JakeHillion/object-introspection
1
0
Fork 0
mirror of https://github.com/JakeHillion/object-introspection.git synced 2024-09-19 11:09:05 +01:00
Code Issues Packages Projects Releases Wiki Activity
main
object-introspection/test/test_drgn_parser.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

707 lines
20 KiB
C++
Raw Permalink Normal View History

TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
#include "test_drgn_parser.h"
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
#include <gtest/gtest.h>
#include <regex>
#include "oi/OIParser.h"
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
#include "oi/SymbolService.h"
TypeGraph: Remove NodeTracker from the TypeGraph class The TypeGraph class should only be responsible for storing Type nodes. Traversing the graph and tracking which nodes have been visited should not be included there. Passes now take a NodeTrackerHolder as an input parameter, which provides access to a zeroed-out NodeTracker.
2023-08-15 15:27:20 +01:00
#include "oi/type_graph/NodeTracker.h"
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
#include "oi/type_graph/Printer.h"
#include "oi/type_graph/TypeGraph.h"
#include "oi/type_graph/Types.h"
using namespace type_graph;
// TODO setup google logging for tests so it doesn't appear on terminal by
// default
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
SymbolService* DrgnParserTest::symbols_ = nullptr;
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
void DrgnParserTest::SetUpTestSuite() {
symbols_ = new SymbolService{TARGET_EXE_PATH};
}
void DrgnParserTest::TearDownTestSuite() {
delete symbols_;
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
DrgnParser DrgnParserTest::getDrgnParser(TypeGraph& typeGraph,
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options) {
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
DrgnParser drgnParser{typeGraph, options};
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
return drgnParser;
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
drgn_type* DrgnParserTest::getDrgnRoot(std::string_view function) {
irequest req{"entry", std::string{function}, "arg0"};
auto* drgnRoot = symbols_->getRootType(req)->type.type;
return drgnRoot;
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
std::string DrgnParserTest::run(std::string_view function,
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options) {
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
TypeGraph typeGraph;
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
auto drgnParser = getDrgnParser(typeGraph, options);
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
auto* drgnRoot = getDrgnRoot(function);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
DrgnParserTest: Split into reusable components
2023-07-06 16:20:46 +01:00
Type& type = drgnParser.parse(drgnRoot);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
std::stringstream out;
TypeGraph: Remove NodeTracker from the TypeGraph class The TypeGraph class should only be responsible for storing Type nodes. Traversing the graph and tracking which nodes have been visited should not be included there. Passes now take a NodeTrackerHolder as an input parameter, which provides access to a zeroed-out NodeTracker.
2023-08-15 15:27:20 +01:00
NodeTracker tracker;
Printer printer{out, tracker, typeGraph.size()};
TypeGraph: Switch from pointers to references References must always have a value, so are semantically clearer than pointers for variables which must always be set. No functional changes.
2023-07-05 10:33:44 +01:00
printer.print(type);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
return out.str();
}
void DrgnParserTest::test(std::string_view function,
std::string_view expected,
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options) {
auto actual = run(function, options);
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
expected.remove_prefix(1); // Remove initial '\n'
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
EXPECT_EQ(expected, actual);
}
DrgnParser: Options should default to false We only want to do the extra work if it's explicitly requested. chaseRawPointers is already explicitly requested whenever it's needed and readEnumValues currently isn't needed at all.
