Simplify reader interface

8 files changed, 113 insertions, 167 deletions

diff --git a/include/serd/serd.h b/include/serd/serd.h
index 9c205d26..d51adebe 100644
--- a/include/serd/serd.h
+++ b/include/serd/serd.h
@@ -954,11 +954,12 @@ void
 serd_reader_set_default_graph(SerdReader* SERD_NONNULL      reader,
                               const SerdNode* SERD_NULLABLE graph);
 
-/// Read a file at a given `uri`
+/// Prepare to read from the file at a local file `uri`
 SERD_API
 SerdStatus
-serd_reader_read_file(SerdReader* SERD_NONNULL reader,
-                      const char* SERD_NONNULL uri);
+serd_reader_start_file(SerdReader* SERD_NONNULL reader,
+                       const char* SERD_NONNULL uri,
+                       bool                     bulk);
 
 /**
    Prepare to read from a stream.
@@ -1004,18 +1005,20 @@ SERD_API
 SerdStatus
 serd_reader_read_document(SerdReader* SERD_NONNULL reader);
 
-/// Finish reading from the source
-SERD_API
-SerdStatus
-serd_reader_end_stream(SerdReader* SERD_NONNULL reader);
+/**
+   Finish reading from the source.
 
-/// Read `utf8`
+   This should be called before starting to read from another source.
+*/
 SERD_API
 SerdStatus
-serd_reader_read_string(SerdReader* SERD_NONNULL reader,
-                        const char* SERD_NONNULL utf8);
+serd_reader_finish(SerdReader* SERD_NONNULL reader);
 
-/// Free `reader`
+/**
+   Free `reader`.
+
+   The reader will be finished via `serd_reader_finish()` if necessary.
+*/
 SERD_API
 void
 serd_reader_free(SerdReader* SERD_NULLABLE reader);
diff --git a/src/byte_source.c b/src/byte_source.c
index 04836984..2f09d634 100644
--- a/src/byte_source.c
+++ b/src/byte_source.c
@@ -50,6 +50,7 @@ SerdStatus
 serd_byte_source_open_source(SerdByteSource*     source,
                              SerdReadFunc        read_func,
                              SerdStreamErrorFunc error_func,
+                             SerdStreamCloseFunc close_func,
                              void*               stream,
                              const char*         name,
                              size_t              page_size)
@@ -57,13 +58,14 @@ serd_byte_source_open_source(SerdByteSource*     source,
   const Cursor cur = {name, 1, 1};
 
   memset(source, '\0', sizeof(*source));
+  source->read_func   = read_func;
+  source->error_func  = error_func;
+  source->close_func  = close_func;
   source->stream      = stream;
-  source->from_stream = true;
   source->page_size   = page_size;
   source->buf_size    = page_size;
   source->cur         = cur;
-  source->error_func  = error_func;
-  source->read_func   = read_func;
+  source->from_stream = true;
 
   if (page_size > 1) {
     source->file_buf = (uint8_t*)serd_allocate_buffer(page_size);
@@ -103,10 +105,15 @@ serd_byte_source_open_string(SerdByteSource* source, const char* utf8)
 SerdStatus
 serd_byte_source_close(SerdByteSource* source)
 {
+  SerdStatus st = SERD_SUCCESS;
+  if (source->close_func) {
+    st = source->close_func(source->stream) ? SERD_ERR_UNKNOWN : SERD_SUCCESS;
+  }
+
   if (source->page_size > 1) {
     serd_free_aligned(source->file_buf);
   }
 
   memset(source, '\0', sizeof(*source));
-  return SERD_SUCCESS;
+  return st;
 }
diff --git a/src/byte_source.h b/src/byte_source.h
index f0df9bf3..d2c19de3 100644
--- a/src/byte_source.h
+++ b/src/byte_source.h
@@ -23,7 +23,8 @@
 #include <stdbool.h>
 #include <stddef.h>
 #include <stdint.h>
-#include <stdio.h>
+
+typedef int (*SerdStreamCloseFunc)(void*);
 
 typedef struct {
   const char* filename;
@@ -34,6 +35,7 @@ typedef struct {
 typedef struct {
   SerdReadFunc        read_func;   ///< Read function (e.g. fread)
   SerdStreamErrorFunc error_func;  ///< Error function (e.g. ferror)
+  SerdStreamCloseFunc close_func;  ///< Function for closing stream
   void*               stream;      ///< Stream (e.g. FILE)
   size_t              page_size;   ///< Number of bytes to read at a time
   size_t              buf_size;    ///< Number of bytes in file_buf
@@ -48,15 +50,13 @@ typedef struct {
 } SerdByteSource;
 
