Merge branch 'master' into blender2.8

# Conflicts:
#	source/blender/alembic/intern/abc_exporter.h
#	source/blender/alembic/intern/abc_util.cc

7 files changed, 37 insertions, 33 deletions

diff --git a/intern/cycles/blender/blender_mesh.cpp b/intern/cycles/blender/blender_mesh.cpp
index e0e89cec65c..54571b1fea1 100644
--- a/intern/cycles/blender/blender_mesh.cpp
+++ b/intern/cycles/blender/blender_mesh.cpp
@@ -560,6 +560,9 @@ static void attr_create_pointiness(Scene *scene,
 		return;
 	}
 	const int num_verts = b_mesh.vertices.length();
+	if(num_verts == 0) {
+		return;
+	}
 	/* STEP 1: Find out duplicated vertices and point duplicates to a single
 	 *         original vertex.
 	 */
@@ -1164,8 +1167,8 @@ void BlenderSync::sync_mesh_motion(BL::Object& b_ob,
 	}
 
 	/* skip empty meshes */
-	size_t numverts = mesh->verts.size();
-	size_t numkeys = mesh->curve_keys.size();
+	const size_t numverts = mesh->verts.size();
+	const size_t numkeys = mesh->curve_keys.size();
 
 	if(!numverts && !numkeys)
 		return;
@@ -1223,13 +1226,12 @@ void BlenderSync::sync_mesh_motion(BL::Object& b_ob,
 
 	/* TODO(sergey): Perform preliminary check for number of verticies. */
 	if(numverts) {
-		/* find attributes */
+		/* Find attributes. */
 		Attribute *attr_mP = mesh->attributes.find(ATTR_STD_MOTION_VERTEX_POSITION);
 		Attribute *attr_mN = mesh->attributes.find(ATTR_STD_MOTION_VERTEX_NORMAL);
 		Attribute *attr_N = mesh->attributes.find(ATTR_STD_VERTEX_NORMAL);
 		bool new_attribute = false;
-
-		/* add new attributes if they don't exist already */
+		/* Add new attributes if they don't exist already. */
 		if(!attr_mP) {
 			attr_mP = mesh->attributes.add(ATTR_STD_MOTION_VERTEX_POSITION);
 			if(attr_N)
@@ -1237,22 +1239,21 @@ void BlenderSync::sync_mesh_motion(BL::Object& b_ob,
 
 			new_attribute = true;
 		}
-
-		/* load vertex data from mesh */
+		/* Load vertex data from mesh. */
 		float3 *mP = attr_mP->data_float3() + time_index*numverts;
 		float3 *mN = (attr_mN)? attr_mN->data_float3() + time_index*numverts: NULL;
-
+		/* NOTE: We don't copy more that existing amount of vertices to prevent
+		 * possible memory corruption.
+		 */
 		BL::Mesh::vertices_iterator v;
 		int i = 0;
-
 		for(b_mesh.vertices.begin(v); v != b_mesh.vertices.end() && i < numverts; ++v, ++i) {
 			mP[i] = get_float3(v->co());
 			if(mN)
 				mN[i] = get_float3(v->normal());
 		}
-
-		/* in case of new attribute, we verify if there really was any motion */
 		if(new_attribute) {
+			/* In case of new attribute, we verify if there really was any motion. */
 			if(b_mesh.vertices.length() != numverts ||
 			   memcmp(mP, &mesh->verts[0], sizeof(float3)*numverts) == 0)
 			{
@@ -1275,7 +1276,6 @@ void BlenderSync::sync_mesh_motion(BL::Object& b_ob,
 				 * they had no motion, but we need them anyway now */
 				float3 *P = &mesh->verts[0];
 				float3 *N = (attr_N)? attr_N->data_float3(): NULL;
-
 				for(int step = 0; step < time_index; step++) {
 					memcpy(attr_mP->data_float3() + step*numverts, P, sizeof(float3)*numverts);
 					if(attr_mN)
@@ -1283,6 +1283,16 @@ void BlenderSync::sync_mesh_motion(BL::Object& b_ob,
 				}
 			}
 		}
+		else {
+			if(b_mesh.vertices.length() != numverts) {
+				VLOG(1) << "Topology differs, discarding motion blur for object "
+				        << b_ob.name() << " at time " << time_index;
+				memcpy(mP, &mesh->verts[0], sizeof(float3)*numverts);
+				if(mN != NULL) {
+					memcpy(mN, attr_N->data_float3(), sizeof(float3)*numverts);
+				}
+			}
+		}
 	}
 
