diffblue-cbmc/unit/path_strategies.cpp

507 lines
16 KiB
C++
Raw Normal View History

/*******************************************************************\
Module: Path Strategy Tests
Author: Kareem Khazem <karkhaz@karkhaz.com>, 2018
\*******************************************************************/
#include <testing-utils/use_catch.h>
#include <path_strategies.h>
#include <fstream>
#include <functional>
#include <string>
#include <ansi-c/ansi_c_language.h>
#include <cbmc/cbmc_parse_options.h>
#include <goto-checker/bmc_util.h>
#include <goto-checker/goto_symex_property_decider.h>
#include <goto-checker/symex_bmc.h>
#include <langapi/mode.h>
#include <util/cmdline.h>
#include <util/config.h>
#include <util/tempfile.h>
// The actual test suite.
//
// Whenever you add a new path exploration strategy 'my-cool-strategy', for each
// of the test programs (under the GIVEN macros), add a new test using a call to
// `check_with_strategy("my-cool-strategy", c, event_list)` where `event_list`
// is a list of the events that you expect to see during symbolic execution.
// Events are either resumes or results.
//
// Whenever symbolic execution pauses and picks a path to resume from, you
// should note the line number of the path you expect that path strategy to
// resume from. A resume is either a JUMP, meaning that it's the target of a
// `goto` instruction, or a NEXT, meaning that it's the instruction following a
// `goto` instruction.
//
// Whenever symbolic execution reaches the end of a path, you should expect a
// result. Results are either DONE, meaning that verification of that path
// succeeded, or FOUND_FAIL, meaning that there was an assertion failure on that
// path.
//
// To figure out what the events should be, run CBMC on the test program with
// your strategy with `--verbosity 10` and look out for lines like
//
// Resuming from jump target 'file nested-if/main.c line 13 function main'
//
// Resuming from next instruction 'file nested-if/main.c line 14 function main'
//
// VERIFICATION SUCCESSFUL
//
// VERIFICATION FAILED
//
// And note the order in which they occur.
SCENARIO("path strategies")
{
std::string c;
GIVEN("a simple conditional program")
{
std::function<void(optionst &)> opts_callback = [](optionst &) {};
c =
"/* 1 */ int main() \n"
"/* 2 */ { \n"
"/* 3 */ int x; \n"
"/* 4 */ if(x) \n"
"/* 5 */ x = 1; \n"
"/* 6 */ else \n"
"/* 7 */ x = 0; \n"
"/* 8 */ } \n";
check_with_strategy(
"lifo",
opts_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
symex_eventt::resume(symex_eventt::enumt::JUMP, 7),
symex_eventt::resume(symex_eventt::enumt::NEXT, 5),
symex_eventt::result(symex_eventt::enumt::SUCCESS)});
check_with_strategy(
"fifo",
opts_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
symex_eventt::resume(symex_eventt::enumt::NEXT, 5),
symex_eventt::resume(symex_eventt::enumt::JUMP, 7),
symex_eventt::result(symex_eventt::enumt::SUCCESS)});
}
GIVEN("a program with nested conditionals")
{
std::function<void(optionst &)> opts_callback = [](optionst &) {};
c =
"/* 1 */ int main() \n"
"/* 2 */ { \n"
"/* 3 */ int x, y; \n"
"/* 4 */ if(x) \n"
"/* 5 */ { \n"
"/* 6 */ if(y) \n"
"/* 7 */ y = 1; \n"
"/* 8 */ else \n"
"/* 9 */ y = 0; \n"
"/* 10 */ } \n"
"/* 11 */ else \n"
"/* 12 */ { \n"
"/* 13 */ if(y) \n"
"/* 14 */ y = 1; \n"
"/* 15 */ else \n"
"/* 16 */ y = 0; \n"
"/* 17 */ } \n"
"/* 18 */ } \n";
check_with_strategy(
"lifo",
opts_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
// Outer else, inner else
symex_eventt::resume(symex_eventt::enumt::JUMP, 13),
symex_eventt::resume(symex_eventt::enumt::JUMP, 16),
// Outer else, inner if
symex_eventt::resume(symex_eventt::enumt::NEXT, 14),
