win32: add (some) long path supportlong_paths

Use the longPathAware manifest entry to handle path lengths above
260.  This requires the LongPathsEnabled registry key to be set
at run-time.

Modify (some) WIN32 API calls to use the wide-character variant.

These features are controlled by the build-time configuration
option FEATURE_LONG_PATHS, which is not enabled by default.

5 files changed, 109 insertions, 14 deletions

diff --git a/Config.in b/Config.in
index 51de85284..39d9b0589 100644
--- a/Config.in
+++ b/Config.in
@@ -494,6 +494,15 @@ config FEATURE_PNG_ALL
 
 endchoice
 
+config FEATURE_LONG_PATHS
+        bool "Support for long paths"
+        default n
+        depends on FEATURE_RESOURCES
+        help
+        Allow paths longer than 260 characters. This is only supported
+        on Windows 10+. The registry key LongPathsEnabled must be set
+        to 1.
+
 config FEATURE_EURO
         bool "Support the euro currency symbol"
         default y
diff --git a/libbb/appletlib.c b/libbb/appletlib.c
index 481cc278e..12de2b610 100644
--- a/libbb/appletlib.c
+++ b/libbb/appletlib.c
@@ -928,6 +928,9 @@ int busybox_main(int argc UNUSED_PARAM, char **argv)
 #   if ENABLE_FEATURE_UTF8_MANIFEST
                 full_write1_str("; Unicode");
 #   endif
+#   if ENABLE_FEATURE_LONG_PATHS
+                full_write1_str("; long paths");
+#   endif
                 full_write1_str(")\n\n");
 #  else
                 full_write1_str(" multi-call binary.\n"); /* reuse */
diff --git a/win32/dirent.c b/win32/dirent.c
index f0e8deae2..872f44fbc 100644
--- a/win32/dirent.c
+++ b/win32/dirent.c
@@ -8,26 +8,44 @@ struct DIR {
         int got_dotdot;
 };
 
+static inline unsigned char get_dtype(DWORD attr, DWORD reserved0)
+{
+        if ((attr & FILE_ATTRIBUTE_REPARSE_POINT) &&
+                        (reserved0 == IO_REPARSE_TAG_SYMLINK ||
+                        reserved0 == IO_REPARSE_TAG_MOUNT_POINT ||
+                        reserved0 == IO_REPARSE_TAG_APPEXECLINK))
+                return DT_LNK;
+        if (attr & FILE_ATTRIBUTE_DIRECTORY)
+                return DT_DIR;
+        return DT_REG;
+}
+
+#if !ENABLE_FEATURE_LONG_PATHS
 static inline void finddata2dirent(struct dirent *ent, WIN32_FIND_DATAA *fdata)
 {
-        /* copy file name from WIN32_FIND_DATA to dirent */
+    /* copy file name from WIN32_FIND_DATA to dirent */
         strcpy(ent->d_name, fdata->cFileName);
-
-        if ((fdata->dwFileAttributes & FILE_ATTRIBUTE_REPARSE_POINT) &&
-                        (fdata->dwReserved0 == IO_REPARSE_TAG_SYMLINK ||
-                        fdata->dwReserved0 ==  IO_REPARSE_TAG_MOUNT_POINT ||
-                        fdata->dwReserved0 ==  IO_REPARSE_TAG_APPEXECLINK))
-                ent->d_type = DT_LNK;
-        else if (fdata->dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY)
-                ent->d_type = DT_DIR;
-        else
-                ent->d_type = DT_REG;
+        ent->d_type = get_dtype(fdata->dwFileAttributes, fdata->dwReserved0);
 }
+#else
+static inline void finddata2dirent(struct dirent *ent, WIN32_FIND_DATAW *fdata)
+{
+        /* copy file name from WIN32_FIND_DATAW to dirent */
+        WideCharToMultiByte(CP_ACP, 0, fdata->cFileName, -1,
+                        ent->d_name, PATH_MAX, NULL, NULL);
+        ent->d_type = get_dtype(fdata->dwFileAttributes, fdata->dwReserved0);
+}
+#endif
 
 DIR *opendir(const char *name)
 {
+#if !ENABLE_FEATURE_LONG_PATHS
         char pattern[MAX_PATH];
         WIN32_FIND_DATAA fdata;
+#else
+        wchar_t wpattern[32768];
+        WIN32_FIND_DATAW fdata;
+#endif
         HANDLE h;
         int len;
         DIR *dir;
@@ -37,6 +55,8 @@ DIR *opendir(const char *name)
                 errno = EINVAL;
                 return NULL;
         }
+
+#if !ENABLE_FEATURE_LONG_PATHS
         /* check that the pattern won't be too long for FindFirstFileA */
         len = strlen(name);
         if (len + 2 >= MAX_PATH) {
@@ -59,6 +79,30 @@ DIR *opendir(const char *name)
                 errno = (err == ERROR_DIRECTORY) ? ENOTDIR : err_win_to_posix();
                 return NULL;
         }
+#else /* ENABLE_FEATURE_LONG_PATHS */
+        /* Convert to wide string.  CP_ACP is CP_UTF8 when the UTF-8 manifest
+         * is active, else the system ANSI code page. */
+        len = MultiByteToWideChar(CP_ACP, 0, name, -1, wpattern, 32760);
+        if (len == 0) {
+                errno = EINVAL;
+                return NULL;
+        }
+        len--;  /* exclude null terminator from length */
+
+        /* append optional '\' and wildcard '*' */
+        if (len && wpattern[len - 1] != L'\\' && wpattern[len - 1] != L'/')
+                wpattern[len++] = L'\\';
+        wpattern[len++] = L'*';
+        wpattern[len] = 0;
+
+        /* open find handle using wide API */
+        h = FindFirstFileW(wpattern, &fdata);
+        if (h == INVALID_HANDLE_VALUE) {
+                DWORD err = GetLastError();
+                errno = (err == ERROR_DIRECTORY) ? ENOTDIR : err_win_to_posix();
+                return NULL;
+        }
+#endif
 
