Remove STL associative containers from builder.hpp and builder.cpp

.pupignore

@@ -7,3 +7,4 @@ test/e2e/fixtures/
 
 # Example projects - require external dependencies
 examples/
+.claude/

COMPATIBILITY.md

@@ -205,13 +205,7 @@ Features in putup that extend beyond tup:
 
 **Affected**: MSVC 19.38+ (Visual Studio 2022 17.8) with `/Zc:nrvo /std:c++20 /O2`
 
-`Graph::make_graph()` returns a `Graph` by value. The `Graph` owns a `StringPool` and an `unordered_map` whose hash/equal functors hold a `StringPool*` pointing into the same object. Under NRVO the returned object is constructed directly in the caller's storage, so the pointers remain valid.
-
-MSVC does not apply NRVO here. It constructs the `Graph` locally, then move-constructs into the caller. After the move, `StringPool` lives at a new address but `DirNameKeyHash`/`DirNameKeyEqual` still hold the old pointer — any subsequent map lookup dereferences freed memory.
-
-The MS docs list only two NRVO exclusions (multiple return variables, `throw` in scope), neither of which applies. This appears to be an undocumented limitation triggered when a function re-assigns an `unordered_map` member whose hash/equal functors contain pointers to the local object's own fields.
-
-**Status**: MSVC CI jobs are disabled. MinGW (GCC on Windows) is unaffected and remains in CI.
+**Status**: Fixed. The root cause was `DirNameKeyHash`/`DirNameKeyEqual` functors holding raw `StringPool*` pointers into the same `Graph` object. When MSVC failed to apply NRVO on `make_graph()`, the move left dangling pointers. The `DirNameKey` hash map has been replaced with `std::vector<SortedPairVec> dir_children` (per-directory sorted arrays), eliminating the pointer coupling entirely. MSVC CI jobs remain disabled for other reasons (MinGW/GCC is the Windows CI target).
 
 ## Reporting Issues
 

DESIGN.md

@@ -458,7 +458,7 @@ struct CommandNode {
     StringId instruction_id = StringId::Empty; // Instruction pattern (e.g., "gcc -c %f -o %o")
     std::vector<NodeId> inputs = {};           // Operand file NodeIds for %f expansion
     std::vector<NodeId> outputs = {};          // Operand file NodeIds for %o expansion
-    std::set<StringId> exported_vars = {};     // Env vars to export (interned)
+    SortedIdVec exported_vars = {};            // Env vars to export (interned StringIds)
     std::optional<GeneratedOutput> generated_output = {};  // Output specification
     OutputAction output_action = {};           // What to do with output
     NodeId parent_command = INVALID_NODE_ID;   // Parent command for InjectImplicitDeps
@@ -546,9 +546,10 @@ struct Graph {
     std::unordered_map<NodeId, std::vector<NodeId>> order_only_to_index;
     std::unordered_map<NodeId, std::vector<NodeId>> order_only_dependents;
 
-    // Node lookup indices (with transparent lookup support)
-    std::unordered_map<DirNameKey, NodeId, ...> dir_name_index;  // (parent, name) -> NodeId
-    std::unordered_map<std::string, NodeId, ...> command_str_index;  // command -> NodeId
+    // Node lookup indices
+    std::vector<SortedPairVec> dir_children;  // Per-directory name -> NodeId (indexed by parent dir)
+    StringPool command_strings;               // Interned expanded command strings
+    SortedPairVec command_index;              // StringId(command) -> NodeId
 
     // ID generators (next available ID for each type)
     NodeId next_file_id = 2;                               // Starts at 2 (BUILD_ROOT is 1)
@@ -558,7 +559,10 @@ struct Graph {
 };
 
 // Path cache is stored externally in BuildGraph for const-correctness
-using PathCache = std::unordered_map<NodeId, std::string>;
+struct PathCache {
+    NodeIdMap32 ids;    // NodeId -> StringId (path interned in pool)
+    StringPool pool;    // Owns the full path strings
+};
 
 class BuildGraph {
     Graph graph_;
@@ -999,7 +1003,7 @@ struct BuildJob {
     std::vector<std::string> inputs = {};
     std::vector<std::string> outputs = {};
     std::vector<std::string> order_only_inputs = {};  // Order-only dependencies
-    std::set<std::string> exported_vars = {};         // Env vars to export to command
+    std::vector<std::string> exported_vars = {};      // Env vars to export to command
 
     // For auto-generated rules (from pattern matching)
     bool capture_stdout = false;             // Capture stdout for depfile parsing

