gpedic / ex_shortuuid / 4b63d038a23d68f7157c2dcd251b86707cd6f83b

100%

push

github

Decode ASCII alphabets with a byte lookup table (#21)

When every alphabet symbol is a single ASCII character (all predefined
alphabets), build_decode_index/1 returns a 256-entry byte table instead of
the codepoint trie. Decoding checks the byte length, then looks up each
byte; every byte of a non-ASCII character is >= 128 and never in the
table, so no separate UTF-8 or ASCII check is needed. Other alphabets keep
the codepoint decoder.

About 25-30% faster decode for ASCII alphabets, ~2.6x faster rejection of
wrong-length input. Tested against a reference decoder with random
symbols, bytes and Unicode.

Benchmarks now run compiled code; top-level script closures were
interpreted and measured mostly interpreter overhead.

17 of 17 new or added lines in 1 file covered. (100.0%)

145 of 145 relevant lines covered (100.0%)

34121.8 hits per line