Add crate benchmark by White-Oak plus makefile

Various cleanups

Author: smt923

collector/benchmarks/issue-31157-compilation-time-ram-regression/.gitignore

@@ -0,0 +1,2 @@
+*.lock
+target/

collector/benchmarks/issue-31157-compilation-time-ram-regression/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "parser"
+version = "0.1.0"
+authors = ["White-Oak <[email protected]>"]
+
+[dependencies]
+
+[dependencies.peruse]
+git = "https://github.com/DanSimon/peruse.git"

collector/benchmarks/issue-31157-compilation-time-ram-regression/makefile

@@ -0,0 +1,6 @@
+.PHONY: all
+
+all:
+	$(CARGO) rustc $(CARGO_OPTS) -- $(CARGO_RUSTC_OPTS)
+patches:
+	@echo ''

collector/benchmarks/issue-31157-compilation-time-ram-regression/src/grammar.rs

@@ -0,0 +1,70 @@
+
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(PartialEq)]
+pub enum AddOp {
+  Add,
+  Subtract,
+  Start,
+}
+
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(PartialEq)]
+pub enum MultOp {
+  Multiply,
+  Divide,
+  Modulo,
+  Start,
+}
+
+#[derive(Clone)]
+#[derive(Debug)]
+#[derive(PartialEq)]
+pub struct AddTerm(pub AddOp, pub Expr);
+
+#[derive(Clone)]
+#[derive(Debug)]
+#[derive(PartialEq)]
+pub struct MultTerm(pub MultOp, pub Expr);
+
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(PartialEq)]
+pub enum Expr {
+  Variable(String),
+  Num(i32),
+  AddSub(Vec<AddTerm>), //a + b - c + d becomes [(+ a) (+ b) (- c) (+ d)]
+  MultDiv(Vec<MultTerm>),
+}
+
+
+//for now this is it's own type and not a statement
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(PartialEq)]
+pub struct Block(pub Vec<Statement>);
+
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(PartialEq)]
+pub enum Statement {
+  Assign(String, Expr),
+  Output(Expr),
+  If(Expr, Comparator, Expr, Block, Option<Block>),
+  While(Expr, Comparator, Expr, Block),
+  Loop(Expr, Block),
+}
+
+#[derive(Debug)]
+#[derive(Clone)]
+#[derive(Eq)]
+#[derive(PartialEq)]
+pub enum Comparator {
+  CEq,  // ==
+  CGt,  // >
+  CLt,  // <
+  CNeq, // !=
+  CGeq, // >=
+  CLeq, // <=
+}

collector/benchmarks/issue-31157-compilation-time-ram-regression/src/grammar_lexer.rs

@@ -0,0 +1,27 @@
+use grammar::*;
+
+#[derive(Debug)]
+#[derive(Eq)]
+#[derive(PartialEq)]
+#[derive(Clone)]
+pub enum Token {
+  Equals,
+  Ident(String),
+  Number(i32),
+  PlusSign,
+  MinusSign,
+  MultSign,
+  DivideSign,
+  ModuloSign,
+  OutputCmd,
+  NewLine,
+  OpenParen,
+  CloseParen,
+  OpenBrace,
+  CloseBrace,
+  IfKeyword,
+  ElseKeyword,
+  WhileKeyword,
+  LoopKeyword,
+  Cmp(Comparator),
+}

collector/benchmarks/issue-31157-compilation-time-ram-regression/src/lexer.rs

@@ -0,0 +1,49 @@
+use peruse::string_parsers::*;
+use peruse::parsers::*;
+
+use grammar_lexer::*;
+use grammar::*;
+use std::str::FromStr;
+
+//type Lexer = SliceParser<I=str, O=Token>;
+
+
+pub fn token() -> Box<Parser<I=str, O=Vec<Token>>> {
+
+  fn lt(s: &str, t: Token) -> RegexLiteralParser<Token> {
+    str_lit(&(String::from_str(r"^[ \t]*").unwrap() + s), t)
+  }
+
+  let ident = capture(r"^[ \t]*([a-zA-Z]\w*)[ \t]*", |caps| Token::Ident(String::from_str(caps.at(1).unwrap()).unwrap()));
+
+  let number = capture(r"^[ \t]*(\d+)[ \t]*", |caps| Token::Number(FromStr::from_str(caps.at(1).unwrap()).unwrap()));
+
+  let lits = one_of(vec![
+    lt("out",         Token::OutputCmd),
+    lt("if",          Token::IfKeyword),
+    lt("else",        Token::ElseKeyword),
+    lt("while",       Token::WhileKeyword),
+    lt("loop",       Token::LoopKeyword),
+    lt(r"\r?\n\s*",   Token::NewLine),
+    lt(r"\(\s*",      Token::OpenParen),
+    lt(r"\)",         Token::CloseParen),
+    lt(r"\}",         Token::CloseBrace),
+    lt("==",          Token::Cmp(Comparator::CEq)),
+    lt("!=",          Token::Cmp(Comparator::CNeq)),
+    lt(">=",          Token::Cmp(Comparator::CGeq)),
+    lt(r"\{\s*",      Token::OpenBrace),
+    lt("<=",          Token::Cmp(Comparator::CLeq)),
+    lt(">",           Token::Cmp(Comparator::CGt)),
+    lt("<",           Token::Cmp(Comparator::CLt)),
+    lt(r"\+",         Token::PlusSign),
+    lt("-",           Token::MinusSign),
+    lt("=",           Token::Equals),
+    lt(r"\*",         Token::MultSign),
+    lt("/",           Token::DivideSign),
+    lt(r"%",           Token::ModuloSign)
+  ]);
+
+  let options = lits.or(number).or(ident).repeat();
+  Box::new(options)
+
+}

