GoogleCodeExporter
commented
9 years ago I believe a fix to this defect would be...
NOTE: I think for 17 (A.1.15. State Creation_4) & 18 (A.1.15. State Creation_5)
you need to set XFAIL_YES rather than XFAIL_NO; based on RFC4465
//static void checkResult(const struct_torture_test* test, const tcomp_result_t
*result, const void* output_ptr, tsk_size_t output_size)
//{
// if(tsk_striequals(test->xoutput.ptr, "decompression_memory_size")){
// printf("xoutput (decompression_memory_size): %s\n",
(HostToNetworkShort(DMS) == *((uint16_t*)output_ptr)) ? "YES" : "NO");
// }
// else{
// printf("xoutput: %s\n", startsWith(output_ptr, output_size,
test->xoutput.ptr, test->xoutput.size) ? "YES" : "NO");
// }
// printf("xcycles: %s\n", (result->consumed_cycles == test->xcycles) ? "YES" :
"NO");
//}
static int checkResult(const struct_torture_test* test, const tcomp_result_t
*result, const void* output_ptr, tsk_size_t output_size)
{
int test_status = 0;
// if not expecting a failure
if (!test->xfail)
{
if (result->isNack)
{
printf("FAIL: UNEXPECTED NACK\n");
test_status += 1;
}
else
{
if(tsk_striequals(test->xoutput.ptr, "decompression_memory_size")){
printf("xoutput (decompression_memory_size): %s\n", (HostToNetworkShort(DMS) == *((uint16_t*)output_ptr)) ? "YES" : "NO");
if (HostToNetworkShort(DMS) != *((uint16_t*)output_ptr))
test_status += 1;
}
else{
printf("xoutput: %s\n", startsWith(output_ptr, output_size, test->xoutput.ptr, test->xoutput.size) ? "YES" : "NO");
if (!startsWith(output_ptr, output_size, test->xoutput.ptr, test->xoutput.size))
test_status += 1;
}
if (result->consumed_cycles == test->xcycles)
{
printf("xcycles: YES\n");
}
else
{
printf("xcycles: NO (exp: %d, got: %d)\n", test->xcycles, result->consumed_cycles);
test_status +=1;
}
}
}
else // bit expecting a failure
{
printf("Expected failure: %s\n", test->xoutput.ptr);
// fail: expecting a NACK and did not get one
if (!(result->isNack))
{
printf("FAIL: MISSING NACK\n");
test_status += 1;
}
}
return test_status;
}
FROM test_tortures() lines I’ve added/modified have >>> above and <<< below
them
#if RUN_TEST_LOOP
for(;;)
#endif
{
int test_result = 0;
#if RUN_TEST_ALL
start = 0, end = sizeof(tests)/sizeof(tests[0]);
#else
start = RUN_TEST_NO, end = RUN_TEST_NO + 1;
#endif
for(i = start; i<end; i++){
>>>
test_result = 0;
<<<
printf("=== Testing %s ===\n\n", tests[i].section_name);
if(tests[i].stream){
tcomp_result_setOutputTCPBuffer(result, buffer, OUTPUT_BUFFER_SIZE, STREAM_ID);
}
else{
tcomp_result_setOutputUDPBuffer(result, buffer, OUTPUT_BUFFER_SIZE);
}
/* Performs decompression */
res_size = tcomp_manager_decompress(manager, tests[i].bytecode, tests[i].bytecode_size, result);
if(result->isNack){
printf("\n==WE GOT A NACK\n\n");
//sendto(tcomp_buffer_getBuffer(result->nack_info), tcomp_buffer_getSize(result->nack_info));
}
else{
tcomp_result_setCompartmentId(result, COMPARTMENT_ID, strlen(COMPARTMENT_ID));
tcomp_manager_provideCompartmentId(manager, result);
}
/* Checks result */
>>>
test_result += checkResult(&tests[i], result, buffer, sizeof(buffer));
<<<
/* Get sub-sequent stream messages */
if(tests[i].stream && (res_size || result->isNack)){
uint64_t consumed_cycles = result->consumed_cycles; // save "consumed_cycles" (shared by all sub-sequent messages)
for(;;){
res_size = tcomp_manager_getNextStreamMessage(manager, result);
if(!res_size && !result->isNack){
break;
}
else if(res_size){
tcomp_result_setCompartmentId(result, COMPARTMENT_ID, strlen(COMPARTMENT_ID));
tcomp_manager_provideCompartmentId(manager, result);
}
else{
printf("\n==WE GOT A NACK\n\n");
//sendto(tcomp_buffer_getBuffer(result->nack_info), tcomp_buffer_getSize(result->nack_info));
}
result->consumed_cycles = consumed_cycles; // restore cycles just for checking. Had been cleared by "tcomp_manager_getNextStreamMessage()"
>>>
test_result += checkResult(&tests[i], result, buffer, sizeof(buffer));
<<<
}
}
>>>
printf("=== %s STATUS: %s ===\n\n", tests[i].section_name, test_result ? "FAIL" : "PASS");
<<<
}
}/* LOOP */
Original comment by alistair...@gmail.com on 11 Jan 2013 at 3:14
Original issue reported on code.google.com by
boss...@yahoo.fron 8 Jan 2013 at 11:43