修复书签 HTML 导入并发布 v2.4.1

CMakeLists.txt

@@ -1,7 +1,7 @@
 cmake_minimum_required(VERSION 3.25)
 
 project(UltraLightBrowser
-    VERSION 2.4.0
+    VERSION 2.4.1
     DESCRIPTION "Minimal, Ultra-Fast Windows 11 Native Browser Shell (macOS Safari UI Edition)"
     LANGUAGES CXX C
 )

README.md

@@ -294,6 +294,7 @@ tests/                       单元测试、页面资源检查、Windows 界面
 
 ## 最近更新
 
+- **v2.4.1**：修复书签 HTML 导入：正确解析属性空格、制表符和完整属性名，避免把 DATA-HREF 或标题中的文字误当网址；按标准标记识别不同语言的收藏栏；超长网址不再虚报导入成功。新增覆盖以上情况的回归测试。
 - **v2.4.0**：新增「显卡视频增强」（设置 → 高级，默认开）：让 NVIDIA RTX 显卡做视频超分辨率和 SDR 转 HDR、Intel 显卡做超分；需在 NVIDIA App 里开启，接电源、画面被放大时生效。「播放诊断」显示视频放大倍数。
 - **v2.3.0**：音频增强的最后一级换成真峰值限幅器：采样点和采样点之间的峰值都压在 -1 dBTP 以下，音量放大到 300% 也不会削波破音；约增加 4ms 延迟。少数安全策略严格、不允许加载的网站自动改用原来的压缩保护，「播放诊断」会显示用的是哪一种。
 - **v2.2.3**：我的线路未连接时显示原因（本机没有 IPv6 / 连不到入口服务器 / 链接失效），便于排查。

resources/resource.rc

@@ -34,8 +34,8 @@ IDI_APP_ICON ICON "app.ico"
 
 // Version information
 VS_VERSION_INFO VERSIONINFO
-FILEVERSION     2,4,0,0
-PRODUCTVERSION  2,4,0,0
+FILEVERSION     2,4,1,0
+PRODUCTVERSION  2,4,1,0
 FILEFLAGSMASK   VS_FFI_FILEFLAGSMASK
 FILEFLAGS       0
 FILEOS          VOS_NT_WINDOWS32
@@ -48,12 +48,12 @@ BEGIN
         BEGIN
             VALUE "CompanyName",      "UltraLightBrowser"
             VALUE "FileDescription",  "UltraLight Browser - Minimal, Ultra-Fast Windows 11 Native Browser (Safari-style UI)"
-            VALUE "FileVersion",      "2.4.0.0"
+            VALUE "FileVersion",      "2.4.1.0"
             VALUE "InternalName",     "UltraLightBrowser"
             VALUE "LegalCopyright",   "Copyright (C) 2026 Freecode100Year"
             VALUE "OriginalFilename", "UltraLightBrowser.exe"
             VALUE "ProductName",      "UltraLight Browser"
-            VALUE "ProductVersion",   "2.4.0.0"
+            VALUE "ProductVersion",   "2.4.1.0"
         END
     END
     BLOCK "VarFileInfo"

src/Library.cpp

@@ -133,18 +133,31 @@ std::size_t FindCI(const std::string& hay, const std::string& needle, std::size_
 }
 
