No paragraphs for html (#27)

* htmlparser added
* Added ParserConfig
    * added option to disable the emphasized parser
    * added option to not wrap HTML in markdown within a paragraph in output
* Updated docs
* Version update 1.1.1

README.md

@@ -1,7 +1,7 @@
 # maddy
 
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Version: 1.1.0](https://img.shields.io/badge/Version-1.1.0-brightgreen.svg)](https://semver.org/)
+[![Version: 1.1.1](https://img.shields.io/badge/Version-1.1.1-brightgreen.svg)](https://semver.org/)
 [![Travis Build Status](https://travis-ci.org/progsource/maddy.svg?branch=master)](https://travis-ci.org/progsource/maddy)
 [![Appveyor Build Status](https://ci.appveyor.com/api/projects/status/04m0lg27kigv1pg8/branch/master?svg=true)](https://ci.appveyor.com/project/progsource/maddy/branch/master)
 
@@ -42,7 +42,13 @@ and in the code, you can then do the following:
 #include "maddy/parser.h"
 
 std::stringstream markdownInput("");
-std::shared_ptr<maddy::Parser> parser = std::make_shared<maddy::Parser>();
+
+// config is optional
+std::shared_ptr<maddy::ParserConfig> config = std::make_shared<maddy::ParserConfig>();
+config->isEmphasizedParserEnabled = true; // default
+config->isHTMLWrappedInParagraph = true; // default
+
+std::shared_ptr<maddy::Parser> parser = std::make_shared<maddy::Parser>(config);
 std::string htmlOutput = parser->Parse(markdownInput);
 ```
 

docs/definitions.md

@@ -7,6 +7,10 @@ destroy the output, if there was HTML in your markdown.
 The Parser expects you to use spaces and not tabs for indentation in the
 markdown.
 
+If a line starts with `<` and `config->isHTMLWrappedInParagraph` is false, it
+expects that the upcoming line is HTML and therefor will not be surrounded by a
+paragraph.
+
 ## Headlines
 
 ```
@@ -241,6 +245,8 @@ results in
 
 ## emphasized
 
+This can be disabled by setting `config->isEmphasizedParserEnabled = false`.
+
 ```
 _emphasized text_
 ```

include/maddy/htmlparser.h

@@ -0,0 +1,127 @@
+/*
+ * This project is licensed under the MIT license. For more information see the
+ * LICENSE file.
+ */
+#pragma once
+
+// -----------------------------------------------------------------------------
+
+#include <functional>
+#include <string>
+
+#include "maddy/blockparser.h"
+
+// -----------------------------------------------------------------------------
+
+namespace maddy {
+
+// -----------------------------------------------------------------------------
+
+/**
+ * HtmlParser
+ *
+ * @class
+ */
+class HtmlParser : public BlockParser
+{
+public:
+  /**
+   * ctor
+   *
+   * @method
+   * @param {std::function<void(std::string&)>} parseLineCallback
+   * @param {std::function<std::shared_ptr<BlockParser>(const std::string& line)>} getBlockParserForLineCallback
+   */
+   HtmlParser(
+    std::function<void(std::string&)> parseLineCallback,
+    std::function<std::shared_ptr<BlockParser>(const std::string& line)> getBlockParserForLineCallback
+  )
+    : BlockParser(parseLineCallback, getBlockParserForLineCallback)
+    , isStarted(false)
+    , isFinished(false)
+    , isGreaterThanFound(false)
+  {}
+
+  /**
+   * IsStartingLine
+   *
+   * If the line is starting with `<`, HTML is expected to follow.
+   * Nothing after that will be parsed, it only is copied.
+   *
+   * @method
+   * @param {const std::string&} line
+   * @return {bool}
+   */
+  static bool
+  IsStartingLine(const std::string& line)
+  {
+    return line[0] == '<';
+  }
+
+  /**
+   * IsFinished
+   *
+   * `>` followed by an empty line will end the HTML block.
+   *
+   * @method
+   * @return {bool}
+   */
+  bool
+  IsFinished() const override
+  {
+    return this->isFinished;
+  }
+
+protected:
+  bool
+  isInlineBlockAllowed() const override
+  {
+    return false;
+  }
+
+  bool
+  isLineParserAllowed() const override
+  {
+    return false;
+  }
+
+  void
+  parseBlock(std::string& line) override
+  {
+    if (!this->isStarted)
+    {
+      this->isStarted = true;
+    }
+
+    if (!line.empty() && line[line.size() - 1] == '>')
+    {
+      this->isGreaterThanFound = true;
+      return;
+    }
+
+    if (line.empty() && this->isGreaterThanFound)
+    {
+      this->isFinished = true;
+      return;
+    }
+
+    if (!line.empty() && this->isGreaterThanFound)
+    {
+      this->isGreaterThanFound = false;
+    }
+
+    if (!line.empty())
+    {
+      line += " ";
+    }
+  }
+
+private:
+  bool isStarted;
+  bool isFinished;
+  bool isGreaterThanFound;
+}; // class HtmlParser
+
+// -----------------------------------------------------------------------------
+
+} // namespace maddy

