AskaEth
c775f8558c
P2-1（必修）：解压页缓存与缩略图 LRU 在 *.bmp 上相互踩踏 - ui/library 的 cache_prune 先按 page_cache 的共用判据 mr_cache_is_page_name 跳过 mrpage_*，不再把 BMP 解压页算进缩略图额度或 unlink（含 Vol10.cbz 这类 BMP 档不再出现「被缩略图 LRU 删掉 → 首页解码失败」）； - mr_cache_touch 命中时 stat 校验文件存在与大小，缺失/尺寸不符即移出索引并按 未命中返回，调用方重新解压——外部清理/掉电残留/误删都能自愈，不再返回死路径。 P2-2（文档化一次性溢出）： - page_cache.h 与 make_room_locked 写明：单条目本身超过 32MiB 时只保留该条、 其它全部淘汰，占用上限受页源单页 64MiB（MR_ZIP_PAGE_MAX_BYTES）约束； 大页仍可读，不因预算拒绝落缓存。 - 新增 33MiB 页夹具（PNG 头 + 可压零填充，deflate 后仅数十 KB）与断言： 大页可读、cache_bytes 允许超 32MiB、小页被淘汰、再来小页时大页按预算淘汰。 顺手：page_cache.h 注释包名 os.meow.picobrowser → os.meow.picbrowser（测试路径同步）。 测试：test_cbz 267 → 296 项全 PASS；run_tests.sh（zip 106）与 run_webtoon_tests.sh 全绿；交叉构建 sdk/tuku 零警告，产物静态 MIPS32r2 ELF。
241 lines
9.9 KiB
Python
Executable File
241 lines
9.9 KiB
Python
Executable File
#!/usr/bin/env python3
|
||
"""生成 tuku 档案册（CBZ）集成测试夹具：真实 zipfile 产物 + 定点字节修补。
|
||
|
||
用法: python3 make_cbz_samples.py <out_dir>
|
||
|
||
产出：
|
||
<out>/library/ 书库夹具（文件夹册 + 档案册混排）
|
||
bookA/001.png 002.png 文件夹册（2 页，PNG）
|
||
vol2.cbz 3 页：PNG(41x31) / PNG(40x30) / 中文名 PNG(12x34)
|
||
+ 非图片、目录项、0 字节条目（都应被过滤）
|
||
Vol10.cbz 1 页 BMP(7x9)，stored
|
||
box/box1.cbz box2.cbz 只含档案册的子目录（书库行应写 "2 books"）
|
||
notabook.zip 只有文本 → 不是书
|
||
emptypages.cbz 目录项 + 文本 → 不是书
|
||
broken.cbz 非 zip 字节 → 不是书
|
||
truncated.cbz vol2 砍半（EOCD 丢失） → 不是书
|
||
.hidden.cbz 隐藏名 → 不参与
|
||
<out>/features/
|
||
pathsafe.cbz 安全/不安全条目混排（只应收 2 个安全图片页）
|
||
cjk.cbz 中文名自然序（第1页 < 第2页 < 第10页）
|
||
big.cbz 80 页 8x8 PNG（解压缓存 LRU 上限用）
|
||
bomb.cbz CD 声明单页 96MiB（应被过滤，防 zip 炸弹）
|
||
<out>/manifest.txt 每档案每条目一行：<档案相对路径>\t<条目名>\t<crc32>\t<size>
|
||
|
||
图片内容都是结构合法的 PNG/BMP/JPEG 头（PNG 是真图），C 测试用假解码器，
|
||
只校验解压字节与 manifest 的 CRC/尺寸一致、以及 meta 探测出的原始尺寸。
|
||
"""
|
||
import os
|
||
import struct
|
||
import sys
|
||
import zipfile
|
||
import zlib
|
||
|
||
DATE = (2024, 1, 1, 0, 0, 0)
|
||
HUGE_PAGE_LEN = 33 * 1024 * 1024 # > 32MiB 预算：锁定「单条目一次性溢出」行为
|
||
|
||
|
||
# ---------------------------------------------------------------- 图片构造
|
||
def png_bytes(w, h, seed=0):
|
||
raw = bytearray()
|
||
for y in range(h):
|
||
raw.append(0) # filter: none
|
||
for x in range(w):
|
||
raw += bytes(((x * 7 + y * 13 + seed) & 0xFF,
|
||
(x * 3 + y * 5 + seed) & 0xFF,
|
||
(x + y + seed) & 0xFF))
|
||
|
||
def chunk(tag, data):
|
||
return (struct.pack(">I", len(data)) + tag + data +
|
||
struct.pack(">I", zlib.crc32(tag + data) & 0xFFFFFFFF))
