From 380c2bca89cd0190f8fa94753869b8c0ba45d921 Mon Sep 17 00:00:00 2001 From: David Robillard Date: Sun, 1 Apr 2018 20:33:33 +0200 Subject: Merge datatype and/or language into node This moves closer to the sord API, and is more convenient in most cases. --- src/n3.c | 23 ++++++---- src/node.c | 129 +++++++++++++++++++++++++++++++++++++++++++++++----- src/reader.c | 4 +- src/serd_internal.h | 2 +- src/writer.c | 48 ++++++++----------- 5 files changed, 156 insertions(+), 50 deletions(-) (limited to 'src') diff --git a/src/n3.c b/src/n3.c index 5ffa3413..e6e21d58 100644 --- a/src/n3.c +++ b/src/n3.c @@ -697,7 +697,11 @@ read_0_9(SerdReader* reader, Ref str, bool at_least_one) } static bool -read_number(SerdReader* reader, Ref* dest, Ref* datatype, bool* ate_dot) +read_number(SerdReader* reader, + Ref* dest, + Ref* datatype, + SerdNodeFlags* flags, + bool* ate_dot) { #define XSD_DECIMAL NS_XSD "decimal" #define XSD_DOUBLE NS_XSD "double" @@ -752,6 +756,7 @@ read_number(SerdReader* reader, Ref* dest, Ref* datatype, bool* ate_dot) *datatype = push_node(reader, SERD_URI, XSD_INTEGER, sizeof(XSD_INTEGER) - 1); } + *flags |= SERD_HAS_DATATYPE; *dest = ref; return true; except: @@ -785,11 +790,13 @@ read_literal(SerdReader* reader, Ref* dest, switch (peek_byte(reader)) { case '@': eat_byte_safe(reader, '@'); + *flags |= SERD_HAS_LANGUAGE; TRY_THROW(*lang = read_LANGTAG(reader)); break; case '^': eat_byte_safe(reader, '^'); eat_byte_check(reader, '^'); + *flags |= SERD_HAS_DATATYPE; TRY_THROW(read_iri(reader, datatype, ate_dot)); break; } @@ -929,7 +936,7 @@ read_anon(SerdReader* reader, ReadContext ctx, bool subject, Ref* dest) *dest = blank_id(reader); } if (ctx.subject) { - TRY_RET(emit_statement(reader, ctx, *dest, 0, 0)); + TRY_RET(emit_statement(reader, ctx, *dest)); } ctx.subject = *dest; @@ -999,7 +1006,7 @@ read_object(SerdReader* reader, ReadContext* ctx, bool emit, bool* ate_dot) break; case '+': case '-': case '.': case '0': case '1': case '2': case '3': case '4': case '5': case '6': case '7': case '8': case '9': - TRY_THROW(ret = read_number(reader, &o, &datatype, ate_dot)); + TRY_THROW(ret = read_number(reader, &o, &datatype, &flags, ate_dot)); break; case '\"': case '\'': @@ -1016,6 +1023,7 @@ read_object(SerdReader* reader, ReadContext* ctx, bool emit, bool* ate_dot) !memcmp(serd_node_get_string(node), "true", 4)) || (node->n_bytes == 5 && !memcmp(serd_node_get_string(node), "false", 5))) { + flags = flags | SERD_HAS_DATATYPE; node->type = SERD_LITERAL; datatype = push_node( reader, SERD_URI, XSD_BOOLEAN, XSD_BOOLEAN_LEN); @@ -1034,7 +1042,7 @@ read_object(SerdReader* reader, ReadContext* ctx, bool emit, bool* ate_dot) } if (ret && emit && simple) { - ret = emit_statement(reader, *ctx, o, datatype, lang); + ret = emit_statement(reader, *ctx, o); } else if (ret && !emit) { ctx->object = o; ctx->datatype = datatype; @@ -1120,7 +1128,7 @@ read_collection(SerdReader* reader, ReadContext ctx, Ref* dest) if (ctx.subject) { // subject predicate _:head *ctx.flags |= (end ? 0 : SERD_LIST_O_BEGIN); - TRY_RET(emit_statement(reader, ctx, *dest, 0, 0)); + TRY_RET(emit_statement(reader, ctx, *dest)); *ctx.flags |= SERD_LIST_CONT; } else { *ctx.flags |= (end ? 