// Outer if, inner else
symex_eventt::resume(symex_eventt::enumt::NEXT, 6),
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
// Outer if, inner if
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
symex_eventt::result(symex_eventt::enumt::SUCCESS)});
check_with_strategy(
"fifo",
opts_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
// Expand outer if, but don't continue depth-first
symex_eventt::resume(symex_eventt::enumt::NEXT, 6),
// Jump to outer else, but again don't continue depth-first
symex_eventt::resume(symex_eventt::enumt::JUMP, 13),
// Expand inner if of the outer if
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
// No more branch points, so complete the path. Then continue BFSing
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
symex_eventt::resume(symex_eventt::enumt::NEXT, 14),
symex_eventt::resume(symex_eventt::enumt::JUMP, 16),
symex_eventt::result(symex_eventt::enumt::SUCCESS)});
}
GIVEN("a loop program to test functional correctness")
{
std::function<void(optionst &)> opts_callback = [](optionst &opts) {
opts.set_option("unwind", 2U);
};
// clang-format off
c =
"/* 1 */ int main() \n"
"/* 2 */ { \n"
"/* 3 */ int x; \n"
"/* 4 */ if(x == 1) { \n"
"/* 5 */ \n"
"/* 6 */ while(x) \n"
"/* 7 */ --x; \n"
"/* 8 */ \n"
"/* 9 */ assert(x); \n"
"/* 10 */ } \n"
"/* 11 */ } \n";
// clang-format on
check_with_strategy(
"lifo",
opts_callback,
c,
{
// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
// The path where x != 1 and we have nothing to check:
symex_eventt::resume(symex_eventt::enumt::JUMP, 11),
// The path where we skip the loop body. Successful because the path is
// implausible, we cannot have skipped the body if x == 1.
symex_eventt::resume(symex_eventt::enumt::NEXT, 6),
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
symex_eventt::result(symex_eventt::enumt::SUCCESS),
// Enter the loop body once. Since we decrement x, the assertion should
// fail.
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
symex_eventt::result(symex_eventt::enumt::FAILURE),
// The path where we enter the loop body twice. Successful because
// infeasible.
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
symex_eventt::result(symex_eventt::enumt::SUCCESS),
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
// Overall we fail.
symex_eventt::result(symex_eventt::enumt::FAILURE),
});
check_with_strategy(
"fifo",
opts_callback,
c,
{
// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
// First the path where we enter the if-block at line 4 then on hitting
// the branch at line 6 consider skipping straight to the end, but find
// nothing there to investigate:
symex_eventt::resume(symex_eventt::enumt::NEXT, 6),
symex_eventt::resume(symex_eventt::enumt::JUMP, 11),
// The path where we skip the loop body. Successful because the path is
// implausible, we cannot have skipped the body if x == 1.
//
// In this case, although we resume from line 7, we don't proceed until
// the end of the path after executing line 7.
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
symex_eventt::result(symex_eventt::enumt::SUCCESS),
// Pop line 7 that we saved from above, and execute the loop a second
// time. Successful, since this path is infeasible.
symex_eventt::resume(symex_eventt::enumt::NEXT, 7),
symex_eventt::result(symex_eventt::enumt::SUCCESS),
// Pop line 7 that we saved from above and bail out. That corresponds to
// executing the loop once, decrementing x to 0; assert(x) should fail.
symex_eventt::resume(symex_eventt::enumt::JUMP, 9),
symex_eventt::result(symex_eventt::enumt::FAILURE),
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
// Overall we fail.
