Clean up and separate internal headers

18 files changed, 767 insertions, 598 deletions

diff --git a/src/byte_sink.h b/src/byte_sink.h
new file mode 100644
index 00000000..dca08825
--- /dev/null
+++ b/src/byte_sink.h
@@ -0,0 +1,93 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_BYTE_SINK_H
+#define SERD_BYTE_SINK_H
+
+#include <stddef.h>
+#include <string.h>
+
+#include "serd/serd.h"
+
+typedef struct SerdByteSinkImpl {
+	SerdSink sink;
+	void*    stream;
+	char*    buf;
+	size_t   size;
+	size_t   block_size;
+} SerdByteSink;
+
+static inline SerdByteSink
+serd_byte_sink_new(SerdSink sink, void* stream, size_t block_size)
+{
+	SerdByteSink bsink;
+	bsink.sink       = sink;
+	bsink.stream     = stream;
+	bsink.size       = 0;
+	bsink.block_size = block_size;
+	bsink.buf        = ((block_size > 1)
+	                    ? (char*)serd_bufalloc(block_size)
+	                    : NULL);
+	return bsink;
+}
+
+static inline void
+serd_byte_sink_flush(SerdByteSink* bsink)
+{
+	if (bsink->block_size > 1 && bsink->size > 0) {
+		bsink->sink(bsink->buf, bsink->size, bsink->stream);
+		bsink->size = 0;
+	}
+}
+
+static inline void
+serd_byte_sink_free(SerdByteSink* bsink)
+{
+	serd_byte_sink_flush(bsink);
+	free(bsink->buf);
+	bsink->buf = NULL;
+}
+
+static inline size_t
+serd_byte_sink_write(const void* buf, size_t len, SerdByteSink* bsink)
+{
+	if (len == 0) {
+		return 0;
+	} else if (bsink->block_size == 1) {
+		return bsink->sink(buf, len, bsink->stream);
+	}
+
+	const size_t orig_len = len;
+	while (len) {
+		const size_t space = bsink->block_size - bsink->size;
+		const size_t n     = MIN(space, len);
+
+		// Write as much as possible into the remaining buffer space
+		memcpy(bsink->buf + bsink->size, buf, n);
+		bsink->size += n;
+		buf          = (const char*)buf + n;
+		len         -= n;
+
+		// Flush page if buffer is full
+		if (bsink->size == bsink->block_size) {
+			bsink->sink(bsink->buf, bsink->block_size, bsink->stream);
+			bsink->size = 0;
+		}
+	}
+	return orig_len;
+}
+
+#endif  // SERD_BYTE_SINK_H
diff --git a/src/byte_source.c b/src/byte_source.c
index a2bbc4cf..7cd613a0 100644
--- a/src/byte_source.c
+++ b/src/byte_source.c
@@ -14,6 +14,8 @@
   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */
 
+#include "byte_source.h"
+
 #include "serd_internal.h"
 
 SerdStatus
diff --git a/src/byte_source.h b/src/byte_source.h
new file mode 100644
index 00000000..0ef8f9a6
--- /dev/null
+++ b/src/byte_source.h
@@ -0,0 +1,109 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_BYTE_SOURCE_H
+#define SERD_BYTE_SOURCE_H
+
+#include "serd_internal.h"
+
+#include <assert.h>
+#include <stddef.h>
+#include <stdio.h>
+
+typedef struct {
+	const char* filename;
+	unsigned    line;
+	unsigned    col;
+} Cursor;
+
+typedef struct {
+	SerdSource          read_func;    ///< Read function (e.g. fread)
+	SerdStreamErrorFunc error_func;   ///< Error function (e.g. ferror)
+	void*               stream;       ///< Stream (e.g. FILE)
+	size_t              page_size;    ///< Number of bytes to read at a time
+	Cursor              cur;          ///< Cursor for error reporting
+	uint8_t*            file_buf;     ///< Buffer iff reading pages from a file
+	const uint8_t*      read_buf;     ///< Pointer to file_buf or read_byte
+	size_t              read_head;    ///< Offset into read_buf
+	uint8_t             read_byte;    ///< 1-byte 'buffer' used when not paging
+	bool                from_stream;  ///< True iff reading from `stream`
+	bool                prepared;     ///< True iff prepared for reading
+	bool                eof;          ///< True iff end of file reached
+} 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,
+                             SerdSource          read_func,
+                             SerdStreamErrorFunc error_func,
+                             void*               stream,
+                             const char*         name,
+                             size_t              page_size);
+
+SerdStatus
+serd_byte_source_close(SerdByteSource* source);
+
+SerdStatus
+serd_byte_source_prepare(SerdByteSource* source);
+
+SerdStatus
+serd_byte_source_page(SerdByteSource* source);
+
+static inline uint8_t
+serd_byte_source_peek(SerdByteSource* source)
+{
+	assert(source->prepared);
+	return source->read_buf[source->read_head];
+}
+
+static inline SerdStatus
+serd_byte_source_advance(SerdByteSource* source)
+{
+	SerdStatus st = SERD_SUCCESS;
+
+	switch (serd_byte_source_peek(source)) {
+	case '\0': break;
+	case '\n': ++source->cur.line; source->cur.col = 0; break;
+	default:   ++source->cur.col;
+	}
+
+	if (source->from_stream) {
+		source->eof = false;
+		if (source->page_size > 1) {
+			if (++source->read_head == source->page_size) {
+				st = serd_byte_source_page(source);
+			}
+		} else {
+			if (!source->read_func(&source->read_byte, 1, 1, source->stream)) {
+				st = source->error_func(source->stream) ? SERD_ERR_UNKNOWN
+				                                        : SERD_FAILURE;
+			}
+		}
+	} else if (!source->eof) {
+		++source->read_head; // Move to next character in string
+	}
+
+	return source->eof ? SERD_FAILURE : st;
+}
+
+#endif  // SERD_BYTE_SOURCE_H
diff --git a/src/env.c b/src/env.c
index f9bfe13d..a05ad1a0 100644
--- a/src/env.c
+++ b/src/env.c
@@ -19,6 +19,8 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "node.h"
+
 typedef struct {
 	SerdNode* name;
 	SerdNode* uri;
diff --git a/src/n3.c b/src/n3.c
index a80285db..19dd7e2f 100644
--- a/src/n3.c
+++ b/src/n3.c
@@ -25,6 +25,8 @@
 #include <string.h>
 
