diff options
Diffstat (limited to 'src/n3.c')
| -rw-r--r-- | src/n3.c | 215 |
1 files changed, 105 insertions, 110 deletions
@@ -1,9 +1,8 @@ -// Copyright 2011-2023 David Robillard <d@drobilla.net> +// Copyright 2011-2025 David Robillard <d@drobilla.net> // SPDX-License-Identifier: ISC #include "reader.h" #include "serd_internal.h" -#include "stack.h" #include "string_utils.h" #include "try.h" #include "uri_utils.h" @@ -46,7 +45,7 @@ read_HEX(SerdReader* const reader) return (uint8_t)eat_byte_safe(reader, c); } - r_err(reader, SERD_ERR_BAD_SYNTAX, "invalid hexadecimal digit '%c'\n", c); + r_err_char(reader, "hexadecimal", c); return 0; } @@ -354,11 +353,11 @@ read_STRING_LITERAL_LONG(SerdReader* const reader, push_byte(reader, ref, c); st = read_character(reader, ref, flags, (uint8_t)q2); } - } else if (c == EOF) { - st = r_err(reader, SERD_ERR_BAD_SYNTAX, "end of file in long string\n"); - } else { + } else if (c > 0) { st = read_character(reader, ref, flags, (uint8_t)eat_byte_safe(reader, c)); + } else { + return r_err(reader, SERD_ERR_BAD_SYNTAX, "end of file in long string\n"); } } @@ -457,7 +456,7 @@ read_PN_CHARS_BASE(SerdReader* const reader, const Ref dest) return push_byte(reader, dest, eat_byte_safe(reader, c)); } - if (c == EOF || !(c & 0x80)) { + if (c < 0x80) { return SERD_FAILURE; } @@ -465,11 +464,7 @@ read_PN_CHARS_BASE(SerdReader* const reader, const Ref dest) read_utf8_code(reader, dest, &code, (uint8_t)c); if (!is_PN_CHARS_BASE(code)) { - r_err( - reader, SERD_ERR_BAD_SYNTAX, "invalid character U+%04X in name\n", code); - if (reader->strict) { - return SERD_ERR_BAD_SYNTAX; - } + st = r_err_char(reader, "name", (int)code); } return st; @@ -493,7 +488,7 @@ read_PN_CHARS(SerdReader* const reader, const Ref dest) return push_byte(reader, dest, eat_byte_safe(reader, c)); } - if (c == EOF || !(c & 0x80)) { + if (c < 0x80) { return SERD_FAILURE; } @@ -501,8 +496,7 @@ read_PN_CHARS(SerdReader* const reader, const Ref dest) TRY(st, read_utf8_code(reader, dest, &code, (uint8_t)c)); if (!is_PN_CHARS(code)) { - return r_err( - reader, SERD_ERR_BAD_SYNTAX, "invalid character U+%04X in name\n", code); + st = r_err_char(reader, "name", (int)code); } return st; @@ -531,7 +525,7 @@ read_PN_LOCAL_ESC(SerdReader* const reader, const Ref dest) return ((c == '!') || in_range(c, '#', '/') || (c == ';') || (c == '=') || (c == '?') || (c == '@') || (c == '_') || (c == '~')) ? push_byte(reader, dest, eat_byte_safe(reader, c)) - : r_err(reader, SERD_ERR_BAD_SYNTAX, "invalid escape\n"); + : r_err(reader, SERD_ERR_BAD_SYNTAX, "bad escape\n"); } static SerdStatus @@ -589,9 +583,7 @@ read_PN_LOCAL(SerdReader* const reader, const Ref dest, bool* const ate_dot) SerdNode* const n = deref(reader, dest); if (trailing_unescaped_dot) { // Ate trailing dot, pop it from stack/node and inform caller - --n->n_bytes; - serd_stack_pop(&reader->stack, 1); - *ate_dot = true; + *ate_dot = pop_last_node_char(reader, n); } return (st > SERD_FAILURE) ? st : SERD_SUCCESS; @@ -599,31 +591,37 @@ read_PN_LOCAL(SerdReader* const reader, const Ref dest, bool* const ate_dot) // Read the remainder of a PN_PREFIX after some initial characters static SerdStatus -read_PN_PREFIX_tail(SerdReader* const reader, const Ref dest) +read_PN_PREFIX_tail(SerdReader* const reader, + const Ref dest, + bool* const ate_dot) { - int c = 0; - while ((c = peek_byte(reader)) > 0) { // Middle: (PN_CHARS | '.')