Optimize enum parsing. In a small enum-heavy benchmark, this improved the time
of the benchmark (parsing a repeated message, where each message had 5 "small" enum values and 5 "large" enum values) from ~39s to ~11s. There is a small memory cost per enum used, but I expect this to be trivial compared with other per-type costs. Fixes issue 97.
Showing
Please
register
or
sign in
to comment