Find a compatibility intersection between chrono and GNU

Author: dhilst

Cargo.toml

@@ -20,9 +20,3 @@ anyhow = "1.0.86"
 
 [features]
 debug = ["winnow/debug"]
-
-[[bin]]
-path = "bin/main.rs"
-name = "parse_datetime"
-test = false
-bench = false

bin/main.rs

@@ -2,7 +2,6 @@
 #![allow(dead_code)]
 
 use std::fmt::{Debug, Display};
-use std::io::{self, Write};
 
 use libfuzzer_sys::arbitrary::{self, Arbitrary};
 

fuzz/fuzz_targets/parse_datetime.rs

@@ -35,11 +35,7 @@ pub struct DateTime {
 }
 
 pub fn parse(input: &mut &str) -> PResult<DateTime> {
-    alt((
-        parse_basic,
-        //parse_8digits
-    ))
-    .parse_next(input)
+    alt((parse_basic, parse_8digits)).parse_next(input)
 }
 
 fn parse_basic(input: &mut &str) -> PResult<DateTime> {
@@ -52,7 +48,6 @@ fn parse_basic(input: &mut &str) -> PResult<DateTime> {
     .parse_next(input)
 }
 
-#[allow(dead_code)]
 fn parse_8digits(input: &mut &str) -> PResult<DateTime> {
     s((
         take(2usize).and_then(dec_uint),

src/items/combined.rs

@@ -30,6 +30,8 @@ use winnow::{
     ascii::{alpha1, dec_uint},
     combinator::{alt, opt, preceded},
     seq,
+    stream::AsChar,
+    token::take_while,
     trace::trace,
     PResult, Parser,
 };
@@ -97,14 +99,26 @@ fn literal2(input: &mut &str) -> PResult<Date> {
 }
 
 pub fn year(input: &mut &str) -> PResult<u32> {
+    // 2147485547 is the maximum value accepted
+    // by GNU, but chrono only behave like GNU
+    // for years in the range: [0, 9999], so we
+    // keep in the range [0, 9999]
     trace(
         "year",
-        dec_uint.try_map(|x| {
-            (0..=2147485547)
-                .contains(&x)
-                .then_some(x)
-                .ok_or(ParseDateTimeError::InvalidInput)
-        }),
+        s(
+            take_while(1..=4, AsChar::is_dec_digit).map(|number_str: &str| {
+                let year = number_str.parse::<u32>().unwrap();
+                if number_str.len() == 2 {
+                    if year <= 68 {
+                        year + 2000
+                    } else {
+                        year + 1900
+                    }
+                } else {
+                    year
+                }
+            }),
+        ),
     )
     .parse_next(input)
 }
@@ -233,9 +247,13 @@ mod tests {
         use super::year;
 
         // the minimun input length is 2
-        assert!(year(&mut "0").is_err());
+        // assert!(year(&mut "0").is_err());
+        // -> GNU accepts year 0
+        // test $(date -d '1-1-1' '+%Y') -eq '0001'
+
+        // test $(date -d '68-1-1' '+%Y') -eq '2068'
         // 2-characters are converted to 19XX/20XX
-        assert_eq!(year(&mut "00").unwrap(), 2000u32);
+        assert_eq!(year(&mut "10").unwrap(), 2010u32);
         assert_eq!(year(&mut "68").unwrap(), 2068u32);
         assert_eq!(year(&mut "69").unwrap(), 1969u32);
         assert_eq!(year(&mut "99").unwrap(), 1999u32);
@@ -245,6 +263,6 @@ mod tests {
         assert_eq!(year(&mut "1568").unwrap(), 1568u32);
         assert_eq!(year(&mut "1569").unwrap(), 1569u32);
         // consumes at most 4 characters from the input
-        assert_eq!(year(&mut "1234567").unwrap(), 1234u32);
+        //assert_eq!(year(&mut "1234567").unwrap(), 1234u32);
     }
 }

src/items/date.rs

@@ -53,8 +53,8 @@ mod timezone {
 use chrono::NaiveDate;
 use chrono::{DateTime, Datelike, FixedOffset, TimeZone, Timelike};
 
-use winnow::error::ParseError;
 use winnow::error::ParserError;
+use winnow::error::{ContextError, ErrMode, ParseError};
 use winnow::trace::trace;
 use winnow::{
     ascii::multispace0,
@@ -100,6 +100,17 @@ where
     separated(0.., multispace0, alt((comment, ignored_hyphen_or_plus))).parse_next(input)
 }
 
+/// Check for the end of a token, without consuming the input
+/// succeedes if the next character in the input is a space or
+/// if the input is empty
+pub(crate) fn eotoken(input: &mut &str) -> PResult<()> {
+    if input.is_empty() || input.chars().next().unwrap().is_space() {
+        return Ok(());
+    }
+
+    Err(ErrMode::Backtrack(ContextError::new()))
+}
+
 /// A hyphen or plus is ignored when it is not followed by a digit
 ///
 /// This includes being followed by a comment! Compare these inputs:
@@ -141,7 +152,7 @@ where
 
 // Parse an item
 pub fn parse_one(input: &mut &str) -> PResult<Item> {
-    //eprintln!("parsing_one -> {input}");
+    // eprintln!("parsing_one -> {input}");
     let result = trace(
         "parse_one",
         alt((
@@ -152,11 +163,11 @@ pub fn parse_one(input: &mut &str) -> PResult<Item> {
             weekday::parse.map(Item::Weekday),
             epoch::parse.map(Item::Timestamp),
             timezone::parse.map(Item::TimeZone),
-            s(date::year).map(Item::Year),
+            date::year.map(Item::Year),
         )),
     )
     .parse_next(input)?;
-    //eprintln!("parsing_one <- {input} {result:?}");
+    // eprintln!("parsing_one <- {input} {result:?}");
 
