ohhhhhhhhhhh (exclude linux-extra-rootfs.img build artifact)

CODE/font/_test_glyph_ref.py

@@ -0,0 +1,55 @@
+# -*- coding: utf-8 -*-
+"""_test_glyph_ref.py — dbf 字形与 TTF 直接渲染逐字节比对（码点映射 + 数据无损验证）"""
+import struct
+from PIL import Image, ImageFont, ImageDraw
+
+def load(path):
+    data = open(path, "rb").read()
+    count = struct.unpack_from("<I", data, 4)[0]
+    w, h, bpp = struct.unpack_from("<HHH", data, 8)
+    idx = [struct.unpack_from("<HHII", data, 14 + i * 12) for i in range(count)]
+    return data, w, h, idx
+
+def lookup(idx, cp):
+    lo, hi = 0, len(idx) - 1
+    while lo <= hi:
+        m = (lo + hi) // 2
+        if idx[m][0] == cp:
+            return idx[m]
+        if idx[m][0] < cp:
+            lo = m + 1
+        else:
+            hi = m - 1
+    return None
+
+SAMPLES = "中文测试开始菜单任务管理器设置搜索通知中心关机重启睡眠亮暗主题浅深任务栏窗口图标网络音量电源时间日期锁屏用户名睡眠退出Deshab 0123"
+
+for name, size in [("simhei_16.dbf", 16), ("simhei_24.dbf", 24)]:
+    data, w, h, idx = load(name)
+    font = ImageFont.truetype("simhei.ttf", size)
+    ascii_w = size * 3 // 4
+    ok = bad = 0
+    diffs = []
+    for ch in SAMPLES:
+        cp = ord(ch)
+        e = lookup(idx, cp)
+        if not e:
+            diffs.append((ch, "MISSING"))
+            bad += 1
+            continue
+        _, gw, off, sz = e
+        dbf_glyph = data[off:off + sz]
+        ref_w = ascii_w if cp < 128 else size
+        img = Image.new("L", (ref_w, h), 0)
+        ImageDraw.Draw(img).text((0, 0), ch, font=font, fill=255)
+        img = img.resize((ref_w, h), Image.LANCZOS)
+        ref = img.tobytes()
+        if dbf_glyph == ref and gw == ref_w:
+            ok += 1
+        else:
+            bad += 1
+            nd = sum(1 for a, b in zip(dbf_glyph, ref) if a != b)
+            diffs.append((hex(cp), "bytes-diff", nd, "/", len(ref)))
+    print(name, "exact-match:", ok, "mismatch:", bad)
+    for d in diffs[:6]:
+        print("  diff:", d)

CODE/font/mkfont.py

@@ -1,6 +1,16 @@
 """mkfont.py — Deshab Bitmap Font generator
-Usage: python mkfont.py <ttf_path> <size> <output.dbf>
-Renders GB2312 charset + ASCII to a packed bitmap font file (.dbf)
+Usage: python mkfont.py <ttf_path> <size> <output.dbf> [--charset charset.txt]
+
+Renders GB2312 charset + ASCII (default) or a custom charset file to a packed
+bitmap font file (.dbf).
+
+Index format (see dbf.h):
+  - Key is the Unicode BMP codepoint of the glyph (NOT the GB2312 byte pair).
+  - Index entries are sorted ascending by codepoint — required by the
+    binary-search reader dbf_lookup().
+
+--charset: UTF-8 text file; every unique character in it becomes a glyph
+(ASCII or CJK). Use to build small UI-subset fonts.
 """
 
 import struct
@@ -9,51 +19,60 @@ from PIL import Image, ImageFont, ImageDraw
 
 DBF_MAGIC = b"DBF\x10"
 
