c791a23456
* remove unneeded loop in `SpliceMutator::mutate` previously we searched for the first and the last difference between exactly the same 2 inputs 3 times in a loop * remove unused struct fields * avoid allocating strings for `Transition`s * avoid allocating `String`s for `Stack`s * avoid allocating Strings for `Element`s * apply some clippy lints * some more clippy lints * simplify regex * remove superflous if condition * remove the Rc<_> in `Element` * small cleanups and regex fix * avoid allocating a vector for the culled pda * bug fix * bug fix * reintroduce the Rc, but make it use the *one* alloced VecDeque this time * slim down dependencies * use Box<[&str]> for storted state stacks this saves us a whopping 8 bytes ;), since we don't have to store the capacity * revert the changes from 9ffa715c10089f157e4e20563143a2df890c8ffe fixes a bug * apply clippy lint --------- Co-authored-by: Andrea Fioraldi <andreafioraldi@gmail.com>
Gramatron preprocessing scripts
In this folder live the scripts to convert a grammar (some examples in the grammars/ subfolder) into a serialized Automaton.
You need as first to convert the grammar to the GNF form using the gnf_converter.py Python script.
Then use the output as input of the construct_automata crate.
Here an example using the Ruby grammar:
./gnf_converter.py --gf grammars/ruby_grammar.json --out ruby_gnf.json --start PROGRAM
cd construct_automata
RUSTFLAGS="-C target-cpu=native" cargo run --release -- --gf ../ruby_gnf.json --out ../ruby_automaton.postcard
You can add the --limit flag to limit the stack size, as described in the Gramatron paper.