0 : SERD_LIST_S_BEGIN); @@ -1159,8 +1167,7 @@ read_collection(SerdReader* reader, ReadContext ctx, Ref* dest) // _:node rdf:rest _:rest *ctx.flags |= SERD_LIST_CONT; ctx.predicate = reader->rdf_rest; - TRY_RET(emit_statement(reader, ctx, - (end ? reader->rdf_nil : rest), 0, 0)); + TRY_RET(emit_statement(reader, ctx, (end ? reader->rdf_nil : rest))); ctx.subject = rest; // _:node = _:rest rest = node; // _:rest = (old)_:node @@ -1494,7 +1501,7 @@ read_nquadsDoc(SerdReader* reader) eat_byte_check(reader, '.'); } - if (!emit_statement(reader, ctx, ctx.object, ctx.datatype, ctx.lang)) { + if (!emit_statement(reader, ctx, ctx.object)) { break; } diff --git a/src/node.c b/src/node.c index 901d5846..c2fabe0b 100644 --- a/src/node.c +++ b/src/node.c @@ -30,13 +30,36 @@ # endif #endif +static const size_t serd_node_align = sizeof(SerdNode); + +static size_t +serd_node_pad_size(const size_t n_bytes) +{ + const size_t pad = serd_node_align - (n_bytes + 2) % serd_node_align; + return n_bytes + 2 + pad; +} + +static size_t +serd_node_total_size(const SerdNode* node) +{ + const size_t len = sizeof(SerdNode) + serd_node_pad_size(node->n_bytes); + if (node->flags & SERD_HAS_LANGUAGE) { + return len + serd_node_total_size(serd_node_get_language(node)); + } else if (node->flags & SERD_HAS_DATATYPE) { + return len + serd_node_total_size(serd_node_get_datatype(node)); + } + return len; +} + SerdNode* serd_node_malloc(size_t n_bytes, SerdNodeFlags flags, SerdType type) { - SerdNode* node = (SerdNode*)calloc(1, sizeof(SerdNode) + n_bytes + 1); + const size_t size = sizeof(SerdNode) + serd_node_pad_size(n_bytes); + SerdNode* node = (SerdNode*)calloc(1, size); node->n_bytes = 0; node->flags = flags; node->type = type; + assert((intptr_t)node % serd_node_align == 0); return node; } @@ -50,11 +73,12 @@ void serd_node_set(SerdNode** dst, const SerdNode* src) { if (src) { - if (!(*dst) || (*dst)->n_bytes < src->n_bytes) { - (*dst) = (SerdNode*)realloc(*dst, sizeof(SerdNode) + src->n_bytes + 1); + const size_t size = serd_node_total_size(src); + if (!(*dst) || serd_node_total_size(*dst) < size) { + (*dst) = (SerdNode*)realloc(*dst, size); } - memcpy(*dst, src, sizeof(SerdNode) + src->n_bytes + 1); + memcpy(*dst, src, size); } else if (*dst) { (*dst)->type = SERD_NOTHING; } @@ -90,6 +114,51 @@ serd_node_new_substring(SerdType type, const char* str, const size_t len) return node; } +SerdNode* +serd_node_new_literal(const char* str, const char* datatype, const char* lang) +{ + if (!str || (lang && datatype && strcmp(datatype, NS_RDF "#langString"))) { + return NULL; + } + + uint32_t flags = 0; + const size_t n_bytes = serd_strlen(str, &flags); + const size_t len = serd_node_pad_size(n_bytes); + + SerdNode* node = NULL; + if (lang) { + flags |= SERD_HAS_LANGUAGE; + const size_t lang_len = strlen(lang); + const size_t total_len = len + sizeof(SerdNode) + lang_len; + node = serd_node_malloc(total_len, flags, SERD_LITERAL); + memcpy(serd_node_buffer(node), str, n_bytes); + node->n_bytes = n_bytes; + + SerdNode* lang_node = node + 1 + (len / serd_node_align); + lang_node->type = SERD_LITERAL; + lang_node->n_bytes = lang_len; + memcpy(serd_node_buffer(lang_node), lang, lang_len); + } else