make an actual library

9 files changed, 162 insertions, 187 deletions

diff --git a/src/Makefile b/src/Makefile
index af806f3..ed0b4ba 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -1,14 +1,3 @@
-#
+DIR_SOURCES := document.c byte_class.c cos_parse.c cos_object.c pdf_doc.c
 
-CFLAGS+=-g -Wall -Wextra
-
-OBJS=xref.o byte_class.o cos_decode.o cos_object.o pdf_doc.o
-
-.PHONY:all clean
-
-all:xref
-
-xref:$(OBJS)
-
-clean:
-	${RM} xref $(OBJS)
+include $(NSBUILD)/Makefile.subdir
diff --git a/src/cos_object.c b/src/cos_object.c
index 5bfd423..2fa3a93 100644
--- a/src/cos_object.c
+++ b/src/cos_object.c
@@ -14,7 +14,8 @@
 #include <stdio.h>
 #include <string.h>
 
-#include "nspdferror.h"
+#include <nspdf/errors.h>
+
 #include "cos_object.h"
 #include "pdf_doc.h"
 
@@ -110,7 +111,7 @@ cos_extract_dictionary_value(struct cos_object *dict,
  * get a value for a key from a dictionary
  */
 nspdferror
-cos_get_dictionary_value(struct pdf_doc *doc,
+cos_get_dictionary_value(struct nspdf_doc *doc,
                          struct cos_object *dict,
                          const char *key,
                          struct cos_object **value_out)
@@ -140,7 +141,7 @@ cos_get_dictionary_value(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_dictionary_int(struct pdf_doc *doc,
+cos_get_dictionary_int(struct nspdf_doc *doc,
                        struct cos_object *dict,
                        const char *key,
                        int64_t *value_out)
@@ -156,7 +157,7 @@ cos_get_dictionary_int(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_dictionary_name(struct pdf_doc *doc,
+cos_get_dictionary_name(struct nspdf_doc *doc,
                         struct cos_object *dict,
                         const char *key,
                         const char **value_out)
@@ -172,7 +173,7 @@ cos_get_dictionary_name(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_dictionary_dictionary(struct pdf_doc *doc,
+cos_get_dictionary_dictionary(struct nspdf_doc *doc,
                         struct cos_object *dict,
                         const char *key,
                         struct cos_object **value_out)
@@ -188,7 +189,7 @@ cos_get_dictionary_dictionary(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_heritable_dictionary_dictionary(struct pdf_doc *doc,
+cos_heritable_dictionary_dictionary(struct nspdf_doc *doc,
                         struct cos_object *dict,
                         const char *key,
                         struct cos_object **value_out)
@@ -206,7 +207,7 @@ cos_heritable_dictionary_dictionary(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_dictionary_array(struct pdf_doc *doc,
+cos_get_dictionary_array(struct nspdf_doc *doc,
                         struct cos_object *dict,
                         const char *key,
                         struct cos_object **value_out)
@@ -222,7 +223,7 @@ cos_get_dictionary_array(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_heritable_dictionary_array(struct pdf_doc *doc,
+cos_heritable_dictionary_array(struct nspdf_doc *doc,
                         struct cos_object *dict,
                         const char *key,
                         struct cos_object **value_out)
@@ -241,7 +242,7 @@ cos_heritable_dictionary_array(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_int(struct pdf_doc *doc,
+cos_get_int(struct nspdf_doc *doc,
             struct cos_object *cobj,
             int64_t *value_out)
 {
@@ -259,7 +260,7 @@ cos_get_int(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_name(struct pdf_doc *doc,
+cos_get_name(struct nspdf_doc *doc,
             struct cos_object *cobj,
             const char **value_out)
 {
@@ -279,7 +280,7 @@ cos_get_name(struct pdf_doc *doc,
 