2023-09-19 17:42:26 +01:00
void DrgnParserTest::test(std::string_view function,
std::string_view expected) {
// Enable options in unit tests so we get more coverage
DrgnParserOptions options = {
.chaseRawPointers = true,
.readEnumValues = true,
};
test(function, expected, options);
}
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
std::pair<size_t, size_t> globMatch(std::string_view pattern,
std::string_view str) {
size_t i = 0;
size_t j = 0;
size_t prevWildcardIdx = -1;
size_t backtrackIdx = -1;
size_t patternLineStart = 0;
size_t strLineStart = 0;
while (i < str.size()) {
if (i + 1 < str.size() && str[i] == '\n')
strLineStart = i + 1;
if (j + 1 < pattern.size() && pattern[j] == '\n')
patternLineStart = j + 1;
if (j < pattern.size() && str[i] == pattern[j]) {
// Exact character match
i++;
j++;
} else if (j < pattern.size() && pattern[j] == '*') {
// Wildcard
backtrackIdx = i + 1;
prevWildcardIdx = j++;
} else if (prevWildcardIdx != static_cast<size_t>(-1)) {
// No match, backtrack to previous wildcard
i = ++backtrackIdx;
j = prevWildcardIdx + 1;
} else {
// No match
return {patternLineStart, strLineStart};
}
}
while (j < pattern.size() && pattern[j] == '*') {
j++;
}
// If the pattern has been fully consumed then it's a match
return {j, i};
}
std::string prefixLines(std::string_view str,
std::string_view prefix,
size_t maxLen) {
std::string res;
res += prefix;
for (size_t i = 0; i < maxLen && i < str.size(); i++) {
char c = str[i];
res += c;
if (c == '\n') {
res += prefix;
}
}
if (str.size() > maxLen)
res += "...";
return res;
}
void DrgnParserTest::testGlob(std::string_view function,
std::string_view expected,
DrgnParserOptions options) {
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
auto actual = run(function, options);
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
expected.remove_prefix(1); // Remove initial '\n'
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
auto [expectedIdx, actualIdx] = globMatch(expected, actual);
if (expected.size() != expectedIdx) {
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
ADD_FAILURE() << prefixLines(expected.substr(expectedIdx), "-", 10000)
<< "\n"
<< prefixLines(actual.substr(actualIdx), "+", 10000);
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
}
void DrgnParserTest::testMultiCompiler(
std::string_view function,
[[maybe_unused]] std::string_view expectedClang,
[[maybe_unused]] std::string_view expectedGcc,
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options) {
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
#if defined(__clang__)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
test(function, expectedClang, options);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
#else
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
test(function, expectedGcc, options);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
#endif
}
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
void DrgnParserTest::testMultiCompilerGlob(
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
std::string_view function,
[[maybe_unused]] std::string_view expectedClang,
[[maybe_unused]] std::string_view expectedGcc,
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options) {
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
#if defined(__clang__)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
testGlob(function, expectedClang, options);
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
#else
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
testGlob(function, expectedGcc, options);
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
#endif
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
TEST_F(DrgnParserTest, SimpleStruct) {
test("oid_test_case_simple_struct", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: SimpleStruct [ns_simple::SimpleStruct] (size: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 4)
Primitive: int8_t
Member: c (offset: 8)
Primitive: int64_t
)");
}
TEST_F(DrgnParserTest, SimpleClass) {
test("oid_test_case_simple_class", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: SimpleClass [ns_simple::SimpleClass] (size: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 4)
Primitive: int8_t
Member: c (offset: 8)
Primitive: int64_t
)");
}
TEST_F(DrgnParserTest, SimpleUnion) {
test("oid_test_case_simple_union", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Union: SimpleUnion [ns_simple::SimpleUnion] (size: 8)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 0)
Primitive: int8_t
Member: c (offset: 0)
Primitive: int64_t
)");
}
TEST_F(DrgnParserTest, Inheritance) {
test("oid_test_case_inheritance_access_public", R"(
oil v2
2023-08-16 20:40:36 +01:00
[4] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: Public [ns_inheritance_access::Public] (size: 8)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Parent (offset: 0)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[1] Class: Base [ns_inheritance_access::Base] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: base_int (offset: 0)
oil v2
2023-08-16 20:40:36 +01:00
[3] Typedef: int32_t
[2] Typedef: __int32_t
Primitive: int32_t
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: public_int (offset: 4)
oil v2
2023-08-16 20:40:36 +01:00
[3]
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
TEST_F(DrgnParserTest, InheritanceMultiple) {