 #include "reader.h"
+#include "string_utils.h"
+#include "uri_utils.h"
 
 #define TRY_THROW(exp) if (!(exp)) goto except;
 #define TRY_RET(exp)   if (!(exp)) return 0;
@@ -590,7 +592,7 @@ static bool
 read_IRIREF_scheme(SerdReader* reader, Ref dest)
 {
 	uint8_t c = peek_byte(reader);
-	if (!isalpha(c)) {
+	if (!is_alpha(c)) {
 		return r_err(reader, SERD_ERR_BAD_SYNTAX,
 		             "bad IRI scheme start `%c'\n", c);
 	}
diff --git a/src/node.c b/src/node.c
index 0f9ee8ff..8a490a45 100644
--- a/src/node.c
+++ b/src/node.c
@@ -16,11 +16,15 @@
 
 #include "serd_internal.h"
 
+#include <assert.h>
 #include <float.h>
 #include <math.h>
 #include <stdlib.h>
 #include <string.h>
 
+#include "node.h"
+#include "string_utils.h"
+
 #ifdef _WIN32
 #    ifndef isnan
 #        define isnan(x) _isnan(x)
@@ -63,12 +67,6 @@ serd_node_malloc(size_t n_bytes, SerdNodeFlags flags, SerdType type)
 	return node;
 }
 
-char*
-serd_node_buffer(SerdNode* node)
-{
-	return (char*)(node + 1);
-}
-
 void
 serd_node_set(SerdNode** dst, const SerdNode* src)
 {
diff --git a/src/node.h b/src/node.h
new file mode 100644
index 00000000..090faf0c
--- /dev/null
+++ b/src/node.h
@@ -0,0 +1,43 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_NODE_H
+#define SERD_NODE_H
+
+#include "serd_internal.h"
+
+struct SerdNodeImpl {
+	size_t        n_bytes;  /**< Size in bytes (not including null) */
+	SerdNodeFlags flags;    /**< Node flags (e.g. string properties) */
+	SerdType      type;     /**< Node type */
+};
+
+static inline char*
+serd_node_buffer(SerdNode* node)
+{
+	return (char*)(node + 1);
+}
+
+static inline const char*
+serd_node_buffer_c(const SerdNode* node)
+{
+	return (const char*)(node + 1);
+}
+
+SerdNode* serd_node_malloc(size_t n_bytes, SerdNodeFlags flags, SerdType type);
+void      serd_node_set(SerdNode** dst, const SerdNode* src);
+
+#endif  // SERD_NODE_H
diff --git a/src/reader.c b/src/reader.c
index ec1c533d..7d831bbe 100644
--- a/src/reader.c
+++ b/src/reader.c
@@ -24,6 +24,8 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "reader.h"
+
 static SerdStatus serd_reader_prepare(SerdReader* reader);
 
 int
diff --git a/src/reader.h b/src/reader.h
index f3ddde0a..62e6a8c8 100644
--- a/src/reader.h
+++ b/src/reader.h
@@ -14,8 +14,93 @@
   OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */
 
+#ifndef SERD_READER_H
+#define SERD_READER_H
+
 #include "serd_internal.h"
 
+#include "byte_source.h"
+#include "node.h"
+#include "stack.h"
+
+#ifdef SERD_STACK_CHECK
+#    define SERD_STACK_ASSERT_TOP(reader, ref) \
+            assert(ref == reader->allocs[reader->n_allocs - 1]);
+#else
+#    define SERD_STACK_ASSERT_TOP(reader, ref)
+#endif
+
+/* Reference to a node in the stack (we can not use pointers since the
+   stack may be reallocated, invalidating any pointers to elements).
+*/
+typedef size_t Ref;
+
+typedef struct {
+	Ref                 graph;
+	Ref                 subject;
+	Ref                 predicate;
+	Ref                 object;
+	Ref                 datatype;
+	Ref                 lang;
+	SerdStatementFlags* flags;
+} ReadContext;
+
+struct SerdReaderImpl {
+	void*             handle;
+	void              (*free_handle)(void* ptr);
+	SerdBaseSink      base_sink;
+	SerdPrefixSink    prefix_sink;
+	SerdStatementSink statement_sink;
+	SerdEndSink       end_sink;
+	SerdErrorSink     error_sink;
+	void*             error_handle;
+	Ref               rdf_first;
+	Ref               rdf_rest;
+	Ref               rdf_nil;
+	SerdNode*         default_graph;
+	SerdByteSource    source;
+	SerdStack         stack;
+	SerdSyntax        syntax;
+	unsigned          next_id;
+	SerdStatus        status;
+	uint8_t*          buf;
+	char*             bprefix;
+	size_t            bprefix_len;
+	bool              strict;     ///< True iff strict parsing
+	bool              seen_genid;
+#ifdef SERD_STACK_CHECK
+	Ref*              allocs;     ///< Stack of push offsets
+	size_t            n_allocs;   ///< Number of stack pushes
+#endif
+};
+
+int r_err(SerdReader* reader, SerdStatus st, const char* fmt, ...);
+
+Ref push_node_padded(SerdReader* reader,
+                     size_t      maxlen,
+                     SerdType    type,
+                     const char* str,
+                     size_t      n_bytes);
+
+Ref push_node(SerdReader* reader,
+              SerdType    type,
+              const char* str,
+              size_t      n_bytes);
+
+size_t genid_size(SerdReader* reader);
+Ref    blank_id(SerdReader* reader);
+void   set_blank_id(SerdReader* reader, Ref ref, size_t buf_size);
+
+SerdNode* deref(SerdReader* reader, Ref ref);
+
+Ref pop_node(SerdReader* reader, Ref ref);
+
+bool emit_statement(SerdReader* reader, ReadContext ctx, Ref o);
+
+bool read_n3_statement(SerdReader* reader);
+SerdStatus read_nquadsDoc(SerdReader* reader);
+SerdStatus read_turtleTrigDoc(SerdReader* reader);
+
 static inline uint8_t
 peek_byte(SerdReader* reader)
 {
@@ -83,3 +168,5 @@ push_bytes(SerdReader* reader, Ref ref, const uint8_t* bytes, unsigned len)
 		push_byte(reader, ref, bytes[i]);
 	}
 }
+
+#endif // SERD_READER_H
diff --git a/src/serd_internal.h b/src/serd_internal.h
index a9e9c0d7..fbf493e9 100644
--- a/src/serd_internal.h
+++ b/src/serd_internal.h
@@ -19,8 +19,6 @@
 
