Embrace Etc/Unknown timezone (#5656)

#5653

components/datetime/src/neo_marker.rs

@@ -302,7 +302,7 @@
 //!
 //! // "uschi" - has metazone symbol data for generic_non_location_short
 //! let mut time_zone = "-0600".parse::<CustomTimeZone>().unwrap();
-//! time_zone.time_zone_id = mapper.as_borrowed().iana_to_bcp47("America/Chicago").unwrap();
+//! time_zone.time_zone_id = mapper.as_borrowed().iana_to_bcp47("America/Chicago");
 //! time_zone.maybe_calculate_metazone(&mzc, &datetime);
 //! assert_try_writeable_eq!(
 //!     tzf.format(&time_zone),

components/datetime/src/provider/time_zones.rs

@@ -77,6 +77,9 @@ pub struct LocationsV1<'data> {
     /// Per-zone location display name
     #[cfg_attr(feature = "serde", serde(borrow))]
     pub locations: ZeroMap<'data, TimeZoneBcp47Id, str>,
+    /// The display name for an unknown time zone. This is not combined with a pattern.
+    #[cfg_attr(feature = "serde", serde(borrow))]
+    pub unknown: Cow<'data, str>,
     /// The format string for a region's generic time.
     #[cfg_attr(
         feature = "serde",

components/datetime/src/time_zone.rs

@@ -524,20 +524,14 @@ impl FormatTimeZone for GenericLocationFormat {
             return Ok(Err(FormatTimeZoneError::MissingZoneSymbols));
         };
 
-        let location = locations
-            .locations
-            .get(&time_zone_id)
-            .or_else(|| {
-                locations
-                    .locations
-                    .get(&TimeZoneBcp47Id(tinystr::tinystr!(8, "unk")))
-            })
-            .unwrap_or("Unknown");
-
-        locations
-            .pattern_generic
-            .interpolate([location])
-            .write_to(sink)?;
+        if let Some(location) = locations.locations.get(&time_zone_id) {
+            locations
+                .pattern_generic
+                .interpolate([location])
+                .write_to(sink)?;
+        } else {
+            sink.write_str(&locations.unknown)?;
+        }
 
         Ok(Ok(()))
     }
@@ -567,34 +561,21 @@ impl FormatTimeZone for SpecificLocationFormat {
             return Ok(Err(FormatTimeZoneError::MissingZoneSymbols));
         };
 
-        let Some(location) = locations.locations.get(&time_zone_id) else {
-            return Ok(Err(FormatTimeZoneError::Fallback));
-        };
-
-        Ok(if zone_variant == ZoneVariant::daylight() {
-            locations
-                .pattern_daylight
-                .interpolate([location])
-                .write_to(sink)?;
-
-            Ok(())
-        } else if zone_variant == ZoneVariant::standard() {
-            locations
-                .pattern_standard
-                .interpolate([location])
-                .write_to(sink)?;
-
-            Ok(())
+        if let Some(location) = locations.locations.get(&time_zone_id) {
+            if zone_variant == ZoneVariant::daylight() {
+                &locations.pattern_daylight
+            } else if zone_variant == ZoneVariant::standard() {
+                &locations.pattern_standard
+            } else {
+                &locations.pattern_generic
+            }
+            .interpolate([location])
+            .write_to(sink)?;
         } else {
-            sink.with_part(writeable::Part::ERROR, |sink| {
-                locations
-                    .pattern_generic
-                    .interpolate([location])
-                    .write_to(sink)
-            })?;
-
-            Err(FormatTimeZoneError::MissingInputField("zone_offset"))
-        })
+            sink.write_str(&locations.unknown)?;
+        }
+
+        Ok(Ok(()))
     }
 }
 

components/timezone/src/ids.rs

@@ -67,19 +67,27 @@ use crate::{
 /// // The IANA zone "Australia/Melbourne" is the BCP-47 zone "aumel":
 /// assert_eq!(
 ///     mapper.iana_to_bcp47("Australia/Melbourne"),
-///     Some(TimeZoneBcp47Id(tinystr!(8, "aumel")))
+///     TimeZoneBcp47Id(tinystr!(8, "aumel"))
 /// );
 ///
 /// // Lookup is ASCII-case-insensitive:
 /// assert_eq!(
 ///     mapper.iana_to_bcp47("australia/melbourne"),
-///     Some(TimeZoneBcp47Id(tinystr!(8, "aumel")))
+///     TimeZoneBcp47Id(tinystr!(8, "aumel"))
 /// );
 ///
 /// // The IANA zone "Australia/Victoria" is an alias:
 /// assert_eq!(
 ///     mapper.iana_to_bcp47("Australia/Victoria"),
-///     Some(TimeZoneBcp47Id(tinystr!(8, "aumel")))
+///     TimeZoneBcp47Id(tinystr!(8, "aumel"))
+/// );
+///
+/// // The IANA zone "Australia/Boing_Boing" does not exist
+/// // (maybe not *yet*), so it produces the special unknown
+/// // timezone in order for this operation to be infallible:
+/// assert_eq!(
+///     mapper.iana_to_bcp47("Australia/Boing_Boing"),
+///     TimeZoneBcp47Id::unknown()
 /// );
 ///
 /// // We can recover the canonical identifier from the mapper:
@@ -175,22 +183,24 @@ impl TimeZoneIdMapperBorrowed<'_> {
     /// let mapper = TimeZoneIdMapper::new();
     /// let mapper = mapper.as_borrowed();
     ///
-    /// let result = mapper.iana_to_bcp47("Asia/CALCUTTA").unwrap();
+    /// let result = mapper.iana_to_bcp47("Asia/CALCUTTA");
     ///
     /// assert_eq!(*result, "inccu");
     ///
     /// // Unknown IANA time zone ID:
-    /// assert_eq!(mapper.iana_to_bcp47("America/San_Francisco"), None);
+    /// assert_eq!(mapper.iana_to_bcp47("America/San_Francisco"), TimeZoneBcp47Id::unknown());
     /// ```
-    pub fn iana_to_bcp47(&self, iana_id: &str) -> Option<TimeZoneBcp47Id> {
+    pub fn iana_to_bcp47(&self, iana_id: &str) -> TimeZoneBcp47Id {
         self.iana_lookup_quick(iana_id)
             .and_then(|trie_value| self.data.bcp47_ids.get(trie_value.index()))
+            .unwrap_or(TimeZoneBcp47Id::unknown())
     }
 