include/maddy/paragraphparser.h

@@ -44,8 +44,9 @@ public:
   /**
    * IsStartingLine
    *
-   * If the line is not empty, it will be a paragraph. So this block parser has
+   * If the line is not empty, it will be a paragraph.
-   * to always run as the last one!
+   *
+   * This block parser has to always run as the last one!
    *
    * @method
    * @param {const std::string&} line

include/maddy/parser.h

@@ -10,11 +10,14 @@
 #include <functional>
 #include <string>
 
+#include "maddy/parserconfig.h"
+
 // BlockParser
 #include "maddy/checklistparser.h"
 #include "maddy/codeblockparser.h"
 #include "maddy/headlineparser.h"
 #include "maddy/horizontallineparser.h"
+#include "maddy/htmlparser.h"
 #include "maddy/orderedlistparser.h"
 #include "maddy/paragraphparser.h"
 #include "maddy/quoteparser.h"
@@ -54,8 +57,9 @@ public:
    *
    * @method
    */
-  Parser()
+  Parser(std::shared_ptr<ParserConfig> config = nullptr)
-    : breakLineParser(std::make_shared<BreakLineParser>())
+    : config(config)
+    , breakLineParser(std::make_shared<BreakLineParser>())
     , emphasizedParser(std::make_shared<EmphasizedParser>())
     , imageParser(std::make_shared<ImageParser>())
     , inlineCodeParser(std::make_shared<InlineCodeParser>())
@@ -113,6 +117,7 @@ public:
   }
 
 private:
+  std::shared_ptr<ParserConfig> config;
   std::shared_ptr<BreakLineParser> breakLineParser;
   std::shared_ptr<EmphasizedParser> emphasizedParser;
   std::shared_ptr<ImageParser> imageParser;
@@ -132,7 +137,11 @@ private:
 
     // Attention! StrongParser has to be before EmphasizedParser
     this->strongParser->Parse(line);
+
+    if (!this->config || this->config->isEmphasizedParserEnabled)
+    {
       this->emphasizedParser->Parse(line);
+    }
 
     this->strikeThroughParser->Parse(line);
 
@@ -195,6 +204,14 @@ private:
     {
       parser = this->createUnorderedListParser();
     }
+    else if (
+      this->config &&
+      !this->config->isHTMLWrappedInParagraph &&
+      maddy::HtmlParser::IsStartingLine(line)
+    )
+    {
+      parser = std::make_shared<maddy::HtmlParser>(nullptr, nullptr);
+    }
     else if (maddy::ParagraphParser::IsStartingLine(line))
     {
       parser = std::make_shared<maddy::ParagraphParser>(

include/maddy/parserconfig.h

@@ -0,0 +1,31 @@
+/*
+ * This project is licensed under the MIT license. For more information see the
+ * LICENSE file.
+ */
+#pragma once
+
+// -----------------------------------------------------------------------------
+
+namespace maddy {
+
+// -----------------------------------------------------------------------------
+
+/**
+ * ParserConfig
+ *
+ * @class
+ */
+struct ParserConfig
+{
+  bool isEmphasizedParserEnabled;
+  bool isHTMLWrappedInParagraph;
+
+  ParserConfig()
+    : isEmphasizedParserEnabled(true)
+    , isHTMLWrappedInParagraph(true)
+  {}
+}; // class ParserConfig
+
+// -----------------------------------------------------------------------------
+
+} // namespace maddy

