Jakub Kotur | 35e48d3 | 2020-12-22 09:58:48 +0100 | [diff] [blame] | 1 | * Remove the `empty` constructors for DFAs and replace them with |
| 2 | `never_match` and `always_match` constructors. |
| 3 | * Consider refactoring the NFA representation such that it can be instantly |
| 4 | loaded from a `&[u8]`, just like a sparse DFA. Main downside is that this |
| 5 | could negatively impact using the NFA with deserialization costs. Before |
| 6 | doing this, we should write PikeVM and backtracking implementations so that |
| 7 | they can be benchmarked. |
| 8 | * Add captures and anchors to NFA. |
| 9 | * Once we're happy, re-organize the public API such that NFAs are exported |
| 10 | and usable on their own. |