-# GB2312 codepoints for simplified Chinese (3755 level-1 + 3008 level-2)
-def gb2312_codepoints():
-    cps = []
-    # ASCII 32-126
-    for c in range(32, 127):
-        cps.append(c)
-    # GB2312 level 1: 0xB0A1 - 0xB0FE, ... , 0xD7F9
+
+def gb2312_chars():
+    """All GB2312 characters (level 1 + level 2) as a list of str."""
+    chars = []
+    # Level 1: 0xB0A1 - 0xD7F9 (pinyin ordered)
     for hi in range(0xB0, 0xD8):
-        for lo in range(0xA1, 0xFF):
-            cps.append((hi << 8) | lo)
-    # GB2312 level 2: skip duplicates, 0xD8A1 - 0xF7FE
-    for hi in range(0xD8, 0xF8):
         for lo in range(0xA1, 0xFF):
             if hi == 0xD7 and lo > 0xF9:
                 break
-            cps.append((hi << 8) | lo)
-    return cps
+            try:
+                chars.append(bytes([hi, lo]).decode("gb2312"))
+            except UnicodeDecodeError:
+                pass
+    # Level 2: 0xD8A1 - 0xF7FE
+    for hi in range(0xD8, 0xF8):
+        for lo in range(0xA1, 0xFF):
+            try:
+                chars.append(bytes([hi, lo]).decode("gb2312"))
+            except UnicodeDecodeError:
+                pass
+    return chars
 
 
-def render_font(ttf_path, size, output_path, bpp=8):
+def charset_from_file(path):
+    """Unique characters from a UTF-8 text file."""
+    with open(path, "r", encoding="utf-8") as f:
+        text = f.read()
+    # 保留出现顺序去重；排序交给索引排序阶段
+    return list(dict.fromkeys(text))
+
+
+def render_font(ttf_path, size, output_path, charset=None, bpp=8):
     font = ImageFont.truetype(ttf_path, size)
-    cps = gb2312_codepoints()
 
-    # Build index and data
+    if charset is None:
+        # 默认：ASCII 32-126 + 全 GB2312
+        chars = [chr(c) for c in range(32, 127)] + gb2312_chars()
+    else:
+        chars = list(charset)
+        # 去重
+        chars = list(dict.fromkeys(chars))
+
     glyph_w = size
     glyph_h = size
-    data_parts = []
-    index_entries = []
-
     ascii_w = size * 3 // 4  # narrower for ASCII
 
-    for cp in cps:
-        if cp < 128:
-            # ASCII — use narrower glyph
-            ch = chr(cp)
-            w = ascii_w
-        else:
-            # GB2312 byte pair → Unicode
-            hi = (cp >> 8) & 0xFF
-            lo = cp & 0xFF
-            try:
-                ch = bytes([hi, lo]).decode("gb2312")
-            except:
-                continue
-            w = glyph_w
+    data_parts = []
+    index_entries = []  # (unicode_cp, width, offset, size)
+
+    for ch in chars:
+        cp = ord(ch)
+        if cp > 0xFFFF:
+            continue  # dbf index key is u16 — BMP only
+        w = ascii_w if cp < 128 else glyph_w
 
         img = Image.new("L", (w, glyph_h), 0)
         draw = ImageDraw.Draw(img)
@@ -82,25 +101,49 @@ def render_font(ttf_path, size, output_path, bpp=8):
         data_parts.append(glyph_data)
         index_entries.append((cp, w, offset, len(glyph_data)))
 
+    # 关键：按 Unicode 码点升序排序（dbf_lookup 二分查找的前提），
+    # 同码点去重（自定义字表可能重复包含同一字符）。
+    seen = {}
+    for cp, w, off, sz in index_entries:
+        if cp not in seen:
+            seen[cp] = (cp, w, off, sz)
+    index_entries = sorted(seen.values(), key=lambda e: e[0])
+
     # Write file
     total_data = sum(len(d) for d in data_parts)
+    # dbf.h 规范：data_offset 相对文件头（magic+header+index），写入时换算
+    data_start = 14 + len(index_entries) * 12
     print(f"Glyphs: {len(index_entries)}, Data: {total_data} bytes")
 
     with open(output_path, "wb") as f:
         f.write(DBF_MAGIC)
         f.write(struct.pack("<I", len(index_entries)))
         f.write(struct.pack("<HHH", glyph_w, glyph_h, bpp))
-        for cp, w, off, sz in index_entries:
-            f.write(struct.pack("<HHII", cp, w, off, sz))
+        for cp, w, rel_off, sz in index_entries:
+            f.write(struct.pack("<HHII", cp, w, data_start + rel_off, sz))
         for d in data_parts:
             f.write(d)
 
