uutils · wanderinglethe · Jun 24, 2023 · Jun 26, 2023 · Jun 26, 2023 · sylvestre
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -10,3 +10,4 @@ readme = "README.md"
 [dependencies]
 regex = "1.8"
 chrono = { version="0.4", default-features=false, features=["std", "alloc", "clock"] }
+nom = "7.1"
-nom = "7.1"
+nom = "7.1"
+
-nom = "7.1"
+nom = "7.1"
+
diff --git a/fuzz/fuzz_targets/from_str.rs b/fuzz/fuzz_targets/from_str.rs
@@ -1,3 +1,6 @@
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
 #![no_main]
 
 use libfuzzer_sys::fuzz_target;

diff --git a/fuzz/fuzz_targets/parse_datetime_from_str.rs b/fuzz/fuzz_targets/parse_datetime_from_str.rs
@@ -1,3 +1,6 @@
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
 #![no_main]
 
 use libfuzzer_sys::fuzz_target;

diff --git a/src/lib.rs b/src/lib.rs
@@ -3,6 +3,7 @@
 
 // Expose parse_datetime
 pub mod parse_datetime;
+pub mod parse_items;
 
 use chrono::{Duration, Local, NaiveDate, Utc};
 use regex::{Error as RegexError, Regex};

diff --git a/src/parse_items.rs b/src/parse_items.rs
@@ -0,0 +1,46 @@
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use nom::error::Error;
+use nom::{IResult, Parser};
+
+pub mod items;
+pub(self) mod fixed_number;
+pub(self) mod nano_seconds;
+
+type PError<'i> = Error<&'i str>;
+type PResult<'i, O> = IResult<&'i str, O, PError<'i>>;
+
+fn singleton_list<'i, O>(mut inner: impl Parser<&'i str, O, PError<'i>>) -> impl Parser<&'i str, Vec<O>, PError<'i>> {
+    move |input: &'i str| {
+        let (tail, result) = inner.parse(input)?;
+        Ok((tail, vec![result]))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    macro_rules! ptest {
+    ($name:ident : $parser:ident($input:literal) => $out:expr, $tail:literal) => {
+        #[test]
+        fn $name() {
+            assert_eq!(
+                $parser.parse($input),
+                Ok((
+                    $tail,
+                    $out
+                ))
+            );
+        }
+        };
+        ($name:ident : $parser:ident($input:literal) => X) => {
+            #[test]
+            fn $name() {
+                let result = $parser.parse($input);
+                assert!(result.is_err(), "{:?}", result);
+            }
+        };
+    }
+
+    pub(super) use ptest;
+}
diff --git a/src/parse_items/fixed_number.rs b/src/parse_items/fixed_number.rs
@@ -0,0 +1,72 @@
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use nom::combinator::all_consuming;
+use nom::{bytes::complete::take, character::complete, combinator::map_parser, Parser};
+
+use crate::parse_items::PError;
+
+macro_rules! fixed_number_impl {
+        ($($t:ident),+) => {$(
+            #[allow(dead_code)]
+            pub fn $t<'i>(width: usize) -> impl Parser<&'i str, $t, PError<'i>> {
+                move |input: &'i str| {
+                    map_parser(take(width), all_consuming(complete::$t)).parse(input)
+                }
+            }
+        )+};
+    }
+
+fixed_number_impl! { u8, u16, u32, u64, u128 }
+
+#[cfg(test)]
+mod tests {
+    use crate::parse_items::{tests::ptest, PResult};
+
+    use super::*;
+
+    #[test]
+    fn zero_width() {
+        let result = u32(0).parse("1234");
+        assert!(result.is_err(), "{:?}", result);
+    }
+
+    #[test]
+    fn one_width() {
+        assert_eq!(u32(1).parse("1234"), Ok(("234", 1)));
+    }
+
+    #[test]
+    fn does_not_fit_type() {
+        let result = u8(4).parse("1234");
+        assert!(result.is_err(), "{:?}", result);
+    }
+
+    #[test]
+    fn does_not_fit_negative() {
+        let result = u8(3).parse("-123");
+        assert!(result.is_err(), "{:?}", result);
+    }
+
+    #[test]
+    fn input_too_short() {
+        let result = u32(6).parse("1234");
+        assert!(result.is_err(), "{:?}", result);
+    }
+
+    #[test]
+    fn three() {
+        assert_eq!(u32(3).parse("123abc"), Ok(("abc", 123)));
+    }
+
+    #[test]
+    fn leading_zeroes() {
+        assert_eq!(u32(3).parse("00123"), Ok(("23", 1)));
+    }
+
+    #[test]
+    fn non_digits() {
+        let result = u32(4).parse("123abc");
+        assert!(result.is_err(), "{:?}", result);
+    }
+}
diff --git a/src/parse_items/gnu-items.md b/src/parse_items/gnu-items.md
@@ -0,0 +1,27 @@
+<!--
+For the full copyright and license information, please view the LICENSE
+file that was distributed with this source code.
+-->
+
+## General date syntax
+https://www.gnu.org/software/coreutils/manual/html_node/General-date-syntax.html
+
+A date string can have different flavours (items):
+- calendar date
+- time of day
+- time zone
+- combined date and time of day
+- day of the week
+- relative
+- numbers
+- empty string (beginning of the day)
+
+Some properties:
+- the order of items should not matter
+- whitespace may be omitted when unambiguous
+- ordinal numbers may be written out in some items
+- comments between parentheses '(', ')'
+- alphabetic case is ignored
+- hyphens not followed by digit are ignored
+- leading zeros on numbers are ignored
+- leap seconds on supported systems