Don't cache resolved objects in the indirect reference object (rely on the xref cache instead).

3 files changed, 20 insertions, 34 deletions

diff --git a/fitz/fitz.h b/fitz/fitz.h
index 150c3e56..89a1cc63 100644
--- a/fitz/fitz.h
+++ b/fitz/fitz.h
@@ -371,7 +371,6 @@ struct fz_obj_s
 			int num;
 			int gen;
 			struct pdf_xref_s *xref;
-			fz_obj *obj;
 		} r;
 	} u;
 };
diff --git a/fitz/obj_simple.c b/fitz/obj_simple.c
index fdd18ce1..f279bd55 100644
--- a/fitz/obj_simple.c
+++ b/fitz/obj_simple.c
@@ -74,7 +74,6 @@ fz_newindirect(int num, int gen, pdf_xref *xref)
 	o->u.r.num = num;
 	o->u.r.gen = gen;
 	o->u.r.xref = xref;
-	o->u.r.obj = nil;
 	return o;
 }
 
@@ -93,23 +92,11 @@ fz_dropobj(fz_obj *o)
 	if (--o->refs == 0)
 	{
 		if (o->kind == FZ_ARRAY)
-		{
 			fz_freearray(o);
-		}
 		else if (o->kind == FZ_DICT)
-		{
 			fz_freedict(o);
-		}
-		else if (o->kind == FZ_INDIRECT)
-		{
-			if (o->u.r.obj)
-				fz_dropobj(o->u.r.obj);
-			fz_free(o);
-		}
 		else
-		{
 			fz_free(o);
-		}
 	}
 }
 
@@ -234,22 +221,23 @@ int fz_togen(fz_obj *obj)
 
 fz_obj *fz_resolveindirect(fz_obj *ref)
 {
-	int error;
-
 	if (fz_isindirect(ref))
 	{
-		if (!ref->u.r.obj && ref->u.r.xref)
+		pdf_xref *xref = ref->u.r.xref;
+		int num = fz_tonum(ref);
+		int gen = fz_togen(ref);
+		if (xref)
 		{
-			error = pdf_loadobject(&ref->u.r.obj, ref->u.r.xref, fz_tonum(ref), fz_togen(ref));
+			fz_error error = pdf_cacheobject(xref, num, gen);
 			if (error)
 			{
-				fz_catch(error, "cannot resolve reference (%d %d R); ignoring error", fz_tonum(ref), fz_togen(ref));
-				ref->u.r.obj = fz_keepobj(ref);
+				fz_catch(error, "cannot load object (%d %d R) into cache", num, gen);
+				return ref;
 			}
+			if (xref->table[num].obj)
+				return xref->table[num].obj;
 		}
-		return ref->u.r.obj;
 	}
-
 	return ref;
 }
 
diff --git a/mupdf/pdf_xref.c b/mupdf/pdf_xref.c
index 9dec068c..196b633e 100644
--- a/mupdf/pdf_xref.c
+++ b/mupdf/pdf_xref.c
@@ -888,13 +888,12 @@ pdf_cacheobject(pdf_xref *xref, int num, int gen)
 	if (x->obj)
 		return fz_okay;
 
-	if (x->type == 'f' || x->type == 'd')
+	if (x->type == 'f')
 	{
 		x->obj = fz_newnull();
 		return fz_okay;
 	}
-
-	if (x->type == 'n')
+	else if (x->type == 'n')
 	{
 		error = fz_seek(xref->file, x->ofs, 0);
 		if (error)
@@ -911,7 +910,6 @@ pdf_cacheobject(pdf_xref *xref, int num, int gen)
 		if (xref->crypt)
 			pdf_cryptobj(xref->crypt, x->obj, num, gen);
 	}
-
 	else if (x->type == 'o')
 	{
 		if (!x->obj)
@@ -919,8 +917,14 @@ pdf_cacheobject(pdf_xref *xref, int num, int gen)
 			error = pdf_loadobjstm(xref, x->ofs, 0, xref->scratch, sizeof xref->scratch);
 			if (error)
 				return fz_rethrow(error, "cannot load object stream containing object (%d %d R)", num, gen);
+			if (!x->obj)
+				return fz_throw("object (%d %d R) was not found in its object stream", num, gen);
 		}
 	}
+	else
+	{
+		return fz_throw("assert: corrupt xref struct");
+	}
 
 	return fz_okay;
 }
@@ -934,14 +938,9 @@ pdf_loadobject(fz_obj **objp, pdf_xref *xref, int num, int gen)
 	if (error)
 		return fz_rethrow(error, "cannot load object (%d %d R) into cache", num, gen);
 
-	if (xref->table[num].obj)
-		*objp = fz_keepobj(xref->table[num].obj);
-	else
-	{
-		fz_warn("cannot load missing object (%d %d R), assuming null object", num, gen);
-		xref->table[num].obj = fz_newnull();
-		*objp = fz_keepobj(xref->table[num].obj);
-	}
+	assert(xref->table[num].obj);
+
+	*objp = fz_keepobj(xref->table[num].obj);
 
 	return fz_okay;
 }