Step 400: add self-hosting test harness with coverage tracking

Infrastructure to feed Whetstone .h files through the C++ parser pipeline.
Tracks construct coverage (classes, functions, namespaces, type aliases),
graceful degradation for unparseable sections, and coverage reporting.
12/12 tests passing.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Bill
2026-02-16 14:11:44 -07:00
parent 0f4b7c8270
commit 22ee1f75c3
3 changed files with 575 additions and 0 deletions

View File

@@ -2479,4 +2479,13 @@ target_link_libraries(step399_test PRIVATE
tree_sitter_javascript tree_sitter_typescript
tree_sitter_java tree_sitter_rust tree_sitter_go)
add_executable(step400_test tests/step400_test.cpp)
target_include_directories(step400_test PRIVATE src)
target_link_libraries(step400_test PRIVATE
nlohmann_json::nlohmann_json
unofficial::tree-sitter::tree-sitter
tree_sitter_python tree_sitter_cpp tree_sitter_elisp
tree_sitter_javascript tree_sitter_typescript
tree_sitter_java tree_sitter_rust tree_sitter_go)
# Step 12: Dear ImGui shell scaffolding created (main.cpp exists but not built due to dependencies)

View File

@@ -0,0 +1,308 @@
#pragma once
// Step 400: Self-Hosting Test Harness
//
// Infrastructure to feed actual Whetstone .h files through the pipeline.
// Compare parsed AST against expected structure. Track coverage.
// Graceful degradation: unparseable sections become opaque blocks, not errors.
#include <string>
#include <vector>
#include <map>
#include <fstream>
#include <sstream>
#include <filesystem>
#include "ast/ASTNode.h"
#include "ast/Module.h"
#include "ast/Function.h"
#include "ast/Parser.h"
#include "ast/ClassDeclaration.h"
#include "ast/EnumNamespaceNodes.h"
#include "ast/CppAdvancedNodes.h"
// --- Coverage tracking ---
struct ConstructCoverage {
std::string constructType; // e.g. "class", "function", "struct", "enum"
int expected = 0;
int parsed = 0;
int opaque = 0; // unparseable sections that became opaque blocks
double ratio() const {
return expected > 0 ? static_cast<double>(parsed) / expected : 1.0;
}
};
struct SelfHostResult {
std::string filePath;
bool parseSuccess = false;
std::unique_ptr<Module> ast;
std::vector<ConstructCoverage> coverage;
std::vector<std::string> warnings;
int totalExpected = 0;
int totalParsed = 0;
int totalOpaque = 0;
double overallCoverage() const {
return totalExpected > 0 ? static_cast<double>(totalParsed) / totalExpected : 1.0;
}
};
// --- Expected structure declaration ---
struct ExpectedConstruct {
std::string type; // "class", "function", "struct", "namespace", "type_alias"
std::string name; // expected name (empty = don't check)
};
struct ExpectedStructure {
std::string filePath;
std::vector<ExpectedConstruct> constructs;
};
// --- Self-Hosting Harness ---
class SelfHostHarness {
public:
// Read a file from disk and return its contents
static std::string readFile(const std::string& path) {
std::ifstream ifs(path);
if (!ifs.is_open()) return "";
std::ostringstream ss;
ss << ifs.rdbuf();
return ss.str();
}
// Parse a Whetstone .h file through the C++ parser
static SelfHostResult parseFile(const std::string& path) {
SelfHostResult result;
result.filePath = path;
std::string source = readFile(path);
if (source.empty()) {
result.warnings.push_back("File not found or empty: " + path);
return result;
}
result.ast = TreeSitterParser::parseCpp(source);
result.parseSuccess = (result.ast != nullptr);
if (!result.parseSuccess) {
result.warnings.push_back("Parse returned null for: " + path);
}
return result;
}
// Parse source string directly (for testing without file I/O)
static SelfHostResult parseSource(const std::string& source,
const std::string& label = "<inline>") {
SelfHostResult result;
result.filePath = label;
if (source.empty()) {
result.warnings.push_back("Empty source provided");
return result;
}
result.ast = TreeSitterParser::parseCpp(source);
result.parseSuccess = (result.ast != nullptr);
return result;
}
// Compare parsed AST against expected structure
static void checkExpected(SelfHostResult& result,
const ExpectedStructure& expected) {
if (!result.ast) {
result.totalExpected = static_cast<int>(expected.constructs.size());
result.totalOpaque = result.totalExpected;
return;
}
// Group expected constructs by type
std::map<std::string, std::vector<std::string>> byType;
for (const auto& c : expected.constructs) {
byType[c.type].push_back(c.name);
}
result.totalExpected = static_cast<int>(expected.constructs.size());
for (const auto& [type, names] : byType) {
