CppCon 2016

Regular expressions are widely used in application development and data processing, yet it is challenging to design and implement a regular expression library that is expressive, efficient and safe.

In this talk, Tim Shen, the current maintainer of libstdc++'s <regex>, will introduce the basics of implementing regular expressions in C++, the status of existing implementations, and what is expected from the standardization process.

For the implementation, several data structures and algorithms will be introduced, with pros and cons listed; we will show how several popular implementations (Boost.Regex, Boost.Xpressive, <regex> from standard library implementations, RE2, etc) pick their algorithms. Several popular features/patterns that hurt performance will be explained, with a "safe" regex usage suggested. Finally a wishlist of features will be presented, in order to deliver a more efficient and usable regex library.