|
||
|
||
ihdr = struct.pack(">IIBBBBB", w, h, 8, 2, 0, 0, 0) # 8-bit RGB
|
||
return (b"\x89PNG\r\n\x1a\n" + chunk(b"IHDR", ihdr) +
|
||
chunk(b"IDAT", zlib.compress(bytes(raw), 9)) + chunk(b"IEND", b""))
|
||
|
||
|
||
def bmp_bytes(w, h):
|
||
stride = (w * 3 + 3) & ~3
|
||
px = bytes(((i * 11) & 0xFF) for i in range(stride * h))
|
||
hdr = b"BM" + struct.pack("<IHHI", 14 + 40 + len(px), 0, 0, 54)
|
||
ih = struct.pack("<IiiHHIIiiII", 40, w, h, 1, 24, 0, len(px), 2835, 2835, 0, 0)
|
||
return hdr + ih + px
|
||
|
||
|
||
def jpeg_header_bytes(w, h):
|
||
"""SOI + APP0 + DQT + SOF0 + EOI：头段合法（探测只看 SOF0），不做熵编码。"""
|
||
app0 = (b"\xFF\xE0" + struct.pack(">H", 16) +
|
||
b"JFIF\x00\x01\x01\x00\x00\x01\x00\x01\x00\x00")
|
||
dqt = b"\xFF\xDB" + struct.pack(">H", 67) + b"\x00" + bytes(64)
|
||
sof = b"\xFF\xC0" + struct.pack(">HBHHB", 11, 8, h, w, 1) + b"\x01\x11\x00"
|
||
return b"\xFF\xD8" + app0 + dqt + sof + b"\xFF\xD9"
|
||
|
||
|
||
# ---------------------------------------------------------------- zip 构造
|
||
def write_zip(path, entries):
|
||
"""entries: [(name, data, zipfile 压缩方式)]"""
|
||
with zipfile.ZipFile(path, "w") as z:
|
||
for name, data, method in entries:
|
||
zi = zipfile.ZipInfo(name, DATE)
|
||
zi.compress_type = method
|
||
z.writestr(zi, data)
|
||
|
||
|
||
def find_cd_record(data, name):
|
||
eocd = data.rfind(b"PK\x05\x06")
|
||
if eocd < 0:
|
||
raise SystemExit("generator: no EOCD")
|
||
total, _cd_size, cd_off = struct.unpack_from("<HII", data, eocd + 10)
|
||
p = cd_off
|
||
for _ in range(total):
|
||
if bytes(data[p:p + 4]) != b"PK\x01\x02":
|
||
raise SystemExit("generator: bad CD signature")
|
||
nlen, elen, clen = struct.unpack_from("<HHH", data, p + 28)
|
||
got = bytes(data[p + 46:p + 46 + nlen]).decode("utf-8")
|
||
if got == name:
|
||
return p
|
||
p += 46 + nlen + elen + clen
|
||
raise SystemExit("generator: entry not found: " + name)
|
||
|
||
|
||
# ---------------------------------------------------------------- 夹具
|
||
def build_library(ro):
|
||
lib = os.path.join(ro, "library")
|
||
os.makedirs(lib, exist_ok=True)
|
||
|
||
os.makedirs(os.path.join(lib, "bookA"), exist_ok=True)
|
||
with open(os.path.join(lib, "bookA", "001.png"), "wb") as f:
|
||
f.write(png_bytes(30, 20, 1))
|
||
with open(os.path.join(lib, "bookA", "002.png"), "wb") as f:
|
||
f.write(png_bytes(30, 21, 2))
|
||
|
||
write_zip(os.path.join(lib, "vol2.cbz"), [