Tupfile

@@ -4,10 +4,15 @@ include_rules
 c-srcs-y  = third_party/sha256/sha256.c
 
 # Core
-srcs-y  = src/core/hash.cpp
+srcs-y  = src/core/arena.cpp
+srcs-y += src/core/hash.cpp
+srcs-y += src/core/id_array.cpp
+srcs-y += src/core/id_bitset.cpp
 srcs-y += src/core/layout.cpp
 srcs-y += src/core/metrics.cpp
+srcs-y += src/core/path.cpp
 srcs-y += src/core/path_utils.cpp
+srcs-y += src/core/sorted_id_vec.cpp
 srcs-y += src/core/string_pool.cpp
 srcs-y += src/core/string_utils.cpp
 srcs-y += src/core/terminal.cpp

bootstrap-linux.sh

@@ -47,28 +47,40 @@ mkdir -p "build/test/unit"
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/terminal.cpp -o build/terminal.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/string_utils.cpp -o build/string_utils.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/string_pool.cpp -o build/string_pool.o)
+(cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/sorted_id_vec.cpp -o build/sorted_id_vec.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/path_utils.cpp -o build/path_utils.o)
+(cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/path.cpp -o build/path.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/metrics.cpp -o build/metrics.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/layout.cpp -o build/layout.o)
+(cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/id_bitset.cpp -o build/id_bitset.o)
+(cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/id_array.cpp -o build/id_array.o)
 (cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/hash.cpp -o build/hash.o)
+(cd "." && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I./include -I./third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c src/core/arena.cpp -o build/arena.o)
 (cd "." && gcc -std=c11 -Wall -Wextra -Werror -fPIC -I./third_party  -O2 -ffunction-sections -fdata-sections -Wno-error -c third_party/sha256/sha256.c -o build/sha256.o)
-(cd "." && g++ build/sha256.o build/hash.o build/layout.o build/metrics.o build/path_utils.o build/string_pool.o build/string_utils.o build/terminal.o build/ast.o build/config.o build/depfile.o build/eval.o build/glob.o build/ignore.o build/lexer.o build/parser.o build/var_tracking.o build/builder.o build/dag.o build/dep_scanner.o build/rule_pattern.o build/gcc.o build/topo.o build/entry.o build/reader.o build/writer.o build/progress_display.o build/runner.o build/scheduler.o build/cmd_build.o build/cmd_clean.o build/cmd_configure.o build/cmd_show.o build/cmd_parse.o build/config_commands.o build/context.o build/multi_variant.o build/options.o build/output.o build/target.o build/process-posix.o build/file_io-posix.o build/env-posix.o build/path-posix.o build/main.o -o build/putup -Wl,--gc-sections )
-(cd "." && ar rcs build/libputup.a build/sha256.o build/hash.o build/layout.o build/metrics.o build/path_utils.o build/string_pool.o build/string_utils.o build/terminal.o build/ast.o build/config.o build/depfile.o build/eval.o build/glob.o build/ignore.o build/lexer.o build/parser.o build/var_tracking.o build/builder.o build/dag.o build/dep_scanner.o build/rule_pattern.o build/gcc.o build/topo.o build/entry.o build/reader.o build/writer.o build/progress_display.o build/runner.o build/scheduler.o build/cmd_build.o build/cmd_clean.o build/cmd_configure.o build/cmd_show.o build/cmd_parse.o build/config_commands.o build/context.o build/multi_variant.o build/options.o build/output.o build/target.o build/process-posix.o build/file_io-posix.o build/env-posix.o build/path-posix.o)
+(cd "." && g++ build/sha256.o build/arena.o build/hash.o build/id_array.o build/id_bitset.o build/layout.o build/metrics.o build/path.o build/path_utils.o build/sorted_id_vec.o build/string_pool.o build/string_utils.o build/terminal.o build/ast.o build/config.o build/depfile.o build/eval.o build/glob.o build/ignore.o build/lexer.o build/parser.o build/var_tracking.o build/builder.o build/dag.o build/dep_scanner.o build/rule_pattern.o build/gcc.o build/topo.o build/entry.o build/reader.o build/writer.o build/progress_display.o build/runner.o build/scheduler.o build/cmd_build.o build/cmd_clean.o build/cmd_configure.o build/cmd_show.o build/cmd_parse.o build/config_commands.o build/context.o build/multi_variant.o build/options.o build/output.o build/target.o build/process-posix.o build/file_io-posix.o build/env-posix.o build/path-posix.o build/main.o -o build/putup -Wl,--gc-sections )
+(cd "." && ar rcs build/libputup.a build/sha256.o build/arena.o build/hash.o build/id_array.o build/id_bitset.o build/layout.o build/metrics.o build/path.o build/path_utils.o build/sorted_id_vec.o build/string_pool.o build/string_utils.o build/terminal.o build/ast.o build/config.o build/depfile.o build/eval.o build/glob.o build/ignore.o build/lexer.o build/parser.o build/var_tracking.o build/builder.o build/dag.o build/dep_scanner.o build/rule_pattern.o build/gcc.o build/topo.o build/entry.o build/reader.o build/writer.o build/progress_display.o build/runner.o build/scheduler.o build/cmd_build.o build/cmd_clean.o build/cmd_configure.o build/cmd_show.o build/cmd_parse.o build/config_commands.o build/context.o build/multi_variant.o build/options.o build/output.o build/target.o build/process-posix.o build/file_io-posix.o build/env-posix.o build/path-posix.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_var_tracking.cpp -o ../../build/test/unit/test_var_tracking.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_types.cpp -o ../../build/test/unit/test_types.