test("oid_test_case_inheritance_multiple_a", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[6] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Derived_2 [ns_inheritance_multiple::Derived_2] (size: 24)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Parent (offset: 0)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[1] Struct: Base_1 [ns_inheritance_multiple::Base_1] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Parent (offset: 4)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[2] Struct: Derived_1 [ns_inheritance_multiple::Derived_1] (size: 12)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Parent (offset: 0)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[3] Struct: Base_2 [ns_inheritance_multiple::Base_2] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: b (offset: 0)
Primitive: int32_t
Parent (offset: 4)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[4] Struct: Base_3 [ns_inheritance_multiple::Base_3] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: c (offset: 0)
Primitive: int32_t
Member: d (offset: 8)
Primitive: int32_t
Parent (offset: 16)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[5] Struct: Base_4 [ns_inheritance_multiple::Base_4] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: e (offset: 0)
Primitive: int32_t
Member: f (offset: 20)
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, Container) {
clang-format: disable bin packing Bin packing often makes code hard to read. Disable it entirely. Test plan: - CI
2023-11-13 16:50:09 +00:00
testMultiCompilerGlob("oid_test_case_std_vector_int_empty",
R"(
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
[13] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: vector<int, std::allocator<int> > [std::vector<int, std::allocator<int> >] (size: 24)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Param
Primitive: int32_t
Param
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[1] Class: allocator<int> [std::allocator<int>] (size: 1)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Param
Primitive: int32_t
Parent (offset: 0)
[3] Typedef: __allocator_base<int>
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[2] Class: new_allocator<int> [__gnu_cxx::new_allocator<int>] (size: 1)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Param
Primitive: int32_t
Function: new_allocator
Function: new_allocator
Function: allocate
Function: deallocate
Function: _M_max_size
Function: allocator
Function: allocator
Function: operator=
Function: ~allocator
Function: allocate
Function: deallocate
Parent (offset: 0)
*
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)",
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
R"(
[9] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: vector<int, std::allocator<int> > [std::vector<int, std::allocator<int> >] (size: 24)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Param
Primitive: int32_t
Param
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[1] Class: allocator<int> [std::allocator<int>] (size: 1)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Parent (offset: 0)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[2] Class: new_allocator<int> [__gnu_cxx::new_allocator<int>] (size: 1)
TypeGraph: Stop identifying containers in DrgnParser Leave it to the new mutator pass IdentifyContainers to replace Class nodes with Container nodes where appropriate. This will allow us to run passes over the type graph before identifying containers, and therefore before we have lost information about the internal details of the container (e.g. alignment of member variables).
2023-11-02 13:18:32 +00:00
Param
Primitive: int32_t
Function: new_allocator
Function: new_allocator
Function: allocate
Function: deallocate
Function: _M_max_size
Function: allocator
Function: allocator
Function: operator=
Function: ~allocator
Function: allocate
Function: deallocate
Parent (offset: 0)
*
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
// TODO test vector with custom allocator
TEST_F(DrgnParserTest, Enum) {
test("oid_test_case_enums_scoped", R"(
Enum: ScopedEnum (size: 4)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enumerator: 0:CaseA
Enumerator: 1:CaseB
Enumerator: 2:CaseC
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
codegen: generate enums as enum class
2023-09-19 17:25:58 +01:00
TEST_F(DrgnParserTest, EnumUint8) {
test("oid_test_case_enums_scoped_uint8", R"(
Enum: ScopedEnumUint8 (size: 1)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enumerator: 2:CaseA
Enumerator: 3:CaseB
Enumerator: 4:CaseC
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
TEST_F(DrgnParserTest, UnscopedEnum) {
test("oid_test_case_enums_unscoped", R"(
Enum: UNSCOPED_ENUM (size: 4)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enumerator: -2:CASE_B
Enumerator: 5:CASE_A
Enumerator: 20:CASE_C
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
TEST_F(DrgnParserTest, EnumNoValues) {
DrgnParserOptions options{
.readEnumValues = false,
};
clang-format: disable bin packing Bin packing often makes code hard to read. Disable it entirely. Test plan: - CI
2023-11-13 16:50:09 +00:00
test("oid_test_case_enums_scoped",
R"(
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enum: ScopedEnum (size: 4)
)",
options);
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
TEST_F(DrgnParserTest, Typedef) {
test("oid_test_case_typedefs_c_style", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Typedef: TdUInt64