 std::string AttrValue(const std::string& tag, const std::string& name) {
-    auto pos = FindCI(tag, name + "=", 0);
-    if (pos == std::string::npos) return {};
-    pos += name.size() + 1;
-    if (pos >= tag.size()) return {};
-    const char quote = tag[pos];
-    if (quote == '"' || quote == '\'') {
-        const auto end = tag.find(quote, pos + 1);
-        if (end == std::string::npos) return {};
-        return tag.substr(pos + 1, end - pos - 1);
+    constexpr const char* whitespace = " \t\r\n\f";
+    // Walk complete attributes, skipping quoted values so DATA-HREF and text
+    // such as TITLE="href=..." cannot be mistaken for the actual link.
+    auto pos = tag.find_first_of(whitespace);
+    while (pos != std::string::npos && pos < tag.size()) {
+        pos = tag.find_first_not_of(whitespace, pos);
+        if (pos == std::string::npos) break;
+        const auto start = pos;
+        pos = tag.find_first_of("= \t\r\n\f", pos);
+        if (pos == std::string::npos) break;
+        const auto attribute = Lower(tag.substr(start, pos - start));
+        pos = tag.find_first_not_of(whitespace, pos);
+        if (pos == std::string::npos) break;
+        if (tag[pos] != '=') continue; // Boolean attribute; parse the next one.
+        pos = tag.find_first_not_of(whitespace, pos + 1);
+        if (pos == std::string::npos) break;
+        const char quote = tag[pos];
+        const bool quoted = quote == '"' || quote == '\'';
+        const auto valueStart = pos + (quoted ? 1 : 0);
+        const auto end = quoted ? tag.find(quote, valueStart) : tag.find_first_of(whitespace, valueStart);
+        if (quoted && end == std::string::npos) break;
+        if (attribute == name) return tag.substr(valueStart, end == std::string::npos ? end : end - valueStart);
+        pos = end == std::string::npos ? end : end + (quoted ? 1 : 0);
     }
-    const auto end = tag.find_first_of(" >", pos);
-    return tag.substr(pos, end == std::string::npos ? std::string::npos : end - pos);
+    return {};
 }
 
 } // namespace
@@ -394,15 +407,33 @@ int Library::ImportNetscapeHtml(const std::string& html, std::int64_t now) {
     while (pos < html.size()) {
         const auto lt = html.find('<', pos);
         if (lt == std::string::npos) break;
-        const auto gt = html.find('>', lt);
-        if (gt == std::string::npos) break;
+        if (html.compare(lt, 4, "<!--") == 0) {
+            const auto end = html.find("-->", lt + 4);
+            if (end == std::string::npos) break;
+            pos = end + 3;
+            continue;
+        }
+        auto gt = lt + 1;
+        char quote = 0;
+        for (; gt < html.size(); ++gt) {
+            const char c = html[gt];
+            if (quote) {
+                if (c == quote) quote = 0;
+            } else if (c == '"' || c == '\'') {
+                quote = c;
+            } else if (c == '>') {
+                break;
+            }
+        }
+        if (gt == html.size()) break;
         std::string tag = html.substr(lt + 1, gt - lt - 1);
         std::string name = Lower(tag.substr(0, tag.find_first_of(" \t\r\n")));
         pos = gt + 1;
         if (name == "h3") {
             const auto close = FindCI(html, "</h3>", pos);
             if (close == std::string::npos) break;
             pendingFolder = Clip(Trim(HtmlUnescape(html.substr(pos, close - pos))), 128);
+            if (Lower(AttrValue(tag, "personal_toolbar_folder")) == "true") pendingFolder = kFavoritesFolder;
             pos = close + 5;
         } else if (name == "dl") {
             folderStack.push_back(pendingFolder);
@@ -415,7 +446,7 @@ int Library::ImportNetscapeHtml(const std::string& html, std::int64_t now) {
             const std::string url = HtmlUnescape(AttrValue(tag, "href"));
             const std::string title = Trim(HtmlUnescape(html.substr(pos, close - pos)));
             pos = close + 4;
-            if (!IsWebUrl(url)) continue;
+            if (!IsWebUrl(url) || url.size() > kMaxUrl) continue;
             std::string folder = kBookmarksFolder;
             for (auto it = folderStack.rbegin(); it != folderStack.rend(); ++it) {
                 if (!it->empty()) { folder = *it; break; }
@@ -427,8 +458,7 @@ int Library::ImportNetscapeHtml(const std::string& html, std::int64_t now) {
             }
             AddFolder(folder);
             if (!existing.insert(folder + "\n" + url).second) continue;
-            AddBookmark(title, url, folder, now);
-            ++imported;
+            if (!AddBookmark(title, url, folder, now).empty()) ++imported;
         }
     }
     return imported;

tests/library_test.cpp

@@ -7,10 +7,41 @@
 
 using namespace UltraLight;
 
