LibMarkdown: Implement hard and soft line breaks

Hard line breaks insert a <br /> when two spaces are at the end of a
line. soft line breaks are just regular newlines, but whitespace is now
stripped before and after them

3 files changed, 75 insertions, 3 deletions

diff --git a/Userland/Libraries/LibMarkdown/Document.cpp b/Userland/Libraries/LibMarkdown/Document.cpp
index 3e91b9f67f..059bbc97c9 100644
--- a/Userland/Libraries/LibMarkdown/Document.cpp
+++ b/Userland/Libraries/LibMarkdown/Document.cpp
@@ -92,6 +92,7 @@ OwnPtr<Document> Document::parse(const StringView& str)
 
         if ((*lines).is_empty()) {
             ++lines;
+
             flush_paragraph();
             continue;
         }
@@ -108,8 +109,9 @@ OwnPtr<Document> Document::parse(const StringView& str)
             continue;
         }
 
+        if (!paragraph_text.is_empty())
+            paragraph_text.append("\n");
         paragraph_text.append(*lines++);
-        paragraph_text.append("\n");
     }
 
     flush_paragraph();
diff --git a/Userland/Libraries/LibMarkdown/Text.cpp b/Userland/Libraries/LibMarkdown/Text.cpp
index 583b62bff4..71995fc7fb 100644
--- a/Userland/Libraries/LibMarkdown/Text.cpp
+++ b/Userland/Libraries/LibMarkdown/Text.cpp
@@ -58,6 +58,20 @@ size_t Text::CodeNode::terminal_length() const
     return code->terminal_length();
 }
 
+void Text::BreakNode::render_to_html(StringBuilder& builder) const
+{
+    builder.append("<br />");
+}
+
+void Text::BreakNode::render_for_terminal(StringBuilder&) const
+{
+}
+
+size_t Text::BreakNode::terminal_length() const
+{
+    return 0;
+}
+
 void Text::TextNode::render_to_html(StringBuilder& builder) const
 {
     builder.append(escape_html_entities(text));
@@ -212,6 +226,8 @@ Vector<Text::Token> Text::tokenize(StringView const& str)
         flush_run(false, false, false, false, false);
     };
 
+    bool in_space = false;
+
     for (size_t offset = 0; offset < str.length(); ++offset) {
         auto has = [&](StringView const& seq) {
             if (offset + seq.length() > str.length())
@@ -229,6 +245,10 @@ Vector<Text::Token> Text::tokenize(StringView const& str)
         };
 
         char ch = str[offset];
+        if (ch != ' ' && in_space) {
+            flush_token();
+            in_space = false;
+        }
 
         if (ch == '\\' && offset + 1 < str.length()) {
             current_token.append(str[offset + 1]);
@@ -249,6 +269,12 @@ Vector<Text::Token> Text::tokenize(StringView const& str)
                 true);
             offset = run_offset - 1;
 
+        } else if (ch == ' ') {
+            if (!in_space) {
+                flush_token();
+                in_space = true;
+            }
+            current_token.append(ch);
         } else if (has("\n")) {
             expect("\n");
         } else if (has("[")) {
@@ -272,7 +298,11 @@ NonnullOwnPtr<Text::MultiNode> Text::parse_sequence(Vector<Token>::ConstIterator
     auto node = make<MultiNode>();
 
     for (; !tokens.is_end(); ++tokens) {
-        if (tokens->is_run) {
+        if (tokens->is_space()) {
+            node->children.append(parse_break(tokens));
+        } else if (*tokens == "\n") {
+            node->children.append(parse_newline(tokens));
+        } else if (tokens->is_run) {
             switch (tokens->run_char()) {
             case '*':
             case '_':
@@ -299,6 +329,29 @@ NonnullOwnPtr<Text::MultiNode> Text::parse_sequence(Vector<Token>::ConstIterator
     return node;
 }
 
+NonnullOwnPtr<Text::Node> Text::parse_break(Vector<Token>::ConstIterator& tokens)
+{
+    auto next_tok = tokens + 1;
+    if (next_tok.is_end() || *next_tok != "\n")
+        return make<TextNode>(tokens->data);
+
+    if (tokens->data.length() >= 2)
+        return make<BreakNode>();
+
+    return make<MultiNode>();
+}
+
+NonnullOwnPtr<Text::Node> Text::parse_newline(Vector<Token>::ConstIterator& tokens)
+{
+    auto node = make<TextNode>(tokens->data);
+    auto next_tok = tokens + 1;
+    if (!next_tok.is_end() && next_tok->is_space())
+        // Skip whitespace after newline.
+        ++tokens;
+
+    return node;
+}
+
 bool Text::can_open(Token const& opening)
 {
     return (opening.run_char() == '*' && opening.left_flanking) || (opening.run_char() == '_' && opening.left_flanking && (!opening.right_flanking || opening.punct_before));
@@ -325,7 +378,11 @@ NonnullOwnPtr<Text::Node> Text::parse_emph(Vector<Token>::ConstIterator& tokens,
 
     auto child = make<MultiNode>();
     for (++tokens; !tokens.is_end(); ++tokens) {
-        if (tokens->is_run) {
+        if (tokens->is_space()) {
+            child->children.append(parse_break(tokens));
+        } else if (*tokens == "\n") {
+            child->children.append(parse_newline(tokens));
+        } else if (tokens->is_run) {
             if (can_close_for(opening, *tokens)) {
                 return make<EmphasisNode>(opening.run_length() >= 2, move(child));
             }
diff --git a/Userland/Libraries/LibMarkdown/Text.h b/Userland/Libraries/LibMarkdown/Text.h
index fdb7f3ed63..d7715412cf 100644
--- a/Userland/Libraries/LibMarkdown/Text.h
+++ b/Userland/Libraries/LibMarkdown/Text.h
@@ -55,6 +55,13 @@ public:
         virtual size_t terminal_length() const override;
     };
 
+    class BreakNode : public Node {
+    public:
+        virtual void render_to_html(StringBuilder& builder) const override;
+        virtual void render_for_terminal(StringBuilder& builder) const override;
+        virtual size_t terminal_length() const override;
+    };
+
     class TextNode : public Node {
     public:
         String text;
@@ -128,6 +135,10 @@ private:
             VERIFY(is_run);
             return data.length();
         }
+        bool is_space() const
+        {
+            return data[0] == ' ';
+        }
         bool operator==(StringView const& str) const { return str == data; }
     };
 
@@ -137,6 +148,8 @@ private:
     static bool can_close_for(Token const& opening, Token const& closing);
 
     static NonnullOwnPtr<MultiNode> parse_sequence(Vector<Token>::ConstIterator& tokens, bool in_link);
+    static NonnullOwnPtr<Node> parse_break(Vector<Token>::ConstIterator& tokens);
+    static NonnullOwnPtr<Node> parse_newline(Vector<Token>::ConstIterator& tokens);
     static NonnullOwnPtr<Node> parse_emph(Vector<Token>::ConstIterator& tokens, bool in_link);
     static NonnullOwnPtr<Node> parse_code(Vector<Token>::ConstIterator& tokens);
     static NonnullOwnPtr<Node> parse_link(Vector<Token>::ConstIterator& tokens);