Merge: Faster hex parsing
authorJean Privat <jean@pryen.org>
Fri, 18 Dec 2015 20:27:41 +0000 (15:27 -0500)
committerJean Privat <jean@pryen.org>
Fri, 18 Dec 2015 20:27:41 +0000 (15:27 -0500)
commit7e303172862324a8aa68db5162fc32d605d4f69e
treec6be5d1568549f9ac133d514e3c0cb0844341876
parent33f944d802136c0aeb750cb4e7f5de9a65c76f09
parentf375b5f7a7d29c166be41ed7edbdcfa0f2b7a209
Merge: Faster hex parsing

As said in #1895, we need faster parsing of UTF-16 escaping sequences, this PR is the answer.

It makes the runtime of the `large_escaped` benchmark go down from ~5s to ~3.5s, and with valgrind, from 26GIr to 20GIr

Note: based on #1886, only the 4 last commits are of interest here

Pull-Request: #1896
Reviewed-by: Jean Privat <jean@pryen.org>
lib/core/text/abstract_text.nit