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
[1] Typedef: uint64_t
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[0] Typedef: __uint64_t
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: uint64_t
)");
}
TEST_F(DrgnParserTest, Using) {
test("oid_test_case_typedefs_using", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Typedef: UsingUInt64
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
[1] Typedef: uint64_t
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[0] Typedef: __uint64_t
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: uint64_t
)");
}
TEST_F(DrgnParserTest, ArrayMember) {
test("oid_test_case_arrays_member_int10", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Foo10 [ns_arrays::Foo10] (size: 40)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: arr (offset: 0)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Array: (length: 10)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, ArrayRef) {
test("oid_test_case_arrays_ref_int10", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
[0] Array: (length: 10)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, ArrayDirect) {
test("oid_test_case_arrays_direct_int10", R"(
[0] Pointer
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, Pointer) {
test("oid_test_case_pointers_struct_primitive_ptrs", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[3] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: PrimitivePtrs [ns_pointers::PrimitivePtrs] (size: 24)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 8)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: int32_t
Member: c (offset: 16)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Pointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: void
)");
}
TEST_F(DrgnParserTest, PointerNoFollow) {
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
DrgnParserOptions options{
.chaseRawPointers = false,
};
clang-format: disable bin packing Bin packing often makes code hard to read. Disable it entirely. Test plan: - CI
2023-11-13 16:50:09 +00:00
test("oid_test_case_pointers_struct_primitive_ptrs",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: PrimitivePtrs [ns_pointers::PrimitivePtrs] (size: 24)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 8)
make StubbedPointer an explicit C++ type
2023-09-21 16:57:08 +01:00
Primitive: StubbedPointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: c (offset: 16)
make StubbedPointer an explicit C++ type
2023-09-21 16:57:08 +01:00
Primitive: StubbedPointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)",
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
options);
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
}
TEST_F(DrgnParserTest, PointerIncomplete) {
test("oid_test_case_pointers_incomplete_raw", R"(
incomplete: name type in compiler errors Summary: We have a good type representation in the Type Graph of an incomplete type and the underlying type that represents. However, this incomplete type still ends up in the generated code as `void` which loses information. For example, a container that can't contain void may fail to compile because it was initialised with `void` but really its because the type it was supposed to be initialised with (say, `Foo`) had incomplete debug information. This change identifies that a type is incomplete in the output by generating it as an incomplete type `struct Incomplete<struct Foo>`. This allows us to name the type correctly in the TreeBuilder output and filter for incomplete types, as well as getting appropriate compiler errors if it mustn't be incomplete. Test Plan: - CI - Added a unit test to namegen. - Enabled and added an extra pointers_incomplete test. This change is tricky to test because it isn't really user visible. The types still use their `inputName` which is unchanged in any successful output - this change is used so the compiler fails with a more detailed error.
2024-01-04 17:01:52 +00:00
[1] Pointer
[0] Incomplete: [IncompleteType]
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
TEST_F(DrgnParserTest, Cycle) {
test("oid_test_case_cycles_raw_ptr", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: RawNode [ns_cycles::RawNode] (size: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: value (offset: 0)
Primitive: int32_t
Member: next (offset: 8)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
[0]
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
)");
}
TEST_F(DrgnParserTest, ClassTemplateInt) {
test("oid_test_case_templates_int", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: TemplatedClass1<int> [ns_templates::TemplatedClass1<int>] (size: 4)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Param
Primitive: int32_t
Member: val (offset: 0)
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, ClassTemplateTwo) {
test("oid_test_case_templates_two", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[3] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: TemplatedClass2<ns_templates::Foo, int> [ns_templates::TemplatedClass2<ns_templates::Foo, int>] (size: 12)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Param
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[1] Struct: Foo [ns_templates::Foo] (size: 8)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 4)
Primitive: int32_t
Param
Primitive: int32_t
Member: tc1 (offset: 0)
