From: Eric Biggers <ebiggers3@gmail.com>
Date: Sun, 7 Jun 2015 14:56:41 +0000 (-0500)
Subject: Add new helper function for ignoring scanned directory entries
X-Git-Tag: v1.8.2~63
X-Git-Url: https://wimlib.net/git/?p=wimlib;a=commitdiff_plain;h=b87b3beaacd32c045ab77dc56c6b90677d882830

Add new helper function for ignoring scanned directory entries
---

diff --git a/include/wimlib/capture.h b/include/wimlib/capture.h
index 7372c052..a731bdb0 100644
--- a/include/wimlib/capture.h
+++ b/include/wimlib/capture.h
@@ -117,4 +117,7 @@ report_capture_error(struct capture_params *params, int error_code,
 	return report_error(params->progfunc, params->progctx, error_code, path);
 }
 
+extern bool
+should_ignore_filename(const tchar *name, int name_nchars);
+
 #endif /* _WIMLIB_CAPTURE_H */
diff --git a/src/capture_common.c b/src/capture_common.c
index 0fa71adc..ee051aea 100644
--- a/src/capture_common.c
+++ b/src/capture_common.c
@@ -307,3 +307,36 @@ try_exclude(const tchar *full_path, const struct capture_params *params)
 
 	return 0;
 }
+
+/*
+ * Determine whether a directory entry of the specified name should be ignored.
+ * This is a lower level function which runs prior to try_exclude().  It handles
+ * the standard '.' and '..' entries, which show up in directory listings but
+ * should not be archived.  It also checks for odd filenames that usually should
+ * not exist but could cause problems if archiving them were to be attempted.
+ */
+bool
+should_ignore_filename(const tchar *name, const int name_nchars)
+{
+	if (name_nchars <= 0) {
+		WARNING("Ignoring empty filename");
+		return true;
+	}
+
+	if (name[0] == T('.') &&
+	    (name_nchars == 1 || (name_nchars == 2 && name[1] == T('.'))))
+		return true;
+
+	for (int i = 0; i < name_nchars; i++) {
+		if (name[i] == T('\0')) {
+			WARNING("Ignoring filename containing embedded null character");
+			return true;
+		}
+		if (name[i] == OS_PREFERRED_PATH_SEPARATOR) {
+			WARNING("Ignoring filename containing embedded path separator");
+			return true;
+		}
+	}
+
+	return false;
+}
diff --git a/src/ntfs-3g_capture.c b/src/ntfs-3g_capture.c
index 1365c503..de472327 100644
--- a/src/ntfs-3g_capture.c
+++ b/src/ntfs-3g_capture.c
@@ -591,17 +591,13 @@ filldir(void *_ctx, const ntfschar *name, const int name_nchars,
 			goto out;
 	}
 
-	/* Ignore . and .. entries  */
-	ret = 0;
-	if ((name_nchars == 1 && name[0] == cpu_to_le16('.')) ||
-	    (name_nchars == 2 && name[0] == cpu_to_le16('.') &&
-				 name[1] == cpu_to_le16('.')))
-		goto out;
-
 	ret = utf16le_to_tstr(name, name_nbytes, &mbs_name, &mbs_name_nbytes);
 	if (ret)
 		goto out;
 
+	if (should_ignore_filename(mbs_name, mbs_name_nbytes))
+		goto out_free_mbs_name;
+
 	path_len = ctx->path_len;
 	if (path_len != 1)
 		ctx->path[path_len++] = '/';
@@ -613,6 +609,7 @@ filldir(void *_ctx, const ntfschar *name, const int name_nchars,
 						  ctx->volume, ctx->params);
 	if (child)
 		dentry_add_child(ctx->parent, child);
+out_free_mbs_name:
 	FREE(mbs_name);
 out:
 	ctx->ret = ret;
diff --git a/src/unix_capture.c b/src/unix_capture.c
index f7dd5074..9b3ea4fe 100644
--- a/src/unix_capture.c
+++ b/src/unix_capture.c
@@ -180,13 +180,12 @@ unix_scan_directory(struct wim_dentry *dir_dentry,
 			break;
 		}
 
-		if (entry->d_name[0] == '.' &&
-		    (entry->d_name[1] == '\0' ||
-		     (entry->d_name[1] == '.' && entry->d_name[2] == '\0')))
+		name_len = strlen(entry->d_name);
+
+		if (should_ignore_filename(entry->d_name, name_len))
 			continue;
 
 		full_path[full_path_len] = '/';
-		name_len = strlen(entry->d_name);
 		memcpy(&full_path[full_path_len + 1], entry->d_name, name_len + 1);
 		ret = unix_build_dentry_tree_recursive(&child,
 						       full_path,
diff --git a/src/win32_capture.c b/src/win32_capture.c
index 62855668..854428e4 100644
--- a/src/win32_capture.c
+++ b/src/win32_capture.c
@@ -415,9 +415,8 @@ winnt_recurse_directory(HANDLE h,
 	{
 		const FILE_NAMES_INFORMATION *info = buf;
 		for (;;) {
-			if (!(info->FileNameLength == 2 && info->FileName[0] == L'.') &&
-			    !(info->FileNameLength == 4 && info->FileName[0] == L'.' &&
-							   info->FileName[1] == L'.'))
+			if (!should_ignore_filename(info->FileName,
+						    info->FileNameLength / 2))
 			{
 				wchar_t *p;
 				wchar_t *filename;