tests/maddy/test_maddy_htmlparser.cpp

@@ -0,0 +1,82 @@
+/*
+ * This project is licensed under the MIT license. For more information see the
+ * LICENSE file.
+ */
+#include <memory>
+
+#include "gmock/gmock.h"
+
+#include "maddy/htmlparser.h"
+
+// -----------------------------------------------------------------------------
+
+class MADDY_HTMLPARSER : public ::testing::Test
+{
+protected:
+  std::shared_ptr<maddy::HtmlParser> pParser;
+
+  void
+  SetUp() override
+  {
+    this->pParser = std::make_shared<maddy::HtmlParser>(
+      nullptr,
+      nullptr
+    );
+  }
+};
+
+// -----------------------------------------------------------------------------
+
+TEST_F(MADDY_HTMLPARSER, IsFinishedReturnsFalseInTheBeginning)
+{
+  ASSERT_FALSE(pParser->IsFinished());
+}
+
+TEST_F(MADDY_HTMLPARSER, IsStartingLineReturnsFalseWhenFacedWithNoSmallerThan)
+{
+  const std::vector<std::string> markdown = {
+    "> quote"
+    , "some text"
+    , "* list"
+    , "1. numbered list"
+    , "|table>"
+  };
+
+  for (size_t i = 0; i < markdown.size(); ++i)
+  {
+    ASSERT_FALSE(maddy::HtmlParser::IsStartingLine(markdown[i]));
+  }
+}
+
+TEST_F(MADDY_HTMLPARSER, IsStartingLineReturnsTrueWhenFacedWithSmallerThan)
+{
+  const std::string markdown = "<div id=\"test\">test element</div>";
+
+  ASSERT_TRUE(maddy::HtmlParser::IsStartingLine(markdown));
+}
+
+TEST_F(MADDY_HTMLPARSER, ItReplacesNoHtml)
+{
+  const std::vector<std::string> markdown {
+    "some text in a paragraph"
+    , ""
+    , "<div> some HTML</div>"
+    , ""
+    , "<div>more"
+    , "HTML"
+    , "</div>"
+    , ""
+  };
+  const std::string expected = "some text in a paragraph <div> some HTML</div><div>more HTML </div>";
+
+  for (std::string md : markdown)
+  {
+    pParser->AddLine(md);
+  }
+  ASSERT_TRUE(pParser->IsFinished());
+
+  std::stringstream& output(pParser->GetResult());
+  const std::string& outputString = output.str();
+
+  ASSERT_EQ(expected, outputString);
+}

tests/maddy/test_maddy_parser.cpp

@@ -19,3 +19,18 @@ TEST(MADDY_PARSER, ItShouldParse)
 
   ASSERT_EQ(testHtml, output);
 }
+
+TEST(MADDY_PARSER, ItShouldParseWithConfig)
+{
+  auto config = std::make_shared<maddy::ParserConfig>();
+  config->isEmphasizedParserEnabled = false;
+  config->isHTMLWrappedInParagraph = false;
+
+  auto parser = std::make_shared<maddy::Parser>(config);
+
+  std::stringstream markdown(testMarkdown);
+
+  const std::string output = parser->Parse(markdown);
+
+  ASSERT_EQ(testHtml2, output);
+}

tests/maddy/test_maddy_parser.h

@@ -42,6 +42,8 @@ And well - let's see how an image would be shown:\n\
 \n\
 ---\n\
 \n\
+<a name=\"to top\"></a>\n\
+\n\
 ### and more headlines\n\
 \n\
 - [ ] how\n\
@@ -67,4 +69,5 @@ foot a|foot b|foot c\n\
 \n\
 ";
 