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[2] Class: TemplatedClass1<ns_templates::Foo> [ns_templates::TemplatedClass1<ns_templates::Foo>] (size: 8)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Param
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1]
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: val (offset: 0)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1]
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: val2 (offset: 8)
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, ClassTemplateValue) {
test("oid_test_case_templates_value", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[2] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: TemplatedClassVal<3> [ns_templates::TemplatedClassVal<3>] (size: 12)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Param
Value: 3
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Primitive: int32_t
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: arr (offset: 0)
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Array: (length: 3)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Primitive: int32_t
)");
}
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
TEST_F(DrgnParserTest, TemplateEnumValue) {
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
testMultiCompilerGlob("oid_test_case_enums_params_scoped_enum_val",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: MyClass<ns_enums_params::MyNS::ScopedEnum::One> [ns_enums_params::MyClass<ns_enums_params::MyNS::ScopedEnum::One>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::ScopedEnum::One
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: ScopedEnum (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)",
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: MyClass<(ns_enums_params::MyNS::ScopedEnum)1> [ns_enums_params::MyClass<(ns_enums_params::MyNS::ScopedEnum)1>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::ScopedEnum::One
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: ScopedEnum (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)");
}
TEST_F(DrgnParserTest, TemplateEnumValueGaps) {
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
testMultiCompilerGlob("oid_test_case_enums_params_scoped_enum_val_gaps",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: ClassGaps<ns_enums_params::MyNS::EnumWithGaps::Twenty> [ns_enums_params::ClassGaps<ns_enums_params::MyNS::EnumWithGaps::Twenty>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::EnumWithGaps::Twenty
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: EnumWithGaps (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)",
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: ClassGaps<(ns_enums_params::MyNS::EnumWithGaps)20> [ns_enums_params::ClassGaps<(ns_enums_params::MyNS::EnumWithGaps)20>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::EnumWithGaps::Twenty
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: EnumWithGaps (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)");
}
TEST_F(DrgnParserTest, TemplateEnumValueNegative) {
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
testMultiCompilerGlob("oid_test_case_enums_params_scoped_enum_val_negative",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: ClassGaps<ns_enums_params::MyNS::EnumWithGaps::MinusTwo> [ns_enums_params::ClassGaps<ns_enums_params::MyNS::EnumWithGaps::MinusTwo>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::EnumWithGaps::MinusTwo
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: EnumWithGaps (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)",
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: ClassGaps<(ns_enums_params::MyNS::EnumWithGaps)-2> [ns_enums_params::ClassGaps<(ns_enums_params::MyNS::EnumWithGaps)-2>] (size: 4)
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
Param
Value: ns_enums_params::MyNS::EnumWithGaps::MinusTwo
DrgnParser: store type of template param value
2023-08-16 20:44:58 +01:00
Enum: EnumWithGaps (size: 4)
DrgnParserTest: Replace testContains with testGlob It now uses '*' as a wildcard character for more powerful matching of type graphs.
2023-11-02 15:54:56 +00:00
*
DrgnParser: Handle enum values in template params We want to use the fully qualified name for scoped enums to keep the C++ compiler happy. When a parameter expects an enum value, we must supply an enum value and not its underlying integer value. Before: isset_bitset<1, 0> After: isset_bitset<1, apache::thrift::detail::IssetBitsetOption::Unpacked>
2023-06-23 16:56:39 +01:00
)");
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
// TODO
// TEST_F(DrgnParserTest, ClassFunctions) {
// test("TestClassFunctions", R"(
//[0] Pointer
//[1] Class: ClassFunctions (size: 4)
// Member: memberA (offset: 0)
// Primitive: int32_t
// Function: foo (virtuality: 0)
// Function: bar (virtuality: 0)
//)");
//}
TEST_F(DrgnParserTest, StructAlignment) {
GTEST_SKIP() << "Alignment not reported by drgn yet";
test("oid_test_case_alignment_struct", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Align16 [ns_alignment::Align16] (size: 16, align: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: c (offset: 0)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, MemberAlignment) {
GTEST_SKIP() << "Alignment not reported by drgn yet";
test("oid_test_case_alignment_member_alignment", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: MemberAlignment [ns_alignment::MemberAlignment] (size: 64)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: c (offset: 0)