 #define _POSIX_C_SOURCE 200809L /* for posix_memalign and posix_fadvise */
 
-#include <assert.h>
-#include <ctype.h>
 #include <errno.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -42,22 +40,6 @@
 #    define MIN(a, b) (((a) < (b)) ? (a) : (b))
 #endif
 
-static const uint8_t replacement_char[] = { 0xEF, 0xBF, 0xBD };
-
-struct SerdNodeImpl {
-	size_t        n_bytes;  /**< Size in bytes (not including null) */
-	SerdNodeFlags flags;    /**< Node flags (e.g. string properties) */
-	SerdType      type;     /**< Node type */
-};
-
-static const SerdNode SERD_NODE_NULL = { 0, 0, SERD_NOTHING };
-
-SerdNode* serd_node_malloc(size_t n_bytes, SerdNodeFlags flags, SerdType type);
-void      serd_node_set(SerdNode** dst, const SerdNode* src);
-char*     serd_node_buffer(SerdNode* node);
-
-/* File and Buffer Utilities */
-
 static inline FILE*
 serd_fopen(const char* path, const char* mode)
 {
@@ -85,452 +67,6 @@ serd_bufalloc(size_t size)
 #endif
 }
 
-/* Byte source */
-
-typedef struct {
-	const char* filename;
-	unsigned    line;
-	unsigned    col;
-} Cursor;
-
-typedef struct {
-	SerdSource          read_func;    ///< Read function (e.g. fread)
-	SerdStreamErrorFunc error_func;   ///< Error function (e.g. ferror)
-	void*               stream;       ///< Stream (e.g. FILE)
-	size_t              page_size;    ///< Number of bytes to read at a time
-	Cursor              cur;          ///< Cursor for error reporting
-	uint8_t*            file_buf;     ///< Buffer iff reading pages from a file
-	const uint8_t*      read_buf;     ///< Pointer to file_buf or read_byte
-	size_t              read_head;    ///< Offset into read_buf
-	uint8_t             read_byte;    ///< 1-byte 'buffer' used when not paging
-	bool                from_stream;  ///< True iff reading from `stream`
-	bool                prepared;     ///< True iff prepared for reading
-	bool                eof;          ///< True iff end of file reached
-} 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,
-                             SerdSource          read_func,
-                             SerdStreamErrorFunc error_func,
-                             void*               stream,
-                             const char*         name,
-                             size_t              page_size);
-
-SerdStatus
-serd_byte_source_close(SerdByteSource* source);
-
-SerdStatus
-serd_byte_source_prepare(SerdByteSource* source);
-
-SerdStatus
-serd_byte_source_page(SerdByteSource* source);
-
-static inline uint8_t
-serd_byte_source_peek(SerdByteSource* source)
-{
-	assert(source->prepared);
-	return source->read_buf[source->read_head];
-}
-
-static inline SerdStatus
-serd_byte_source_advance(SerdByteSource* source)
-{
-	SerdStatus st = SERD_SUCCESS;
-
-	switch (serd_byte_source_peek(source)) {
-	case '\0': break;
-	case '\n': ++source->cur.line; source->cur.col = 0; break;
-	default:   ++source->cur.col;
-	}
-
-	if (source->from_stream) {
-		source->eof = false;
-		if (source->page_size > 1) {
-			if (++source->read_head == source->page_size) {
-				st = serd_byte_source_page(source);
-			}
-		} else {
-			if (!source->read_func(&source->read_byte, 1, 1, source->stream)) {
-				st = source->error_func(source->stream) ? SERD_ERR_UNKNOWN
-				                                        : SERD_FAILURE;
-			}
-		}
-	} else if (!source->eof) {
-		++source->read_head; // Move to next character in string
-	}
-
-	return source->eof ? SERD_FAILURE : st;
-}
-
-/* Stack */
-
-/** A dynamic stack in memory. */
-typedef struct {
-	char*    buf;       ///< Stack memory
-	size_t   buf_size;  ///< Allocated size of buf (>= size)
-	size_t   size;      ///< Conceptual size of stack in buf
-} SerdStack;
-
-/** An offset to start the stack at. Note 0 is reserved for NULL. */
-#define SERD_STACK_BOTTOM sizeof(void*)
-
-static inline SerdStack
-serd_stack_new(size_t size)
-{
-	SerdStack stack;
-	stack.buf       = (char*)calloc(size, 1);
-	stack.buf_size  = size;
-	stack.size      = SERD_STACK_BOTTOM;
-	return stack;
-}
-
-static inline bool
-serd_stack_is_empty(SerdStack* stack)
-{
-	return stack->size <= SERD_STACK_BOTTOM;
-}
-
-static inline void
-serd_stack_free(SerdStack* stack)
-{
-	free(stack->buf);
-	stack->buf      = NULL;
-	stack->buf_size = 0;
-	stack->size     = 0;
-}
-
-static inline void*
-serd_stack_push(SerdStack* stack, size_t n_bytes)
-{
-	const size_t new_size = stack->size + n_bytes;
-	if (stack->buf_size < new_size) {
-		stack->buf_size += (stack->buf_size >> 1); // *= 1.5
-		stack->buf = (char*)realloc(stack->buf, stack->buf_size);
-	}
-	char* const ret = (stack->buf + stack->size);
-	stack->size = new_size;
-	return ret;
-}
-
-static inline void
-serd_stack_pop(SerdStack* stack, size_t n_bytes)
-{
-	assert(stack->size >= n_bytes);
-	stack->size -= n_bytes;
-}
-
-static inline void*
-serd_stack_push_aligned(SerdStack* stack, size_t n_bytes, size_t align)
-{
-	// Push one byte to ensure space for a pad count
-	serd_stack_push(stack, 1);
-
-	// Push padding if necessary
-	const uint8_t pad = align - stack->size % align;
-	if (pad > 0) {
-		serd_stack_push(stack, pad);
-	}
-
-	// Set top of stack to pad count so we can properly pop later
-	stack->buf[stack->size - 1] = (char)pad;
-
-	// Push requested space at aligned location
-	return serd_stack_push(stack, n_bytes);
-}
-
-static inline void
-serd_stack_pop_aligned(SerdStack* stack, size_t n_bytes)
-{
-	// Pop requested space down to aligned location
-	serd_stack_pop(stack, n_bytes);
-
-	// Get amount of padding from top of stack
-	const uint8_t pad = (uint8_t)stack->buf[stack->size - 1];
-
-	// Pop padding and pad count
-	serd_stack_pop(stack, pad + 1);
-}
-
-/* Byte Sink */
-
-typedef struct SerdByteSinkImpl {
-	SerdSink sink;
-	void*    stream;
-	char*    buf;
-	size_t   size;
-	size_t   block_size;
-} SerdByteSink;
-
-static inline SerdByteSink
-serd_byte_sink_new(SerdSink sink, void* stream, size_t block_size)
-{
-	SerdByteSink bsink;
-	bsink.sink       = sink;
-	bsink.stream     = stream;
-	bsink.size       = 0;
-	bsink.block_size = block_size;
-	bsink.buf        = ((block_size > 1)