if (datatype) { + flags |= SERD_HAS_DATATYPE; + const size_t datatype_len = strlen(datatype); + const size_t total_len = len + sizeof(SerdNode) + datatype_len; + node = serd_node_malloc(total_len, flags, SERD_LITERAL); + memcpy(serd_node_buffer(node), str, n_bytes); + node->n_bytes = n_bytes; + + SerdNode* datatype_node = node + 1 + (len / serd_node_align); + datatype_node->type = SERD_URI; + datatype_node->n_bytes = datatype_len; + memcpy(serd_node_buffer(datatype_node), datatype, datatype_len); + } else { + node = serd_node_malloc(n_bytes, flags, SERD_LITERAL); + memcpy(serd_node_buffer(node), str, n_bytes); + node->n_bytes = n_bytes; + } + + return node; +} + SerdNode* serd_node_copy(const SerdNode* node) { @@ -97,8 +166,8 @@ serd_node_copy(const SerdNode* node) return NULL; } - const size_t size = sizeof(SerdNode) + node->n_bytes + 1; - SerdNode* copy = (SerdNode*)malloc(size); + const size_t size = serd_node_total_size(node); + SerdNode* copy = (SerdNode*)calloc(1, size + 3); memcpy(copy, node, size); return copy; } @@ -106,11 +175,17 @@ serd_node_copy(const SerdNode* node) bool serd_node_equals(const SerdNode* a, const SerdNode* b) { - return (a == b) || - (a && b && a->type == b->type && a->n_bytes == b->n_bytes && - !memcmp(serd_node_get_string(a), - serd_node_get_string(b), - a->n_bytes)); + if (a == b) { + return true; + } else if (!a || !b) { + return false; + } + + const size_t a_size = serd_node_total_size(a); + if (serd_node_total_size(b) == a_size) { + return !memcmp(a, b, a_size); + } + return false; } static size_t @@ -428,6 +503,38 @@ serd_node_get_length(const SerdNode* node) return node ? node->n_bytes : 0; } +const SerdNode* +serd_node_get_datatype(const SerdNode* node) +{ + if (!node || !(node->flags & SERD_HAS_DATATYPE)) { + return NULL; + } + + const size_t len = serd_node_pad_size(node->n_bytes); + assert((intptr_t)node % serd_node_align == 0); + assert(len % serd_node_align == 0); + + const SerdNode* const datatype = node + 1 + (len / serd_node_align); + assert(datatype->type == SERD_URI || datatype->type == SERD_CURIE); + return datatype; +} + +const SerdNode* +serd_node_get_language(const SerdNode* node) +{ + if (!node || !(node->flags & SERD_HAS_LANGUAGE)) { + return NULL; + } + + const size_t len = serd_node_pad_size(node->n_bytes); + assert((intptr_t)node % serd_node_align == 0); + assert(len % serd_node_align == 0); + + const SerdNode* const lang = node + 1 + (len / serd_node_align); + assert(lang->type == SERD_LITERAL); + return lang; +} + SerdNodeFlags serd_node_get_flags(const SerdNode* node) { diff --git a/src/reader.c b/src/reader.c index 41fe4848..ec1c533d 100644 --- a/src/reader.c +++ b/src/reader.c @@ -113,7 +113,7 @@ pop_node(SerdReader* reader, Ref ref) } bool -emit_statement(SerdReader* reader, ReadContext ctx, Ref o, Ref d, Ref l) +emit_statement(SerdReader* reader, ReadContext ctx, Ref o) { SerdNode* graph = deref(reader, ctx.graph); if (!graph && reader->default_graph) { @@ -123,7 +123,7 @@ emit_statement(SerdReader* reader, ReadContext ctx, Ref o, Ref d, Ref l) !reader->statement_sink( reader->handle, *ctx.flags, graph, deref(reader, ctx.subject), deref(reader, ctx.predicate), - deref(reader, o), deref(reader, d), deref(reader, l)); + deref(reader, o)); *ctx.flags &= SERD_ANON_CONT|SERD_LIST_CONT; // Preserve only cont flags return ret; } diff --git a/src/serd_internal.h b/src/serd_internal.h index ed22ae5f..63d7e62e 100644 --- a/src/serd_internal.h +++ b/src/serd_internal.h @@ -617,7 +617,7 @@ SerdNode* deref(SerdReader* reader, Ref ref); Ref pop_node(SerdReader* reader, Ref ref); -bool emit_statement(SerdReader* reader, ReadContext ctx, Ref o, Ref d, Ref l); +bool emit_statement(SerdReader* reader, ReadContext ctx, Ref o); bool read_n3_statement(SerdReader* reader); SerdStatus read_nquadsDoc(SerdReader* reader); diff --git a/src/writer.c b/src/writer.c index 79987b3d..6e47ee2a 100644 --- a/src/writer.c +++ b/src/writer.c @@ -103,8 +103,6 @@ typedef enum { static bool write_node(SerdWriter* writer, const SerdNode* node, - const SerdNode* datatype, - const SerdNode* lang, Field field, SerdStatementFlags flags); @@ -428,12 +426,12 @@ is_inline_start(const SerdWriter* writer, Field field, SerdStatementFlags flags) static bool write_literal(SerdWriter* writer, const SerdNode* node, - const SerdNode* datatype, - const SerdNode* lang, SerdStatementFlags flags) { - const char* node_str = serd_node_get_string(node); - const char* type_uri = serd_node_get_string(datatype); + const SerdNode* datatype = serd_node_get_datatype(node); + const SerdNode* lang = serd_node_get_language(node); + const char* node_str = serd_node_get_string(node); + const char* type_uri = serd_node_get_string(datatype); if (supports_abbrev(writer) && type_uri) { if (!strncmp(type_uri, NS_XSD, sizeof(NS_XSD) - 1) && ( !strcmp(type_uri + sizeof(NS_XSD) - 1, "boolean") || @@ -468,7 +466,7 @@ write_literal(SerdWriter* writer, sink(serd_node_get_string(lang), lang->n_bytes, writer); } else if (type_uri) { sink("^^", 2, writer); - return write_node(writer, datatype, NULL, NULL, FIELD_NONE, flags); + return write_node(writer, datatype, FIELD_NONE, flags); } return true; } @@ -626,15 +624,13 @@ write_blank(SerdWriter* const writer, static bool write_node(SerdWriter* writer, const SerdNode* node, - const SerdNode* datatype, - const SerdNode* lang, Field field, SerdStatementFlags flags) { bool ret = false; switch (node->type) { case SERD_LITERAL: - ret = write_literal(writer, node, datatype, lang, flags); + ret = write_literal(writer, node, flags); break; case SERD_URI: ret = write_uri_node(writer, node, field, flags); @@ -659,7 +655,7 @@ is_resource(const SerdNode* node) static void write_pred(SerdWriter* writer, SerdStatementFlags flags, const SerdNode* pred) { - write_node(writer, pred, NULL, NULL, FIELD_PREDICATE, flags); + write_node(writer, pred, FIELD_PREDICATE, flags); write_sep(writer, SEP_P_O); serd_node_set(&writer->context.predicate, pred); } @@ -668,9 +664,7 @@ static bool write_list_obj(SerdWriter* writer, SerdStatementFlags flags, const SerdNode* predicate, - const SerdNode* object, - const SerdNode* datatype, - const SerdNode* lang) + const SerdNode* object) { if (!strcmp(serd_node_get_string(object), NS_RDF "nil")) { --writer->indent; @@ -678,7 +672,7 @@ write_list_obj(SerdWriter* writer, return true; } else if (!strcmp(serd_node_get_string(predicate), NS_RDF "first")) { write_sep(writer, SEP_LIST_SEP); - write_node(writer, object, datatype, lang, FIELD_OBJECT, flags); + write_node(writer, object, FIELD_OBJECT, flags); } return false; } @@ -689,9 +683,7 @@ serd_writer_write_statement(SerdWriter* writer, const SerdNode* graph, const SerdNode* subject, const SerdNode* predicate, - const SerdNode* object, - const SerdNode* datatype, - const SerdNode* lang) + const SerdNode* object) { if (!subject || !predicate || !object || !is_resource(subject) || !is_resource(predicate)) { @@ -706,14 +698,14 @@ serd_writer_write_statement(SerdWriter* writer, switch (writer->syntax) { case SERD_NTRIPLES: case SERD_NQUADS: - TRY(write_node(writer, subject, NULL, NULL, FIELD_SUBJECT, flags)); + TRY(write_node(writer, subject, FIELD_SUBJECT, flags)); sink(" ", 1, writer); - TRY(write_node(writer, predicate, NULL, NULL, FIELD_PREDICATE, flags)); + TRY(write_node(writer, predicate, FIELD_PREDICATE, flags)); sink(" ", 1, writer); - TRY(write_node(writer, object, datatype, lang, FIELD_OBJECT, flags)); + TRY(write_node(writer, object, FIELD_OBJECT, flags)); if (writer->syntax == SERD_NQUADS && graph) { sink(" ", 1, writer); - TRY(write_node(writer, graph, datatype, lang, FIELD_GRAPH, flags)); + TRY(write_node(writer, graph, FIELD_GRAPH, flags)); } sink(" .\n", 3, writer); return SERD_SUCCESS; @@ -733,7 +725,7 @@ serd_writer_write_statement(SerdWriter* writer, reset_context(writer, true); if (graph) { - TRY(write_node(writer, graph, datatype, lang, FIELD_GRAPH, flags)); + TRY(write_node(writer, graph, FIELD_GRAPH, flags)); ++writer->indent; write_sep(writer, SEP_GRAPH_BEGIN); serd_node_set(&writer->context.graph, graph); @@ -741,7 +733,7 @@ serd_writer_write_statement(SerdWriter* writer, } if ((flags & SERD_LIST_CONT)) { - if (write_list_obj(writer, flags, predicate, object, datatype, lang)) { + if (write_list_obj(writer, flags, predicate, object)) { // Reached end of list if (--writer->list_depth == 0 && writer->list_subj) { reset_context(writer, false); @@ -758,7 +750,7 @@ serd_writer_write_statement(SerdWriter* writer, ++writer->indent; } write_sep(writer, SEP_END_O); - write_node(writer, object, datatype, lang, FIELD_OBJECT, flags); + write_node(writer, object, FIELD_OBJECT, flags); if (!(flags & SERD_ANON_O_BEGIN)) { --writer->indent; } @@ -767,7 +759,7 @@ serd_writer_write_statement(SerdWriter* writer, Sep sep = ctx(writer, FIELD_PREDICATE) ? SEP_END_P : SEP_S_P; write_sep(writer, sep); write_pred(writer, flags, predicate); - write_node(writer, object, datatype, lang, FIELD_OBJECT, flags); + write_node(writer, object, FIELD_OBJECT, flags); } } else { // No abbreviation @@ -782,7 +774,7 @@ serd_writer_write_statement(SerdWriter* writer, } if (!(flags & SERD_ANON_CONT)) { - write_node(writer, subject, NULL, NULL, FIELD_SUBJECT, flags); + write_node(writer, subject, FIELD_SUBJECT, flags); ++writer->indent; write_sep(writer, SEP_S_P); } else { @@ -796,7 +788,7 @@ serd_writer_write_statement(SerdWriter* writer, write_pred(writer, flags, predicate); } - write_node(writer, object, datatype, lang, FIELD_OBJECT, flags); + write_node(writer, object, FIELD_OBJECT, flags); } if (flags & (SERD_ANON_S_BEGIN|SERD_ANON_O_BEGIN)) { -- cgit v1.2.1