     /// Same as [`Self::iana_to_bcp47()`] but works with potentially ill-formed UTF-8.
-    pub fn iana_bytes_to_bcp47(&self, iana_id: &[u8]) -> Option<TimeZoneBcp47Id> {
+    pub fn iana_bytes_to_bcp47(&self, iana_id: &[u8]) -> TimeZoneBcp47Id {
         self.iana_lookup_quick(iana_id)
             .and_then(|trie_value| self.data.bcp47_ids.get(trie_value.index()))
+            .unwrap_or(TimeZoneBcp47Id::unknown())
     }
 
     /// Normalizes the syntax of an IANA time zone ID.

components/timezone/src/ixdtf.rs

@@ -3,8 +3,8 @@
 // (online at: https://github.com/unicode-org/icu4x/blob/main/LICENSE ).
 
 use crate::{
-    CustomTimeZone, CustomZonedDateTime, InvalidOffsetError, MetazoneCalculator, TimeZoneBcp47Id,
-    TimeZoneIdMapper, UtcOffset, ZoneOffsetCalculator,
+    CustomTimeZone, CustomZonedDateTime, InvalidOffsetError, MetazoneCalculator, TimeZoneIdMapper,
+    UtcOffset, ZoneOffsetCalculator,
 };
 use alloc::str::FromStr;
 use icu_calendar::{AnyCalendar, Date, DateError, DateTime, Iso, RangeError, Time};
@@ -127,10 +127,7 @@ impl CustomTimeZone {
         match record {
             TimeZoneRecord::Name(iana_identifier) => {
                 let mapper = TimeZoneIdMapper::new();
-                let time_zone_id = mapper
-                    .as_borrowed()
-                    .iana_bytes_to_bcp47(iana_identifier)
-                    .unwrap_or(TimeZoneBcp47Id::unknown());
+                let time_zone_id = mapper.as_borrowed().iana_bytes_to_bcp47(iana_identifier);
 
                 let mut tz = Self {
                     time_zone_id,

components/timezone/src/lib.rs

@@ -97,7 +97,7 @@
 //! let mut time_zone = CustomTimeZone::new_with_offset("-0600".parse().unwrap());
 //! let mapper = TimeZoneIdMapper::new();
 //! time_zone.time_zone_id =
-//!     mapper.as_borrowed().iana_to_bcp47("America/Chicago").unwrap_or(TimeZoneBcp47Id::unknown());
+//!     mapper.as_borrowed().iana_to_bcp47("America/Chicago");
 //!
 //! // Alternatively, set it directly from the BCP-47 ID
 //! assert_eq!(time_zone.time_zone_id, TimeZoneBcp47Id(tinystr!(8, "uschi")));

components/timezone/src/time_zone.rs

@@ -149,16 +149,12 @@ impl CustomTimeZone {
             };
         }
         let mapper = TimeZoneIdMapper::new();
-        if let Some(time_zone_id) = mapper.as_borrowed().iana_bytes_to_bcp47(code_units) {
-            return Self {
-                offset: None,
-                time_zone_id,
-                metazone_id: None,
-                zone_variant: None,
-            };
+        Self {
+            offset: None,
+            time_zone_id: mapper.as_borrowed().iana_bytes_to_bcp47(code_units),
+            metazone_id: None,
+            zone_variant: None,
         }
-
-        Self::unknown()
     }
 
     /// Infer the metazone ID.

ffi/capi/src/timezone.rs

@@ -167,11 +167,7 @@ pub mod ffi {
             mapper: &crate::timezone_mapper::ffi::TimeZoneIdMapper,
             id: &DiplomatStr,
         ) {
-            self.0.time_zone_id = mapper
-                .0
-                .as_borrowed()
-                .iana_bytes_to_bcp47(id)
-                .unwrap_or(TimeZoneBcp47Id::unknown());
+            self.0.time_zone_id = mapper.0.as_borrowed().iana_bytes_to_bcp47(id);
         }
 
         /// Writes the value of the `time_zone_id` field as a string.

ffi/capi/src/timezone_mapper.rs

@@ -47,11 +47,10 @@ pub mod ffi {
             &self,
             value: &DiplomatStr,
             write: &mut diplomat_runtime::DiplomatWrite,
-        ) -> Option<()> {
+        ) {
             let handle = self.0.as_borrowed();
-            let bcp47 = handle.iana_bytes_to_bcp47(value)?;
+            let bcp47 = handle.iana_bytes_to_bcp47(value);
             let _infallible = bcp47.0.write_to(write);
-            Some(())
         }
 
         #[diplomat::rust_link(icu::timezone::TimeZoneIdMapperBorrowed::normalize_iana, FnInStruct)]