 
 nspdferror
-cos_get_dictionary(struct pdf_doc *doc,
+cos_get_dictionary(struct nspdf_doc *doc,
                    struct cos_object *cobj,
                    struct cos_object **value_out)
 {
@@ -297,7 +298,7 @@ cos_get_dictionary(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_array(struct pdf_doc *doc,
+cos_get_array(struct nspdf_doc *doc,
                    struct cos_object *cobj,
                    struct cos_object **value_out)
 {
@@ -318,7 +319,7 @@ cos_get_array(struct pdf_doc *doc,
  * get a value for a key from a dictionary
  */
 nspdferror
-cos_get_array_value(struct pdf_doc *doc,
+cos_get_array_value(struct nspdf_doc *doc,
                     struct cos_object *array,
                     unsigned int index,
                     struct cos_object **value_out)
@@ -350,7 +351,7 @@ cos_get_array_value(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_array_dictionary(struct pdf_doc *doc,
+cos_get_array_dictionary(struct nspdf_doc *doc,
                         struct cos_object *array,
                         unsigned int index,
                         struct cos_object **value_out)
@@ -366,7 +367,7 @@ cos_get_array_dictionary(struct pdf_doc *doc,
 }
 
 nspdferror
-cos_get_array_size(struct pdf_doc *doc,
+cos_get_array_size(struct nspdf_doc *doc,
                    struct cos_object *cobj,
                    unsigned int *size_out)
 {
diff --git a/src/cos_object.h b/src/cos_object.h
index 48241c6..a40c691 100644
--- a/src/cos_object.h
+++ b/src/cos_object.h
@@ -1,4 +1,4 @@
-struct pdf_doc;
+struct nspdf_doc;
 
 enum cos_type {
     COS_TYPE_NULL,
@@ -83,13 +83,6 @@ struct cos_object {
     } u;
 };
 
-/**
- * Decode input stream into an object
- *
- * lex and parse a byte stream to generate a COS object.
- */
-nspdferror cos_decode_object(struct pdf_doc *doc, uint64_t *offset_out, struct cos_object **cosobj_out);
-
 nspdferror cos_free_object(struct cos_object *cos_obj);
 
 /**
@@ -117,35 +110,35 @@ nspdferror cos_extract_dictionary_value(struct cos_object *dict, const char *key
  *         NSPDFERROR_TYPE if the object passed in \p dict is not a dictionary.
  *         NSPDFERROR_NOTFOUND if the key is not present in the dictionary.
  */
-nspdferror cos_get_dictionary_value(struct pdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
+nspdferror cos_get_dictionary_value(struct nspdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
 
 
-nspdferror cos_get_dictionary_int(struct pdf_doc *doc, struct cos_object *dict, const char *key, int64_t *value_out);
+nspdferror cos_get_dictionary_int(struct nspdf_doc *doc, struct cos_object *dict, const char *key, int64_t *value_out);
 
 
-nspdferror cos_get_dictionary_name(struct pdf_doc *doc, struct cos_object *dict, const char *key, const char **value_out);
+nspdferror cos_get_dictionary_name(struct nspdf_doc *doc, struct cos_object *dict, const char *key, const char **value_out);
 
-nspdferror cos_get_dictionary_dictionary(struct pdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
+nspdferror cos_get_dictionary_dictionary(struct nspdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
 
-nspdferror cos_heritable_dictionary_dictionary(struct pdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
+nspdferror cos_heritable_dictionary_dictionary(struct nspdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
 
-nspdferror cos_get_dictionary_array(struct pdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
+nspdferror cos_get_dictionary_array(struct nspdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
 
-nspdferror cos_heritable_dictionary_array(struct pdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
+nspdferror cos_heritable_dictionary_array(struct nspdf_doc *doc, struct cos_object *dict, const char *key, struct cos_object **value_out);
 
 
-nspdferror cos_get_int(struct pdf_doc *doc, struct cos_object *cobj, int64_t *value_out);
+nspdferror cos_get_int(struct nspdf_doc *doc, struct cos_object *cobj, int64_t *value_out);
 
