Setup initial repo

.gitignore

@@ -0,0 +1,7 @@
+result
+p.vim
+compile_commands.json
+
+bin/
+build/
+.cache/

Makefile

@@ -0,0 +1,103 @@
+SRC_DIR := src
+BIN_DIR := bin
+OBJ_DIR := build
+INCLUDE_DIR := include
+
+TARGET_DEBUG := $(BIN_DIR)/crib-dbg
+TARGET_RELEASE := $(BIN_DIR)/crib
+
+PCH_DEBUG := $(OBJ_DIR)/debug/pch.h.gch
+PCH_RELEASE := $(OBJ_DIR)/release/pch.h.gch
+
+CXX ?= g++
+CC ?= gcc
+
+PCRE_CFLAGS := $(shell pkg-config --cflags libpcre2-8)
+PCRE_LIBS := $(shell pkg-config --static --libs libpcre2-8)
+
+LIBGRAPHEME_CFLAGS := $(shell pkg-config --cflags libgrapheme)
+LIBGRAPHEME_LIBS := $(shell pkg-config --static --libs libgrapheme)
+
+MRUBY_CFLAGS ?= ./libs/mruby/build/host/include
+MRUBY_LIBS ?= ./libs/mruby/build/host/lib/libmruby.a
+
+CFLAGS_DEBUG :=\
+	-std=c++20 -Wall -Wextra -static \
+	-O0 -fno-inline -gsplit-dwarf \
+	-g -fno-omit-frame-pointer \
+	-I./include -I./libs/unicode_width
+
+CFLAGS_RELEASE :=\
+	-std=c++20 -O3 -march=x86-64 -mtune=generic -fno-rtti \
+	-fvisibility=hidden -static \
+	-fomit-frame-pointer -DNDEBUG -s \
+	-I./include -I./libs/unicode_width
+
+CFLAGS_DEBUG += $(PCRE_CFLAGS) $(LIBGRAPHEME_CFLAGS) $(MRUBY_CFLAGS) $(C_SANITIZER)
+CFLAGS_RELEASE += $(PCRE_CFLAGS) $(LIBGRAPHEME_CFLAGS) $(MRUBY_CFLAGS)
+
+PCH_CFLAGS_DEBUG := $(CFLAGS_DEBUG) -x c++-header
+PCH_CFLAGS_RELEASE := $(CFLAGS_RELEASE) -x c++-header
+
+UNICODE_SRC := $(wildcard libs/unicode_width/*.c)
+
+UNICODE_OBJ_DEBUG := $(patsubst libs/unicode_width/%.c,$(OBJ_DIR)/debug/unicode_width/%.o,$(UNICODE_SRC))
+UNICODE_OBJ_RELEASE := $(patsubst libs/unicode_width/%.c,$(OBJ_DIR)/release/unicode_width/%.o,$(UNICODE_SRC))
+
+LIBS := $(PCRE_LIBS) $(LIBGRAPHEME_LIBS) $(MRUBY_LIBS)
+
+SRC := $(wildcard $(SRC_DIR)/**/*.cc) $(wildcard $(SRC_DIR)/*.cc)
+OBJ_DEBUG := $(patsubst $(SRC_DIR)/%.cc,$(OBJ_DIR)/debug/%.o,$(SRC))
+OBJ_RELEASE := $(patsubst $(SRC_DIR)/%.cc,$(OBJ_DIR)/release/%.o,$(SRC))
+
+DEP_DEBUG := $(OBJ_DEBUG:.o=.d)
+DEP_RELEASE := $(OBJ_RELEASE:.o=.d)
+
+.PHONY: all debug release clean
+
+all: release
+
+debug: $(TARGET_DEBUG)
+
+release: $(TARGET_RELEASE)
+
+$(PCH_DEBUG): $(INCLUDE_DIR)/pch.h
+	mkdir -p $(dir $@)
+	$(CXX) $(PCH_CFLAGS_DEBUG) -o $@ $<
+
+$(PCH_RELEASE): $(INCLUDE_DIR)/pch.h
+	mkdir -p $(dir $@)
+	$(CXX) $(PCH_CFLAGS_RELEASE) -o $@ $<
+
+$(TARGET_DEBUG): $(PCH_DEBUG) $(OBJ_DEBUG) $(UNICODE_OBJ_DEBUG)
+	mkdir -p $(BIN_DIR)
+	$(CXX) $(CFLAGS_DEBUG) -o $@ $(OBJ_DEBUG) $(UNICODE_OBJ_DEBUG) $(LIBS)
+
+$(TARGET_RELEASE): $(PCH_RELEASE) $(OBJ_RELEASE) $(UNICODE_OBJ_RELEASE)
+	mkdir -p $(BIN_DIR)
+	$(CXX) $(CFLAGS_RELEASE) -o $@ $(OBJ_RELEASE) $(UNICODE_OBJ_RELEASE) $(LIBS)
+
+$(OBJ_DIR)/debug/%.o: $(SRC_DIR)/%.cc $(PCH_DEBUG) $(GENERATED_HEADER)
+	mkdir -p $(dir $@)
+	$(CXX) $(CFLAGS_DEBUG) -include $(INCLUDE_DIR)/pch.h -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/release/%.o: $(SRC_DIR)/%.cc $(PCH_RELEASE) $(GENERATED_HEADER)
+	mkdir -p $(dir $@)
+	$(CXX) $(CFLAGS_RELEASE) -include $(INCLUDE_DIR)/pch.h -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/debug/unicode_width/%.o: libs/unicode_width/%.c
+	mkdir -p $(dir $@)
+	$(CC) -MMD -MP -c $< -o $@
+
+$(OBJ_DIR)/release/unicode_width/%.o: libs/unicode_width/%.c
+	mkdir -p $(dir $@)
+	$(CC) -MMD -MP -c $< -o $@
+
+DEP_DEBUG += $(UNICODE_OBJ_DEBUG:.o=.d)
+DEP_RELEASE += $(UNICODE_OBJ_RELEASE:.o=.d)
+
+-include $(DEP_DEBUG)
+-include $(DEP_RELEASE)
+
+clean:
+	rm -rf $(OBJ_DIR) $(BIN_DIR)

