Ticket #4842: chimerax.cpp

#include <assert.h>
#include <iostream>
#include <peglib.h>
#include <ctime>

using namespace peg;
using namespace std;

int main(void) {
  // (2) Make a parser
  parser parser;
  parser.set_logger([](size_t line, size_t col, const string& msg, const string &rule) {
          cerr << line << ":" << col << ": " << msg << "\n";
        });
  auto grammar = (R"---(
                command_list <- command  (";" command)*
                command <- "color" atom_specifier < "#" [a-f0-9]{6} >
                atom_specifier <- as_term "&" atom_specifier / as_term "|" atom_specifier / as_term
                as_term <- "(" atom_specifier ")" zone_selector? / "~" as_term zone_selector? / SELECTOR_NAME / model_list
                model_list <- model+
                model <- ("#!" / "#") model_hierarchy ("##" attribute_list)? model_parts* zone_selector? / attribute_list model_parts* zone_selector* / model_parts zone_selector*
                # should be negative lookbehind for white space before the '.' in the below...
                model_hierarchy <- < model_range_list (!Space "."  !Space model_hierarchy)* >
                model_range_list <- model_range ("," model_range_list)*
                model_range <- MODEL_SPEC_START "-" MODEL_SPEC_END / MODEL_SPEC_ANY
                model_parts <- chain+
                chain <- "/" part_list ("//" attribute_list)? residue* / "//" attribute_list residue* / residue+
                residue <- ":" part_list ("::" attribute_list)? atom* / "::" attribute_list atom* / atom+
                part_list <- PART_RANGE_LIST "," part_list / PART_RANGE_LIST
                # atom ranges are not allowed
                atom <- "@" atom_list ("@@" attribute_list)? / "@@" attribute_list
                atom_list <- ATOM_NAME "," atom_list / ATOM_NAME
                attribute_list <- attr_test ("," attr_test)*
                attr_test <- ATTR_NAME ATTR_OPERATOR ATTR_VALUE / ATTR_NAME / "~" ATTR_NAME
                zone_selector <- ZONE_OPERATOR < [0-9]* "." [0-9]+ >
                # think about ranges in these character sets
                ATOM_NAME <- < [^#/:@; \t\n]+ >
                ATTR_NAME <- < [a-zA-Z_] [a-zA-Z0-9]* >
                ATTR_OPERATOR <- ">=" | ">" | "<=" | "<" | "==" | "=" | "!==" | "!=" | "<>"
                EndOfLine <- "\r\n" / "\n" / "\r"
                Space <- ' ' / '\t' / EndOfLine
                # Outer token delimiters to prevent automatic whitespace elimination inside quotes
                ATTR_VALUE <- < '"' < [^"]+ > '"' > / < "'" < [^']+ > "'" > / < [^#/:@,;"']+ >
                # limit model numbers to 5 digits to avoid conflicts with hex colors
                MODEL_SPEC <- < < [0-9]{1,5} > ![0-9A-Fa-f] >
                MODEL_SPEC_ANY <- MODEL_SPEC / "*"
                MODEL_SPEC_END <- MODEL_SPEC / "end" / "*"
                MODEL_SPEC_START <- MODEL_SPEC / "start" / "*"
                RANGE_CHAR <- [^#/:@,;- \t\n]
                PART_RANGE_LIST <- < "-"? RANGE_CHAR+ ("-" RANGE_CHAR+)? >
                SELECTOR_NAME <- < [a-zA-Z_][-+a-zA-Z0-9_]* >
                ZONE_OPERATOR <- "@>" | "@<" | ":>" | ":<" | "/>" | "/<" | "#>" | "#<"
                %whitespace  <-  [ \t\r\n]*
    )---");
auto t0 = clock();
        auto ok = parser.load_grammar(grammar);
auto t1 = clock();
std::cerr << "load grammar: " << (t1-t0) / (float)CLOCKS_PER_SEC << " seconds\n";
        assert(ok);

  assert(static_cast<bool>(parser) == true);

  // (4) Parse
  auto do_packrat = true;
  if (do_packrat)
    parser.enable_packrat_parsing(); // Enable packrat parsing.

  string test_cmd = "";
  test_cmd += "color @@serial_number=1";
  for (int i = 2; i <= 8956; ++i) {
        test_cmd += "|@@serial_number=";
        test_cmd += to_string(i);
  }
  test_cmd += " #ff8654";
  std::cerr << "command string is " << test_cmd.size() << " characters\n";
        

t0 = clock();
  parser.parse(test_cmd);
t1 = clock();
std::cerr << "parse (" << (do_packrat ? "" : "no ") << "packrat): " << (t1-t0) / (float)CLOCKS_PER_SEC << " seconds\n";
}