 
-nspdferror cos_get_name(struct pdf_doc *doc, struct cos_object *cobj, const char **value_out);
+nspdferror cos_get_name(struct nspdf_doc *doc, struct cos_object *cobj, const char **value_out);
 
 
-nspdferror cos_get_dictionary(struct pdf_doc *doc, struct cos_object *cobj, struct cos_object **value_out);
+nspdferror cos_get_dictionary(struct nspdf_doc *doc, struct cos_object *cobj, struct cos_object **value_out);
 
-nspdferror cos_get_array(struct pdf_doc *doc, struct cos_object *cobj, struct cos_object **value_out);
+nspdferror cos_get_array(struct nspdf_doc *doc, struct cos_object *cobj, struct cos_object **value_out);
 
-nspdferror cos_get_array_size(struct pdf_doc *doc, struct cos_object *cobj, unsigned int *size_out);
+nspdferror cos_get_array_size(struct nspdf_doc *doc, struct cos_object *cobj, unsigned int *size_out);
 
-nspdferror cos_get_array_value(struct pdf_doc *doc, struct cos_object *array, unsigned int index, struct cos_object **value_out);
+nspdferror cos_get_array_value(struct nspdf_doc *doc, struct cos_object *array, unsigned int index, struct cos_object **value_out);
 
-nspdferror cos_get_array_dictionary(struct pdf_doc *doc, struct cos_object *arrau, unsigned int index, struct cos_object **value_out);
+nspdferror cos_get_array_dictionary(struct nspdf_doc *doc, struct cos_object *arrau, unsigned int index, struct cos_object **value_out);
diff --git a/src/cos_decode.c b/src/cos_parse.c
index 8873060..ca3d802 100644
--- a/src/cos_decode.c
+++ b/src/cos_parse.c
@@ -5,8 +5,10 @@
 #include <stdio.h>
 #include <string.h>
 
+#include <nspdf/errors.h>
+
+#include "cos_parse.h"
 #include "byte_class.h"
-#include "nspdferror.h"
 #include "cos_object.h"
 #include "pdf_doc.h"
 
@@ -46,7 +48,7 @@ static uint8_t xtoi(uint8_t x)
 }
 
 static nspdferror
-cos_decode_number(struct pdf_doc *doc,
+cos_decode_number(struct nspdf_doc *doc,
                       uint64_t *offset_out,
                       struct cos_object **cosobj_out)
 {
@@ -101,7 +103,7 @@ cos_decode_number(struct pdf_doc *doc,
  *
  */
 static nspdferror
-cos_decode_string(struct pdf_doc *doc,
+cos_decode_string(struct nspdf_doc *doc,
                   uint64_t *offset_out,
                   struct cos_object **cosobj_out)
 {
@@ -230,7 +232,7 @@ cos_decode_string(struct pdf_doc *doc,
  * decode hex encoded string
  */
 static nspdferror
-cos_decode_hex_string(struct pdf_doc *doc,
+cos_decode_hex_string(struct nspdf_doc *doc,
                       uint64_t *offset_out,
                       struct cos_object **cosobj_out)
 {
@@ -293,7 +295,7 @@ cos_decode_hex_string(struct pdf_doc *doc,
  * decode a dictionary object
  */
 static nspdferror
-cos_decode_dictionary(struct pdf_doc *doc,
+cos_decode_dictionary(struct nspdf_doc *doc,
                       uint64_t *offset_out,
                       struct cos_object **cosobj_out)
 {
@@ -324,7 +326,7 @@ cos_decode_dictionary(struct pdf_doc *doc,
     while ((DOC_BYTE(doc, offset) != '>') &&
            (DOC_BYTE(doc, offset + 1) != '>')) {
 
-        res = cos_decode_object(doc, &offset, &key);
+        res = cos_parse_object(doc, &offset, &key);
         if (res != NSPDFERROR_OK) {
             /* todo free up any dictionary entries already created */
             printf("key object decode failed\n");
@@ -337,7 +339,7 @@ cos_decode_dictionary(struct pdf_doc *doc,
         }
         //printf("key: %s\n", key->u.n);
 