-const std::string testHtml = "<h1>This is a test</h1><p>This should result in a praragraph it's that simple. </p><ul><li>an <i>unordered</i> list<ul><li>with some <strong>hierarchy</strong><ol><li>and an <em>ordered</em></li><li>list</li><li>directly</li></ol></li><li>inside</li></ul></li></ul><pre><code>\nvar c = 'blub';\n</code></pre><blockquote><p>A Quote  </p><p>With some <s>text</s>  blocks inside  </p><ul><li>even a list </li><li>should be </li><li>possible </li></ul></blockquote><p>And well <code>inline code</code> should also work. </p><h2>Another Headline</h2><p>And not to forget <a href=\"http://progsource.de\">link to progsource</a> should work. And well - let's see how an image would be shown: </p><p><img src=\"http://progsource.de/img/progsource.png\" alt=\"an image\"/> </p><hr/><h3>and more headlines</h3><ul class=\"checklist\"><li><label><input type=\"checkbox\"/> how</label></li><li><label><input type=\"checkbox\"/> about<ul class=\"checklist\"><li><label><input type=\"checkbox\"/> a</label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> nice</label></li></ul></label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> check</label></li><li><label><input type=\"checkbox\"/> list</label></li></ul><h4>even a table</h4><table><thead><tr><th>Left header</th><th>middle header</th><th>last header</th></tr></thead><tbody><tr><td>cell 1</td><td>cell <strong>2</strong></td><td>cell 3</td></tr><tr><td>cell 4</td><td>cell 5</td><td>cell 6</td></tr></tbody><tfoot><tr><td>foot a</td><td>foot b</td><td>foot c</td></tr></tfoot></table><h5>h5</h5><h6>h6</h6>";
+const std::string testHtml = "<h1>This is a test</h1><p>This should result in a praragraph it's that simple. </p><ul><li>an <i>unordered</i> list<ul><li>with some <strong>hierarchy</strong><ol><li>and an <em>ordered</em></li><li>list</li><li>directly</li></ol></li><li>inside</li></ul></li></ul><pre><code>\nvar c = 'blub';\n</code></pre><blockquote><p>A Quote  </p><p>With some <s>text</s>  blocks inside  </p><ul><li>even a list </li><li>should be </li><li>possible </li></ul></blockquote><p>And well <code>inline code</code> should also work. </p><h2>Another Headline</h2><p>And not to forget <a href=\"http://progsource.de\">link to progsource</a> should work. And well - let's see how an image would be shown: </p><p><img src=\"http://progsource.de/img/progsource.png\" alt=\"an image\"/> </p><hr/><p><a name=\"to top\"></a> </p><h3>and more headlines</h3><ul class=\"checklist\"><li><label><input type=\"checkbox\"/> how</label></li><li><label><input type=\"checkbox\"/> about<ul class=\"checklist\"><li><label><input type=\"checkbox\"/> a</label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> nice</label></li></ul></label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> check</label></li><li><label><input type=\"checkbox\"/> list</label></li></ul><h4>even a table</h4><table><thead><tr><th>Left header</th><th>middle header</th><th>last header</th></tr></thead><tbody><tr><td>cell 1</td><td>cell <strong>2</strong></td><td>cell 3</td></tr><tr><td>cell 4</td><td>cell 5</td><td>cell 6</td></tr></tbody><tfoot><tr><td>foot a</td><td>foot b</td><td>foot c</td></tr></tfoot></table><h5>h5</h5><h6>h6</h6>";
+const std::string testHtml2 = "<h1>This is a test</h1><p>This should result in a praragraph it's that simple. </p><ul><li>an <i>unordered</i> list<ul><li>with some <strong>hierarchy</strong><ol><li>and an _ordered_</li><li>list</li><li>directly</li></ol></li><li>inside</li></ul></li></ul><pre><code>\nvar c = 'blub';\n</code></pre><blockquote><p>A Quote  </p><p>With some <s>text</s>  blocks inside  </p><ul><li>even a list </li><li>should be </li><li>possible </li></ul></blockquote><p>And well <code>inline code</code> should also work. </p><h2>Another Headline</h2><p>And not to forget <a href=\"http://progsource.de\">link to progsource</a> should work. And well - let's see how an image would be shown: </p><p><img src=\"http://progsource.de/img/progsource.png\" alt=\"an image\"/> </p><hr/><a name=\"to top\"></a><h3>and more headlines</h3><ul class=\"checklist\"><li><label><input type=\"checkbox\"/> how</label></li><li><label><input type=\"checkbox\"/> about<ul class=\"checklist\"><li><label><input type=\"checkbox\"/> a</label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> nice</label></li></ul></label></li><li><label><input type=\"checkbox\" checked=\"checked\"/> check</label></li><li><label><input type=\"checkbox\"/> list</label></li></ul><h4>even a table</h4><table><thead><tr><th>Left header</th><th>middle header</th><th>last header</th></tr></thead><tbody><tr><td>cell 1</td><td>cell <strong>2</strong></td><td>cell 3</td></tr><tr><td>cell 4</td><td>cell 5</td><td>cell 6</td></tr></tbody><tfoot><tr><td>foot a</td><td>foot b</td><td>foot c</td></tr></tfoot></table><h5>h5</h5><h6>h6</h6>";