diff --git a/pdf.c b/pdf.c
index 703cf5f9f18dd5f5d3b0eeead822247ead90c650..f6dd414ffd223250d832fe06e97f258a4c71770c 100644
--- a/pdf.c
+++ b/pdf.c
@@ -338,6 +338,25 @@ init_parser(void)
p_xref = xr_td;
}
+const HParsedToken *
+dictentry(const HCountedArray *dict, const char *key)
+{
+ HParsedToken *ent;
+ HBytes k;
+ size_t len;
+
+ len = strlen(key);
+ for (size_t i = 0; i < dict->used; i++) {
+ ent = dict->elements[i];
+ k = H_INDEX_BYTES(ent, 0);
+
+ if (k.len == len && strncmp(key, k.token, k.len) == 0)
+ return H_INDEX_TOKEN(ent, 1);
+ }
+
+ return NULL;
+}
+
/*
* This continuation takes the stream dictionary (as first element of x) and
* should return a parser that consumes exactly the bytes that make up the
@@ -346,21 +365,12 @@ init_parser(void)
HParser *
kstream(HAllocator *mm__, const HParsedToken *x, void *env)
{
- HCountedArray *dict = H_INDEX_SEQ(x, 0);
- HParsedToken *ent, *v = NULL;
- HBytes k;
+ const HCountedArray *dict = H_INDEX_SEQ(x, 0);
+ const HParsedToken *v = NULL;
size_t sz;
/* look for the Length entry */
- for (size_t i = 0; i < dict->used; i++) {
- ent = dict->elements[i];
- k = H_INDEX_BYTES(ent, 0);
-
- if (k.len == 6 && strncmp("Length", k.token, k.len) == 0) {
- v = H_INDEX_TOKEN(ent, 1);
- break;
- }
- }
+ v = dictentry(dict, "Length");
if (v == NULL || v->token_type != TT_SINT || v->sint < 0)
goto fail;
sz = (size_t)v->sint;
@@ -385,22 +395,113 @@ fail:
/*
- * minimal main program
+ * main program
*/
#include <stdio.h>
#include <inttypes.h>
#include <err.h>
+#include <errno.h>
+#include <stdlib.h> /* realloc() */
#include <fcntl.h> /* open() */
#include <unistd.h> /* lseek() */
#include <sys/mman.h> /* mmap() */
-int main(int argc, char *argv[])
+const char *infile = NULL;
+
+/*
+ * This helper implements the standard backwards parsing strategy to read all
+ * cross-reference sections and trailer dictionaries, starting from the
+ * 'startxref' offset found at the very end of the input.
+ *
+ * Allocates and returns an array of HParsedTokens, each containing the result
+ * of a successful 'p_xref' parse. Sets the output parameter 'nxrefs' to the
+ * number of elements.
+ *
+ * A return value of NULL indicates an empty result.
+ */
+const HParsedToken **
+parse_xrefs(const char *input, size_t sz, size_t *nxrefs)
+{
+ HParseResult *res = NULL;
+ const HParsedToken **xrefs = NULL; /* empty result */
+ const HParsedToken *tok = NULL;
+ size_t n = 0;
+ size_t offset = 0;
+
+ // XXX try formulating this as a parser using h_seek()
+
+ /* search for the "startxref" section from the back of the file */
+ HParser *p = h_left(p_startxref, h_end_p());
+ for (size_t i = 0; i < sz; i++) {
+ res = h_parse(p, input + sz - i, i);
+ if (res != NULL)
+ break;
+ }
+ if (res == NULL) {
+ fprintf(stderr, "%s: startxref not found\n", infile);
+ goto end;
+ }
+ offset = H_INDEX_UINT(res->ast, 0);
+
+ for (;;) {
+ res = h_parse(p_xref, input + offset, sz - offset);
+ if (res == NULL) {
+ fprintf(stderr, "%s: error parsing xref section at "
+ "position %zu (0x%zx)\n", infile, offset, offset);
+ break;
+ }
+
+ /* save this section in xrefs */
+ if (n >= SIZE_MAX / sizeof(HParsedToken *))
+ errc(1, EOVERFLOW, "overflow");
+ xrefs = realloc(xrefs, (n + 1) * sizeof(HParsedToken *));
+ if (xrefs == NULL)
+ err(1, "realloc");
+ xrefs[n++] = res->ast;
+
+ /* look up the next offset (to the previous xref section) */
+ tok = dictentry(H_INDEX_SEQ(res->ast, 1), "Prev");
+ if (tok == NULL)
+ break;
+ if (tok->token_type != TT_SINT) {
+ fprintf(stderr, "%s: /Prev not an integer\n", infile);
+ break;
+ }
+
+ /*
+ * validate the new offset. we don't want to get caught in a
+ * loop. the offsets should strictly decrease, unless the file
+ * is a "linearized" PDF. in that case there should be exactly
+ * two xref sections in the reverse order, so we allow the
+ * first section to point forward.
+ */
+ if (n > 1 && tok->sint >= offset) {
+ fprintf(stderr, "%s: /Prev pointer of xref section at "
+ "%zu (0x%zx) points forward\n", infile, offset,
+ offset);
+ break;
+ }
+
+ offset = (size_t)tok->sint;
+ }
+ // XXX debug
+ //fprintf(stderr, "%s: %zu xref sections parsed\n", infile, n);
+ //for (size_t i = 0; i < n; i++)
+ // h_pprintln(stderr, xrefs[i]);
+
+end:
+ *nxrefs = n;
+ return xrefs;
+}
+
+int
+main(int argc, char *argv[])
{
HParseResult *res = NULL;
- const char *infile = NULL;
+ const HParsedToken **xrefs;
const uint8_t *input;
- size_t sz, startxref;
+ size_t sz, nxrefs;
int fd;
/* command line handling */
@@ -424,28 +525,8 @@ int main(int argc, char *argv[])
/* build parsers */
init_parser();
- /* search for the "startxref" section from the back of the file */
- HParser *p = h_left(p_startxref, h_end_p());
- for (size_t i = 0; i < sz; i++) {
- res = h_parse(p, input + sz - i, i);
- if (res) break;
- }
- if (res == NULL) {
- fprintf(stderr, "%s: startxref not found\n", infile);
- return 1;
- }
- startxref = H_INDEX_UINT(res->ast, 0);
-
- /* parse cross-references and trailer dictionary */
- res = h_parse(p_xref, input + startxref, sz - startxref);
- if (!res) {
- fprintf(stderr, "%s: error parsing xref/trailer at "
- "position %zu (0x%zx)\n", infile, startxref, startxref);
- // continue anyway...
- }
- // XXX debug
- //h_pprintln(stderr, res->ast);
- //return 0;
+ /* parse all cross-reference sections and trailer dictionaries */
+ xrefs = parse_xrefs(input, sz, &nxrefs);
/* run the main parser */
res = h_parse(p_pdf, input, sz);