* + SerdStatus st = SERD_SUCCESS; + bool trailing_unescaped_dot = false; + + while (!st) { // Middle: (PN_CHARS | '.')* + const int c = peek_byte(reader); if (c == '.') { push_byte(reader, dest, eat_byte_safe(reader, c)); - } else if (read_PN_CHARS(reader, dest)) { - break; + trailing_unescaped_dot = true; + } else if (!(st = read_PN_CHARS(reader, dest))) { + trailing_unescaped_dot = false; } } - const SerdNode* const n = deref(reader, dest); - if (n->buf[n->n_bytes - 1] == '.' && read_PN_CHARS(reader, dest)) { - return r_err(reader, SERD_ERR_BAD_SYNTAX, "prefix ends with '.'\n"); + if (trailing_unescaped_dot) { + SerdNode* const n = deref(reader, dest); + *ate_dot = pop_last_node_char(reader, n); } - return SERD_SUCCESS; + return st; } static SerdStatus -read_PN_PREFIX(SerdReader* const reader, const Ref dest) +read_PN_PREFIX(SerdReader* const reader, const Ref dest, bool* const ate_dot) { const SerdStatus st = read_PN_CHARS_BASE(reader, dest); - return st ? st : read_PN_PREFIX_tail(reader, dest); + return st ? st : read_PN_PREFIX_tail(reader, dest, ate_dot); } static SerdStatus @@ -631,7 +629,7 @@ read_LANGTAG(SerdReader* const reader, Ref* const dest) { int c = peek_byte(reader); if (!is_alpha(c)) { - return r_err(reader, SERD_ERR_BAD_SYNTAX, "unexpected '%c'\n", c); + return r_err_char(reader, "language", c); } *dest = push_node(reader, SERD_LITERAL, "", 0); @@ -657,7 +655,7 @@ read_IRIREF_scheme(SerdReader* const reader, const Ref dest) { int c = peek_byte(reader); if (!is_alpha(c)) { - return r_err(reader, SERD_ERR_BAD_SYNTAX, "bad IRI scheme start '%c'\n", c); + return r_err_char(reader, "IRI scheme start", c); } while ((c = peek_byte(reader)) > 0) { @@ -666,11 +664,7 @@ read_IRIREF_scheme(SerdReader* const reader, const Ref dest) } if (!is_uri_scheme_char(c)) { - return r_err(reader, - SERD_ERR_BAD_SYNTAX, - "bad IRI scheme char U+%04X (%c)\n", - (unsigned)c, - (char)c); + return r_err_char(reader, "IRI scheme", c); } push_byte(reader, dest, eat_byte_safe(reader, c)); @@ -704,8 +698,7 @@ read_IRIREF(SerdReader* const reader, Ref* const dest) case '"': case '<': *dest = pop_node(reader, *dest); - return r_err( - reader, SERD_ERR_BAD_SYNTAX, "invalid IRI character '%c'\n", c); + return r_err_char(reader, "IRI", c); case '>': return SERD_SUCCESS; @@ -713,7 +706,7 @@ read_IRIREF(SerdReader* const reader, Ref* const dest) case '\\': if (read_UCHAR(reader, *dest, &code)) { *dest = pop_node(reader, *dest); - return r_err(reader, SERD_ERR_BAD_SYNTAX, "invalid IRI escape\n"); + return r_err_char(reader, "IRI escape", c); } if (code == ' ' || code == '<' || code == '>') { @@ -731,11 +724,12 @@ read_IRIREF(SerdReader* const reader, Ref* const dest) case '|': case '}': *dest = pop_node(reader, *dest); - return r_err( - reader, SERD_ERR_BAD_SYNTAX, "invalid IRI character '%c'\n", c); + return r_err_char(reader, "IRI", c); default: - if (c <= 0x20) { + if (c <= 0) { + st = r_err(reader, SERD_ERR_BAD_SYNTAX, "unexpected end of file\n"); + } else if (c <= 0x20) { st = r_err(reader, SERD_ERR_BAD_SYNTAX, "invalid IRI character (escape %%%02X)\n", @@ -743,8 +737,6 @@ read_IRIREF(SerdReader* const reader, Ref* const dest) if (!reader->strict) { st = SERD_FAILURE; push_byte(reader, *dest, c); - } else { - break; } } else if (!