-	                    ? (char*)serd_bufalloc(block_size)
-	                    : NULL);
-	return bsink;
-}
-
-static inline void
-serd_byte_sink_flush(SerdByteSink* bsink)
-{
-	if (bsink->block_size > 1 && bsink->size > 0) {
-		bsink->sink(bsink->buf, bsink->size, bsink->stream);
-		bsink->size = 0;
-	}
-}
-
-static inline void
-serd_byte_sink_free(SerdByteSink* bsink)
-{
-	serd_byte_sink_flush(bsink);
-	free(bsink->buf);
-	bsink->buf = NULL;
-}
-
-static inline size_t
-serd_byte_sink_write(const void* buf, size_t len, SerdByteSink* bsink)
-{
-	if (len == 0) {
-		return 0;
-	} else if (bsink->block_size == 1) {
-		return bsink->sink(buf, len, bsink->stream);
-	}
-
-	const size_t orig_len = len;
-	while (len) {
-		const size_t space = bsink->block_size - bsink->size;
-		const size_t n     = MIN(space, len);
-
-		// Write as much as possible into the remaining buffer space
-		memcpy(bsink->buf + bsink->size, buf, n);
-		bsink->size += n;
-		buf          = (const char*)buf + n;
-		len         -= n;
-
-		// Flush page if buffer is full
-		if (bsink->size == bsink->block_size) {
-			bsink->sink(bsink->buf, bsink->block_size, bsink->stream);
-			bsink->size = 0;
-		}
-	}
-	return orig_len;
-}
-
-/* Character utilities */
-
-/** Return true if `c` lies within [`min`...`max`] (inclusive) */
-static inline bool
-in_range(const char c, const char min, const char max)
-{
-	return (c >= min && c <= max);
-}
-
-/** RFC2234: ALPHA ::= %x41-5A / %x61-7A  ; A-Z / a-z */
-static inline bool
-is_alpha(const char c)
-{
-	return in_range(c, 'A', 'Z') || in_range(c, 'a', 'z');
-}
-
-/** RFC2234: DIGIT ::= %x30-39  ; 0-9 */
-static inline bool
-is_digit(const char c)
-{
-	return in_range(c, '0', '9');
-}
-
-/* RFC2234: HEXDIG ::= DIGIT / "A" / "B" / "C" / "D" / "E" / "F" */
-static inline bool
-is_hexdig(const char c)
-{
-	return is_digit(c) || in_range(c, 'A', 'F');
-}
-
-/* Turtle / JSON / C: XDIGIT ::= DIGIT / A-F / a-f */
-static inline bool
-is_xdigit(const char c)
-{
-	return is_hexdig(c) || in_range(c, 'a', 'f');
-}
-
-static inline bool
-is_space(const char c)
-{
-	switch (c) {
-	case ' ': case '\f': case '\n': case '\r': case '\t': case '\v':
-		return true;
-	default:
-		return false;
-	}
-}
-
-static inline bool
-is_base64(const char c)
-{
-	return is_alpha(c) || is_digit(c) || c == '+' || c == '/' || c == '=';
-}
-
-static inline bool
-is_windows_path(const char* path)
-{
-	return is_alpha(path[0]) && (path[1] == ':' || path[1] == '|')
-		&& (path[2] == '/' || path[2] == '\\');
-}
-
-/* String utilities */
-
-size_t
-serd_substrlen(const char* str, size_t len, SerdNodeFlags* flags);
-
-static inline int
-serd_strncasecmp(const char* s1, const char* s2, size_t n)
-{
-	for (; n > 0 && *s2; s1++, s2++, --n) {
-		if (toupper(*s1) != toupper(*s2)) {
-			return ((*(const uint8_t*)s1 < *(const uint8_t*)s2) ? -1 : +1);
-		}
-	}
-	return 0;
-}
-
-static inline uint32_t
-utf8_num_bytes(const uint8_t c)
-{
-	if ((c & 0x80) == 0) {  // Starts with `0'
-		return 1;
-	} else if ((c & 0xE0) == 0xC0) {  // Starts with `110'
-		return 2;
-	} else if ((c & 0xF0) == 0xE0) {  // Starts with `1110'
-		return 3;
-	} else if ((c & 0xF8) == 0xF0) {  // Starts with `11110'
-		return 4;
-	}
-	return 0;
-}
-
-/// Return the code point of a UTF-8 character with known length
-static inline uint32_t
-parse_counted_utf8_char(const uint8_t* utf8, size_t size)
-{
-	uint32_t c = utf8[0] & ((1 << (8 - size)) - 1);
-	for (size_t i = 1; i < size; ++i) {
-		const uint8_t in = utf8[i] & 0x3F;
-		c = (c << 6) | in;
-	}
-	return c;
-}
-
-/// Parse a UTF-8 character, set *size to the length, and return the code point
-static inline uint32_t
-parse_utf8_char(const uint8_t* utf8, size_t* size)
-{
-	switch (*size = utf8_num_bytes(utf8[0])) {
-	case 1: case 2: case 3: case 4:
-		return parse_counted_utf8_char(utf8, *size);
-	default:
-		return *size = 0;
-	}
-}
-
-/* URI utilities */
-
-static inline bool
-slice_equals(const SerdStringView* a, const SerdStringView* b)
-{
-	return a->len == b->len
-		&& !strncmp((const char*)a->buf, (const char*)b->buf, a->len);
-}
-
-static inline size_t
-uri_path_len(const SerdURI* uri)
-{
-	return uri->path_base.len + uri->path.len;
-}
-
-static inline char
-uri_path_at(const SerdURI* uri, size_t i)
-{
-	if (i < uri->path_base.len) {
-		return uri->path_base.buf[i];
-	} else {
-		return uri->path.buf[i - uri->path_base.len];
-	}
-}
-
-/**
-   Return the index of the first differing character after the last root slash,
-   or zero if `uri` is not under `root`.
-*/
-static inline size_t
-uri_rooted_index(const SerdURI* uri, const SerdURI* root)
-{
-	if (!root || !root->scheme.len ||
-	    !slice_equals(&root->scheme, &uri->scheme) ||
-	    !slice_equals(&root->authority, &uri->authority)) {
-		return 0;
-	}
-
-	bool         differ          = false;
-	const size_t path_len        = uri_path_len(uri);
-	const size_t root_len        = uri_path_len(root);
-	size_t       last_root_slash = 0;
-	for (size_t i = 0; i < path_len && i < root_len; ++i) {
-		const char u = uri_path_at(uri, i);
-		const char r = uri_path_at(root, i);
-
-		differ = differ || u != r;
-		if (r == '/') {
-			last_root_slash = i;
-			if (differ) {
-				return 0;
-			}
-		}
-	}
-
-	return last_root_slash + 1;
-}
-
-/** Return true iff `uri` shares path components with `root` */
-static inline bool
-uri_is_related(const SerdURI* uri, const SerdURI* root)
-{
-	return uri_rooted_index(uri, root) > 0;
-}
-
-/** Return true iff `uri` is within the base of `root` */
-static inline bool
-uri_is_under(const SerdURI* uri, const SerdURI* root)
-{
-	const size_t index = uri_rooted_index(uri, root);
-	return index > 0 && uri->path.len > index;
-}
-
-static inline bool
-is_uri_scheme_char(const char c)
-{
-	switch (c) {
-	case ':': case '+': case '-': case '.':
-		return true;
-	default:
-		return is_alpha(c) || is_digit(c);
-	}
-}
-
-/* Error reporting */
-
 static inline void
 serd_error(SerdErrorSink error_sink, void* handle, const SerdError* e)
 {
@@ -542,93 +78,4 @@ serd_error(SerdErrorSink error_sink, void* handle, const SerdError* e)
 	}
 }
 
