DetailNode getNearestEmptyLine(DetailNode node) {
372        DetailNode newLine = node;
373        while (newLine != null) {
374            final DetailNode previousSibling = newLine.getPreviousSibling();
375            if (newLine.getType() == JavadocCommentsTokenTypes.NEWLINE && isEmptyLine(newLine)) {
376                break;
377            }
378            newLine = previousSibling;
379        }
380        return newLine;
381    }
382
383    /**
384     * Tests whether the paragraph tag is immediately followed by the text.
385     *
386     * @param tag html tag.
387     * @return true, if the paragraph tag is immediately followed by the text.
388     */
389    private static boolean isImmediatelyFollowedByText(DetailNode tag) {
390        final DetailNode nextSibling = getNextSibling(tag);
391
392        return nextSibling == null || nextSibling.getText().startsWith(" ");
393    }
394
395    /**
396     * Tests whether the paragraph tag is immediately followed by the new line.
397     *
398     * @param tag html tag.
399     * @return true, if the paragraph tag is immediately followed by the new line.
400     */
401    private static boolean isImmediatelyFollowedByNewLine(DetailNode tag) {
402        final DetailNode sibling = getNextSibling(tag);
403        return sibling != null && sibling.getType() == JavadocCommentsTokenTypes.NEWLINE;
404    }
405
406    /**
407     * Custom getNextSibling method to handle different types of paragraph tag.
408     * It works for both {@code <p>} and {@code <p></p>} tags.
409     *
410     * @param tag HTML_ELEMENT tag.
411     * @return next sibling of the tag.
412     */
413    private static DetailNode getNextSibling(DetailNode tag) {
414        DetailNode nextSibling;
415        final DetailNode paragraphStartTagToken = tag.getFirstChild();
416        final DetailNode nextNode = paragraphStartTagToken.getNextSibling();
417
418        if (nextNode == null) {
419            nextSibling = tag.getNextSibling();
420        }
421        else if (nextNode.getType() == JavadocCommentsTokenTypes.HTML_CONTENT) {
422            nextSibling = nextNode.getFirstChild();
423        }
424        else {
425            nextSibling = nextNode;
426        }
427
428        if (nextSibling != null
429                && nextSibling.getType() == JavadocCommentsTokenTypes.HTML_COMMENT) {
430            nextSibling = nextSibling.getNextSibling();
431        }
432        return nextSibling;
433    }
434}