 	/* hair motion */
diff --git a/intern/cycles/device/device_cpu.cpp b/intern/cycles/device/device_cpu.cpp
index 2761d9488ca..3c481bb2b39 100644
--- a/intern/cycles/device/device_cpu.cpp
+++ b/intern/cycles/device/device_cpu.cpp
@@ -856,7 +856,7 @@ int2 CPUSplitKernel::split_kernel_local_size()
 }
 
 int2 CPUSplitKernel::split_kernel_global_size(device_memory& /*kg*/, device_memory& /*data*/, DeviceTask * /*task*/) {
-	return make_int2(64, 1);
+	return make_int2(1, 1);
 }
 
 uint64_t CPUSplitKernel::state_buffer_size(device_memory& kernel_globals, device_memory& /*data*/, size_t num_threads) {
diff --git a/intern/cycles/device/device_split_kernel.cpp b/intern/cycles/device/device_split_kernel.cpp
index ae462a560b7..fa641161c05 100644
--- a/intern/cycles/device/device_split_kernel.cpp
+++ b/intern/cycles/device/device_split_kernel.cpp
@@ -151,7 +151,8 @@ bool DeviceSplitKernel::path_trace(DeviceTask *task,
 		/* Calculate max groups */
 
 		/* Denotes the maximum work groups possible w.r.t. current requested tile size. */
-		unsigned int max_work_groups = num_global_elements / WORK_POOL_SIZE + 1;
+		unsigned int work_pool_size = (device->info.type == DEVICE_CPU) ? WORK_POOL_SIZE_CPU : WORK_POOL_SIZE_GPU;
+		unsigned int max_work_groups = num_global_elements / work_pool_size + 1;
 
 		/* Allocate work_pool_wgs memory. */
 		work_pool_wgs.resize(max_work_groups * sizeof(unsigned int));
@@ -256,10 +257,8 @@ bool DeviceSplitKernel::path_trace(DeviceTask *task,
 			activeRaysAvailable = false;
 
 			for(int rayStateIter = 0; rayStateIter < global_size[0] * global_size[1]; ++rayStateIter) {
-				int8_t state = ray_state.get_data()[rayStateIter];
-
-				if(state != RAY_INACTIVE) {
-					if(state == RAY_INVALID) {
+				if(!IS_STATE(ray_state.get_data(), rayStateIter, RAY_INACTIVE)) {
+					if(IS_STATE(ray_state.get_data(), rayStateIter, RAY_INVALID)) {
 						/* Something went wrong, abort to avoid looping endlessly. */
 						device->set_error("Split kernel error: invalid ray state");
 						return false;
diff --git a/intern/cycles/device/opencl/opencl_util.cpp b/intern/cycles/device/opencl/opencl_util.cpp
index 8128fcee09b..6dca642f3f3 100644
--- a/intern/cycles/device/opencl/opencl_util.cpp
+++ b/intern/cycles/device/opencl/opencl_util.cpp
@@ -281,6 +281,7 @@ void OpenCLDeviceBase::OpenCLProgram::add_log(string msg, bool debug)
 	}
 	else if(!debug) {
 		printf("%s\n", msg.c_str());
+		fflush(stdout);
 	}
 	else {
 		VLOG(2) << msg;
diff --git a/intern/cycles/kernel/kernel_shadow.h b/intern/cycles/kernel/kernel_shadow.h
index 0426e0a62c9..db6f839d9ed 100644
--- a/intern/cycles/kernel/kernel_shadow.h
+++ b/intern/cycles/kernel/kernel_shadow.h
@@ -422,9 +422,9 @@ ccl_device_inline bool shadow_blocked(KernelGlobals *kg,
 		return false;
 	}
 #ifdef __SHADOW_TRICKS__
-    const int skip_object = state->catcher_object;
+	const int skip_object = state->catcher_object;
 #else
-    const int skip_object = OBJECT_NONE;
+	const int skip_object = OBJECT_NONE;
 #endif
 	/* Do actual shadow shading. */
 	/* First of all, we check if integrator requires transparent shadows.
diff --git a/intern/cycles/kernel/kernel_types.h b/intern/cycles/kernel/kernel_types.h
index 19c91248922..623f3728c69 100644
--- a/intern/cycles/kernel/kernel_types.h
+++ b/intern/cycles/kernel/kernel_types.h
@@ -56,7 +56,13 @@ CCL_NAMESPACE_BEGIN
 
 #define VOLUME_STACK_SIZE		16
 
-#define WORK_POOL_SIZE 64
+#define WORK_POOL_SIZE_GPU 64
+#define WORK_POOL_SIZE_CPU 1
+#ifdef __KERNEL_GPU__
+#  define WORK_POOL_SIZE WORK_POOL_SIZE_GPU
+#else
+#  define WORK_POOL_SIZE WORK_POOL_SIZE_CPU
+#endif
 
 /* device capabilities */
 #ifdef __KERNEL_CPU__
diff --git a/intern/locale/msgfmt.cc b/intern/locale/msgfmt.cc
index 6ee1ee14781..02c58ebc5bc 100644
--- a/intern/locale/msgfmt.cc
+++ b/intern/locale/msgfmt.cc
@@ -42,18 +42,6 @@ bool starts_with(const std::string &str,
   }
 }
 
-std::string ltrim(const std::string &str) {
-  std::string result = str;
-  result.erase(0, result.find_first_not_of(" \t\r\n"));
-  return result;
-}
-
-std::string rtrim(const std::string &str) {
-  std::string result = str;
-  result.erase(result.find_last_not_of(" \t\r\n") + 1);
-  return result;
-}
-
 std::string trim(const std::string &str) {
   std::string result = str;
   result.erase(0, result.find_first_not_of(" \t\r\n"));