From a382841dc7d554b2ba3920d52f73e4cec428743e Mon Sep 17 00:00:00 2001
From: Marvin Borner
Date: Mon, 12 Aug 2019 16:01:51 +0200
Subject: Basic parser functions

---
 .gitignore       |  3 +++
 example.run      |  1 +
 src/DataTypes.kt |  3 +++
 src/Lexical.kt   | 64 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 src/Loader.kt    | 12 +++++++++++
 src/Main.kt      |  5 +++++
 src/Testing.kt   | 13 ++++++++++++
 7 files changed, 101 insertions(+)
 create mode 100644 .gitignore
 create mode 100644 example.run
 create mode 100644 src/DataTypes.kt
 create mode 100644 src/Lexical.kt
 create mode 100644 src/Loader.kt
 create mode 100644 src/Main.kt
 create mode 100644 src/Testing.kt

diff --git a/.gitignore b/.gitignore
new file mode 100644
index 0000000..5f05927
--- /dev/null
+++ b/.gitignore
@@ -0,0 +1,3 @@
+.idea
+out
+*.iml
\ No newline at end of file
diff --git a/example.run b/example.run
new file mode 100644
index 0000000..f7d418f
--- /dev/null
+++ b/example.run
@@ -0,0 +1 @@
+print string(2 + 2);
\ No newline at end of file
diff --git a/src/DataTypes.kt b/src/DataTypes.kt
new file mode 100644
index 0000000..290481d
--- /dev/null
+++ b/src/DataTypes.kt
@@ -0,0 +1,3 @@
+class Int {
+
+}
\ No newline at end of file
diff --git a/src/Lexical.kt b/src/Lexical.kt
new file mode 100644
index 0000000..bd8bc3a
--- /dev/null
+++ b/src/Lexical.kt
@@ -0,0 +1,64 @@
+class Lexical {
+    fun analyze(source: String): MutableList<Pair<String, TokenType>> {
+        var buffer = ""
+        val tokens = mutableListOf<Pair<String, TokenType>>()
+        for (i in source.indices) {
+            if (source.length > i + 1 && getType(source[i]) == getType(source[i + 1])) {
+                buffer += source[i]
+            } else {
+                buffer += source[i]
+                if (getType(source[i]) != Type.Empty)
+                    tokens += buffer to getTokenType(buffer to getType(source[i]))
+                buffer = ""
+            }
+        }
+        return tokens
+    }
+
+    /**
+     * Matches the characters to groups ([Type])
+     */
+    private fun getType(char: Char): Type {
+        return when (char) {
+            in 'a'..'z', in 'A'..'Z' -> Type.Alphabetic
+            in '1'..'9' -> Type.Numeric
+            ' ' -> Type.Empty
+            else -> Type.Unknown
+        }
+    }
+
+    /**
+     * Matches the tokens to a [TokenType]
+     */
+    private fun getTokenType(token: Pair<String, Type>): TokenType {
+        var type = when (token.first) {
+            "print" -> TokenType.Keyword // TODO: DataType matching
+            "=", "+", "-", "*", "/", "==", ">", "<" -> TokenType.Operator
+            ";", ":" -> TokenType.Symbol
+            else -> TokenType.Unknown
+        }
+        if (type == TokenType.Unknown)
+            type = when (token.second) {
+                Type.Numeric -> TokenType.Constant
+                Type.Alphabetic -> TokenType.Identifier
+                else -> TokenType.Unknown
+            }
+        return type
+    }
+}
+
+private enum class Type {
+    Alphabetic,
+    Numeric,
+    Empty,
+    Unknown
+}
+
+enum class TokenType {
+    Keyword,
+    Identifier,
+    Operator,
+    Constant,
+    Symbol,
+    Unknown
+}
\ No newline at end of file
diff --git a/src/Loader.kt b/src/Loader.kt
new file mode 100644
index 0000000..2161c51
--- /dev/null
+++ b/src/Loader.kt
@@ -0,0 +1,12 @@
+import java.io.File
+import java.io.InputStream
+
+class Loader(path: String) {
+    private val inputStream: InputStream = File(path).inputStream()
+    private val inputString = inputStream.bufferedReader().use { it.readText() }
+
+    // TODO: Add preprocessor managing imports
+    fun load(): String {
+        return inputString
+    }
+}
\ No newline at end of file
diff --git a/src/Main.kt b/src/Main.kt
new file mode 100644
index 0000000..24a2e08
--- /dev/null
+++ b/src/Main.kt
@@ -0,0 +1,5 @@
+import Testing
+
+fun main() {
+    Testing()
+}
diff --git a/src/Testing.kt b/src/Testing.kt
new file mode 100644
index 0000000..0b16145
--- /dev/null
+++ b/src/Testing.kt
@@ -0,0 +1,13 @@
+class Testing {
+    init {
+        val source = Loader("/home/melvin/coding/run/example.run").load()
+        val tokens = Lexical().analyze(source)
+        for (token in tokens) {
+            print(token.first)
+            print("\n")
+            print(token.second)
+            print("\n")
+        }
+        print(source)
+    }
+}
\ No newline at end of file
-- 
cgit v1.2.3