-int
-r_err(SerdReader* reader, SerdStatus st, const char* fmt, ...);
-
-/* Reader */
-
-#ifdef SERD_STACK_CHECK
-#    define SERD_STACK_ASSERT_TOP(reader, ref) \
-            assert(ref == reader->allocs[reader->n_allocs - 1]);
-#else
-#    define SERD_STACK_ASSERT_TOP(reader, ref)
-#endif
-
-/* Reference to a node in the stack (we can not use pointers since the
-   stack may be reallocated, invalidating any pointers to elements).
-*/
-typedef size_t Ref;
-
-typedef struct {
-	Ref                 graph;
-	Ref                 subject;
-	Ref                 predicate;
-	Ref                 object;
-	Ref                 datatype;
-	Ref                 lang;
-	SerdStatementFlags* flags;
-} ReadContext;
-
-struct SerdReaderImpl {
-	void*             handle;
-	void              (*free_handle)(void* ptr);
-	SerdBaseSink      base_sink;
-	SerdPrefixSink    prefix_sink;
-	SerdStatementSink statement_sink;
-	SerdEndSink       end_sink;
-	SerdErrorSink     error_sink;
-	void*             error_handle;
-	Ref               rdf_first;
-	Ref               rdf_rest;
-	Ref               rdf_nil;
-	SerdNode*         default_graph;
-	SerdByteSource    source;
-	SerdStack         stack;
-	SerdSyntax        syntax;
-	unsigned          next_id;
-	SerdStatus        status;
-	uint8_t*          buf;
-	char*             bprefix;
-	size_t            bprefix_len;
-	bool              strict;     ///< True iff strict parsing
-	bool              seen_genid;
-#ifdef SERD_STACK_CHECK
-	Ref*              allocs;     ///< Stack of push offsets
-	size_t            n_allocs;   ///< Number of stack pushes
-#endif
-};
-
-Ref push_node_padded(SerdReader* reader,
-                     size_t      maxlen,
-                     SerdType    type,
-                     const char* str,
-                     size_t      n_bytes);
-
-Ref push_node(SerdReader* reader,
-              SerdType    type,
-              const char* str,
-              size_t      n_bytes);
-
-size_t genid_size(SerdReader* reader);
-Ref    blank_id(SerdReader* reader);
-void   set_blank_id(SerdReader* reader, Ref ref, size_t buf_size);
-
-SerdNode* deref(SerdReader* reader, Ref ref);
-
-Ref pop_node(SerdReader* reader, Ref ref);
-
-bool emit_statement(SerdReader* reader, ReadContext ctx, Ref o);
-
-bool read_n3_statement(SerdReader* reader);
-SerdStatus read_nquadsDoc(SerdReader* reader);
-SerdStatus read_turtleTrigDoc(SerdReader* reader);
-
-typedef enum {
-	FIELD_NONE,
-	FIELD_SUBJECT,
-	FIELD_PREDICATE,
-	FIELD_OBJECT,
-	FIELD_GRAPH
-} Field;
-
 #endif  // SERD_INTERNAL_H
diff --git a/src/serdi.c b/src/serdi.c
index 98ef99c8..692ec56c 100644
--- a/src/serdi.c
+++ b/src/serdi.c
@@ -26,6 +26,8 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "string_utils.h"
+
 #define SERDI_ERROR(msg)       fprintf(stderr, "serdi: " msg);
 #define SERDI_ERRORF(fmt, ...) fprintf(stderr, "serdi: " fmt, __VA_ARGS__);
 