+static void TestBookmarkImport(const std::filesystem::path& dir) {
+    struct Case { const char* name; std::string html; int count; std::string url; std::string folder; };
+    const Case cases[] = {
+        {"attribute whitespace", R"(<A HREF = 'https://space.example/'>Space</A>)", 1, "https://space.example/", kBookmarksFolder},
+        {"attribute boundary", R"(<A DATA-HREF="https://wrong.example/" HREF="https://right.example/">Right</A>)", 1, "https://right.example/", kBookmarksFolder},
+        {"quoted attribute content", R"(<A TITLE="href=https://wrong.example/" HREF="https://right.example/">Right</A>)", 1, "https://right.example/", kBookmarksFolder},
+        {"quoted tag delimiter", R"(<A TITLE="a > b" HREF="https://right.example/">Right</A>)", 1, "https://right.example/", kBookmarksFolder},
+        {"comment text is not markup", R"(<!-- User's <A HREF="https://wrong.example/">old link</A> --><A HREF="https://right.example/">Right</A>)", 1, "https://right.example/", kBookmarksFolder},
+        {"tab-delimited unquoted URL", "<A HREF=https://tab.example/\tADD_DATE=1>Tab</A>", 1, "https://tab.example/", kBookmarksFolder},
+        {"boolean and mixed-case attributes", R"(<A DISABLED HrEf="https://case.example/">Case</A>)", 1, "https://case.example/", kBookmarksFolder},
+        {"data attribute alone is not a link", R"(<A DATA-HREF="https://wrong.example/">Not a link</A>)", 0, "", kBookmarksFolder},
+        {"unterminated attribute rejected", R"(<A HREF="https://bad.example/>Bad</A>)", 0, "", kBookmarksFolder},
+        {"localized toolbar", R"(<H3 PERSONAL_TOOLBAR_FOLDER="true">Meine Favoriten</H3><DL><A HREF="https://bar.example/">Bar</A></DL>)", 1, "https://bar.example/", kFavoritesFolder},
+        {"non-toolbar folder preserved", R"(<H3 PERSONAL_TOOLBAR_FOLDER="false">Meine Favoriten</H3><DL><A HREF="https://folder.example/">Folder</A></DL>)", 1, "https://folder.example/", "Meine Favoriten"},
+        {"URL size boundary accepted", "<A HREF=\"https://large.example/" + std::string(8170, 'a') + "\">Large</A>", 1, "https://large.example/" + std::string(8170, 'a'), kBookmarksFolder},
+        {"oversized URL rejected", "<A HREF=\"https://large.example/" + std::string(8192, 'a') + "\">Large</A>", 0, "", kBookmarksFolder},
+    };
+    int failures = 0;
+    for (const auto& c : cases) {
+        Library lib(dir / "import-cases");
+        const int count = lib.ImportNetscapeHtml(c.html, 1);
+        if (count != c.count || lib.Bookmarks().size() != static_cast<std::size_t>(c.count) ||
+            (c.count && (lib.Bookmarks().front().url != c.url || lib.Bookmarks().front().folder != c.folder))) {
+            std::cerr << "FAIL: bookmark import " << c.name << '\n';
+            ++failures;
+        }
+    }
+    assert(failures == 0);
+}
+
 int main() {
     const auto dir = std::filesystem::temp_directory_path() / "ulb-library-test";
     std::filesystem::remove_all(dir);
     std::filesystem::create_directories(dir);
+    TestBookmarkImport(dir);
     std::ofstream(dir / "history.json") << R"([{"url":"https://old.example/","title":"x","last":1,"visits":1}])";
     std::ofstream(dir / "session.json") << "[]";
     {