 SerdStatus
-serd_byte_source_open_file(SerdByteSource* source, FILE* file, bool bulk);
-
-SerdStatus
 serd_byte_source_open_string(SerdByteSource* source, const char* utf8);
 
 SerdStatus
 serd_byte_source_open_source(SerdByteSource*     source,
                              SerdReadFunc        read_func,
                              SerdStreamErrorFunc error_func,
+                             SerdStreamCloseFunc close_func,
                              void*               stream,
                              const char*         name,
                              size_t              page_size);
diff --git a/src/reader.c b/src/reader.c
index 805d41a7..ab9e54b5 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -16,11 +16,10 @@
 
 #include "reader.h"
 #include "byte_source.h"
+#include "serd_internal.h"
 #include "stack.h"
 #include "system.h"
 
-#include "serd_internal.h"
-
 #include <errno.h>
 #include <stdarg.h>
 #include <stdio.h>
@@ -207,6 +206,7 @@ serd_reader_free(SerdReader* reader)
   pop_node(reader, reader->rdf_nil);
   pop_node(reader, reader->rdf_rest);
   pop_node(reader, reader->rdf_first);
+  serd_reader_finish(reader);
   serd_node_free(reader->default_graph);
 
 #ifdef SERD_STACK_CHECK
@@ -239,39 +239,6 @@ serd_reader_set_default_graph(SerdReader* reader, const SerdNode* graph)
   reader->default_graph = serd_node_copy(graph);
 }
 
-SerdStatus
-serd_reader_read_file(SerdReader* reader, const char* uri)
-{
-  char* const path = serd_parse_file_uri(uri, NULL);
-  if (!path) {
-    return SERD_ERR_BAD_ARG;
-  }
-
-  FILE* fd = serd_fopen(path, "rb");
-  if (!fd) {
-    serd_free(path);
-    return SERD_ERR_UNKNOWN;
-  }
-
-  SerdStatus st = serd_reader_start_stream(reader,
-                                           (SerdReadFunc)fread,
-                                           (SerdStreamErrorFunc)ferror,
-                                           fd,
-                                           path,
-                                           SERD_PAGE_SIZE);
-
-  if (!st) {
-    st = serd_reader_read_document(reader);
-  }
-
-  const SerdStatus est = serd_reader_end_stream(reader);
-
-  fclose(fd);
-  free(path);
-
-  return st ? st : est;
-}
-
 static SerdStatus
 skip_bom(SerdReader* me)
 {
@@ -298,7 +265,31 @@ serd_reader_start_stream(SerdReader*         reader,
                          size_t              page_size)
 {
   return serd_byte_source_open_source(
-    &reader->source, read_func, error_func, stream, name, page_size);
+    &reader->source, read_func, error_func, NULL, stream, name, page_size);
+}
+
+SerdStatus
+serd_reader_start_file(SerdReader* reader, const char* uri, bool bulk)
+{
+  char* const path = serd_parse_file_uri(uri, NULL);
+  if (!path) {
+    return SERD_ERR_BAD_ARG;
+  }
+
+  FILE* fd = serd_fopen(path, "rb");
+  free(path);
+  if (!fd) {
+    return SERD_ERR_UNKNOWN;
+  }
+
+  return serd_byte_source_open_source(&reader->source,
+                                      bulk ? (SerdReadFunc)fread
+                                           : serd_file_read_byte,
+                                      (SerdStreamErrorFunc)ferror,
+                                      (SerdStreamCloseFunc)fclose,
+                                      fd,
+                                      uri,
+                                      bulk ? SERD_PAGE_SIZE : 1);
 }
 