collector/benchmarks/issue-31157-compilation-time-ram-regression/src/lib.rs

@@ -0,0 +1,13 @@
+extern crate peruse;
+
+pub mod grammar;
+pub mod grammar_lexer;
+pub mod lexer;
+pub mod parser;
+
+#[cfg(test)]
+mod test {
+    #[test]
+    fn it_works() {
+    }
+}

collector/benchmarks/issue-31157-compilation-time-ram-regression/src/parser.rs

@@ -0,0 +1,143 @@
+use grammar::*;
+use grammar_lexer::*;
+
+use peruse::slice_parsers::*;
+use peruse::parsers::*;
+
+pub fn program() -> Box<Parser<I=[Token], O=Block>> {
+
+    fn expression() -> Box<Parser<I=[Token], O=Expr>> {
+        let simple_term = matcher(|token| match token {
+            Token::Ident(name) => Some(Expr::Variable(name)),
+            Token::Number(i)   => Some(Expr::Num(i)),
+            _ => None
+        });
+
+        let paren_expr = || lit(Token::OpenParen).then_r(recursive(|| expression())).then_l(lit(Token::CloseParen));
+
+        let factor = paren_expr().or(simple_term);
+
+        let multop = matcher(|token| match token {
+            Token::MultSign     => Some(MultOp::Multiply),
+            Token::DivideSign   => Some(MultOp::Divide),
+            Token::ModuloSign   => Some(MultOp::Modulo),
+            _ => None
+        });
+
+        let addop = matcher(|token| match token {
+            Token::PlusSign     => Some(AddOp::Add),
+            Token::MinusSign    => Some(AddOp::Subtract),
+            _ => None
+        });
+
+        let mult =  {
+            let p = factor.clone().then(multop.then(factor).repeat()).map(|(first, seq)| {
+                let mut ops = Vec::new();
+                ops.push(MultTerm(MultOp::Start, first));
+                for &(ref op, ref value) in seq.iter() {
+                    ops.push(MultTerm(op.clone(), value.clone())); //maybe box the value instead
+                }
+                if ops.len() == 1 {
+                    ops[0].1.clone()
+                } else{
+                    Expr::MultDiv(ops)
+                }
+            });
+            boxed(p)
+        };
+
+        let plus = {
+            let p = mult.clone().then(addop.then(mult).repeat()).map(|(first, seq)| {
+                let mut ops = Vec::new();
+                    ops.push(AddTerm(AddOp::Start, first));
+                    for &(ref op, ref value) in seq.iter() {
+                        ops.push(AddTerm(op.clone(), value.clone()));
+                    }
+                    if ops.len() == 1 {
+                        ops[0].1.clone()
+                    } else{
+                        Expr::AddSub(ops)
+                    }
+            });
+            boxed(p)
+        };
+
+        Box::new(plus)
+    }
+
+    let assignment = {
+        let target = matcher(|token| match token {
+            Token::Ident(name) => Some(name),
+            _ => None
+        });
+        boxed(target
+            .then_l(lit(Token::Equals))
+            .then(recursive(|| expression()))
+            .map(|(target, expr)| Statement::Assign(target, expr)))
+        };
+
+        let comparator = || matcher(|token| match token {
+            Token::Cmp(cmp) => Some(cmp),
+            _ => None
+        });
+
+        fn code_block() -> Box<Parser<I=[Token], O=Block>> {
+            Box::new(lit(Token::OpenBrace)
+            .then_r(recursive(|| program()))
+            .then_l(lit(Token::CloseBrace)))
+        }
+
+        fn if_stmt() -> Box<Parser<I=[Token], O=Statement>> {
+
+            let comparator = || matcher(|token| match token {
+                Token::Cmp(cmp) => Some(cmp),
+                _ => None
+            });
+
+
+            let else_block = lit(Token::ElseKeyword)
+            .then_r(
+                recursive(if_stmt).map(|i| Block(vec![i]))
+                .or(boxed(recursive(|| code_block())))
+            );
+
+            let p = lit(Token::IfKeyword)
+            .then_r(recursive(|| expression()))
+            .then(comparator())
+            .then(recursive(expression))
+            .then_l(lit(Token::OpenBrace))
+            .then(recursive(program))
+            .then_l(lit(Token::CloseBrace))
+            .then(opt(else_block))
+            .map(|((((l, cmp), r), block), else_opt)| Statement::If(l, cmp, r, block, else_opt));
+
+            Box::new(p)
+        };
+
+        let while_stmt = {
+            let p = lit(Token::WhileKeyword)
+            .then_r(recursive(expression))
+            .then(comparator())
+            .then(recursive(expression))
+            .then_l(lit(Token::OpenBrace))
+            .then(recursive(program))
+            .then_l(lit(Token::CloseBrace))
+            .map(|(((l, cmp), r), block)| Statement::While(l, cmp, r, block));
+            boxed(p)
+        };
+
+        let loop_stmt = {
+            let p = lit(Token::LoopKeyword)
+            .then_r(recursive(expression))
+            .then_l(lit(Token::OpenBrace))
+            .then(recursive(program))
+            .then_l(lit(Token::CloseBrace))
+            .map(|(l, block)| Statement::Loop(l, block));
+            boxed(p)
+        };
+        let output = boxed(lit(Token::OutputCmd).then_r(recursive(|| expression())).map(|e| Statement::Output(e)));
+
+        let statements = one_of(vec![assignment, output, boxed(recursive(|| if_stmt())), while_stmt, loop_stmt]).then_l(lit(Token::NewLine)).repeat();
+
+        Box::new(statements.map(|v| Block(v)))
+    }