Move user ID localpart classification into a public function

CHANGELOG.md

@@ -27,6 +27,9 @@ Improvements:
 * Add support for historical uppercase MXIDs
 * Made all dependencies optional
   * `serde` is the only one that is enabled by default
+* The `user_id` module is now public and contains `fn localpart_is_fully_conforming`
+  * This function can be used to determine whether a user name (the localpart of a user ID) is valid
+    without actually constructing a full user ID first
 
 # 0.14.1
 

src/lib.rs

@@ -20,10 +20,9 @@ use std::num::NonZeroU8;
 use serde::de::{self, Deserialize as _, Deserializer, Unexpected};
 
 #[doc(inline)]
-pub use crate::device_id::DeviceId;
 pub use crate::{
-    error::Error, event_id::EventId, room_alias_id::RoomAliasId, room_id::RoomId,
-    room_id_or_room_alias_id::RoomIdOrAliasId, room_version_id::RoomVersionId,
+    device_id::DeviceId, error::Error, event_id::EventId, room_alias_id::RoomAliasId,
+    room_id::RoomId, room_id_or_room_alias_id::RoomIdOrAliasId, room_version_id::RoomVersionId,
     server_name::is_valid_server_name, user_id::UserId,
 };
 
@@ -40,7 +39,7 @@ mod room_id;
 mod room_id_or_room_alias_id;
 mod room_version_id;
 mod server_name;
-mod user_id;
+pub mod user_id;
 
 /// All identifiers must be 255 bytes or less.
 const MAX_BYTES: usize = 255;

src/user_id.rs

@@ -84,29 +84,43 @@ impl TryFrom<Cow<'_, str>> for UserId {
         let colon_idx = parse_id(&user_id, &['@'])?;
         let localpart = &user_id[1..colon_idx.get() as usize];
 
-        // See https://matrix.org/docs/spec/appendices#user-identifiers
-        let is_fully_conforming = localpart.bytes().all(|b| match b {
-            b'0'..=b'9' | b'a'..=b'z' | b'-' | b'.' | b'=' | b'_' | b'/' => true,
-            _ => false,
-        });
-
-        // If it's not fully conforming, check if it contains characters that are also disallowed
-        // for historical user IDs. If there are, return an error.
-        // See https://matrix.org/docs/spec/appendices#historical-user-ids
-        if !is_fully_conforming && localpart.bytes().any(|b| b < 0x21 || b == b':' || b > 0x7E) {
-            return Err(Error::InvalidCharacters);
-        }
+        let is_historical = localpart_is_fully_comforming(localpart)?;
 
         Ok(Self {
             full_id: user_id.into_owned(),
             colon_idx,
-            is_historical: !is_fully_conforming,
+            is_historical: !is_historical,
         })
     }
 }
 
 common_impls!(UserId, "a Matrix user ID");
 
+/// Check whether the given user id localpart is valid and fully conforming
+///
+/// Returns an `Err` for invalid user ID localparts, `Ok(false)` for historical user ID localparts
+/// and `Ok(true)` for fully conforming user ID localparts.
+pub fn localpart_is_fully_comforming(localpart: &str) -> Result<bool, Error> {
+    if localpart.is_empty() {
+        return Err(Error::InvalidLocalPart);
+    }
+
+    // See https://matrix.org/docs/spec/appendices#user-identifiers
+    let is_fully_conforming = localpart.bytes().all(|b| match b {
+        b'0'..=b'9' | b'a'..=b'z' | b'-' | b'.' | b'=' | b'_' | b'/' => true,
+        _ => false,
+    });
+
+    // If it's not fully conforming, check if it contains characters that are also disallowed
+    // for historical user IDs. If there are, return an error.
+    // See https://matrix.org/docs/spec/appendices#historical-user-ids
+    if !is_fully_conforming && localpart.bytes().any(|b| b < 0x21 || b == b':' || b > 0x7E) {
+        Err(Error::InvalidCharacters)
+    } else {
+        Ok(is_fully_conforming)
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use std::convert::TryFrom;