|
||
("ch01/p10.png", png_bytes(40, 30, 3), zipfile.ZIP_DEFLATED),
|
||
("readme.txt", b"not a page\n", zipfile.ZIP_STORED),
|
||
("ch01/p2.png", png_bytes(41, 31, 4), zipfile.ZIP_DEFLATED),
|
||
("onlydir/", b"", zipfile.ZIP_STORED),
|
||
("\u4e2d\u6587\u540d.png", png_bytes(12, 34, 5), zipfile.ZIP_DEFLATED),
|
||
("empty.bin", b"", zipfile.ZIP_DEFLATED),
|
||
])
|
||
|
||
write_zip(os.path.join(lib, "Vol10.cbz"), [
|
||
("cover.bmp", bmp_bytes(7, 9), zipfile.ZIP_STORED),
|
||
])
|
||
|
||
os.makedirs(os.path.join(lib, "box"), exist_ok=True)
|
||
write_zip(os.path.join(lib, "box", "box1.cbz"), [
|
||
("a.png", png_bytes(8, 8, 6), zipfile.ZIP_DEFLATED),
|
||
("b.png", png_bytes(8, 9, 7), zipfile.ZIP_DEFLATED),
|
||
])
|
||
write_zip(os.path.join(lib, "box", "box2.cbz"), [
|
||
("c.jpg", jpeg_header_bytes(12, 34), zipfile.ZIP_DEFLATED),
|
||
])
|
||
|
||
write_zip(os.path.join(lib, "notabook.zip"), [
|
||
("readme.txt", b"nothing to read\n", zipfile.ZIP_STORED),
|
||
])
|
||
write_zip(os.path.join(lib, "emptypages.cbz"), [
|
||
("onlydir/", b"", zipfile.ZIP_STORED),
|
||
("readme.txt", b"x\n", zipfile.ZIP_STORED),
|
||
])
|
||
with open(os.path.join(lib, "broken.cbz"), "wb") as f:
|
||
f.write(b"NOTAZIP!" * 100)
|
||
|
||
raw = open(os.path.join(lib, "vol2.cbz"), "rb").read()
|
||
with open(os.path.join(lib, "truncated.cbz"), "wb") as f:
|
||
f.write(raw[:len(raw) // 2])
|
||
|
||
write_zip(os.path.join(lib, ".hidden.cbz"), [
|
||
("h.png", png_bytes(8, 8, 8), zipfile.ZIP_DEFLATED),
|
||
])
|
||
|
||
|
||
def build_features(ro):
|
||
feat = os.path.join(ro, "features")
|
||
os.makedirs(feat, exist_ok=True)
|
||
|
||
write_zip(os.path.join(feat, "pathsafe.cbz"), [
|
||
("ok1.png", png_bytes(10, 10, 9), zipfile.ZIP_DEFLATED),
|
||
("../evil.png", png_bytes(10, 10, 10), zipfile.ZIP_DEFLATED),
|
||
("/abs.png", png_bytes(10, 10, 11), zipfile.ZIP_DEFLATED),
|
||
("a/../../b.png", png_bytes(10, 10, 12), zipfile.ZIP_DEFLATED),
|
||
("sub/ok2.jpg", jpeg_header_bytes(13, 17), zipfile.ZIP_DEFLATED),
|
||
("sub/../ok3.png", png_bytes(10, 10, 13), zipfile.ZIP_DEFLATED),
|
||
(".hidden/x.png", png_bytes(10, 10, 14), zipfile.ZIP_DEFLATED),
|
||
("._mac.png", png_bytes(10, 10, 15), zipfile.ZIP_DEFLATED),
|
||
("sub\\win.png", png_bytes(10, 10, 16), zipfile.ZIP_DEFLATED),
|
||
("text.txt", b"nope\n", zipfile.ZIP_STORED),
|
||
("dir/", b"", zipfile.ZIP_STORED),
|
||
])
|
||
|
||
