Update seeking behaviour of null streams.

In order to (hopefully) allow page content streams to be interpreted
without having to preload them all into memory before we run them, we
need to make the stream reading code cope with other users moving
the stream pointer.

For example: Consider the case where we are midway through
interpreting a contents stream, and us hitting an operator that
requires something to be read from Resources. This will move the
underlying stream file pointer, and cause the contents stream to
read incorrectly when control returns to the interpreter.

The solution to this seems to be fairly simple; whenever we create
a filter out of the file stream, the existing code puts in a 'null'
filter first, to enforce a length limit on the stream. This null
filter already does most of the work we need it to, in that by it
being there, the buffering of data is done in the null filter rather
than in the underlying stream layer.

All we need to do is to keep track of where in the underlying stream
the null filter thinks it is, and ensure that it seeks there before
each read (in case anyone else has moved it).

We move the setting of the offset to be explicit in the pdf_open_filter
(and associated) call(s), rather than requiring fz_seeks elsewhere.

3 files changed, 17 insertions, 19 deletions

diff --git a/fitz/filt_basic.c b/fitz/filt_basic.c
index 09d63402..ac6a5903 100644
--- a/fitz/filt_basic.c
+++ b/fitz/filt_basic.c
@@ -14,6 +14,7 @@ struct null_filter
 {
 	fz_stream *chain;
 	int remain;
+	int pos;
 };
 
 static int
@@ -21,8 +22,12 @@ read_null(fz_stream *stm, unsigned char *buf, int len)
 {
 	struct null_filter *state = stm->state;
 	int amount = MIN(len, state->remain);
-	int n = fz_read(state->chain, buf, amount);
+	int n;
+
+	fz_seek(state->chain, state->pos, 0);
+	n = fz_read(state->chain, buf, amount);
 	state->remain -= n;
+	state->pos += n;
 	return n;
 }
 
@@ -36,7 +41,7 @@ close_null(fz_context *ctx, void *state_)
 }
 
 fz_stream *
-fz_open_null(fz_stream *chain, int len)
+fz_open_null(fz_stream *chain, int len, int offset)
 {
 	struct null_filter *state;
 	fz_context *ctx = chain->ctx;
@@ -46,6 +51,7 @@ fz_open_null(fz_stream *chain, int len)
 		state = fz_malloc_struct(ctx, struct null_filter);
 		state->chain = chain;
 		state->remain = len;
+		state->pos = offset;
 	}
 	fz_catch(ctx)
 	{
diff --git a/fitz/fitz-internal.h b/fitz/fitz-internal.h
index 1e54c900..197d0279 100644
--- a/fitz/fitz-internal.h
+++ b/fitz/fitz-internal.h
@@ -520,7 +520,7 @@ static inline int fz_is_eof_bits(fz_stream *stm)
  */
 
 fz_stream *fz_open_copy(fz_stream *chain);
-fz_stream *fz_open_null(fz_stream *chain, int len);
+fz_stream *fz_open_null(fz_stream *chain, int len, int offset);
 fz_stream *fz_open_arc4(fz_stream *chain, unsigned char *key, unsigned keylen);
 fz_stream *fz_open_aesd(fz_stream *chain, unsigned char *key, unsigned keylen);
 fz_stream *fz_open_a85d(fz_stream *chain);
diff --git a/pdf/pdf_stream.c b/pdf/pdf_stream.c
index 84f966ec..5338d81c 100644
--- a/pdf/pdf_stream.c
+++ b/pdf/pdf_stream.c
@@ -226,7 +226,7 @@ build_filter_chain(fz_stream *chain, pdf_document *xref, pdf_obj *fs, pdf_obj *p
  * stream length, followed by a decryption filter.
  */
 static fz_stream *
-pdf_open_raw_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int num, int gen)
+pdf_open_raw_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int num, int gen, int offset)
 {
 	int hascrypt;
 	int len;
@@ -236,7 +236,7 @@ pdf_open_raw_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int n
 	fz_keep_stream(chain);
 