-        res = cos_decode_object(doc, &offset, &value);
+        res = cos_parse_object(doc, &offset, &value);
         if (res != NSPDFERROR_OK) {
             printf("Unable to decode value object in dictionary\n");
             /* todo free up any dictionary entries already created */
@@ -371,7 +373,7 @@ cos_decode_dictionary(struct pdf_doc *doc,
  * decode a list
  */
 static nspdferror
-cos_decode_list(struct pdf_doc *doc,
+cos_decode_list(struct nspdf_doc *doc,
                 uint64_t *offset_out,
                 struct cos_object **cosobj_out)
 {
@@ -406,7 +408,7 @@ cos_decode_list(struct pdf_doc *doc,
 
     while (DOC_BYTE(doc, offset) != ']') {
 
-        res = cos_decode_object(doc, &offset, &value);
+        res = cos_parse_object(doc, &offset, &value);
         if (res != NSPDFERROR_OK) {
             cos_free_object(cosobj);
             printf("Unable to decode value object in list\n");
@@ -442,7 +444,7 @@ cos_decode_list(struct pdf_doc *doc,
  * \todo deal with # symbols on pdf versions 1.2 and later
  */
 static nspdferror
-cos_decode_name(struct pdf_doc *doc,
+cos_decode_name(struct nspdf_doc *doc,
                 uint64_t *offset_out,
                 struct cos_object **cosobj_out)
 {
@@ -498,7 +500,7 @@ cos_decode_name(struct pdf_doc *doc,
  * decode a cos boolean object
  */
 static int
-cos_decode_boolean(struct pdf_doc *doc,
+cos_decode_boolean(struct nspdf_doc *doc,
                    uint64_t *offset_out,
                    struct cos_object **cosobj_out)
 {
@@ -574,7 +576,7 @@ cos_decode_boolean(struct pdf_doc *doc,
  * decode the null object.
  */
 static nspdferror
-cos_decode_null(struct pdf_doc *doc,
+cos_decode_null(struct nspdf_doc *doc,
                 uint64_t *offset_out,
                 struct cos_object **cosobj_out)
 {
@@ -640,7 +642,7 @@ cos_decode_null(struct pdf_doc *doc,
  * integer
  */
 static nspdferror
-cos_attempt_decode_reference(struct pdf_doc *doc,
+cos_attempt_decode_reference(struct nspdf_doc *doc,
                              uint64_t *offset_out,
                              struct cos_object **cosobj_out)
 {
@@ -712,7 +714,7 @@ cos_attempt_decode_reference(struct pdf_doc *doc,
         }
         //printf("decoding\n");
 
-        res = cos_decode_object(doc, &offset, &indirect);
+        res = cos_parse_object(doc, &offset, &indirect);
         if (res != NSPDFERROR_OK) {
             cos_free_object(generation);
             return res;
@@ -752,7 +754,7 @@ cos_attempt_decode_reference(struct pdf_doc *doc,
 