symex_eventt::result(symex_eventt::enumt::FAILURE),
});
}
GIVEN("program to check for stop-on-fail with path exploration")
{
std::function<void(optionst &)> halt_callback = [](optionst &opts) {
opts.set_option("stop-on-fail", true);
};
std::function<void(optionst &)> no_halt_callback = [](optionst &) {};
c =
"/* 1 */ int main() \n"
"/* 2 */ { \n"
"/* 3 */ int x, y; \n"
"/* 4 */ if(x) \n"
"/* 5 */ assert(0); \n"
"/* 6 */ else \n"
"/* 7 */ assert(0); \n"
"/* 8 */ } \n";
GIVEN("no stopping on failure")
{
check_with_strategy(
"lifo",
no_halt_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
symex_eventt::resume(symex_eventt::enumt::JUMP, 7),
symex_eventt::result(symex_eventt::enumt::FAILURE),
symex_eventt::resume(symex_eventt::enumt::NEXT, 5),
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
symex_eventt::result(symex_eventt::enumt::FAILURE),
// Overall result
symex_eventt::result(symex_eventt::enumt::FAILURE)});
}
GIVEN("stopping on failure")
{
check_with_strategy(
"lifo",
halt_callback,
c,
{// Entry state is line 0
symex_eventt::resume(symex_eventt::enumt::NEXT, 0),
symex_eventt::resume(symex_eventt::enumt::JUMP, 7),
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
symex_eventt::result(symex_eventt::enumt::FAILURE),
// Overall result
symex_eventt::result(symex_eventt::enumt::FAILURE)});
}
}
}
// In theory, there should be no need to change the code below when adding new
// test cases...
void symex_eventt::validate_result(
listt &events,
const incremental_goto_checkert::resultt::progresst result,
std::size_t &counter)
{
INFO(
"Expecting result to be '"
<< (result == incremental_goto_checkert::resultt::progresst::DONE
? "success"
: "failure")
<< "' (item at index [" << counter << "] in expected results list");
++counter;
if(result == incremental_goto_checkert::resultt::progresst::DONE)
{
REQUIRE(!events.empty());
REQUIRE(events.front().first == symex_eventt::enumt::SUCCESS);
events.pop_front();
}
else
{
REQUIRE(!events.empty());
REQUIRE(events.front().first == symex_eventt::enumt::FAILURE);
events.pop_front();
}
}
void symex_eventt::validate_resume(
listt &events,
const goto_symex_statet &state,
std::size_t &counter)
{
REQUIRE(!events.empty());
int dst = 0;
if(!state.saved_target->source_location.get_line().empty())
dst = std::stoi(state.saved_target->source_location.get_line().c_str());
if(state.has_saved_next_instruction)
{
INFO(
"Expecting resume to be 'next' to line "
<< dst << " (item at index [" << counter
<< "] in expected resumes list)");
REQUIRE(events.front().first == symex_eventt::enumt::NEXT);
}
else if(state.has_saved_jump_target)
{
INFO(
"Expecting resume to be 'jump' to line "
<< dst << " (item at index [" << counter
<< "] in expected resumes list)");
REQUIRE(events.front().first == symex_eventt::enumt::JUMP);
}
else
REQUIRE(false);
++counter;
REQUIRE(events.front().second == dst);
events.pop_front();
}
// This is a simplified copy of single_path_symex_checkert
// because we have to check the worklist every time goto-symex returns.