Primitive: int8_t
Member: c32 (offset: 32, align: 32)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, VirtualFunctions) {
clang-format: disable bin packing Bin packing often makes code hard to read. Disable it entirely. Test plan: - CI
2023-11-13 16:50:09 +00:00
testMultiCompiler("oid_test_case_inheritance_polymorphic_a_as_a",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: A [ns_inheritance_polymorphic::A] (size: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: _vptr$A (offset: 0)
make StubbedPointer an explicit C++ type
2023-09-21 16:57:08 +01:00
Primitive: StubbedPointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: int_a (offset: 8)
Primitive: int32_t
Function: ~A (virtual)
Function: myfunc (virtual)
Function: A
Function: A
)",
R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Class: A [ns_inheritance_polymorphic::A] (size: 16)
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: _vptr.A (offset: 0)
make StubbedPointer an explicit C++ type
2023-09-21 16:57:08 +01:00
Primitive: StubbedPointer
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
Member: int_a (offset: 8)
Primitive: int32_t
Function: operator=
Function: A
Function: A
Function: ~A (virtual)
Function: myfunc (virtual)
)");
}
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
TEST_F(DrgnParserTest, BitfieldsSingle) {
test("oid_test_case_bitfields_single", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Single [ns_bitfields::Single] (size: 4)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: bitfield (offset: 0, bitsize: 3)
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, BitfieldsWithinBytes) {
test("oid_test_case_bitfields_within_bytes", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: WithinBytes [ns_bitfields::WithinBytes] (size: 2)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: a (offset: 0, bitsize: 3)
Primitive: int8_t
Member: b (offset: 0.375, bitsize: 5)
Primitive: int8_t
Member: c (offset: 1, bitsize: 7)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, BitfieldsStraddleBytes) {
test("oid_test_case_bitfields_straddle_bytes", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: StraddleBytes [ns_bitfields::StraddleBytes] (size: 3)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: a (offset: 0, bitsize: 7)
Primitive: int8_t
Member: b (offset: 1, bitsize: 7)
Primitive: int8_t
Member: c (offset: 2, bitsize: 2)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, BitfieldsMixed) {
test("oid_test_case_bitfields_mixed", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Mixed [ns_bitfields::Mixed] (size: 12)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: a (offset: 0)
Primitive: int32_t
Member: b (offset: 4, bitsize: 4)
Primitive: int8_t
Member: c (offset: 4.5, bitsize: 12)
Primitive: int16_t
Member: d (offset: 6)
Primitive: int8_t
Member: e (offset: 8, bitsize: 22)
Primitive: int32_t
)");
}
TEST_F(DrgnParserTest, BitfieldsMoreBitsThanType) {
GTEST_SKIP() << "drgn errors out";
test("oid_test_case_bitfields_more_bits_than_type", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: MoreBitsThanType [ns_bitfields::MoreBitsThanType] (size: 4)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: a (offset: 0, bitsize: 8)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, BitfieldsZeroBits) {
test("oid_test_case_bitfields_zero_bits", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: ZeroBits [ns_bitfields::ZeroBits] (size: 2)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: b1 (offset: 0, bitsize: 3)
Primitive: int8_t
Member: b2 (offset: 1, bitsize: 2)
Primitive: int8_t
)");
}
TEST_F(DrgnParserTest, BitfieldsEnum) {
test("oid_test_case_bitfields_enum", R"(
Printer: Use NodeIds for stable IDs when printing graphs This removes Printer's legacy behaviour of generating an ID for each node as it gets printed. This old method meant that if new nodes were added to or removed from a graph, every ID after the new/removed node would change. Now IDs are stable so it is easier to follow specific nodes through multiple transformation passes in CodeGen.
2023-08-02 11:11:52 +01:00
[1] Pointer
TypeGraph: Make Class types use fully qualified names as their input names This will ensure we continue to get fully qualified names in user-visible output when we switch to CodeGen v2.
2023-12-15 12:25:30 +00:00
[0] Struct: Enum [ns_bitfields::Enum] (size: 4)
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: e (offset: 0, bitsize: 2)
Enum: MyEnum (size: 4)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enumerator: 0:One
Enumerator: 1:Two
Enumerator: 2:Three
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
Member: f (offset: 0.25, bitsize: 4)
Enum: MyEnum (size: 4)
TypeGraph: Add option to record enumerator names
2023-08-16 19:15:09 +01:00
Enumerator: 0:One
Enumerator: 1:Two
Enumerator: 2:Three
TypeGraph: Support bitfields - Change member and parent offsets to work in bits, not bytes - Printer still displays offsets in bytes, with decimals when using bitfields - AddPadding: Don't pad bitfields - CodeGen: Emit code for bitfields
2023-06-29 09:15:03 +01:00
)");
}
TypeGraph: Add core code This code mostly works, but is obviously not complete. This commit just adds the code and tests, but does not enable it in OID or OIL.
2023-05-26 13:52:09 +01:00
// TODO test virtual classes
Reference in New Issue Copy Permalink