 
 /*
- * Decode input stream into an object
+ * Parse input stream into an object
  *
  * lex and parse a byte stream to generate COS objects
  *
@@ -804,7 +806,7 @@ cos_attempt_decode_reference(struct pdf_doc *doc,
  *   ;
  */
 nspdferror
-cos_decode_object(struct pdf_doc *doc,
+cos_parse_object(struct nspdf_doc *doc,
                   uint64_t *offset_out,
                   struct cos_object **cosobj_out)
 {
diff --git a/src/cos_parse.h b/src/cos_parse.h
new file mode 100644
index 0000000..adfb835
--- /dev/null
+++ b/src/cos_parse.h
@@ -0,0 +1,10 @@
+struct nspdf_doc;
+struct cos_object;
+
+/**
+ * Decode input stream into an object
+ *
+ * lex and parse a byte stream to generate a COS object.
+ */
+nspdferror cos_parse_object(struct nspdf_doc *doc, uint64_t *offset_out, struct cos_object **cosobj_out);
+
diff --git a/src/xref.c b/src/document.c
index 452aa19..9be0ab5 100644
--- a/src/xref.c
+++ b/src/document.c
@@ -1,3 +1,12 @@
+/*
+ * Copyright 2018 Vincent Sanders <vince@netsurf-browser.org>
+ *
+ * This file is part of libnspdf.
+ *
+ * Licensed under the MIT License,
+ *                http://www.opensource.org/licenses/mit-license.php
+ */
+
 #include <stdio.h>
 #include <stdint.h>
 #include <inttypes.h>
@@ -5,59 +14,27 @@
 #include <stdbool.h>
 #include <string.h>
 
-#include "nspdferror.h"
+#include <nspdf/document.h>
+
+#include "cos_parse.h"
 #include "byte_class.h"
 #include "cos_object.h"
 #include "pdf_doc.h"
 
 #define SLEN(x) (sizeof((x)) - 1)
 
-
-int
-read_whole_pdf(struct pdf_doc *doc, const char *fname)
-{
-    FILE *f;
-    off_t len;
-    uint8_t *buf;
-    size_t rd;
-
-    f = fopen(fname, "r");
-    if (f == NULL) {
-        perror("pdf open");
-        return 1;
-    }
-
-    fseek(f, 0, SEEK_END);
-    len = ftello(f);
-
-    buf = malloc(len);
-    fseek(f, 0, SEEK_SET);
-
-    rd = fread(buf, len, 1, f);
-    if (rd != 1) {
-        perror("pdf read");
-        free(buf);
-        return 1;
-    }
-
-    fclose(f);
-
-    doc->start = doc->buffer = buf;
-    doc->length = doc->buffer_length = len;
-
-    return 0;
-}
-
-
 #define STARTXREF_TOK "startxref"
-/* Number of bytes to search back from file end to find xref start token, convention says 1024 bytes */
-#define STARTXREF_SEARCH_SIZE 1024
-
 
+/* Number of bytes to search back from file end to find xref start token,
+ * convention says 1024 bytes
+ */
+#define STARTXREF_SEARCH_SIZE 1024
 
 
 static nspdferror
-doc_read_uint(struct pdf_doc *doc, uint64_t *offset_out, uint64_t *result_out)
+doc_read_uint(struct nspdf_doc *doc,
+              uint64_t *offset_out,
+              uint64_t *result_out)
 {
     uint8_t c; /* current byte from source data */
     unsigned int len; /* number of decimal places in number */
@@ -90,10 +67,11 @@ doc_read_uint(struct pdf_doc *doc, uint64_t *offset_out, uint64_t *result_out)
     return -1; /* number too long */
 }
 
+
 /**
  * finds the startxref marker at the end of input
  */
-nspdferror find_startxref(struct pdf_doc *doc, uint64_t *offset_out)
+static nspdferror find_startxref(struct nspdf_doc *doc, uint64_t *offset_out)
 {
     uint64_t offset; /* offset of characters being considered for startxref */
     uint64_t earliest; /* earliest offset to serch for startxref */
@@ -123,10 +101,14 @@ nspdferror find_startxref(struct pdf_doc *doc, uint64_t *offset_out)
     return NSPDFERROR_SYNTAX;
 }
 
+
 /**
  * decodes a startxref field
  */
-nspdferror decode_startxref(struct pdf_doc *doc, uint64_t *offset_out, uint64_t *start_xref_out)
+static nspdferror
+decode_startxref(struct nspdf_doc *doc,
+                 uint64_t *offset_out,
+                 uint64_t *start_xref_out)
 {
     uint64_t offset; /* offset of characters being considered for startxref */
     uint64_t start_xref;
@@ -181,7 +163,7 @@ nspdferror decode_startxref(struct pdf_doc *doc, uint64_t *offset_out, uint64_t
 /**
  * finds the next trailer
  */
-nspdferror find_trailer(struct pdf_doc *doc, uint64_t *offset_out)
+static nspdferror find_trailer(struct nspdf_doc *doc, uint64_t *offset_out)
 {
     uint64_t offset; /* offset of characters being considered for trailer */
 
