feat(formatter): adding a first version of an ArkScript code formatter

CMakeLists.txt

@@ -184,6 +184,7 @@ if (ARK_TESTS)
     add_executable(unittests ${UT_SOURCES})
 
     add_subdirectory(${ark_SOURCE_DIR}/lib/ut)
+    target_include_directories(unittests PUBLIC ${ark_SOURCE_DIR}/include)
     target_link_libraries(unittests PUBLIC ArkReactor termcolor ut)
 
     add_compile_definitions(BOOST_UT_DISABLE_MODULE)

README.md

@@ -186,20 +186,35 @@ SYNOPSIS
         arkscript --dev-info 
         arkscript -e <expression> 
         arkscript -c <file> [-d] 
+        arkscript <file> [-d] [-L <lib_dir>]  
+        arkscript -f <file> [--dry-run] 
+        arkscript --ast <file> [-d] [-L <lib_dir>] 
         arkscript -bcr <file> -on 
         arkscript -bcr <file> -a [-s <start> <end>] 
         arkscript -bcr <file> -st [-s <start> <end>] 
         arkscript -bcr <file> -vt [-s <start> <end>] 
         arkscript -bcr <file> [-cs] [-p <page>] [-s <start> <end>] 
-        arkscript <file> [-d] [-L <lib_dir>]  
 
 OPTIONS
         -h, --help                  Display this message
         -v, --version               Display ArkScript version and exit
         --dev-info                  Display development information and exit
         -e, --eval                  Evaluate ArkScript expression
+
         -c, --compile               Compile the given program to bytecode, but do not run
         -d, --debug...              Increase debug level (default: 0)
+
+        -L, --lib                   Set the location of the ArkScript standard library. Paths can be
+                                    delimited by ';'
+
+        -f, --format                Format the given source file in place
+        --dry-run                   Do not modify the file, only print out the changes
+
+        --ast                       Compile the given program and output its AST as JSON to stdout
+        -d, --debug...              Increase debug level (default: 0)
+        -L, --lib                   Set the location of the ArkScript standard library. Paths can be
+                                    delimited by ';'
+
         -bcr, --bytecode-reader     Launch the bytecode reader
         -on, --only-names           Display only the bytecode segments names and sizes
         -a, --all                   Display all the bytecode segments (default)
@@ -208,8 +223,9 @@ OPTIONS
         -cs, --code                 Display only the code segments
         -p, --page                  Set the bytecode reader code segment to display
         -s, --slice                 Select a slice of instructions in the bytecode
-        -L, --lib                   Set the location of the ArkScript standard library. Paths can be
-                                    delimited by ';'
+
+VERSION
+        4.0.0-86587c14
 
 LICENSE
         Mozilla Public License 2.0

include/CLI/Formatter.hpp

@@ -0,0 +1,79 @@
+#ifndef ARK_FORMATTER_HPP
+#define ARK_FORMATTER_HPP
+
+#include <string>
+
+#include <Ark/Compiler/AST/Parser.hpp>
+
+constexpr struct FormatterConfig
+{
+    static constexpr std::size_t SpacePerIndent = 2;   ///< Indentation level of each node
+    static constexpr std::size_t LongLineLength = 32;  ///< Max number of characters per line segment to consider splitting
+} FormatterConfig;
+
+class Formatter final
+{
+public:
+    Formatter(std::string filename, bool dry_run);
+
+    void run();
+
+    [[nodiscard]] const std::string& output() const;
+
+private:
+    const std::string m_filename;
+    bool m_dry_run;  ///< If true, only prints the formatted file instead of saving it to disk
+    Ark::internal::Parser m_parser;
+    std::string m_output;
+
+    bool isListStartingWithKeyword(const Ark::internal::Node& node, Ark::internal::Keyword keyword);
+    bool isBeginBlock(const Ark::internal::Node& node);
+    bool isFuncDef(const Ark::internal::Node& node);
+    bool isFuncCall(const Ark::internal::Node& node);
+
+    /**
+     * @param node
+     * @return true if the node is a String|Number|Symbol|Field
+     * @return false
+     */
+    bool isPlainValue(const Ark::internal::Node& node);
+
+    /**
+     * @brief Compute the line on which the deepest right most node of node is at
+     * @param node
+     * @return
+     */
+    std::size_t lineOfLastNodeIn(const Ark::internal::Node& node);
+
+    bool should_split_on_newline(const Ark::internal::Node& node);
+
+    inline constexpr std::string prefix(std::size_t indent) const
+    {
+        return std::string(indent * FormatterConfig.SpacePerIndent, ' ');
+    }
+
+    /**
+     * @brief Handles all node formatting
+     * @param node
+     * @param indent indentation level, starting at 0, increment by 1
+     * @param after_newline when false, do not add prefix
+     * @return
+     */
+    std::string format(const Ark::internal::Node& node, std::size_t indent, bool after_newline);
+
+    std::string formatComment(const std::string& comment, std::size_t indent);
+
+    std::string formatBlock(const Ark::internal::Node& node, std::size_t indent, bool after_newline);
+
+    std::string formatFunction(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatVariable(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatCondition(const Ark::internal::Node& node, std::size_t indent, bool is_macro = false);
+    std::string formatLoop(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatBegin(const Ark::internal::Node& node, std::size_t indent, bool after_newline);
+    std::string formatImport(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatDel(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatCall(const Ark::internal::Node& node, std::size_t indent);
+    std::string formatMacro(const Ark::internal::Node& node, std::size_t indent);
+};
+
+#endif  // ARK_FORMATTER_HPP