flake.nix

@@ -62,6 +62,7 @@
 
         buildInputs = with pkgs; [
           libgrapheme
+          pcre2
 
           mruby
         ];
@@ -69,8 +70,8 @@
         buildPhase = ''
           make \
             MRBC=${mruby}/bin/mrbc \
-            MRUBY_INCLUDE=${mruby}/include \
+            MRUBY_CFLAGS=-I${mruby}/include \
-            MRUBY_LIB=${mruby}/lib/libmruby.a
+            MRUBY_LIBS=-L${mruby}/lib -lmruby
         '';
 
         installPhase = ''
@@ -105,12 +106,12 @@
         ];
 
         shellHook = ''
-          export CC="ccache gcc"
+          export CC="ccache $CC"
-          export CXX="ccache g++"
+          export CXX="ccache $CXX"
 
           export MRBC=${mruby}/bin/mrbc
-          export MRUBY_INCLUDE=${mruby}/include
+          export MRUBY_CFLAGS=-I${mruby}/include
-          export MRUBY_LIB=${mruby}/lib/libmruby.a
+          export MRUBY_LIBS="-L${mruby}/lib -lmruby"
         '';
       };
     };

include/pch.h

@@ -0,0 +1,58 @@
+#ifndef PCH_H
+#define PCH_H
+
+#define PCRE2_CODE_UNIT_WIDTH 8
+#define PCRE_WORKSPACE_SIZE 512
+
+#include "mruby.h"
+#include "mruby/array.h"
+#include "mruby/compile.h"
+#include "mruby/hash.h"
+#include "mruby/irep.h"
+#include "mruby/string.h"
+#include <pcre2.h>
+extern "C" {
+#include "unicode_width.h"
+#include <grapheme.h>
+}
+#include <algorithm>
+#include <atomic>
+#include <cctype>
+#include <chrono>
+#include <cmath>
+#include <cstdarg>
+#include <cstdint>
+#include <cstdio>
+#include <cstdlib>
+#include <cstring>
+#include <deque>
+#include <fcntl.h>
+#include <filesystem>
+#include <fstream>
+#include <functional>
+#include <immintrin.h>
+#include <limits.h>
+#include <map>
+#include <mutex>
+#include <optional>
+#include <poll.h>
+#include <queue>
+#include <set>
+#include <shared_mutex>
+#include <signal.h>
+#include <stack>
+#include <string.h>
+#include <string>
+#include <sys/ioctl.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <termios.h>
+#include <thread>
+#include <unistd.h>
+#include <unordered_map>
+#include <unordered_set>
+#include <vector>
+
+using namespace std::chrono_literals;
+
+#endif