 SerdStatus
@@ -340,18 +331,7 @@ serd_reader_read_chunk(SerdReader* reader)
 }
 
 SerdStatus
-serd_reader_end_stream(SerdReader* reader)
+serd_reader_finish(SerdReader* reader)
 {
   return serd_byte_source_close(&reader->source);
 }
-
-SerdStatus
-serd_reader_read_string(SerdReader* reader, const char* utf8)
-{
-  serd_reader_start_string(reader, utf8);
-
-  const SerdStatus st  = serd_reader_read_document(reader);
-  const SerdStatus est = serd_byte_source_close(&reader->source);
-
-  return st ? st : est;
-}
diff --git a/src/serdi.c b/src/serdi.c
index e7b8d7ba..6049c31b 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -14,10 +14,9 @@
   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */
 
-#define _POSIX_C_SOURCE 200809L /* for fileno and posix_fadvise */
-
 #include "serd_config.h"
 #include "string_utils.h"
+#include "system.h"
 
 #include "serd/serd.h"
 
@@ -29,15 +28,8 @@
 #  include <io.h>
 #endif
 
-#if USE_POSIX_FADVISE && USE_FILENO
-#  include <fcntl.h>
-#endif
-
-#include <errno.h>
 #include <stdbool.h>
-#include <stdint.h>
 #include <stdio.h>
-#include <stdlib.h>
 #include <string.h>
 
 #define SERDI_ERROR(msg) fprintf(stderr, "serdi: " msg)
@@ -135,38 +127,6 @@ quiet_error_sink(void* handle, const SerdError* e)
   return SERD_SUCCESS;
 }
 
-static inline FILE*
-serd_fopen(const char* path, const char* mode)
-{
-  FILE* fd = fopen(path, mode);
-  if (!fd) {
-    SERDI_ERRORF("failed to open file %s (%s)\n", path, strerror(errno));
-    return NULL;
-  }
-
-#if USE_POSIX_FADVISE && USE_FILENO
-  posix_fadvise(fileno(fd), 0, 0, POSIX_FADV_SEQUENTIAL | POSIX_FADV_NOREUSE);
-#endif
-
-  return fd;
-}
-
-/** fread-like wrapper for getc (which is faster). */
-static size_t
-serd_file_read_byte(void* buf, size_t size, size_t nmemb, void* stream)
-{
-  (void)size;
-  (void)nmemb;
-
-  const int c = getc((FILE*)stream);
-  if (c == EOF) {
-    *((uint8_t*)buf) = 0;
-    return 0;
-  }
-  *((uint8_t*)buf) = (uint8_t)c;
-  return 1;
-}
-
 static SerdWriterFlags
 choose_style(const SerdSyntax input_syntax,
              const SerdSyntax output_syntax,
@@ -204,25 +164,23 @@ main(int argc, char** argv)
     return print_usage(argv[0], true);
   }
 
-  FILE*       in_fd         = NULL;
   SerdSyntax  input_syntax  = (SerdSyntax)0;
   SerdSyntax  output_syntax = (SerdSyntax)0;
-  bool        from_file     = true;
+  bool        from_string   = false;
+  bool        from_stdin    = false;
   bool        ascii         = false;
   bool        bulk_read     = true;
   bool        bulk_write    = false;
   bool        full_uris     = false;
   bool        lax           = false;
   bool        quiet         = false;
-  const char* in_name       = NULL;
   const char* add_prefix    = NULL;
   const char* chop_prefix   = NULL;
   const char* root_uri      = NULL;
   int         a             = 1;
   for (; a < argc && argv[a][0] == '-'; ++a) {
     if (argv[a][1] == '\0') {
-      in_name = (const char*)"(stdin)";
-      in_fd   = stdin;
+      from_stdin = true;
       break;
     }
 