diff --git a/src/stack.h b/src/stack.h
new file mode 100644
index 00000000..a8298c0d
--- /dev/null
+++ b/src/stack.h
@@ -0,0 +1,113 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_STACK_H
+#define SERD_STACK_H
+
+#include "serd_internal.h"
+
+#include <assert.h>
+#include <stddef.h>
+#include <stdlib.h>
+
+/** An offset to start the stack at. Note 0 is reserved for NULL. */
+#define SERD_STACK_BOTTOM sizeof(void*)
+
+/** A dynamic stack in memory. */
+typedef struct {
+	char*    buf;       ///< Stack memory
+	size_t   buf_size;  ///< Allocated size of buf (>= size)
+	size_t   size;      ///< Conceptual size of stack in buf
+} SerdStack;
+
+static inline SerdStack
+serd_stack_new(size_t size)
+{
+	SerdStack stack;
+	stack.buf       = (char*)calloc(1, size);
+	stack.buf_size  = size;
+	stack.size      = SERD_STACK_BOTTOM;
+	return stack;
+}
+
+static inline bool
+serd_stack_is_empty(SerdStack* stack)
+{
+	return stack->size <= SERD_STACK_BOTTOM;
+}
+
+static inline void
+serd_stack_free(SerdStack* stack)
+{
+	free(stack->buf);
+	stack->buf      = NULL;
+	stack->buf_size = 0;
+	stack->size     = 0;
+}
+
+static inline void*
+serd_stack_push(SerdStack* stack, size_t n_bytes)
+{
+	const size_t new_size = stack->size + n_bytes;
+	if (stack->buf_size < new_size) {
+		stack->buf_size += (stack->buf_size >> 1); // *= 1.5
+		stack->buf = (char*)realloc(stack->buf, stack->buf_size);
+	}
+	char* const ret = (stack->buf + stack->size);
+	stack->size = new_size;
+	return ret;
+}
+
+static inline void
+serd_stack_pop(SerdStack* stack, size_t n_bytes)
+{
+	assert(stack->size >= n_bytes);
+	stack->size -= n_bytes;
+}
+
+static inline void*
+serd_stack_push_aligned(SerdStack* stack, size_t n_bytes, size_t align)
+{
+	// Push one byte to ensure space for a pad count
+	serd_stack_push(stack, 1);
+
+	// Push padding if necessary
+	const uint8_t pad = align - stack->size % align;
+	if (pad > 0) {
+		serd_stack_push(stack, pad);
+	}
+
+	// Set top of stack to pad count so we can properly pop later
+	stack->buf[stack->size - 1] = pad;
+
+	// Push requested space at aligned location
+	return serd_stack_push(stack, n_bytes);
+}
+
+static inline void
+serd_stack_pop_aligned(SerdStack* stack, size_t n_bytes)
+{
+	// Pop requested space down to aligned location
+	serd_stack_pop(stack, n_bytes);
+
+	// Get amount of padding from top of stack
+	const uint8_t pad = stack->buf[stack->size - 1];
+
+	// Pop padding and pad count
+	serd_stack_pop(stack, pad + 1);
+}
+
+#endif  // SERD_STACK_H
diff --git a/src/string.c b/src/string.c
index 01ed25a0..6e0edcd8 100644
--- a/src/string.c
+++ b/src/string.c
@@ -15,6 +15,7 @@
 */
 
 #include "serd_internal.h"
+#include "string_utils.h"
 
 #include <math.h>
 
@@ -41,34 +42,6 @@ serd_strerror(SerdStatus status)
 	return "Unknown error";  // never reached
 }
 