ConstructCoverage cov;
cov.constructType = type;
cov.expected = static_cast<int>(names.size());
if (type == "class" || type == "struct") {
auto& classes = result.ast->getChildren("classes");
for (const auto& name : names) {
bool found = false;
for (auto* c : classes) {
auto* cls = static_cast<ClassDeclaration*>(c);
if (name.empty() || cls->name == name) {
found = true;
break;
}
}
if (found) cov.parsed++;
else cov.opaque++;
}
}
else if (type == "function") {
auto& fns = result.ast->getChildren("functions");
for (const auto& name : names) {
bool found = false;
for (auto* f : fns) {
auto* fn = static_cast<Function*>(f);
if (name.empty() || fn->name == name) {
found = true;
break;
}
}
if (found) cov.parsed++;
else cov.opaque++;
}
}
else if (type == "namespace") {
auto& stmts = result.ast->getChildren("statements");
for (const auto& name : names) {
bool found = false;
for (auto* s : stmts) {
if (s->conceptType == "NamespaceDeclaration") {
auto* ns = static_cast<NamespaceDeclaration*>(s);
if (name.empty() || ns->name == name) {
found = true;
break;
}
}
}
if (found) cov.parsed++;
else cov.opaque++;
}
}
else if (type == "type_alias") {
auto& stmts = result.ast->getChildren("statements");
for (const auto& name : names) {
bool found = false;
for (auto* s : stmts) {
if (s->conceptType == "TypeAlias") {
auto* ta = static_cast<TypeAlias*>(s);
if (name.empty() || ta->aliasName == name) {
found = true;
break;
}
}
}
if (found) cov.parsed++;
else cov.opaque++;
}
}
else if (type == "enum") {
auto& stmts = result.ast->getChildren("statements");
for (const auto& name : names) {
bool found = false;
for (auto* s : stmts) {
if (s->conceptType == "EnumDeclaration") {
auto* en = static_cast<EnumDeclaration*>(s);
if (name.empty() || en->name == name) {
found = true;
break;
}
}
}
if (found) cov.parsed++;
else cov.opaque++;
}
}
else {
// Unknown type — mark all as opaque
cov.opaque = cov.expected;
result.warnings.push_back("Unknown construct type: " + type);
}
result.totalParsed += cov.parsed;
result.totalOpaque += cov.opaque;
result.coverage.push_back(std::move(cov));
}
}
// Count all classes found in parsed AST
static int countClasses(const Module* mod) {
if (!mod) return 0;
return static_cast<int>(mod->getChildren("classes").size());
}
// Count all functions found in parsed AST
static int countFunctions(const Module* mod) {
if (!mod) return 0;
return static_cast<int>(mod->getChildren("functions").size());
}
// Count all statements of a specific concept type
static int countStatements(const Module* mod, const std::string& conceptType) {
if (!mod) return 0;
int count = 0;
for (auto* s : mod->getChildren("statements")) {
if (s->conceptType == conceptType) count++;
}
return count;
}
// Get list of all class names
static std::vector<std::string> getClassNames(const Module* mod) {
std::vector<std::string> names;
if (!mod) return names;
for (auto* c : mod->getChildren("classes")) {
names.push_back(static_cast<ClassDeclaration*>(c)->name);
}
return names;
}
// Get list of all function names
static std::vector<std::string> getFunctionNames(const Module* mod) {
std::vector<std::string> names;
if (!mod) return names;
for (auto* f : mod->getChildren("functions")) {
names.push_back(static_cast<Function*>(f)->name);
}
return names;
}
// Check if a specific construct was parsed (by type and name)
static bool hasParsedConstruct(const Module* mod,
const std::string& type,
const std::string& name) {
if (!mod) return false;
if (type == "class" || type == "struct") {
for (auto* c : mod->getChildren("classes")) {
if (static_cast<ClassDeclaration*>(c)->name == name) return true;
}
} else if (type == "function") {
for (auto* f : mod->getChildren("functions")) {
if (static_cast<Function*>(f)->name == name) return true;
}
}
return false;
}
// Generate a coverage report string
static std::string coverageReport(const SelfHostResult& result) {
std::ostringstream out;
out << "Self-Host Report: " << result.filePath << "\n";
out << " Parse success: " << (result.parseSuccess ? "yes" : "no") << "\n";
out << " Overall coverage: " << (result.overallCoverage() * 100) << "%"
<< " (" << result.totalParsed << "/" << result.totalExpected << ")\n";
for (const auto& cov : result.coverage) {
out << " " << cov.constructType << ": "
<< cov.parsed << "/" << cov.expected
<< " (" << (cov.ratio() * 100) << "%)";
if (cov.opaque > 0) out << " [" << cov.opaque << " opaque]";
out << "\n";
}
for (const auto& w : result.warnings) {
out << " WARNING: " << w << "\n";
}
return out.str();
}
};

