[pbs-devel] [RFC proxmox-backup 09/39] s3 client: add helper to parse http date headers

[Christian Ebner]

Add a helper to parse the preferred date/time format for http `Date`
headers as specified in RFC 2616 [0], which is a fixed-length subset
of the format specified in RFC 1123 [1], itself being a followup to
RFC 822 [2]. Does not implement the format as described in the
obsolete RFC 850 [3].

This allows to parse the `Date` and `Last-Modified` headers of S3 API
responses.

[0] https://datatracker.ietf.org/doc/html/rfc2616#section-3.3
[1] https://datatracker.ietf.org/doc/html/rfc1123#section-5.2.14
[2] https://datatracker.ietf.org/doc/html/rfc822#section-5
[3] https://datatracker.ietf.org/doc/html/rfc850

Signed-off-by: Christian Ebner <c.ebner at proxmox.com>
---
 pbs-s3-client/src/lib.rs | 97 +++++++++++++++++++++++++++++++++++++++-
 1 file changed, 96 insertions(+), 1 deletion(-)

diff --git a/pbs-s3-client/src/lib.rs b/pbs-s3-client/src/lib.rs
index 308db64d8..00fa26455 100644
--- a/pbs-s3-client/src/lib.rs
+++ b/pbs-s3-client/src/lib.rs
@@ -6,7 +6,12 @@ pub use object_key::{S3ObjectKey, S3_CONTENT_PREFIX};
 
 use std::time::Duration;
 
-use anyhow::{bail, Error};
+use anyhow::{bail, Context, Error};
+
+const VALID_DAYS_OF_WEEK: [&str; 7] = ["Mon", "Tue", "Wed", "Thu", "Fri", "Sat", "Sun"];
+const VALID_MONTHS: [&str; 12] = [
+    "Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec",
+];
 
 #[derive(Debug)]
 pub struct LastModifiedTimestamp {
@@ -121,3 +126,93 @@ impl std::str::FromStr for LastModifiedTimestamp {
 }
 
 serde_plain::derive_deserialize_from_fromstr!(LastModifiedTimestamp, "last modified timestamp");
+
+/// Preferred date format specified by RFC2616, given as fixed-length
+/// subset of RFC1123, which itself is a followup to RFC822.
+///
+/// https://datatracker.ietf.org/doc/html/rfc2616#section-3.3
+/// https://datatracker.ietf.org/doc/html/rfc1123#section-5.2.14
+/// https://datatracker.ietf.org/doc/html/rfc822#section-5
+#[derive(Debug)]
+pub struct HttpDate {
+    epoch: i64,
+}
+
+impl HttpDate {
+    pub fn to_duration(&self) -> Result<Duration, Error> {
+        let seconds = u64::try_from(self.epoch)?;
+        Ok(Duration::from_secs(seconds))
+    }
+}
+
+impl std::str::FromStr for HttpDate {
+    type Err = Error;
+
+    fn from_str(timestamp: &str) -> Result<Self, Self::Err> {
+        let input = timestamp.as_bytes();
+        if input.len() != 29 {
+            bail!("unexpected length: got {}", input.len());
+        }
+
+        let expect = |pos: usize, c: u8| {
+            if input[pos] != c {
+                bail!("unexpected char at pos {pos}");
+            }
+            Ok(())
+        };
+
+        let digit = |pos: usize| -> Result<i32, Error> {
+            let digit = input[pos] as i32;
+            if !(48..=57).contains(&digit) {
+                bail!("unexpected char at pos {pos}");
+            }
+            Ok(digit - 48)
+        };
+
+        fn check_max(i: i32, max: i32) -> Result<i32, Error> {
+            if i > max {
+                bail!("value too large ({i} > {max})");
+            }
+            Ok(i)
+        }
+
+        let mut tm = proxmox_time::TmEditor::new(true);
+
+        if !VALID_DAYS_OF_WEEK
+            .iter()
+            .any(|valid| valid.as_bytes() == &input[0..3])
+        {
+            bail!("unexpected day of week, got {:?}", &input[0..3]);
+        }
+
+        expect(3, b',').context("unexpected separator after day of week")?;
+        expect(4, b' ').context("missing space after day of week separator")?;
+        tm.set_mday(check_max(digit(5)? * 10 + digit(6)?, 31)?)?;
+        expect(7, b' ').context("unexpected separator after day")?;
+        if let Some(month) = VALID_MONTHS
+            .iter()
+            .position(|month| month.as_bytes() == &input[8..11])
+        {
+            // valid conversion to i32, position stems from fixed size array of 12 months.
+            tm.set_mon(check_max(month as i32 + 1, 12)?)?;
+        } else {
+            bail!("invalid month");
+        }
+        expect(11, b' ').context("unexpected separator after month")?;
+        tm.set_year(digit(12)? * 1000 + digit(13)? * 100 + digit(14)? * 10 + digit(15)?)?;
+        expect(16, b' ').context("unexpected separator after year")?;
+        tm.set_hour(check_max(digit(17)? * 10 + digit(18)?, 23)?)?;
+        expect(19, b':').context("unexpected separator after hour")?;
+        tm.set_min(check_max(digit(20)? * 10 + digit(21)?, 59)?)?;
+        expect(22, b':').context("unexpected separator after minute")?;
+        tm.set_sec(check_max(digit(23)? * 10 + digit(24)?, 60)?)?;
+        expect(25, b' ').context("unexpected separator after second")?;
+        if !input.ends_with(b"GMT") {
+            bail!("unexpected timezone");
+        }
+
+        let epoch = tm.into_epoch()?;
+
+        Ok(Self { epoch })
+    }
+}
-- 
2.39.5