From 020548fad91b8fd18a99c7408c983ddf4f06266a Mon Sep 17 00:00:00 2001
From: PoiScript <poiscript@gmail.com>
Date: Tue, 14 Apr 2020 17:55:19 +0800
Subject: [PATCH] feat(parse): create combinators module

---
 src/parse/combinators.rs | 159 +++++++++++++++++++++++++++++++++++++++
 src/parse/mod.rs         |   1 +
 2 files changed, 160 insertions(+)
 create mode 100644 src/parse/combinators.rs
 create mode 100644 src/parse/mod.rs

diff --git a/src/parse/combinators.rs b/src/parse/combinators.rs
new file mode 100644
index 0000000..74b48e1
--- /dev/null
+++ b/src/parse/combinators.rs
@@ -0,0 +1,159 @@
+//! Parsers combinators
+
+use memchr::memchr;
+use nom::{
+    bytes::complete::take_while1,
+    combinator::verify,
+    error::{ErrorKind, ParseError},
+    Err, IResult,
+};
+
+// read until the first line_ending, if line_ending is not present, return the input directly
+pub fn line<'a, E>(input: &'a str) -> IResult<&'a str, &'a str, E>
+where
+    E: ParseError<&'a str>,
+{
+    if let Some(i) = memchr(b'\n', input.as_bytes()) {
+        if i > 0 && input.as_bytes()[i - 1] == b'\r' {
+            Ok((&input[i + 1..], &input[0..i - 1]))
+        } else {
+            Ok((&input[i + 1..], &input[0..i]))
+        }
+    } else {
+        Ok(("", input))
+    }
+}
+
+pub fn lines_till<'a, F, E>(predicate: F) -> impl Fn(&'a str) -> IResult<&str, &str, E>
+where
+    F: Fn(&str) -> bool,
+    E: ParseError<&'a str>,
+{
+    move |i| {
+        let mut input = i;
+
+        loop {
+            // TODO: better error kind
+            if input.is_empty() {
+                return Err(Err::Error(E::from_error_kind(input, ErrorKind::Many0)));
+            }
+
+            let (input_, line_) = line(input)?;
+
+            debug_assert_ne!(input, input_);
+
+            if predicate(line_) {
+                let offset = i.len() - input.len();
+                return Ok((input_, &i[0..offset]));
+            }
+
+            input = input_;
+        }
+    }
+}
+
+pub fn lines_while<'a, F, E>(predicate: F) -> impl Fn(&'a str) -> IResult<&str, &str, E>
+where
+    F: Fn(&str) -> bool,
+    E: ParseError<&'a str>,
+{
+    move |i| {
+        let mut input = i;
+
+        loop {
+            // unlike lines_till, line_while won't return error
+            if input.is_empty() {
+                return Ok(("", i));
+            }
+
+            let (input_, line_) = line(input)?;
+
+            debug_assert_ne!(input, input_);
+
+            if !predicate(line_) {
+                let offset = i.len() - input.len();
+                return Ok((input, &i[0..offset]));
+            }
+
+            input = input_;
+        }
+    }
+}
+
+#[test]
+fn test_lines_while() {
+    assert_eq!(
+        lines_while::<_, ()>(|line| line == "foo")("foo"),
+        Ok(("", "foo"))
+    );
+    assert_eq!(
+        lines_while::<_, ()>(|line| line == "foo")("bar"),
+        Ok(("bar", ""))
+    );
+    assert_eq!(
+        lines_while::<_, ()>(|line| line == "foo")("foo\n\n"),
+        Ok(("\n", "foo\n"))
+    );
+    assert_eq!(
+        lines_while::<_, ()>(|line| line.trim().is_empty())("\n\n\n"),
+        Ok(("", "\n\n\n"))
+    );
+}
+
+pub fn eol<'a, E>(input: &'a str) -> IResult<&str, &str, E>
+where
+    E: ParseError<&'a str>,
+{
+    verify(line, |s: &str| {
+        s.as_bytes().iter().all(u8::is_ascii_whitespace)
+    })(input)
+}
+
+pub fn one_word<'a, E>(input: &'a str) -> IResult<&str, &str, E>
+where
+    E: ParseError<&'a str>,
+{
+    take_while1(|c: char| !c.is_ascii_whitespace())(input)
+}
+
+pub fn blank_lines_count<'a, E>(input: &'a str) -> IResult<&str, usize, E>
+where
+    E: ParseError<&'a str>,
+{
+    let mut count = 0;
+    let mut input = input;
+
+    loop {
+        if input.is_empty() {
+            return Ok(("", count));
+        }
+
+        let (input_, line_) = line(input)?;
+
+        debug_assert_ne!(input, input_);
+
+        if !line_.as_bytes().iter().all(u8::is_ascii_whitespace) {
+            return Ok((input, count));
+        }
+
+        count += 1;
+
+        input = input_;
+    }
+}
+
+#[test]
+fn test_blank_lines_count() {
+    assert_eq!(blank_lines_count::<()>("foo"), Ok(("foo", 0)));
+    assert_eq!(blank_lines_count::<()>(" foo"), Ok((" foo", 0)));
+    assert_eq!(blank_lines_count::<()>("  \t\nfoo\n"), Ok(("foo\n", 1)));
+    assert_eq!(
+        blank_lines_count::<()>("\n    \r\n\nfoo\n"),
+        Ok(("foo\n", 3))
+    );
+    assert_eq!(
+        blank_lines_count::<()>("\r\n   \n  \r\n   foo\n"),
+        Ok(("   foo\n", 3))
+    );
+    assert_eq!(blank_lines_count::<()>("\r\n   \n  \r\n   \n"), Ok(("", 4)));
+}
diff --git a/src/parse/mod.rs b/src/parse/mod.rs
new file mode 100644
index 0000000..0c49327
--- /dev/null
+++ b/src/parse/mod.rs
@@ -0,0 +1 @@
+pub mod combinators;