     Ok(result)
 }

src/items/mod.rs

@@ -42,15 +42,15 @@ use std::fmt::Display;
 use chrono::FixedOffset;
 use winnow::{
     ascii::{dec_uint, float},
-    combinator::{alt, opt, preceded},
+    combinator::{alt, opt, preceded, terminated},
     error::{AddContext, ContextError, ErrMode, StrContext},
     seq,
     stream::AsChar,
     token::take_while,
     PResult, Parser,
 };
 
-use super::s;
+use super::{eotoken, s};
 
 #[derive(PartialEq, Debug, Clone, Default)]
 pub struct Offset {
@@ -206,7 +206,11 @@ fn second(input: &mut &str) -> PResult<f64> {
 }
 
 pub(crate) fn timezone(input: &mut &str) -> PResult<Offset> {
-    alt((timezone_num, timezone_name_offset)).parse_next(input)
+    let result =
+        terminated(alt((timezone_num, timezone_name_offset)), eotoken).parse_next(input)?;
+
+    // space_or_eof(input, result)
+    Ok(result)
 }
 
 /// Parse a timezone starting with `+` or `-`

src/items/time.rs

@@ -120,15 +120,20 @@ mod tests {
     mod iso_8601 {
         use std::env;
 
+        use chrono::{TimeZone, Utc};
+
         use crate::ParseDateTimeError;
         use crate::{parse_datetime, tests::TEST_TIME};
 
         #[test]
         fn test_t_sep() {
             env::set_var("TZ", "UTC");
             let dt = "2021-02-15T06:37:47";
-            let actual = parse_datetime(dt);
-            assert_eq!(actual.unwrap().timestamp(), TEST_TIME);
+            let actual = parse_datetime(dt).unwrap();
+            assert_eq!(
+                actual,
+                Utc.timestamp_opt(TEST_TIME, 0).unwrap().fixed_offset()
+            );
         }
 
         #[test]
@@ -400,7 +405,6 @@ mod tests {
         #[test]
         fn test_invalid_input() {
             let result = parse_datetime("foobar");
-            println!("{result:?}");
             assert_eq!(result, Err(ParseDateTimeError::InvalidInput));
 
             let result = parse_datetime("invalid 1");
@@ -618,13 +622,9 @@ mod tests {
     }
 
     mod test_relative {
-        use chrono::NaiveDate;
 
-        use crate::{items, parse_datetime};
-        use std::{
-            env,
-            io::{self, Write},
-        };
+        use crate::parse_datetime;
+        use std::env;
 
         #[test]
         fn test_month() {
@@ -677,45 +677,50 @@ mod tests {
                 "2024-03-29T00:00:00+00:00",
             );
         }
+    }
+
+    mod test_gnu {
+        use crate::parse_datetime;
+
         fn make_gnu_date(input: &str, fmt: &str) -> String {
             std::process::Command::new("date")
                 .arg("-d")
                 .arg(input)
                 .arg(format!("+{fmt}"))
                 .output()
                 .map(|mut output| {
-                    io::stdout().write_all(&output.stdout).unwrap();
+                    //io::stdout().write_all(&output.stdout).unwrap();
                     output.stdout.pop(); // remove trailing \n
                     String::from_utf8(output.stdout).expect("from_utf8")
                 })
                 .unwrap()
         }
 
-        #[test]
-        fn chrono_date() {
-            const FMT: &str = "%Y-%m-%d %H:%M:%S";
-            let year = 262144;
-            let input = format!("{year}-01-01 00:00:00");
-
-            assert!(NaiveDate::from_ymd_opt(year, 1, 1).is_none());
-            assert!(chrono::DateTime::parse_from_str(&input, FMT).is_err());
-            // the parsing works, but hydration fails
-            assert!(items::parse(&mut input.to_string().as_str()).is_ok());
-            assert!(parse_datetime(&input).is_err());
-            // GNU date works
-            assert_eq!(make_gnu_date(&input, FMT), input);
+        fn has_gnu_date() -> bool {
+            std::process::Command::new("date")
+                .arg("--version")
+                .output()
+                .map(|output| String::from_utf8(output.stdout).unwrap())
+                .map(|output| output.starts_with("date (GNU coreutils)"))
+                .unwrap_or(false)
         }
 
         #[test]
         fn gnu_compat() {
+            // skip if GNU date is not present
+            if !has_gnu_date() {
+                eprintln!("GNU date not found, skipping gnu_compat tests");
+                return;
+            }
+
             const FMT: &str = "%Y-%m-%d %H:%M:%S";
             let input = "0000-03-02 00:00:00";
             assert_eq!(
                 make_gnu_date(input, FMT),
                 parse_datetime(input).unwrap().format(FMT).to_string()
             );
 
-            let input = "262144-03-10 00:00:00";
+            let input = "2621-03-10 00:00:00";
             assert_eq!(
                 make_gnu_date(input, FMT),
                 parse_datetime(input)
@@ -724,7 +729,7 @@ mod tests {
                     .to_string()
             );
 
-            let input = "10384-03-10 00:00:00";
+            let input = "1038-03-10 00:00:00";
             assert_eq!(
                 make_gnu_date(input, FMT),
                 parse_datetime(input)