@@ -200,33 +182,9 @@ nspdferror find_trailer(struct pdf_doc *doc, uint64_t *offset_out)
     return NSPDFERROR_SYNTAX;
 }
 
-/**
- * find the PDF comment marker to identify the start of the document
- */
-int check_header(struct pdf_doc *doc)
-{
-    uint64_t offset; /* offset of characters being considered for startxref */
-
-    for (offset = 0; offset < 1024; offset++) {
-        if ((DOC_BYTE(doc, offset) == '%') &&
-            (DOC_BYTE(doc, offset + 1) == 'P') &&
-            (DOC_BYTE(doc, offset + 2) == 'D') &&
-            (DOC_BYTE(doc, offset + 3) == 'F') &&
-            (DOC_BYTE(doc, offset + 4) == '-') &&
-            (DOC_BYTE(doc, offset + 5) == '1') &&
-            (DOC_BYTE(doc, offset + 6) == '.')) {
-            doc->start = doc->buffer + offset;
-            doc->length -= offset;
-            /* read number for minor */
-            return 0;
-        }
-    }
-    return -1;
-}
-
 
-nspdferror
-decode_trailer(struct pdf_doc *doc,
+static nspdferror
+decode_trailer(struct nspdf_doc *doc,
                uint64_t *offset_out,
                struct cos_object **trailer_out)
 {
@@ -249,7 +207,7 @@ decode_trailer(struct pdf_doc *doc,
     offset += 7;
     doc_skip_ws(doc, &offset);
 
-    res = cos_decode_object(doc, &offset, &trailer);
+    res = cos_parse_object(doc, &offset, &trailer);
     if (res != 0) {
         return res;
     }
@@ -265,8 +223,9 @@ decode_trailer(struct pdf_doc *doc,
     return NSPDFERROR_OK;
 }
 
-nspdferror
-decode_xref(struct pdf_doc *doc, uint64_t *offset_out)
+
+static nspdferror
+decode_xref(struct nspdf_doc *doc, uint64_t *offset_out)
 {
     uint64_t offset;
     nspdferror res;
@@ -359,7 +318,8 @@ decode_xref(struct pdf_doc *doc, uint64_t *offset_out)
 /**
  * recursively parse trailers and xref tables
  */
-nspdferror decode_xref_trailer(struct pdf_doc *doc, uint64_t xref_offset)
+static nspdferror
+decode_xref_trailer(struct nspdf_doc *doc, uint64_t xref_offset)
 {
     nspdferror res;
     uint64_t offset; /* the current data offset */
@@ -455,6 +415,7 @@ decode_xref_trailer_failed:
     return res;
 }
 
+
 /**
  * decode non-linear pdf trailer data
  *
@@ -477,7 +438,7 @@ decode_xref_trailer_failed:
  * find the subsequent trailer.
  *
  */
-nspdferror decode_trailers(struct pdf_doc *doc)
+static nspdferror decode_trailers(struct nspdf_doc *doc)
 {
     nspdferror res;
     uint64_t offset; /* the current data offset */
@@ -499,11 +460,12 @@ nspdferror decode_trailers(struct pdf_doc *doc)
     return decode_xref_trailer(doc, startxref);
 }
 
+
 /**
  * recursively decodes a page tree
  */
-nspdferror
-decode_page_tree(struct pdf_doc *doc,
+static nspdferror
+decode_page_tree(struct nspdf_doc *doc,
                  struct cos_object *page_tree_node,
                  unsigned int *page_index)
 {
@@ -594,12 +556,14 @@ decode_page_tree(struct pdf_doc *doc,
             return res;
         }
 