 	len = pdf_to_int(pdf_dict_gets(stmobj, "Length"));
-	chain = fz_open_null(chain, len);
+	chain = fz_open_null(chain, len, offset);
 
 	fz_try(ctx)
 	{
@@ -258,7 +258,7 @@ pdf_open_raw_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int n
  * to stream length and decrypting.
  */
 static fz_stream *
-pdf_open_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int num, int gen, pdf_image_params *imparams)
+pdf_open_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int num, int gen, int offset, pdf_image_params *imparams)
 {
 	pdf_obj *filters;
 	pdf_obj *params;
@@ -266,7 +266,7 @@ pdf_open_filter(fz_stream *chain, pdf_document *xref, pdf_obj *stmobj, int num,
 	filters = pdf_dict_getsa(stmobj, "Filter", "F");
 	params = pdf_dict_getsa(stmobj, "DecodeParms", "DP");
 
-	chain = pdf_open_raw_filter(chain, xref, stmobj, num, gen);
+	chain = pdf_open_raw_filter(chain, xref, stmobj, num, gen, offset);
 
 	if (pdf_is_name(filters))
 		chain = build_filter(chain, xref, filters, params, num, gen, imparams);
@@ -298,7 +298,7 @@ pdf_open_inline_stream(pdf_document *xref, pdf_obj *stmobj, int length, fz_strea
 	if (pdf_array_len(filters) > 0)
 		return build_filter_chain(chain, xref, filters, params, 0, 0, imparams);
 
-	return fz_open_null(chain, length);
+	return fz_open_null(chain, length, fz_tell(chain));
 }
 
 /*
@@ -324,9 +324,8 @@ pdf_open_raw_stream(pdf_document *xref, int num, int gen)
 	if (x->stm_ofs == 0)
 		fz_throw(xref->ctx, "object is not a stream");
 
-	stm = pdf_open_raw_filter(xref->file, xref, x->obj, num, gen);
+	stm = pdf_open_raw_filter(xref->file, xref, x->obj, num, gen, x->stm_ofs);
 	fz_lock_stream(stm);
-	fz_seek(xref->file, x->stm_ofs, 0);
 	return stm;
 }
 
@@ -345,7 +344,6 @@ fz_stream *
 pdf_open_image_stream(pdf_document *xref, int num, int gen, pdf_image_params *params)
 {
 	pdf_xref_entry *x;
-	fz_stream *stm;
 
 	if (num < 0 || num >= xref->len)
 		fz_throw(xref->ctx, "object id out of range (%d %d R)", num, gen);
@@ -358,9 +356,7 @@ pdf_open_image_stream(pdf_document *xref, int num, int gen, pdf_image_params *pa
 	if (x->stm_ofs == 0)
 		fz_throw(xref->ctx, "object is not a stream");
 
-	stm = pdf_open_filter(xref->file, xref, x->obj, num, gen, params);
-	fz_seek(xref->file, x->stm_ofs, 0);
-	return stm;
+	return pdf_open_filter(xref->file, xref, x->obj, num, gen, x->stm_ofs, params);
 }
 
 fz_stream *
@@ -410,14 +406,10 @@ pdf_open_image_decomp_stream(fz_context *ctx, fz_buffer *buffer, pdf_image_param
 fz_stream *
 pdf_open_stream_with_offset(pdf_document *xref, int num, int gen, pdf_obj *dict, int stm_ofs)
 {
-	fz_stream *stm;
-
 	if (stm_ofs == 0)
 		fz_throw(xref->ctx, "object is not a stream");
 
-	stm = pdf_open_filter(xref->file, xref, dict, num, gen, NULL);
-	fz_seek(xref->file, stm_ofs, 0);
-	return stm;
+	return pdf_open_filter(xref->file, xref, dict, num, gen, stm_ofs, NULL);
 }
 
 /*