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_target.cpp -o ../../build/test/unit/test_target.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_string_utils.cpp -o ../../build/test/unit/test_string_utils.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_string_pool.cpp -o ../../build/test/unit/test_string_pool.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_sorted_id_vec.cpp -o ../../build/test/unit/test_sorted_id_vec.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_rule_pattern.cpp -o ../../build/test/unit/test_rule_pattern.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_platform_process.cpp -o ../../build/test/unit/test_platform_process.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_platform_file_io.cpp -o ../../build/test/unit/test_platform_file_io.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_path_utils.cpp -o ../../build/test/unit/test_path_utils.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_path.cpp -o ../../build/test/unit/test_path.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_parser.cpp -o ../../build/test/unit/test_parser.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_node_id_map.cpp -o ../../build/test/unit/test_node_id_map.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_main.cpp -o ../../build/test/unit/test_main.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_lexer.cpp -o ../../build/test/unit/test_lexer.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_layout.cpp -o ../../build/test/unit/test_layout.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_index.cpp -o ../../build/test/unit/test_index.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_ignore.cpp -o ../../build/test/unit/test_ignore.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_id_bitset.cpp -o ../../build/test/unit/test_id_bitset.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_id_array.cpp -o ../../build/test/unit/test_id_array.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_hash.cpp -o ../../build/test/unit/test_hash.o)
+(cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_arena.cpp -o ../../build/test/unit/test_arena.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_graph.cpp -o ../../build/test/unit/test_graph.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_glob.cpp -o ../../build/test/unit/test_glob.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_exec.cpp -o ../../build/test/unit/test_exec.o)
@@ -80,4 +92,4 @@ mkdir -p "build/test/unit"
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c test_bench.cpp -o ../../build/test/unit/test_bench.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -c e2e_fixture.cpp -o ../../build/test/unit/e2e_fixture.o)
 (cd "test/unit" && g++ -std=c++20 -Wall -Wextra -Werror -Wpedantic -fPIC -fno-exceptions -fno-rtti -I../../include -I../../third_party  -O2 -DNDEBUG -ffunction-sections -fdata-sections -Wno-error -c ../../third_party/catch_amalgamated.cpp -o ../../build/test/unit/catch_amalgamated.o)
-(cd "test/unit" && g++ ../../build/test/unit/test_bench.o ../../build/test/unit/test_builder.o ../../build/test/unit/test_dep_scanner.o ../../build/test/unit/test_depfile.o ../../build/test/unit/test_e2e.o ../../build/test/unit/test_eval.o ../../build/test/unit/test_exec.o ../../build/test/unit/test_glob.o ../../build/test/unit/test_graph.o ../../build/test/unit/test_hash.o ../../build/test/unit/test_ignore.o ../../build/test/unit/test_index.o ../../build/test/unit/test_layout.o ../../build/test/unit/test_lexer.o ../../build/test/unit/test_main.o ../../build/test/unit/test_parser.o ../../build/test/unit/test_path_utils.o ../../build/test/unit/test_platform_file_io.o ../../build/test/unit/test_platform_process.o ../../build/test/unit/test_rule_pattern.o ../../build/test/unit/test_string_utils.o ../../build/test/unit/test_target.o ../../build/test/unit/test_types.o ../../build/test/unit/test_var_tracking.o ../../build/test/unit/catch_amalgamated.o ../../build/test/unit/e2e_fixture.o ../../build/libputup.a -o ../../build/test/unit/putup_test -Wl,--gc-sections )
+(cd "test/unit" && g++ ../../build/test/unit/test_bench.o ../../build/test/unit/test_builder.o ../../build/test/unit/test_dep_scanner.o ../../build/test/unit/test_depfile.o ../../build/test/unit/test_e2e.o ../../build/test/unit/test_eval.o ../../build/test/unit/test_exec.o ../../build/test/unit/test_glob.o ../../build/test/unit/test_graph.o ../../build/test/unit/test_arena.o ../../build/test/unit/test_hash.o ../../build/test/unit/test_id_array.o ../../build/test/unit/test_id_bitset.o ../../build/test/unit/test_ignore.o ../../build/test/unit/test_index.o ../../build/test/unit/test_layout.o ../../build/test/unit/test_lexer.o ../../build/test/unit/test_main.o ../../build/test/unit/test_node_id_map.o ../../build/test/unit/test_parser.o ../../build/test/unit/test_path.o ../../build/test/unit/test_path_utils.o ../../build/test/unit/test_platform_file_io.o ../../build/test/unit/test_platform_process.o ../../build/test/unit/test_rule_pattern.o ../../build/test/unit/test_sorted_id_vec.o ../../build/test/unit/test_string_pool.o ../../build/test/unit/test_string_utils.o ../../build/test/unit/test_target.o ../../build/test/unit/test_types.o ../../build/test/unit/test_var_tracking.o ../../build/test/unit/catch_amalgamated.o ../../build/test/unit/e2e_fixture.o ../../build/libputup.a -o ../../build/test/unit/putup_test -Wl,--gc-sections )