Something went wrong. Try again.
A lexicon-driven AppView for ATProto.
Something went wrong. Try again.
1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768697071727374757677use std::time::{SystemTime, UNIX_EPOCH};
/// Base32-sortstring alphabet used by AT Protocol TIDs.const BASE32_SORT: &[u8; 32] = b"234567abcdefghijklmnopqrstuvwxyz";
/// Generate a TID (timestamp identifier) compatible with the AT Protocol spec.////// Layout: 64-bit value = `(microsecond_timestamp << 10) | random_10bit_clock_id`/// Encoded as a 13-character base32-sortstring.pub fn generate_tid() -> String { let us = SystemTime::now() .duration_since(UNIX_EPOCH) .expect("system clock before UNIX epoch") .as_micros() as u64;
// 10-bit random clock ID from UUID v4 bytes let rand_bytes = uuid::Uuid::new_v4(); let clock_id = u16::from_le_bytes([rand_bytes.as_bytes()[0], rand_bytes.as_bytes()[1]]) & 0x3FF;
let val = (us << 10) | clock_id as u64; encode_base32_sort(val)}
/// Encode a u64 into a 13-character base32-sortstring.fn encode_base32_sort(mut val: u64) -> String { let mut buf = [0u8; 13]; for i in (0..13).rev() { buf[i] = BASE32_SORT[(val & 0x1F) as usize]; val >>= 5; } // SAFETY: all bytes come from BASE32_SORT which is ASCII String::from_utf8(buf.to_vec()).unwrap()}
#[cfg(test)]mod tests { use super::*;
#[test] fn tid_is_13_chars() { let tid = generate_tid(); assert_eq!(tid.len(), 13, "TID should be 13 characters, got: {tid}"); }
#[test] fn tid_uses_valid_charset() { let tid = generate_tid(); let valid = "234567abcdefghijklmnopqrstuvwxyz"; for ch in tid.chars() { assert!(valid.contains(ch), "invalid character '{ch}' in TID {tid}"); } }
#[test] fn tids_are_unique() { let a = generate_tid(); let b = generate_tid(); assert_ne!(a, b, "two TIDs should differ"); }
#[test] fn tids_are_sortable() { // TIDs generated later should sort after earlier ones let a = generate_tid(); std::thread::sleep(std::time::Duration::from_millis(2)); let b = generate_tid(); assert!(b > a, "later TID '{b}' should sort after earlier TID '{a}'"); }
#[test] fn encode_base32_sort_known_value() { // Zero should encode to all '2's (the first character in the alphabet) let result = encode_base32_sort(0); assert_eq!(result, "2222222222222"); }}