const HParser *init_parser(void) { // CORE H_RULE (digit, h_ch_range(0x30, 0x39)); H_RULE (alpha, h_choice(h_ch_range(0x41, 0x5a), h_ch_range(0x61, 0x7a), NULL)); H_RULE (space, h_in((uint8_t *)" \t\n\r\f\v", 6)); // AUX. H_RULE (plus, h_ch('+')); H_RULE (slash, h_ch('/')); H_ARULE(equals, h_ch('=')); H_ARULE(bsfdig, h_choice(alpha, digit, plus, slash, NULL)); H_ARULE(bsfdig_4bit, h_in((uint8_t *)"AEIMQUYcgkosw048", 16)); H_ARULE(bsfdig_2bit, h_in((uint8_t *)"AQgw", 4)); H_ARULE(base64_3, h_repeat_n(bsfdig, 4)); H_ARULE(base64_2, h_sequence(bsfdig, bsfdig, bsfdig_4bit, equals, NULL)); H_ARULE(base64_1, h_sequence(bsfdig, bsfdig_2bit, equals, equals, NULL)); H_ARULE(base64, h_sequence(h_many(base64_3), h_optional(h_choice(base64_2, base64_1, NULL)), NULL)); H_ARULE(ws, h_many(space)); H_ARULE(document, h_sequence(ws, base64, ws, h_end_p(), NULL)); // BUG sometimes inputs that should just don't parse. // It *seemed* to happen mostly with things like "bbbbaaaaBA==". // Using less actions seemed to make it less likely. return document; }
// dummy! int test_lalr(void) { HAllocator *mm__ = &system_allocator; /* E -> E '-' T | T T -> '(' E ')' | 'n' -- also try [0-9] for the charset paths */ HParser *n = h_ch('n'); HParser *E = h_indirect(); HParser *T = h_choice(h_sequence(h_ch('('), E, h_ch(')'), NULL), n, NULL); HParser *E_ = h_choice(h_sequence(E, h_ch('-'), T, NULL), T, NULL); h_bind_indirect(E, E_); HParser *p = E; printf("\n==== G R A M M A R ====\n"); HCFGrammar *g = h_cfgrammar_(mm__, h_desugar_augmented(mm__, p)); if (g == NULL) { fprintf(stderr, "h_cfgrammar failed\n"); return 1; } h_pprint_grammar(stdout, g, 0); printf("\n==== D F A ====\n"); HLRDFA *dfa = h_lr0_dfa(g); if (dfa) { h_pprint_lrdfa(stdout, g, dfa, 0); } else { fprintf(stderr, "h_lalr_dfa failed\n"); } printf("\n==== L R ( 0 ) T A B L E ====\n"); HLRTable *table0 = h_lr0_table(g, dfa); if (table0) { h_pprint_lrtable(stdout, g, table0, 0); } else { fprintf(stderr, "h_lr0_table failed\n"); } h_lrtable_free(table0); printf("\n==== L A L R T A B L E ====\n"); if (h_compile(p, PB_LALR, NULL)) { fprintf(stderr, "does not compile\n"); return 2; } h_pprint_lrtable(stdout, g, (HLRTable *)p->backend_data, 0); printf("\n==== P A R S E R E S U L T ====\n"); HParseResult *res = h_parse(p, (uint8_t *)"n-(n-((n)))-n", 13); if (res) { h_pprint(stdout, res->ast, 0, 2); } else { printf("no parse\n"); } return 0; }
static void test_cfg_many_seq(void) { HParser *p = h_many(h_sequence(h_ch('A'), h_ch('B'), NULL)); g_check_parse_match(p, PB_LLk, "ABAB",4, "((u0x41 u0x42) (u0x41 u0x42))"); g_check_parse_match(p, PB_LALR, "ABAB",4, "((u0x41 u0x42) (u0x41 u0x42))"); g_check_parse_match(p, PB_GLR, "ABAB",4, "((u0x41 u0x42) (u0x41 u0x42))"); // these would instead parse as (u0x41 u0x42 u0x41 u0x42) due to a faulty // reshape on h_many. }
static void test_wrong_bit_length(void) { HParseResult *r; HParser *p; p = h_right(h_ch('a'), h_ch('b')); r = h_parse(p, (const uint8_t *)"ab", 2); g_check_cmp_int64(r->bit_length, ==, 16); h_parse_result_free(r); p = h_bind(h_ch('a'), k_test_wrong_bit_length, NULL); r = h_parse(p, (const uint8_t *)"ab", 2); g_check_cmp_int64(r->bit_length, ==, 16); h_parse_result_free(r); }
const HParser* init_domain() { static const HParser *domain = NULL; if (domain) return domain; const HParser *letter = h_choice(h_ch_range('a', 'z'), h_ch_range('A', 'Z'), NULL); const HParser *let_dig = h_choice(letter, h_ch_range('0', '9'), NULL); const HParser *ldh_str = h_many1(h_choice(let_dig, h_ch('-'), NULL)); const HParser *label = h_attr_bool(h_sequence(letter, h_optional(h_sequence(h_optional(ldh_str), let_dig, NULL)), NULL), validate_label); /** * You could write it like this ... * HParser *indirect_subdomain = h_indirect(); * const HParser *subdomain = h_choice(label, * h_sequence(indirect_subdomain, * h_ch('.'), * label, * NULL), * NULL); * h_bind_indirect(indirect_subdomain, subdomain); * * ... but this is easier and equivalent */ const HParser *subdomain = h_sepBy1(label, h_ch('.')); domain = h_choice(subdomain, h_ch(' '), NULL); return domain; }
