From 467daf81489533b301a41da6ffea36f5f612db9f Mon Sep 17 00:00:00 2001
From: Eric Scouten <git@scouten.me>
Date: Sun, 10 Dec 2023 17:05:15 -0800
Subject: [PATCH] Add primitive parser for normalized line

Normalized is defined as having any trailing spaces removed.
---
 src/primitives/line.rs       | 25 +++++++++++++++++
 src/primitives/mod.rs        |  2 +-
 src/tests/primitives/line.rs | 53 ++++++++++++++++++++++++++++++++++++
 3 files changed, 79 insertions(+), 1 deletion(-)

diff --git a/src/primitives/line.rs b/src/primitives/line.rs
index 9fe0975..db2bcec 100644
--- a/src/primitives/line.rs
+++ b/src/primitives/line.rs
@@ -1,5 +1,10 @@
 use nom::{bytes::complete::take_till, IResult};
 
+/// Return a single line from the source.
+///
+/// A line is terminated by end-of-input or a single `\n` character
+/// or a single `\r\n` sequence. The end of line sequence is consumed
+/// but not included in the returned line.
 #[allow(dead_code)] // TEMPORARY
 pub(crate) fn line(input: &str) -> IResult<&str, &str> {
     take_till(|c| c == '\n')(input)
@@ -7,6 +12,21 @@ pub(crate) fn line(input: &str) -> IResult<&str, &str> {
         .map(|ri| trim_rem_end_matches(ri, '\r'))
 }
 
+/// Return a single _normalized_ line from the source.
+///
+/// A line is terminated by end-of-input or a single `\n` character
+/// or a single `\r\n` sequence. The end of line sequence is consumed
+/// but not included in the returned line.
+///
+/// All trailing spaces are removed from the line.
+#[allow(dead_code)] // TEMPORARY
+pub(crate) fn normalized_line(input: &str) -> IResult<&str, &str> {
+    take_till(|c| c == '\n')(input)
+        .map(|ri| trim_rem_start_matches(ri, '\n'))
+        .map(|ri| trim_rem_end_matches(ri, '\r'))
+        .map(trim_trailing_spaces)
+}
+
 #[allow(dead_code)] // TEMPORARY
 fn trim_rem_start_matches<'a>(rem_inp: (&'a str, &'a str), c: char) -> (&'a str, &'a str) {
     if let Some(rem) = rem_inp.0.strip_prefix(c) {
@@ -24,3 +44,8 @@ fn trim_rem_end_matches<'a>(rem_inp: (&'a str, &'a str), c: char) -> (&'a str, &
         rem_inp
     }
 }
+
+#[allow(dead_code)] // TEMPORARY
+fn trim_trailing_spaces<'a>(rem_inp: (&'a str, &'a str)) -> (&'a str, &'a str) {
+    (rem_inp.0, rem_inp.1.trim_end_matches(' '))
+}
diff --git a/src/primitives/mod.rs b/src/primitives/mod.rs
index bfaed2a..e91fe93 100644
--- a/src/primitives/mod.rs
+++ b/src/primitives/mod.rs
@@ -3,4 +3,4 @@
 
 mod line;
 #[allow(unused_imports)]
-pub(crate) use line::line;
+pub(crate) use line::{line, normalized_line};
diff --git a/src/tests/primitives/line.rs b/src/tests/primitives/line.rs
index 2d6288c..922f3b3 100644
--- a/src/tests/primitives/line.rs
+++ b/src/tests/primitives/line.rs
@@ -11,6 +11,11 @@ mod fn_line {
         assert_eq!(line("abc"), Ok(("", "abc")));
     }
 
+    #[test]
+    fn trailing_space() {
+        assert_eq!(line("abc "), Ok(("", "abc ")));
+    }
+
     #[test]
     fn consumes_lf() {
         // Should consume but not return \n.
@@ -35,3 +40,51 @@ mod fn_line {
         assert_eq!(line("abc\rdef"), Ok(("", "abc\rdef")));
     }
 }
+
+mod normalized_line {
+    use crate::primitives::normalized_line;
+
+    #[test]
+    fn empty_source() {
+        assert_eq!(normalized_line(""), Ok(("", "")));
+    }
+
+    #[test]
+    fn simple_line() {
+        assert_eq!(normalized_line("abc"), Ok(("", "abc")));
+    }
+
+    #[test]
+    fn trailing_space() {
+        assert_eq!(normalized_line("abc "), Ok(("", "abc")));
+    }
+
+    #[test]
+    fn trailing_spaces() {
+        assert_eq!(normalized_line("abc   "), Ok(("", "abc")));
+    }
+
+    #[test]
+    fn consumes_lf() {
+        // Should consume but not return \n.
+        assert_eq!(normalized_line("abc  \ndef"), Ok(("def", "abc")));
+    }
+
+    #[test]
+    fn consumes_crlf() {
+        // Should consume but not return \r\n.
+        assert_eq!(normalized_line("abc\r\ndef"), Ok(("def", "abc")));
+    }
+
+    #[test]
+    fn doesnt_consume_lfcr() {
+        // Should consume \n but not a subsequent \r.
+        assert_eq!(normalized_line("abc\n\rdef"), Ok(("\rdef", "abc")));
+    }
+
+    #[test]
+    fn doesnt_consume_standalone_cr() {
+        // Shouldn't terminate normalized_line at \r without \n.
+        assert_eq!(normalized_line("abc\rdef"), Ok(("", "abc\rdef")));
+    }
+}