 
-if __name__ == "__main__":
-    if len(sys.argv) != 4:
-        print(f"Usage: {sys.argv[0]} <ttf_path> <size> <output.dbf>")
+def main():
+    argv = sys.argv[1:]
+    charset_path = None
+    if "--charset" in argv:
+        i = argv.index("--charset")
+        if i + 1 >= len(argv):
+            print("--charset requires a file path")
+            sys.exit(1)
+        charset_path = argv[i + 1]
+        argv = argv[:i] + argv[i + 2:]
+    if len(argv) != 3:
+        print(f"Usage: {sys.argv[0]} <ttf_path> <size> <output.dbf> [--charset file]")
         sys.exit(1)
-    ttf = sys.argv[1]
-    sz = int(sys.argv[2])
-    out = sys.argv[3]
-    render_font(ttf, sz, out)
+    ttf, size, out = argv[0], int(argv[1]), argv[2]
+    charset = charset_from_file(charset_path) if charset_path else None
+    if charset is not None:
+        print(f"Custom charset: {len(charset)} unique chars from {charset_path}")
+    render_font(ttf, size, out, charset=charset)
+
+
+if __name__ == "__main__":
+    main()

CODE/font/mkfont_check.py

@@ -0,0 +1,89 @@
+"""mkfont_check.py — .dbf 字库校验工具
+
+Usage: python mkfont_check.py <font.dbf> [字符...]
+
+校验项：
+  1. magic "DBF\x10"、bpp、字形尺寸
+  2. 索引按 codepoint 升序（dbf_lookup 二分查找前提）
+  3. 抽样码点查找（默认：中文测试开始菜单任务管理器设置搜索通知中心
+     关机重启睡眠亮暗主题浅深任务栏窗口图标网络音量电源时间日期）
+  4. 无字符参数时把「中」渲染为 ASCII art 供肉眼检查
+"""
+
+import struct
+import sys
+
+if hasattr(sys.stdout, "reconfigure"):
+    sys.stdout.reconfigure(encoding="utf-8", errors="replace")
+
+DBF_MAGIC = b"DBF\x10"
+DEFAULT_SAMPLE = "中文测试开始菜单任务管理器设置搜索通知中心关机重启睡眠亮暗主题浅深任务栏窗口图标网络音量电源时间日期"
+
+
+def load(path):
+    with open(path, "rb") as f:
+        data = f.read()
+    if data[:4] != DBF_MAGIC:
+        raise SystemExit(f"BAD MAGIC: {data[:4]!r}")
+    count = struct.unpack_from("<I", data, 4)[0]
+    w, h, bpp = struct.unpack_from("<HHH", data, 8)
+    idx = []
+    for i in range(count):
+        cp, gw, off, sz = struct.unpack_from("<HHII", data, 14 + i * 12)
+        idx.append((cp, gw, off, sz))
+    return data, count, w, h, bpp, idx
+
+
+def lookup(idx, cp):
+    lo, hi = 0, len(idx) - 1
+    while lo <= hi:
+        mid = (lo + hi) // 2
+        if idx[mid][0] == cp:
+            return idx[mid]
+        if idx[mid][0] < cp:
+            lo = mid + 1
+        else:
+            hi = mid - 1
+    return None
+
+
+def ascii_art(data, entry, h):
+    _, w, off, _sz = entry
+    print(f"  +{'-' * w}+")
+    for y in range(h):
+        row = " "
+        for x in range(w):
+            v = data[off + y * w + x]
+            row += "█" if v > 200 else ("▓" if v > 120 else ("░" if v > 50 else " "))
+        print("  |" + row + "|")
+    print(f"  +{'-' * w}+")
+
+
+def main():
+    if len(sys.argv) < 2:
+        print(f"Usage: {sys.argv[0]} <font.dbf> [chars]")
+        sys.exit(1)
+    path = sys.argv[1]
+    data, count, w, h, bpp, idx = load(path)
+    print(f"{path}: count={count} glyph={w}x{h} bpp={bpp}")
+
+    # 校验升序
+    bad = sum(1 for a, b in zip(idx, idx[1:]) if a[0] >= b[0])
+    if bad:
+        raise SystemExit(f"INDEX NOT SORTED: {bad} violations")
+    print("index sorted: OK")
+
+    sample = sys.argv[2] if len(sys.argv) > 2 else DEFAULT_SAMPLE
+    missing = [c for c in sample if lookup(idx, ord(c)) is None]
+    if missing:
+        print(f"MISSING GLYPHS: {''.join(missing)}")
+        sys.exit(1)
+    print(f"all {len(sample)} sample glyphs found: OK")
+
+    # 肉眼检查首字符
+    first = sample[0]
+    ascii_art(data, lookup(idx, ord(first)), h)
+
+
+if __name__ == "__main__":
+    main()