-static inline void
-serd_update_flags(const char c, SerdNodeFlags* const flags)
-{
-	switch (c) {
-	case '\r': case '\n':
-		*flags |= SERD_HAS_NEWLINE;
-		break;
-	case '"':
-		*flags |= SERD_HAS_QUOTE;
-	}
-}
-
-size_t
-serd_substrlen(const char* const    str,
-               const size_t         len,
-               SerdNodeFlags* const flags)
-{
-	if (flags) {
-		size_t i = 0;
-		*flags = 0;
-		for (; i < len && str[i]; ++i) {
-			serd_update_flags(str[i], flags);
-		}
-		return i;
-	}
-	return strlen(str);
-}
-
 size_t
 serd_strlen(const char* str, SerdNodeFlags* flags)
 {
diff --git a/src/string_utils.h b/src/string_utils.h
new file mode 100644
index 00000000..f4620859
--- /dev/null
+++ b/src/string_utils.h
@@ -0,0 +1,165 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_STRING_UTILS_H
+#define SERD_STRING_UTILS_H
+
+#include <ctype.h>
+
+/** Unicode replacement character in UTF-8 */
+static const uint8_t replacement_char[] = { 0xEF, 0xBF, 0xBD };
+
+/** Return true if `c` lies within [`min`...`max`] (inclusive) */
+static inline bool
+in_range(const char c, const char min, const char max)
+{
+	return (c >= min && c <= max);
+}
+
+/** RFC2234: ALPHA ::= %x41-5A / %x61-7A  ; A-Z / a-z */
+static inline bool
+is_alpha(const char c)
+{
+	return in_range(c, 'A', 'Z') || in_range(c, 'a', 'z');
+}
+
+/** RFC2234: DIGIT ::= %x30-39  ; 0-9 */
+static inline bool
+is_digit(const char c)
+{
+	return in_range(c, '0', '9');
+}
+
+/** RFC2234: HEXDIG ::= DIGIT / "A" / "B" / "C" / "D" / "E" / "F" */
+static inline bool
+is_hexdig(const char c)
+{
+	return is_digit(c) || in_range(c, 'A', 'F');
+}
+
+/** Turtle / JSON / C: XDIGIT ::= DIGIT / A-F / a-f */
+static inline bool
+is_xdigit(const char c)
+{
+	return is_hexdig(c) || in_range(c, 'a', 'f');
+}
+
+/** Return true iff `c` is ASCII whitespace. */
+static inline bool
+is_space(const char c)
+{
+	switch (c) {
+	case ' ': case '\f': case '\n': case '\r': case '\t': case '\v':
+		return true;
+	default:
+		return false;
+	}
+}
+
+/** Return true iff `c` is a valid encoded base64 character. */
+static inline bool
+is_base64(const char c)
+{
+	return is_alpha(c) || is_digit(c) || c == '+' || c == '/' || c == '=';
+}
+
+/** Return true iff `path` looks like a Windows path with a drive letter. */
+static inline bool
+is_windows_path(const char* path)
+{
+	return is_alpha(path[0]) && (path[1] == ':' || path[1] == '|')
+		&& (path[2] == '/' || path[2] == '\\');
+}
+
+static inline void
+serd_update_flags(const char c, SerdNodeFlags* const flags)
+{
+	switch (c) {
+	case '\r': case '\n':
+		*flags |= SERD_HAS_NEWLINE;
+		break;
+	case '"':
+		*flags |= SERD_HAS_QUOTE;
+	}
+}
+
+static inline size_t
+serd_substrlen(const char* const    str,
+               const size_t         len,
+               SerdNodeFlags* const flags)
+{
+	if (flags) {
+		size_t i = 0;
+		*flags = 0;
+		for (; i < len && str[i]; ++i) {
+			serd_update_flags(str[i], flags);
+		}
+		return i;
+	}
+	return strlen(str);
+}
+
+static inline int
+serd_strncasecmp(const char* s1, const char* s2, size_t n)
+{
+	for (; n > 0 && *s2; s1++, s2++, --n) {
+		if (toupper(*s1) != toupper(*s2)) {
+			return ((*(const uint8_t*)s1 < *(const uint8_t*)s2) ? -1 : +1);
+		}
+	}
+	return 0;
+}
+
+static inline uint32_t
+utf8_num_bytes(const uint8_t c)
+{
+	if ((c & 0x80) == 0) {  // Starts with `0'
+		return 1;
+	} else if ((c & 0xE0) == 0xC0) {  // Starts with `110'
+		return 2;
+	} else if ((c & 0xF0) == 0xE0) {  // Starts with `1110'
+		return 3;
+	} else if ((c & 0xF8) == 0xF0) {  // Starts with `11110'
+		return 4;
+	}
+	return 0;
+}
+
+/// Return the code point of a UTF-8 character with known length
+static inline uint32_t
+parse_counted_utf8_char(const uint8_t* utf8, size_t size)
+{
+	uint32_t c = utf8[0] & ((1 << (8 - size)) - 1);
+	for (size_t i = 1; i < size; ++i) {
+		const uint8_t in = utf8[i] & 0x3F;
+		c = (c << 6) | in;
+	}
+	return c;
+}
+
+/// Parse a UTF-8 character, set *size to the length, and return the code point
+static inline uint32_t
+parse_utf8_char(const uint8_t* utf8, size_t* size)
+{
+	switch (*size = utf8_num_bytes(utf8[0])) {
+	case 1: case 2: case 3: case 4:
+		return parse_counted_utf8_char(utf8, *size);
+	default:
+		return *size = 0;
+	}
+}
+
+#endif  // SERD_STRING_UTILS_H
diff --git a/src/uri.c b/src/uri.c
index 9cf6a2a6..138ef55c 100644
--- a/src/uri.c
+++ b/src/uri.c
@@ -19,6 +19,8 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "uri_utils.h"
+
 char*
 serd_file_uri_parse(const char* uri, char** hostname)
 {
diff --git a/src/uri_utils.h b/src/uri_utils.h
new file mode 100644
index 00000000..7c19081a
--- /dev/null
+++ b/src/uri_utils.h
@@ -0,0 +1,103 @@
+/*
+  Copyright 2011-2018 David Robillard <http://drobilla.net>
+
+  Permission to use, copy, modify, and/or distribute this software for any
+  purpose with or without fee is hereby granted, provided that the above
+  copyright notice and this permission notice appear in all copies.
+
+  THIS SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+  WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+  MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+  ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+  WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+  ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+  OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+*/
+
+#ifndef SERD_URI_UTILS_H
+#define SERD_URI_UTILS_H
+
+#include "string_utils.h"
+
+static inline bool
+slice_equals(const SerdStringView* a, const SerdStringView* b)
+{
+	return a->len == b->len && !strncmp(a->buf, b->buf, a->len);
+}
+
+static inline size_t
+uri_path_len(const SerdURI* uri)
+{
+	return uri->path_base.len + uri->path.len;
+}
+
+static inline char
+uri_path_at(const SerdURI* uri, size_t i)
+{
+	if (i < uri->path_base.len) {
+		return uri->path_base.buf[i];
+	} else {
+		return uri->path.buf[i - uri->path_base.len];
+	}
+}
+
+/**
+   Return the index of the first differing character after the last root slash,
+   or zero if `uri` is not under `root`.
+*/
+static inline size_t
+uri_rooted_index(const SerdURI* uri, const SerdURI* root)
+{
+	if (!root || !root->scheme.len ||
+	    !slice_equals(&root->scheme, &uri->scheme) ||
+	    !slice_equals(&root->authority, &uri->authority)) {
+		return 0;
+	}
+
+	bool         differ          = false;
+	const size_t path_len        = uri_path_len(uri);
+	const size_t root_len        = uri_path_len(root);
+	size_t       last_root_slash = 0;
+	for (size_t i = 0; i < path_len && i < root_len; ++i) {
+		const uint8_t u = uri_path_at(uri, i);
+		const uint8_t r = uri_path_at(root, i);
+
+		differ = differ || u != r;
+		if (r == '/') {
+			last_root_slash = i;
+			if (differ) {
+				return 0;
+			}
+		}
+	}
+
+	return last_root_slash + 1;
+}
+
+/** Return true iff `uri` shares path components with `root` */
+static inline bool
+uri_is_related(const SerdURI* uri, const SerdURI* root)
+{
+	return uri_rooted_index(uri, root) > 0;
+}
+
+/** Return true iff `uri` is within the base of `root` */
+static inline bool
+uri_is_under(const SerdURI* uri, const SerdURI* root)
+{
+	const size_t index = uri_rooted_index(uri, root);
+	return index > 0 && uri->path.len > index;
+}
+
+static inline bool
+is_uri_scheme_char(const uint8_t c)
+{
+	switch (c) {
+	case ':': case '+': case '-': case '.':
+		return true;
+	default:
+		return is_alpha(c) || is_digit(c);
+	}
+}
+
+#endif  // SERD_URI_UTILS_H
diff --git a/src/writer.c b/src/writer.c
index 9404edfe..174d331a 100644
--- a/src/writer.c
+++ b/src/writer.c
@@ -21,12 +21,26 @@
 #include <stdlib.h>
 #include <string.h>
 