HParser* init_domain() { static HParser *ret = NULL; if (ret) return ret; H_RULE (letter, h_choice(h_ch_range('a','z'), h_ch_range('A','Z'), NULL)); H_RULE (let_dig, h_choice(letter, h_ch_range('0','9'), NULL)); H_RULE (ldh_str, h_many1(h_choice(let_dig, h_ch('-'), NULL))); H_VARULE(label, h_sequence(letter, h_optional(h_sequence(h_optional(ldh_str), let_dig, NULL)), NULL)); H_RULE (subdomain, h_sepBy1(label, h_ch('.'))); H_ARULE (domain, h_choice(subdomain, h_ch(' '), NULL)); ret = domain; return ret; }
static void test_llk_zero_end(void) { HParserBackend be = PB_LLk; HParser *z = h_ch('\x00'); HParser *az = h_sequence(h_ch('a'), z, NULL); HParser *ze = h_sequence(z, h_end_p(), NULL); HParser *aze = h_sequence(h_ch('a'), z, h_end_p(), NULL); // some cases surrounding the bug g_check_parse_match (z, be, "\x00", 1, "u0"); g_check_parse_failed(z, be, "", 0); g_check_parse_match (ze, be, "\x00", 1, "(u0)"); g_check_parse_failed(ze, be, "\x00b", 2); g_check_parse_failed(ze, be, "", 0); g_check_parse_match (az, be, "a\x00", 2, "(u0x61 u0)"); g_check_parse_match (aze, be, "a\x00", 2, "(u0x61 u0)"); g_check_parse_failed(aze, be, "a\x00b", 3); // the following should not parse but did when the LL(k) backend failed to // check for the end of input, mistaking it for a zero character. g_check_parse_failed(az, be, "a", 1); g_check_parse_failed(aze, be, "a", 1); }
static void test_lalr_charset_lhs(void) { HParserBackend be = PB_LALR; HParser *p = h_many(h_choice(h_sequence(h_ch('A'), h_ch('B'), NULL), h_in((uint8_t*)"AB",2), NULL)); // the above would abort because of an unhandled case in trying to resolve // a conflict where an item's left-hand-side was an HCF_CHARSET. // however, the compile should fail - the conflict cannot be resolved. if(h_compile(p, be, NULL) == 0) { g_test_message("LALR compile didn't detect ambiguous grammar"); // it says it compiled it - well, then it should parse it! // (this helps us see what it thinks it should be doing.) g_check_parse_match(p, be, "AA",2, "(u0x41 u0x41)"); g_check_parse_match(p, be, "AB",2, "((u0x41 u0x42))"); g_test_fail(); return; } }
void init_parser(void) { // CORE HParser *digit = h_ch_range(0x30, 0x39); HParser *alpha = h_choice(h_ch_range(0x41, 0x5a), h_ch_range(0x61, 0x7a), NULL); // AUX. HParser *plus = h_ch('+'); HParser *slash = h_ch('/'); HParser *equals = h_ch('='); HParser *bsfdig = h_choice(alpha, digit, plus, slash, NULL); HParser *bsfdig_4bit = h_in((uint8_t *)"AEIMQUYcgkosw048", 16); HParser *bsfdig_2bit = h_in((uint8_t *)"AQgw", 4); HParser *base64_3 = h_repeat_n(bsfdig, 4); HParser *base64_2 = h_sequence(bsfdig, bsfdig, bsfdig_4bit, equals, NULL); HParser *base64_1 = h_sequence(bsfdig, bsfdig_2bit, equals, equals, NULL); HParser *base64 = h_sequence(h_many(base64_3), h_optional(h_choice(base64_2, base64_1, NULL)), NULL); document = h_sequence(h_whitespace(base64), h_whitespace(h_end_p()), NULL); }
static void test_benchmark_1() { HParser *parser = h_sepBy1(h_choice(h_ch('1'), h_ch('2'), h_ch('3'), NULL), h_ch(',')); HBenchmarkResults *res = h_benchmark(parser, testcases); h_benchmark_report(stderr, res); }
void init_parser(void) { // CORE const HParser *digit = h_ch_range(0x30, 0x39); const HParser *alpha = h_choice(h_ch_range(0x41, 0x5a), h_ch_range(0x61, 0x7a), NULL); // AUX. const HParser *plus = h_ch('+'); const HParser *slash = h_ch('/'); const HParser *equals = h_ch('='); const HParser *bsfdig = h_choice(alpha, digit, plus, slash, NULL); const HParser *bsfdig_4bit = h_choice( h_ch('A'), h_ch('E'), h_ch('I'), h_ch('M'), h_ch('Q'), h_ch('U'), h_ch('Y'), h_ch('c'), h_ch('g'), h_ch('k'), h_ch('o'), h_ch('s'), h_ch('w'), h_ch('0'), h_ch('4'), h_ch('8'), NULL); const HParser *bsfdig_2bit = h_choice(h_ch('A'), h_ch('Q'), h_ch('g'), h_ch('w'), NULL); const HParser *base64_2 = h_sequence(bsfdig, bsfdig, bsfdig_4bit, equals, NULL); const HParser *base64_1 = h_sequence(bsfdig, bsfdig_2bit, equals, equals, NULL); const HParser *base64 = h_choice(base64_2, base64_1, NULL); // why does this parse "A=="?! // why does this parse "aaA=" but not "aA=="?! document = base64; }