@@ -243,8 +201,7 @@ main(int argc, char** argv)
     } else if (argv[a][1] == 'v') {
       return print_version();
     } else if (argv[a][1] == 's') {
-      in_name   = (const char*)"(string)";
-      from_file = false;
+      from_string = true;
       ++a;
       break;
     } else if (argv[a][1] == 'i') {
@@ -297,22 +254,9 @@ main(int argc, char** argv)
   _setmode(_fileno(stdout), _O_BINARY);
 #endif
 
-  char*       input_path = NULL;
-  const char* input      = (const char*)argv[a++];
-  if (from_file) {
-    in_name = in_name ? in_name : input;
-    if (!in_fd) {
-      if (!strncmp(input, "file:", 5)) {
-        input_path = serd_parse_file_uri(input, NULL);
-        input      = input_path;
-      }
-      if (!input || !(in_fd = serd_fopen(input, "rb"))) {
-        return 1;
-      }
-    }
-  }
+  const char* input = argv[a++];
 
-  if (!input_syntax && !(input_syntax = guess_syntax(in_name))) {
+  if (!input_syntax && !(input_syntax = guess_syntax(input))) {
     input_syntax = SERD_TRIG;
   }
 
@@ -329,7 +273,7 @@ main(int argc, char** argv)
   SerdNode* base = NULL;
   if (a < argc) { // Base URI given on command line
     base = serd_new_uri(SERD_MEASURE_STRING((const char*)argv[a]));
-  } else if (from_file && in_fd != stdin) { // Use input file URI
+  } else if (!from_string && !from_stdin) { // Use input file URI
     base = serd_new_file_uri(SERD_MEASURE_STRING(input), SERD_EMPTY_STRING());
   }
 
@@ -355,36 +299,31 @@ main(int argc, char** argv)
   serd_node_free(root);
 
   SerdStatus st = SERD_SUCCESS;
-  if (!from_file) {
+  if (from_string) {
     st = serd_reader_start_string(reader, input);
-  } else {
+  } else if (from_stdin) {
     st = serd_reader_start_stream(reader,
-                                  bulk_read ? (SerdReadFunc)fread
-                                            : serd_file_read_byte,
+                                  serd_file_read_byte,
                                   (SerdStreamErrorFunc)ferror,
-                                  in_fd,
-                                  in_name,
-                                  bulk_read ? 4096 : 1);
+                                  stdin,
+                                  "(stdin)",
+                                  1);
+  } else {
+    st = serd_reader_start_file(reader, input, bulk_read);
   }
 
   if (!st) {
     st = serd_reader_read_document(reader);
   }
 
-  serd_reader_end_stream(reader);
-
+  serd_reader_finish(reader);
   serd_reader_free(reader);
   serd_writer_finish(writer);
   serd_writer_free(writer);
   serd_env_free(env);
   serd_node_free(base);
-  free(input_path);
-
-  if (from_file) {
-    fclose(in_fd);
-  }
 
-  if (fclose(out_fd)) {
+  if (fclose(stdout)) {
     perror("serdi: write error");
     st = SERD_ERR_UNKNOWN;
   }
diff --git a/src/system.h b/src/system.h
index fdfab4a4..df5827de 100644
--- a/src/system.h
+++ b/src/system.h
@@ -19,6 +19,7 @@
 
 #include "attributes.h"
 
+#include <stdint.h>
 #include <stdio.h>
 
 /// Open a file configured for fast sequential reading
@@ -37,4 +38,20 @@ serd_allocate_buffer(size_t size);
 void
 serd_free_aligned(void* ptr);
 
+/** fread-like wrapper for getc (which is faster). */
+static inline size_t
+serd_file_read_byte(void* buf, size_t size, size_t nmemb, void* stream)
+{
+  (void)size;
+  (void)nmemb;
+
+  const int c = getc((FILE*)stream);
+  if (c == EOF) {
+    *((uint8_t*)buf) = 0;
+    return 0;
+  }
+  *((uint8_t*)buf) = (uint8_t)c;
+  return 1;
+}
+
 #endif // SERD_SYSTEM_H
diff --git a/test/test_read_chunk.c b/test/test_read_chunk.c
index c4e267f3..df8b43c2 100644
--- a/test/test_read_chunk.c
+++ b/test/test_read_chunk.c
@@ -108,7 +108,7 @@ main(void)
   assert(serd_reader_read_chunk(reader) == SERD_FAILURE);
   assert(n_end == 1);
   assert(serd_reader_read_chunk(reader) == SERD_FAILURE);
-  assert(!serd_reader_end_stream(reader));
+  assert(!serd_reader_finish(reader));
 