+        /*
         printf("page index:%d page:%p resources:%p mediabox:%p contents:%p\n",
                *page_index,
                page,
                page->resources,
                page->mediabox,
                page->contents);
+        */
 
         (*page_index)++;
         res = NSPDFERROR_OK;
@@ -609,7 +573,8 @@ decode_page_tree(struct pdf_doc *doc,
     return res;
 }
 
-nspdferror decode_catalog(struct pdf_doc *doc)
+
+static nspdferror decode_catalog(struct nspdf_doc *doc)
 {
     nspdferror res;
     struct cos_object *catalog;
@@ -645,38 +610,63 @@ nspdferror decode_catalog(struct pdf_doc *doc)
     return res;
 }
 
-nspdferror new_pdf_doc(struct pdf_doc **doc_out)
+/* exported interface documented in nspdf/document.h */
+nspdferror nspdf_document_create(struct nspdf_doc **doc_out)
 {
-    struct pdf_doc *doc;
-    doc = calloc(1, sizeof(struct pdf_doc));
+    struct nspdf_doc *doc;
+    doc = calloc(1, sizeof(struct nspdf_doc));
     if (doc == NULL) {
         return NSPDFERROR_NOMEM;
     }
+
     *doc_out = doc;
+
     return NSPDFERROR_OK;
 }
 
-int main(int argc, char **argv)
+/* exported interface documented in nspdf/document.h */
+nspdferror nspdf_document_destroy(struct nspdf_doc *doc)
 {
-    struct pdf_doc *doc;
-    int res;
+    free(doc);
 
-    if (argc < 2) {
-        fprintf(stderr, "Usage %s <filename>\n", argv[0]);
-        return 1;
-    }
+    return NSPDFERROR_OK;
+}
 
-    res = new_pdf_doc(&doc);
-    if (res != NSPDFERROR_OK) {
-        printf("failed to read file\n");
-        return res;
-    }
 
-    res = read_whole_pdf(doc, argv[1]);
-    if (res != 0) {
-        printf("failed to read file\n");
-        return res;
+/**
+ * find the PDF comment marker to identify the start of the document
+ */
+static nspdferror check_header(struct nspdf_doc *doc)
+{
+    uint64_t offset; /* offset of characters being considered for header */
+    for (offset = 0; offset < 1024; offset++) {
+        if ((DOC_BYTE(doc, offset) == '%') &&
+            (DOC_BYTE(doc, offset + 1) == 'P') &&
+            (DOC_BYTE(doc, offset + 2) == 'D') &&
+            (DOC_BYTE(doc, offset + 3) == 'F') &&
+            (DOC_BYTE(doc, offset + 4) == '-') &&
+            (DOC_BYTE(doc, offset + 5) == '1') &&
+            (DOC_BYTE(doc, offset + 6) == '.')) {
+            doc->start += offset;
+            doc->length -= offset;
+
+            /* \todo read number for minor */
+            return NSPDFERROR_OK;
+        }
     }
+    return NSPDFERROR_NOTFOUND;
+}
+
+/* exported interface documented in nspdf/document.h */
+nspdferror
+nspdf_document_parse(struct nspdf_doc *doc,
+                     const uint8_t *buffer,
+                     uint64_t buffer_length)
+{
+    nspdferror res;
+
+    doc->start = buffer;
+    doc->length = buffer_length;
 
     res = check_header(doc);
     if (res != 0) {
@@ -696,5 +686,5 @@ int main(int argc, char **argv)
         return res;
     }
 