         /* initialize DIR structure and copy first dir entry */
         dir = xzalloc(sizeof(DIR));
@@ -80,8 +124,14 @@ struct dirent *readdir(DIR *dir)
         /* if first entry, dirent has already been set up by opendir */
         if (dir->not_first) {
                 /* get next entry and convert from WIN32_FIND_DATA to dirent */
+#if !ENABLE_FEATURE_LONG_PATHS
                 WIN32_FIND_DATAA fdata;
-                if (FindNextFileA(dir->dd_handle, &fdata)) {
+                if (FindNextFileA(dir->dd_handle, &fdata))
+#else
+                WIN32_FIND_DATAW fdata;
+                if (FindNextFileW(dir->dd_handle, &fdata))
+#endif
+                {
                         finddata2dirent(&dir->dd_dir, &fdata);
                 } else if (!dir->got_dot) {
                         strcpy(dir->dd_dir.d_name, ".");
diff --git a/win32/mingw.c b/win32/mingw.c
index 122337d19..0c28d87fe 100644
--- a/win32/mingw.c
+++ b/win32/mingw.c
@@ -396,6 +396,9 @@ static inline mode_t file_attr_to_st_mode(DWORD attr)
 
 static int get_file_attr(const char *fname, WIN32_FILE_ATTRIBUTE_DATA *fdata)
 {
+#if ENABLE_FEATURE_LONG_PATHS
+        wchar_t wpath[32768];
+#endif
         char *want_dir;
         int dev = get_dev_type(fname);
 
@@ -412,7 +415,18 @@ static int get_file_attr(const char *fname, WIN32_FILE_ATTRIBUTE_DATA *fdata)
         }
 
         want_dir = last_char_is_dir_sep(fname);
-        if (GetFileAttributesExA(fname, GetFileExInfoStandard, fdata)) {
+
+#if !ENABLE_FEATURE_LONG_PATHS
+        if (GetFileAttributesExA(fname, GetFileExInfoStandard, fdata))
+#else
+        /* Convert to wide string for long path support.  CP_ACP is CP_UTF8
+         * when the UTF-8 manifest is active, else the system ANSI code page. */
+        if (MultiByteToWideChar(CP_ACP, 0, fname, -1, wpath, 32768) == 0)
+                return EINVAL;
+
+        if (GetFileAttributesExW(wpath, GetFileExInfoStandard, fdata))
+#endif
+        {
                 if (!(fdata->dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) && want_dir)
                         return ENOTDIR;
                 fdata->dwFileAttributes &= ~FILE_ATTRIBUTE_DEVICE;
@@ -421,9 +435,16 @@ static int get_file_attr(const char *fname, WIN32_FILE_ATTRIBUTE_DATA *fdata)
 
         if (GetLastError() == ERROR_SHARING_VIOLATION) {
                 HANDLE hnd;
+#if !ENABLE_FEATURE_LONG_PATHS
                 WIN32_FIND_DATA fd;
 
-                if ((hnd=FindFirstFile(fname, &fd)) != INVALID_HANDLE_VALUE) {
+                if ((hnd=FindFirstFile(fname, &fd)) != INVALID_HANDLE_VALUE)
+#else
+                WIN32_FIND_DATAW fd;
+
+                if ((hnd=FindFirstFileW(wpath, &fd)) != INVALID_HANDLE_VALUE)
+#endif
+                {
                         fdata->dwFileAttributes =
                                         fd.dwFileAttributes & ~FILE_ATTRIBUTE_DEVICE;
                         fdata->ftCreationTime = fd.ftCreationTime;
diff --git a/win32/resources/resources.rc b/win32/resources/resources.rc
index b76f1af23..50642831c 100644
--- a/win32/resources/resources.rc
+++ b/win32/resources/resources.rc
@@ -52,6 +52,9 @@ BEGIN
     "<asmv3:application>"
       "<asmv3:windowsSettings>"
         "<activeCodePage xmlns=""http://schemas.microsoft.com/SMI/2019/WindowsSettings"">UTF-8</activeCodePage>"
+# if ENABLE_FEATURE_LONG_PATHS
+        "<longPathAware xmlns=""http://schemas.microsoft.com/SMI/2016/WindowsSettings"">true</longPathAware>"
+# endif
       "</asmv3:windowsSettings>"
     "</asmv3:application>"
     "<trustInfo xmlns=""urn:schemas-microsoft-com:asm.v3"">"
@@ -78,7 +81,16 @@ END
 1 24
 BEGIN
   "<?xml version=""1.0"" encoding=""UTF-8"" standalone=""yes""?>"
+# if !ENABLE_FEATURE_LONG_PATHS
   "<assembly xmlns=""urn:schemas-microsoft-com:asm.v1"" manifestVersion=""1.0"">"
+# else
+  "<assembly xmlns=""urn:schemas-microsoft-com:asm.v1"" xmlns:asmv3=""urn:schemas-microsoft-com:asm.v3"" manifestVersion=""1.0"">"
+    "<asmv3:application>"
+      "<asmv3:windowsSettings>"
+        "<longPathAware xmlns=""http://schemas.microsoft.com/SMI/2016/WindowsSettings"">true</longPathAware>"
+      "</asmv3:windowsSettings>"
+    "</asmv3:application>"
+# endif
     "<trustInfo xmlns=""urn:schemas-microsoft-com:asm.v3"">"
       "<security>"
         "<requestedPrivileges>"