View File

@@ -0,0 +1,258 @@
// Step 400: Self-Hosting Test Harness (12 tests)
#include <cassert>
#include <iostream>
#include <string>
#include "SelfHostHarness.h"
int main() {
int passed = 0;
// Test 1: Parse simple struct from source string
{
std::string src = R"(
struct Foo {
int x;
std::string name;
};
)";
auto result = SelfHostHarness::parseSource(src, "test_struct");
assert(result.parseSuccess);
assert(result.ast != nullptr);
std::cout << "PASS: test 1 — parse simple struct\n";
passed++;
}
// Test 2: Parse class + free function from source string
{
std::string src = R"(
class Widget {
public:
void draw() { }
};
void process(Widget* w) {
w->draw();
}
)";
auto result = SelfHostHarness::parseSource(src, "test_class_func");
assert(result.parseSuccess);
assert(SelfHostHarness::countClasses(result.ast.get()) >= 1);
assert(SelfHostHarness::countFunctions(result.ast.get()) >= 1);
std::cout << "PASS: test 2 — parse class + free function\n";
passed++;
}
// Test 3: Coverage tracking — all constructs found
{
std::string src = R"(
struct Point { int x; int y; };
void doSomething() { }
)";
auto result = SelfHostHarness::parseSource(src, "test_coverage");
ExpectedStructure expected;
expected.filePath = "test_coverage";
expected.constructs = {
{"struct", "Point"},
{"function", "doSomething"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalExpected == 2);
assert(result.totalParsed == 2);
assert(result.totalOpaque == 0);
assert(result.overallCoverage() == 1.0);
std::cout << "PASS: test 3 — coverage tracking, all found\n";
passed++;
}
// Test 4: Coverage tracking — missing construct is opaque
{
std::string src = R"(
void foo() { }
)";
auto result = SelfHostHarness::parseSource(src, "test_missing");
ExpectedStructure expected;
expected.filePath = "test_missing";
expected.constructs = {
{"function", "foo"},
{"class", "MissingClass"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalExpected == 2);
assert(result.totalParsed == 1);
assert(result.totalOpaque == 1);
assert(result.overallCoverage() == 0.5);
std::cout << "PASS: test 4 — missing construct is opaque\n";
passed++;
}
// Test 5: Coverage report generation
{
std::string src = R"(
class Foo { };
void bar() { }
)";
auto result = SelfHostHarness::parseSource(src, "report_test");
ExpectedStructure expected;
expected.filePath = "report_test";
expected.constructs = {
{"class", "Foo"},
{"function", "bar"}
};
SelfHostHarness::checkExpected(result, expected);
std::string report = SelfHostHarness::coverageReport(result);
assert(!report.empty());
assert(report.find("report_test") != std::string::npos);
assert(report.find("100%") != std::string::npos);
std::cout << "PASS: test 5 — coverage report generation\n";
passed++;
}
// Test 6: Graceful degradation — empty source doesn't crash
{
auto result = SelfHostHarness::parseSource("", "empty");
assert(!result.parseSuccess || result.ast == nullptr ||
result.warnings.size() > 0);
std::cout << "PASS: test 6 — graceful degradation on empty source\n";
passed++;
}
// Test 7: hasParsedConstruct lookup
{
std::string src = R"(
class Pipeline { };
void run() { }
)";
auto result = SelfHostHarness::parseSource(src, "lookup");
assert(SelfHostHarness::hasParsedConstruct(result.ast.get(), "class", "Pipeline"));
assert(SelfHostHarness::hasParsedConstruct(result.ast.get(), "function", "run"));