-    return 0;
+    return res;
 }
diff --git a/src/nspdferror.h b/src/nspdferror.h
deleted file mode 100644
index 3e26813..0000000
--- a/src/nspdferror.h
+++ /dev/null
@@ -1,10 +0,0 @@
-typedef enum {
-    NSPDFERROR_OK,
-    NSPDFERROR_NOMEM,
-    NSPDFERROR_SYNTAX, /**< syntax error in parse */
-    NSPDFERROR_SIZE, /**< not enough input data */
-    NSPDFERROR_RANGE, /**< value outside type range */
-    NSPDFERROR_TYPE, /**< wrong type error */
-    NSPDFERROR_NOTFOUND, /**< key not found */
-    NSPDFERROR_FORMAT, /**< objects do not cornform to expected format */
-} nspdferror;
diff --git a/src/pdf_doc.c b/src/pdf_doc.c
index 4a5cad1..281025c 100644
--- a/src/pdf_doc.c
+++ b/src/pdf_doc.c
@@ -12,7 +12,9 @@
 #include <stdbool.h>
 #include <stdio.h>
 
-#include "nspdferror.h"
+#include <nspdf/errors.h>
+
+#include "cos_parse.h"
 #include "byte_class.h"
 #include "cos_object.h"
 #include "pdf_doc.h"
@@ -20,7 +22,7 @@
 /**
  * move offset to next non whitespace byte
  */
-nspdferror doc_skip_ws(struct pdf_doc *doc, uint64_t *offset)
+nspdferror doc_skip_ws(struct nspdf_doc *doc, uint64_t *offset)
 {
     uint8_t c;
     /* TODO sort out keeping offset in range */
@@ -43,7 +45,7 @@ nspdferror doc_skip_ws(struct pdf_doc *doc, uint64_t *offset)
 /**
  * move offset to next non eol byte
  */
-nspdferror doc_skip_eol(struct pdf_doc *doc, uint64_t *offset)
+nspdferror doc_skip_eol(struct nspdf_doc *doc, uint64_t *offset)
 {
     uint8_t c;
     /* TODO sort out keeping offset in range */
@@ -60,7 +62,7 @@ static struct cos_object cos_null_obj = {
 };
 
 nspdferror
-xref_get_referenced(struct pdf_doc *doc, struct cos_object **cobj_out)
+xref_get_referenced(struct nspdf_doc *doc, struct cos_object **cobj_out)
 {
     nspdferror res;
     struct cos_object *cobj;
@@ -90,7 +92,7 @@ xref_get_referenced(struct pdf_doc *doc, struct cos_object **cobj_out)
     if (entry->object == NULL) {
         /* indirect object has never been decoded */
         offset = entry->offset;
-        res = cos_decode_object(doc, &offset, &indirect);
+        res = cos_parse_object(doc, &offset, &indirect);
         if (res != NSPDFERROR_OK) {
             printf("failed to decode indirect object\n");
             return res;
diff --git a/src/pdf_doc.h b/src/pdf_doc.h
index 986556f..e9bdc14 100644
--- a/src/pdf_doc.h
+++ b/src/pdf_doc.h
@@ -18,11 +18,9 @@ struct page_table_entry {
 };
 
 /** pdf document */
-struct pdf_doc {
-    uint8_t *buffer;
-    uint64_t buffer_length;
+struct nspdf_doc {
 
-    uint8_t *start; /* start of pdf document in input stream */
+    const uint8_t *start; /* start of pdf document in input stream */
     uint64_t length;
 
     int major;
@@ -47,7 +45,7 @@ struct pdf_doc {
 /* byte data acessory, allows for more complex buffer handling in future */
 #define DOC_BYTE(doc, offset) (doc->start[(offset)])
 
-nspdferror doc_skip_ws(struct pdf_doc *doc, uint64_t *offset);
-nspdferror doc_skip_eol(struct pdf_doc *doc, uint64_t *offset);
+nspdferror doc_skip_ws(struct nspdf_doc *doc, uint64_t *offset);
+nspdferror doc_skip_eol(struct nspdf_doc *doc, uint64_t *offset);
 
-nspdferror xref_get_referenced(struct pdf_doc *doc, struct cos_object **cobj_out);
+nspdferror xref_get_referenced(struct nspdf_doc *doc, struct cos_object **cobj_out);