(c & 0x80)) { push_byte(reader, *dest, c); @@ -762,15 +754,10 @@ read_IRIREF(SerdReader* const reader, Ref* const dest) } static SerdStatus -read_PrefixedName(SerdReader* const reader, - const Ref dest, - const bool read_prefix, - bool* const ate_dot) +read_PrefixedName(SerdReader* const reader, const Ref dest, bool* const ate_dot) { SerdStatus st = SERD_SUCCESS; - if (read_prefix) { - TRY_FAILING(st, read_PN_PREFIX(reader, dest)); - } + TRY_FAILING(st, read_PN_PREFIX(reader, dest, ate_dot)); if (peek_byte(reader) != ':') { return SERD_FAILURE; @@ -825,18 +812,16 @@ read_number(SerdReader* const reader, // all other cases ::= ( '-' | '+' ) [0-9]+ ( . )? ( [0-9]+ )? ... TRY(st, read_0_9(reader, *dest, true)); if ((c = peek_byte(reader)) == '.') { - has_decimal = true; - // Annoyingly, dot can be end of statement, so tentatively eat skip_byte(reader, c); c = peek_byte(reader); if (!is_digit(c) && c != 'e' && c != 'E') { - *ate_dot = true; // Force caller to deal with stupid grammar - return SERD_SUCCESS; // Next byte is not a number character + *ate_dot = true; // Force caller to deal with silly grammar + } else { + has_decimal = true; + push_byte(reader, *dest, '.'); + read_0_9(reader, *dest, false); } - - push_byte(reader, *dest, '.'); - read_0_9(reader, *dest, false); } } c = peek_byte(reader); @@ -868,7 +853,7 @@ read_iri(SerdReader* const reader, Ref* const dest, bool* const ate_dot) } *dest = push_node(reader, SERD_CURIE, "", 0); - return read_PrefixedName(reader, *dest, true, ate_dot); + return read_PrefixedName(reader, *dest, ate_dot); } static SerdStatus @@ -920,29 +905,30 @@ read_verb(SerdReader* const reader, Ref* const dest) return read_IRIREF(reader, dest); } - /* Either a qname, or "a". Read the prefix first, and if it is in fact - "a", produce that instead. - */ - *dest = push_node(reader, SERD_CURIE, "", 0); + Ref p = push_node(reader, SERD_CURIE, "", 0); - SerdStatus st = read_PN_PREFIX(reader, *dest); - bool ate_dot = false; - const SerdNode* const node = deref(reader, *dest); - const int next = peek_byte(reader); - if (!st && node->n_bytes == 1 && node->buf[0] == 'a' && next != ':' && - !is_PN_CHARS_BASE((uint32_t)next)) { - pop_node(reader, *dest); - *dest = push_node(reader, SERD_URI, NS_RDF "type", 47); - return SERD_SUCCESS; + // Try to read as a prefixed name + bool ate_dot = false; + SerdStatus st = read_PrefixedName(reader, p, &ate_dot); + + if (st == SERD_FAILURE) { + // Check if this is actually the "a" shorthand + const SerdNode* const node = deref(reader, p); + if (node->n_bytes == 1 && node->buf[0] == 'a') { + pop_node(reader, p); + p = push_node(reader, SERD_URI, NS_RDF "type", 47); + st = SERD_SUCCESS; + } else { + st = SERD_ERR_BAD_SYNTAX; + } } - if (st > SERD_FAILURE || - (st = read_PrefixedName(reader, *dest, false, &ate_dot)) || ate_dot) { - *dest = pop_node(reader, *dest); - st = st > SERD_FAILURE ? st : SERD_ERR_BAD_SYNTAX; + if (st) { + pop_node(reader, p); return r_err(reader, st, "bad verb\n"); } + *dest = p; return SERD_SUCCESS; } @@ -981,9 +967,7 @@ read_BLANK_NODE_LABEL(SerdReader* const reader, SerdNode* n = deref(reader, ref); if (n->buf[n->n_bytes - 1] == '.' && read_PN_CHARS(reader, ref)) { // Ate trailing dot, pop it from stack/node and inform caller - --n->n_bytes; - serd_stack_pop(&reader->stack, 1); - *ate_dot = true; + *ate_dot = pop_last_node_char(reader, n); } if (fancy_syntax(reader)) { @@ -1078,6 +1062,40 @@ read_anon(SerdReader* const reader, : SERD_ERR_BAD_SYNTAX; } +// Read a "named" object: a boolean literal or a prefixed name +static SerdStatus +read_named_object(SerdReader* const reader, + Ref* const dest, + Ref* const datatype, + bool* const ate_dot) +{ + static const char* const XSD_BOOLEAN = NS_XSD "boolean"; + static const size_t XSD_BOOLEAN_LEN = 40; + + // Try to read as a prefixed name + const Ref o = push_node(reader, SERD_CURIE, "", 0); + SerdStatus st = read_PrefixedName(reader, o, ate_dot); + + if (st == SERD_FAILURE) { + // Check if this is actually a boolean literal + SerdNode* const node = deref(reader, o); + if ((node->n_bytes == 4 && !memcmp(node->buf, "true", 4)) || + (node->n_bytes == 5 && !memcmp(node->buf, "false", 5))) { + node->type = SERD_LITERAL; + *datatype = push_node(reader, SERD_URI, XSD_BOOLEAN, XSD_BOOLEAN_LEN); + st = SERD_SUCCESS; + } + } + + if (st) { + pop_node(reader, o); + return r_err(reader, SERD_ERR_BAD_SYNTAX, "expected prefixed name\n"); + } + + *dest = o; + return SERD_SUCCESS; +} + /* If emit is true: recurses, calling statement_sink for every statement encountered, and leaves stack in original calling state (i.e. pops everything it pushes). */ @@ -1087,9 +1105,6 @@ read_object(SerdReader* const reader, const bool emit, bool* const ate_dot) { - static const char* const XSD_BOOLEAN = NS_XSD "boolean"; - static const size_t XSD_BOOLEAN_LEN = 40; - #ifndef NDEBUG const size_t orig_stack_size = reader->stack.size; #endif @@ -1097,7 +1112,6 @@ read_object(SerdReader* const reader, SerdStatus st = SERD_FAILURE; bool simple = (ctx->subject != 0); - SerdNode* node = NULL; Ref o = 0; Ref datatype = 0; Ref lang = 0; @@ -1147,27 +1161,8 @@ read_object(SerdReader* const reader, st = read_literal(reader, &o, &datatype, &lang, &flags, ate_dot); break; default: - /* Either a boolean literal, or a qname. Read the prefix first, and if - it is in fact a "true" or "false" literal, produce that instead. - */ - o = push_node(reader, SERD_CURIE, "", 0); - while (!read_PN_CHARS_BASE(reader, o)) { - } - node = deref(reader, o); - if ((node->n_bytes == 4 && !memcmp(node->buf, "true", 4)) || - (node->n_bytes == 5 && !memcmp(node->buf, "false", 5))) { - node->type = SERD_LITERAL; - datatype = push_node(reader, SERD_URI, XSD_BOOLEAN, XSD_BOOLEAN_LEN); - st = SERD_SUCCESS; - } else if (read_PN_PREFIX_tail(reader, o) > SERD_FAILURE) { - st = SERD_ERR_BAD_SYNTAX; - } else { - if ((st = read_PrefixedName(reader, o, false, ate_dot))) { - st = st > SERD_FAILURE ? st : SERD_ERR_BAD_SYNTAX; - pop_node(reader, o); - return r_err(reader, st, "expected prefixed name\n"); - } - } + // Either a boolean literal or a prefixed name + st = read_named_object(reader, &o, &datatype, ate_dot); } if (!st && simple && o) { @@ -1444,12 +1439,12 @@ read_prefixID(SerdReader* const reader, const bool sparql, const bool token) } read_ws_star(reader); - Ref name = push_node(reader, SERD_LITERAL, "", 0); - TRY_FAILING(st, read_PN_PREFIX(reader, name)); - - if (eat_byte_check(reader, ':') != ':') { + Ref name = push_node(reader, SERD_LITERAL, "", 0); + bool ate_dot = false; + TRY_FAILING(st, read_PN_PREFIX(reader, name, &ate_dot)); + if (ate_dot || eat_byte_check(reader, ':') != ':') { pop_node(reader, name); - return SERD_ERR_BAD_SYNTAX; + return r_err(reader, SERD_ERR_BAD_SYNTAX, "expected a prefix name\n"); } read_ws_star(reader); |