libs/unicode_width/unicode_width.h

@@ -0,0 +1,96 @@
+/* Generated by unicode-width generator.
+ *
+ * Unicode 16.0.0 data.
+ * For terminal width calculation.
+ *
+ * Copyright 2025 Dair Aidarkhanov
+ * SPDX-License-Identifier: 0BSD
+ */
+
+#ifndef UNICODE_WIDTH_H
+#define UNICODE_WIDTH_H
+
+#include <stddef.h>
+#include <stdint.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Version of Unicode data.
+ */
+#define UNICODE_WIDTH_VERSION_MAJOR 16
+#define UNICODE_WIDTH_VERSION_MINOR 0
+#define UNICODE_WIDTH_VERSION_PATCH 0
+
+/* Width state values.
+ */
+typedef enum {
+  WIDTH_STATE_DEFAULT = 0,
+  WIDTH_STATE_LINE_FEED = 1,
+  WIDTH_STATE_EMOJI_MODIFIER = 2,
+  WIDTH_STATE_REGIONAL_INDICATOR = 3,
+  WIDTH_STATE_SEVERAL_REGIONAL_INDICATOR = 4,
+  WIDTH_STATE_EMOJI_PRESENTATION = 5,
+  WIDTH_STATE_ZWJ_EMOJI_PRESENTATION = 6,
+  WIDTH_STATE_KEYCAP_ZWJ_EMOJI_PRESENTATION = 7,
+  WIDTH_STATE_REGIONAL_INDICATOR_ZWJ_PRESENTATION = 8,
+  WIDTH_STATE_TAG_END_ZWJ_EMOJI_PRESENTATION = 9,
+  WIDTH_STATE_ZWJ_SEQUENCE_MEMBER = 10,
+} width_state_t;
+
+/* State for the width calculation state machine.
+ */
+typedef struct {
+  width_state_t state;
+  uint_least32_t previous_codepoint;
+} unicode_width_state_t;
+
+/* Initialize a unicode width state.
+ * Must be called before any other function.
+ *
+ * @param state Pointer to state to initialize
+ */
+void unicode_width_init(unicode_width_state_t *state);
+
+/* Process a Unicode codepoint and return its width.
+ * Width is 0, 1, 2, or 3, or -1 for control characters.
+ *
+ * Control characters (0x00-0x1F except newlines, 0x7F, and 0x80-0x9F) return
+ * -1, allowing the caller to decide how to display them. For readline-like
+ * applications, control characters are typically displayed using caret notation
+ * (^X) with width 2. Use unicode_width_control_char() to get this width.
+ *
+ * Newlines (LF, CR, CRLF) return width 0 as they don't consume horizontal space
+ * in terminal displays.
+ *
+ * Note that the width of a codepoint may depend on context (preceding/following
+ * codepoints), so this function keeps track of context in the provided state.
+ *
+ * @param state Pointer to state
+ * @param codepoint Unicode codepoint to process
+ * @return Width of the codepoint in columns, or -1 for control characters
+ */
+int unicode_width_process(unicode_width_state_t *state,
+                          uint_least32_t codepoint);
+
+/* Get the display width of a control character in caret notation (e.g., ^A).
+ * This is useful for applications like readline that display control chars.
+ *
+ * @param codepoint The Unicode codepoint to check
+ * @return The display width (usually 2 for ^X notation), or -1 if not a control
+ * char
+ */
+int unicode_width_control_char(uint_least32_t codepoint);
+
+/* Reset a unicode width state to its initial state.
+ *
+ * @param state Pointer to state to reset
+ */
+void unicode_width_reset(unicode_width_state_t *state);
+
+#ifdef __cplusplus
+} /* extern "C" */
+#endif
+
+#endif /* UNICODE_WIDTH_H */

src/main.cc

@@ -0,0 +1,3 @@
+#include "pch.h"
+
+int main() { return 0; }