assert(!SelfHostHarness::hasParsedConstruct(result.ast.get(), "class", "Missing"));
std::cout << "PASS: test 7 — hasParsedConstruct lookup\n";
passed++;
}
// Test 8: getClassNames and getFunctionNames
{
std::string src = R"(
class Alpha { };
class Beta { };
void one() { }
void two() { }
)";
auto result = SelfHostHarness::parseSource(src, "names");
auto classNames = SelfHostHarness::getClassNames(result.ast.get());
auto funcNames = SelfHostHarness::getFunctionNames(result.ast.get());
assert(classNames.size() >= 2);
assert(funcNames.size() >= 2);
std::cout << "PASS: test 8 — getClassNames and getFunctionNames\n";
passed++;
}
// Test 9: Parse Whetstone-style header pattern (struct + inline function)
{
std::string src = R"(
#pragma once
#include <string>
#include <vector>
struct ConflictInfo {
std::string type1;
std::string type2;
std::string message;
};
inline void checkConflicts(const std::vector<ConflictInfo>& items) {
for (const auto& item : items) {
// process
}
}
)";
auto result = SelfHostHarness::parseSource(src, "whetstone_pattern");
assert(result.parseSuccess);
// Should find at least the struct and the function
ExpectedStructure expected;
expected.constructs = {
{"struct", "ConflictInfo"},
{"function", "checkConflicts"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalParsed >= 1); // At least struct or function found
std::cout << "PASS: test 9 — Whetstone-style header pattern\n";
passed++;
}
// Test 10: Type alias coverage tracking
{
std::string src = R"(
using json = nlohmann::json;
using StringVec = std::vector<std::string>;
)";
auto result = SelfHostHarness::parseSource(src, "alias_test");
assert(result.parseSuccess);
int aliases = SelfHostHarness::countStatements(result.ast.get(), "TypeAlias");
assert(aliases == 2);
ExpectedStructure expected;
expected.constructs = {
{"type_alias", "json"},
{"type_alias", "StringVec"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalParsed == 2);
std::cout << "PASS: test 10 — type alias coverage tracking\n";
passed++;
}
// Test 11: Namespace coverage tracking
{
std::string src = R"(
namespace detail {
void helper() { }
}
)";
auto result = SelfHostHarness::parseSource(src, "ns_test");
assert(result.parseSuccess);
ExpectedStructure expected;
expected.constructs = {
{"namespace", "detail"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalParsed == 1);
std::cout << "PASS: test 11 — namespace coverage tracking\n";
passed++;
}
// Test 12: Mixed coverage with partial success
{
std::string src = R"(
class Validator {
public:
void validate() { }
};
void helperFunc() { }
)";
auto result = SelfHostHarness::parseSource(src, "mixed_test");
ExpectedStructure expected;
expected.constructs = {
{"class", "Validator"},
{"function", "helperFunc"},
{"class", "NonExistent"},
{"function", "ghostFunc"}
};
SelfHostHarness::checkExpected(result, expected);
assert(result.totalExpected == 4);
assert(result.totalParsed == 2);
assert(result.totalOpaque == 2);
assert(result.overallCoverage() == 0.5);
std::string report = SelfHostHarness::coverageReport(result);
assert(report.find("50%") != std::string::npos);
std::cout << "PASS: test 12 — mixed coverage with partial success\n";
passed++;
}
std::cout << "\nStep 400 result: " << passed << "/12 tests passed\n";
return (passed == 12) ? 0 : 1;
}