ViperC/App/lib/Stub/StubMerger/StubMergerCollect.vp

import t, c
from stdint import *
import memhub
import string
import stdlib
import mlirlite
import w32.win32file as win32file
import w32.win32base as win32base
import w32.fileio as fileio
import viperlib
import sys
import ast
import lib.core.VLogger as VLogger
import lib.core.IncludesScanner as IncludesScanner
import lib.core.Gen.Translator as HandlesTranslator
import lib.Phase.BuildPipeline as BuildPipeline
import lib.Stub.StubGen.VpiReader as VpiReader
import lib.Projectrans.Utils as Utils
import lib.Projectrans.Config as Config


import lib.Stub.StubMerger.StubMergerStore as StubMergerStore
import lib.Stub.StubMerger.StubMergerTopo as StubMergerTopo
import lib.Stub.StubMerger.StubMergerIR as StubMergerIR
import lib.Stub.StubMerger.StubMergerBuild as StubMergerBuild
import lib.Stub.StubMerger as StubMerger

# ============================================================
# StubMergerCollect - 切片路径构建与 stub 文件 SHA1 收集
#
# 函数实现直接位于本模块，调用方经本模块前缀访问；全局变量跨模块经 StubMerger._xxx 访问。
# ============================================================

# ============================================================
# _sliced_path - 构建切片路径（stdlib.malloc 分配，调用者负责 free；level>0 时自动调 BuildPipeline.ensure_dir 建子目录）
#
# 按 Config.Sha1SliceLevel 分散到 SHA1 前缀子目录：level=0 → {temp_dir}/{sha1}.{ext}；level=1 → {temp_dir}/b7/{sha1}.{ext}；level=2 → {temp_dir}/b7/90/{sha1}.{ext}
#
# Args: temp_dir 基础目录；td_len temp_dir 长度（避免重复 strlen）；sha1 SHA1 字符串；ext 文件扩展名（如 "stub.ll"）
# Returns: 完整路径（stdlib.malloc 分配，需 stdlib.free），None 失败
# ============================================================
def _sliced_path(temp_dir: str, td_len: sizet, sha1: str, ext: str) -> str:
    """构建切片路径（stdlib.malloc 分配，调用者负责 free）

    level>0 时自动调用 ensure_dir 创建子目录。
    """
    if temp_dir is None or sha1 is None or ext is None:
        return None
    sha1_len: sizet = string.strlen(sha1)
    ext_len: sizet = string.strlen(ext)

    # 子目录前缀：复用 Config.slice_subdir（SHA1 切片唯一权威实现，mbuddy 分配无需 free）
    subdir_buf: str = Config.slice_subdir(sha1, Config.Sha1SliceLevel)
    sub_len: sizet = 0
    if subdir_buf is not None:
        sub_len = string.strlen(subdir_buf)

        # 确保子目录存在: {temp_dir}/{subdir}
        # 不创建目录会导致 fileio.File(MODE.W) 静默失败（file.closed==True），
        # 进而读取端读不到文件返回 0（"加载 IR 失败"）。
        dir_path_len: sizet = td_len + sub_len + 2
        dir_path: bytes = stdlib.malloc(dir_path_len)
        if dir_path is not None:
            viperlib.snprintf(dir_path, dir_path_len, "%s/%s", temp_dir, subdir_buf)
            BuildPipeline.ensure_dir(dir_path)
            stdlib.free(dir_path)

    # 构建完整路径
    path: str = None
    path_len: sizet = 0
    if sub_len > 0:
        path_len = td_len + sub_len + sha1_len + ext_len + 4
        path = stdlib.malloc(path_len)
        if path is not None:
            viperlib.snprintf(path, path_len, "%s/%s/%s.%s", temp_dir, subdir_buf, sha1, ext)
    else:
        path_len = td_len + sha1_len + ext_len + 3
        path = stdlib.malloc(path_len)
        if path is not None:
            viperlib.snprintf(path, path_len, "%s/%s.%s", temp_dir, sha1, ext)

