remove regex plugin + rollup + chores #436

These are dev dependencies, so we don't need to worry about the minimum Rust version supported.

The latest update to rand requires a newer version of Rust. Since it's a dev dependency, we shouldn't need to do a semver bump when updating rand. However, CI needs to be told not to run tests. Instead, we merely check that we can build the crate and produce documentation.

The regex_macros crate hasn't been maintained in quite some time, and has been broken. Nobody has complained. Given the fact that there are no immediate plans to improve the situation, and the fact that it is slower than the runtime engine, we simply remove it.

The 0.2.1 release of simd includes a fix so that it can compile on the latest nightly. We needn't worry about semver here because simd is a nightly-only dependency.

There are a few sub-crates in this repository, so sharing a target directory makes sense.

This updates dependencies and makes sure everything compiles and runs. This also simplifies the build script.

Principally, this updates docopt to 0.8, which replaces rustc-serialize with serde.

This commit tweaks the heuristic employed to determine whether to use TBM or not. For the most part, the heuristic was tweaked by combining the actual benchmark results with a bit of hand waving. In particular, the primary change here is that the frequency rank cutoff is no longer a constant, but rather, a function of the pattern length. That is, we guess that TBM will do well with longer patterns, even if it contains somewhat infrequent bytes. We do put a constant cap on this heuristic. That is, regardless of the length of the pattern, if a "very rare" byte is found in the pattern, then we won't use TBM.

As far as I can tell, nobody has actually described a substring search algorithm that used both frequency analysis and vector instructions. So I'm naming it.

See also: #399

4fab6c added the current bench runner script as `benches/run`, and removed the old `run-bench` script. It was later renamed to `bench/run` when `benches` was renamed to `bench` in b217bf. This patch fixes a few references to the old benchmark runner in the hacking guide as well as a few references to the old directory structure. The cargo plugin syntax in the example is also updated.

The DFA can't produce captures, but is still faster than the Pike VM NFA, so the normal approach to finding capture groups is to look for the entire match with the DFA and then run the NFA on the substring of the input that matched. In cases where the regex in anchored, the match always starts at the beginning of the input, so there is never any point to trying the DFA first. The DFA can still be useful for rejecting inputs which are not in the language of the regular expression, but anchored regex with capture groups are most commonly used in a parsing context, so it seems like a fair trade-off. Fixes #348

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

remove regex plugin + rollup + chores #436

remove regex plugin + rollup + chores #436

Commits on Dec 30, 2017