write_zip(os.path.join(feat, "cjk.cbz"), [
|
||
("\u7b2c2\u9875.png", png_bytes(9, 9, 17), zipfile.ZIP_DEFLATED),
|
||
("\u7b2c10\u9875.png", png_bytes(9, 9, 18), zipfile.ZIP_DEFLATED),
|
||
("\u7b2c1\u9875.png", png_bytes(9, 9, 19), zipfile.ZIP_DEFLATED),
|
||
])
|
||
|
||
entries = []
|
||
for i in range(80):
|
||
entries.append(("p%03d.png" % i, png_bytes(8, 8, 20 + i), zipfile.ZIP_DEFLATED))
|
||
write_zip(os.path.join(feat, "big.cbz"), entries)
|
||
|
||
bomb = os.path.join(feat, "bomb.cbz")
|
||
write_zip(bomb, [("huge.png", png_bytes(10, 10, 30), zipfile.ZIP_STORED)])
|
||
data = bytearray(open(bomb, "rb").read())
|
||
rec = find_cd_record(data, "huge.png")
|
||
struct.pack_into("<I", data, rec + 24, 0x06000000) # 96MiB > 64MiB 上限
|
||
with open(bomb, "wb") as f:
|
||
f.write(bytes(data))
|
||
|
||
raw = open(os.path.join(feat, "pathsafe.cbz"), "rb").read()
|
||
with open(os.path.join(feat, "truncated.cbz"), "wb") as f:
|
||
f.write(raw[:len(raw) // 2])
|
||
|
||
# 单页 33MiB（PNG 头 + 可压零填充，deflate 后只有几十 KB）：
|
||
# 页过滤放行（< 64MiB），解压后落缓存时应允许一次性超过 32MiB 预算。
|
||
head = png_bytes(64, 64, 31)
|
||
payload = head + b"\x00" * (HUGE_PAGE_LEN - len(head))
|
||
write_zip(os.path.join(feat, "hugepage.cbz"), [
|
||
("huge.png", payload, zipfile.ZIP_DEFLATED),
|
||
])
|
||
|
||
|
||
def write_manifest(ro):
|
||
"""每档案每条目：<相对路径>\t<条目名>\t<crc32>\t<size>（Tab 分隔，UTF-8 原名）。"""
|
||
lines = []
|
||
for base, _dirs, files in os.walk(ro):
|
||
for name in sorted(files):
|
||
if not name.lower().endswith((".cbz", ".zip")):
|
||
continue
|
||
if name == "hugepage.cbz":
|
||
continue # 33MiB 页只为缓存溢出测试，manifest 不校验它
|
||
path = os.path.join(base, name)
|
||
rel = os.path.relpath(path, ro)
|
||
try:
|
||
with zipfile.ZipFile(path) as z:
|
||
for info in z.infolist():
|
||
if info.is_dir():
|
||
continue
|
||
payload = z.read(info)
|
||
lines.append("%s\t%s\t%08x\t%d" % (
|
||
rel, info.filename, zlib.crc32(payload) & 0xFFFFFFFF,
|
||
len(payload)))
|
||
except zipfile.BadZipFile:
|
||
continue
|
||
with open(os.path.join(ro, "manifest.txt"), "w", encoding="utf-8") as f:
|
||
f.write("\n".join(lines) + "\n")
|
||
|
||
|
||
def main():
|
||
if len(sys.argv) != 2:
|
||
raise SystemExit("usage: make_cbz_samples.py <out_dir>")
|
||
ro = sys.argv[1]
|
||
os.makedirs(ro, exist_ok=True)
|
||
build_library(ro)
|
||
build_features(ro)
|
||
write_manifest(ro)
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main()
|