void _check_with_strategy(
const std::string &strategy,
const std::string &program,
std::function<void(optionst &)> opts_callback,
symex_eventt::listt &events)
{
temporary_filet tmp("path-explore_", ".c");
std::ofstream of(tmp().c_str());
REQUIRE(of.is_open());
of << program << std::endl;
of.close();
register_language(new_ansi_c_language);
cmdlinet cmdline;
cmdline.args.push_back(tmp());
config.main = std::string("main");
config.set(cmdline);
optionst options;
cbmc_parse_optionst::set_default_options(options);
options.set_option("paths", true);
options.set_option("exploration-strategy", strategy);
REQUIRE(is_valid_path_strategy(strategy));
opts_callback(options);
ui_message_handlert ui_message_handler(cmdline, "path-explore");
ui_message_handler.set_verbosity(0);
messaget log(ui_message_handler);
goto_modelt goto_model;
int ret;
ret = cbmc_parse_optionst::get_goto_program(
goto_model, options, cmdline, ui_message_handler);
REQUIRE(ret == -1);
symbol_tablet symex_symbol_table;
namespacet ns(goto_model.get_symbol_table(), symex_symbol_table);
propertiest properties(initialize_properties(goto_model));
std::unique_ptr<path_storaget> worklist = get_path_strategy(strategy);
guard_managert guard_manager;
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
{
// Put initial state into the work list
symex_target_equationt equation(ui_message_handler);
symex_bmct symex(
ui_message_handler,
goto_model.get_symbol_table(),
equation,
options,
*worklist,
guard_manager);
setup_symex(symex, ns, options, ui_message_handler);
symex.initialize_path_storage_from_entry_point_of(
goto_symext::get_goto_function(goto_model), symex_symbol_table);
}
std::size_t expected_results_cnt = 0;
while(!worklist->empty())
{
path_storaget::patht &resume = worklist->peek();
symex_eventt::validate_resume(events, resume.state, expected_results_cnt);
symex_bmct symex(
ui_message_handler,
goto_model.get_symbol_table(),
resume.equation,
options,
*worklist,
guard_manager);
setup_symex(symex, ns, options, ui_message_handler);
symex.resume_symex_from_saved_state(
goto_symext::get_goto_function(goto_model),
resume.state,
&resume.equation,
symex_symbol_table);
postprocess_equation(
symex, resume.equation, options, ns, ui_message_handler);
incremental_goto_checkert::resultt result(
incremental_goto_checkert::resultt::progresst::DONE);
if(symex.get_remaining_vccs() > 0)
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
{
update_properties_status_from_symex_target_equation(
properties, result.updated_properties, resume.equation);
goto_symex_property_decidert property_decider(
options, ui_message_handler, resume.equation, ns);
const auto solver_runtime = prepare_property_decider(
properties, resume.equation, property_decider, ui_message_handler);
run_property_decider(
result,
properties,
property_decider,
ui_message_handler,
solver_runtime,
false);
symex_eventt::validate_result(
events, result.progress, expected_results_cnt);
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
}
worklist->pop();
if(
result.progress ==
incremental_goto_checkert::resultt::progresst::FOUND_FAIL &&
options.get_bool_option("stop-on-fail"))
{
worklist->clear();
}
if(worklist->empty())
{
update_status_of_not_checked_properties(
properties, result.updated_properties);
update_status_of_unknown_properties(
properties, result.updated_properties);
}
}
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
const resultt overall_result = determine_result(properties);
symex_eventt::validate_result(
events,
overall_result == resultt::FAIL
? incremental_goto_checkert::resultt::progresst::FOUND_FAIL
: incremental_goto_checkert::resultt::progresst::DONE,
expected_results_cnt);
Only safety-check paths when new VCCs get added CBMC will no longer bother running safety checks on path segments that did not generate any new VCCs when running in path-exploration mode. A "path segment" is a list of instructions between two program branch points. CBMC would previously pause symbolic execution and run safety checks at every branch point, which is wasteful when no additional VCCs have been generated since the last pause. As of this commit, CBMC will check to see if new VCCs have been added, and only do the safety check if so. To support this, the new member goto_symext::path_segment_vccs records the number of VCCs that were newly generated along this path segment. It differs from the VCC counters in goto_symex_statet, which hold the number of VCCs generated along the current path since the start of the program. goto_symext objects are minted for each path segment, so they don't hold any information about the VCCs previously along the path. This commit also removes the "Starting new path" output, since this no longer makes much sense. On the other hand, there is an additional final status output; CBMC will report failure after the entire program has been explored if even a single path contained a failing assertion, and report success otherwise. The unit tests have been updated to reflect the fact that CBMC no longer emits "SUCCESS" at each branch point and at the end of paths, and also that CBMC makes one final report of the program safety as a whole. This commit fixes #2684.
2018-10-26 00:54:18 +08:00
INFO("The expected results list contains " << events.size() << " items");
REQUIRE(events.empty());
}