   serd_reader_free(reader);
   serd_sink_free(sink);
diff --git a/test/test_reader_writer.c b/test/test_reader_writer.c
index 8f80deee..697469a9 100644
--- a/test/test_reader_writer.c
+++ b/test/test_reader_writer.c
@@ -19,6 +19,7 @@
 #include "serd/serd.h"
 
 #include <assert.h>
+#include <stdbool.h>
 #include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -150,13 +151,14 @@ test_read_string(void)
   serd_sink_set_statement_func(sink, test_sink);
 
   // Test reading a string that ends exactly at the end of input (no newline)
-  const SerdStatus st =
-    serd_reader_read_string(reader,
-                            "<http://example.org/s> <http://example.org/p> "
-                            "<http://example.org/o> .");
+  assert(
+    !serd_reader_start_string(reader,
+                              "<http://example.org/s> <http://example.org/p> "
+                              "<http://example.org/o> ."));
 
-  assert(!st);
+  assert(!serd_reader_read_document(reader));
   assert(rt->n_statements == 1);
+  assert(!serd_reader_finish(reader));
 
   serd_reader_free(reader);
   serd_sink_free(sink);
@@ -257,12 +259,12 @@ test_writer(const char* const path)
 static void
 test_reader(const char* path)
 {
-  ReaderTest*     rt     = (ReaderTest*)calloc(1, sizeof(ReaderTest));
-  SerdSink* const sink   = serd_sink_new(rt, NULL);
+  ReaderTest      rt     = {0, NULL};
+  SerdSink* const sink   = serd_sink_new(&rt, NULL);
   SerdReader*     reader = serd_reader_new(SERD_TURTLE, sink);
 
-  assert(sink);
   assert(reader);
+  assert(sink);
   serd_sink_set_statement_func(sink, test_sink);
 
   SerdNode* g = serd_new_uri(SERD_STATIC_STRING("http://example.org/"));
@@ -280,17 +282,16 @@ test_reader(const char* path)
 
   serd_node_free(g);
 
-  assert(serd_reader_read_file(reader, "http://notafile"));
-  assert(serd_reader_read_file(reader, "file:///better/not/exist"));
-  assert(serd_reader_read_file(reader, "file://"));
+  assert(serd_reader_start_file(reader, "http://notafile", false));
+  assert(serd_reader_start_file(reader, "file://invalid", false));
+  assert(serd_reader_start_file(reader, "file:///nonexistant", false));
 
-  const SerdStatus st = serd_reader_read_file(reader, path);
-  assert(!st);
-  assert(rt->n_statements == 6);
-  assert(rt->graph && serd_node_string(rt->graph) &&
-         !strcmp(serd_node_string(rt->graph), "http://example.org/"));
-
-  assert(serd_reader_read_string(reader, "This isn't Turtle at all."));
+  assert(!serd_reader_start_file(reader, path, true));
+  assert(!serd_reader_read_document(reader));
+  assert(rt.n_statements == 6);
+  assert(rt.graph && serd_node_string(rt.graph) &&
+         !strcmp(serd_node_string(rt.graph), "http://example.org/"));
+  serd_reader_finish(reader);
 
   // A read of a big page hits EOF then fails to read chunks immediately
   {
@@ -311,7 +312,7 @@ test_reader(const char* path)
     assert(serd_reader_read_chunk(reader) == SERD_FAILURE);
     assert(serd_reader_read_chunk(reader) == SERD_FAILURE);
 
-    serd_reader_end_stream(reader);
+    serd_reader_finish(reader);
     fclose(temp);
   }
 
@@ -333,7 +334,6 @@ test_reader(const char* path)
 
   serd_reader_free(reader);
   serd_sink_free(sink);
-  free(rt);
 }
 
 int