    return path

# ============================================================
# _load_includes_sha1_set - 从全局内存存储器收集 includes 的 SHA1
#
# 数据来源是 PopulateSha1MapStore 填充的全局存储器（不读取 _sha1_map.txt）。
# 用于过滤 stub 文件，避免加载 TransPyV 自身源文件（lib/*）的 stub 导致链接失败。
# ============================================================
def _load_includes_sha1_set(pool: memhub.MemBuddy*,
                            temp_dir: str,
                            sha1_set: str) -> int:
    """从全局内存存储器将 includes 的 SHA1 写入 sha1_set

    pool/temp_dir 参数保留以兼容现有调用者，但不再使用。
    数据来源是 PopulateSha1MapStore 填充的全局存储器。
    sha1_set 大小为 MAX_INCLUDES_SHA1 * 17（每个 SHA1 16字符+null）
    返回找到的 includes SHA1 数量，-1 表示错误"""
    if sha1_set is None:
        return -1
    if StubMerger._sha1_store_arr is None:
        return -1
    count: int = 0
    for i in range(StubMerger._sha1_store_count):
        if count >= StubMerger.MAX_INCLUDES_SHA1:
            break
        idx: sizet = sizet(i) * 17
        string.strcpy(sha1_set + sizet(count) * 17, StubMerger._sha1_store_arr + idx)
        count += 1
    return count
# ============================================================
# _is_in_sha1_set - 检查 SHA1 是否在集合中
# ============================================================
def _is_in_sha1_set(sha1: str, sha1_set: str, set_count: int) -> int:
    """检查 sha1 是否在集合中"""
    if sha1 is None or sha1_set is None:
        return 0
    for i in range(set_count):
        if string.strcmp(sha1_set + i * 17, sha1) == 0:
            return 1
    return 0
# ============================================================
# WriteIncludesSha1Map - 将 includes 扫描结果写入 _sha1_map.txt（仅人类可读的调试输出，程序内部不读取，
# 机器分析用 PopulateSha1MapStore 填充的全局内存存储器）
#
# Args: mb 内存池；temp_dir 临时目录（_sha1_map.txt 写入位置）；scan_result IncludesScanner 扫描结果
# Returns: 0 成功，非 0 失败
# ============================================================
def WriteIncludesSha1Map(mb: memhub.MemBuddy*, temp_dir: str,
                         scan_result: rawptr[IncludesScanner.ScanResult],
                         filter_set: str,
                         filter_count: int) -> int:
    """将 includes 扫描结果写入 _sha1_map.txt（人类可读输出，程序内部使用内存存储器）

    若 filter_set 不为 None 且 filter_count > 0，只写入 filter_set 中的条目。
    """
    if temp_dir is None or scan_result is None:
        return 1

    # 构造路径 temp_dir/_sha1_map.txt
    dir_len: sizet = string.strlen(temp_dir)
    map_path: bytes = stdlib.malloc(dir_len + 32)
    if map_path is None:
        return 1
    viperlib.snprintf(map_path, dir_len + 32, "%s/_sha1_map.txt", temp_dir)

    # 打开文件写入（CREATE_ALWAYS）
    f: rawptr[fileio.File] = fileio.File(map_path, fileio.MODE.W)
    if f.closed:
        return 1

    # 写入每个 include 条目: {sha1}:includes/{rel_path}\n
    entry_size_w: sizet = IncludesScanner.FileEntry.__sizeof__()
    line_buf: bytes = stdlib.malloc(512)
    if line_buf is None:
        f.close()
        return 1

    written_count: int = 0
    for i in range(scan_result.Count):
        ea_w: uint64 = uint64(scan_result.Entries) + i * entry_size_w
        ent_w: rawptr[IncludesScanner.FileEntry] = (rawptr[IncludesScanner.FileEntry])(ea_w)
        if ent_w is None or ent_w.Sha1 is None or ent_w.RelPath is None:
            continue
        # 若提供过滤集合，只写入集合中的条目
        if filter_set is not None and filter_count > 0:
            if _is_in_sha1_set(ent_w.Sha1, filter_set, filter_count) == 0:
                continue
        viperlib.snprintf(line_buf, 512, "%s:includes/%s\n", ent_w.Sha1, ent_w.RelPath)
        line_len_w: sizet = string.strlen(line_buf)
        f.write(line_buf, line_len_w)
        written_count += 1

    f.close()
    return 0