+#include "byte_sink.h"
+#include "node.h"
+#include "stack.h"
+#include "string_utils.h"
+#include "uri_utils.h"
+
 typedef struct {
 	SerdNode* graph;
 	SerdNode* subject;
 	SerdNode* predicate;
 } WriteContext;
 
+typedef enum {
+	FIELD_NONE,
+	FIELD_SUBJECT,
+	FIELD_PREDICATE,
+	FIELD_OBJECT,
+	FIELD_GRAPH
+} Field;
+
 static const WriteContext WRITE_CONTEXT_NULL = { NULL, NULL, NULL };
 
 typedef enum {
diff --git a/wscript b/wscript
index f3dd30eb..11a8de68 100644
--- a/wscript
+++ b/wscript
@@ -185,23 +185,35 @@ def lint(ctx):
 def amalgamate(ctx):
     "builds single-file amalgamated source"
     import shutil
+    import re
     shutil.copy('serd/serd.h', 'build/serd.h')
+
+    def include_line(line):
+        return (not re.match('#include "[^/]*\.h"', line) and
+                not re.match('#include "serd/serd.h"', line))
+
     with open('build/serd.c', 'w') as amalgamation:
-        with open('src/serd_internal.h') as serd_internal_h:
-            for l in serd_internal_h:
-                amalgamation.write(l.replace('serd/serd.h', 'serd.h'))
+        amalgamation.write('/* This is amalgamated code, do not edit! */\n')
+        amalgamation.write('#include "serd.h"\n\n')
+
+        for header_path in ['src/serd_internal.h',
+                            'src/byte_sink.h',
+                            'src/byte_source.h',
+                            'src/stack.h',
+                            'src/string_utils.h',
+                            'src/uri_utils.h',
+                            'src/reader.h']:
+            with open(header_path) as header:
+                for l in header:
+                    if include_line(l):
+                        amalgamation.write(l)
 
         for f in lib_source:
             with open(f) as fd:
                 amalgamation.write('\n/**\n   @file %s\n*/' % f)
-                header = True
                 for l in fd:
-                    if header:
-                        if l == '*/\n':
-                            header = False
-                    else:
-                        if l != '#include "serd_internal.h"\n':
-                            amalgamation.write(l)
+                    if include_line(l):
+                        amalgamation.write(l)
 
     for i in ['c', 'h']:
         Logs.info('Wrote build/serd.%s' % i)