From: Duncan Horn <40036384+dunhor@users.noreply.github.com>
Date: Wed, 12 Jun 2024 19:01:40 +0000 (-0700)
Subject: Update ustar creation sanity check to use WCS path on Windows (#2230)
X-Git-Tag: v3.7.5~40
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=6ee1eebefdf41f36ef1a548c9a7000d132c453f3;p=thirdparty%2Flibarchive.git

Update ustar creation sanity check to use WCS path on Windows (#2230)

On Windows, the MBS pathname might be null if the string was set with a
WCS that can't be represented by the current locale. This is handled
properly by the rest of the code, but there's a sanity check that does
not make the proper distinction.

Note: this is a partial cherry-pick from
https://github.com/libarchive/libarchive/pull/2095, which I'm going to
go through and break into smaller pieces in hopes of getting some things
in while discussion of other things can continue.
---

diff --git a/libarchive/archive_write_set_format_ustar.c b/libarchive/archive_write_set_format_ustar.c
index 673487b27..d8f0b4584 100644
--- a/libarchive/archive_write_set_format_ustar.c
+++ b/libarchive/archive_write_set_format_ustar.c
@@ -254,7 +254,11 @@ archive_write_ustar_header(struct archive_write *a, struct archive_entry *entry)
 		sconv = ustar->opt_sconv;
 
 	/* Sanity check. */
+#if defined(_WIN32) && !defined(__CYGWIN__)
+	if (archive_entry_pathname_w(entry) == NULL) {
+#else
 	if (archive_entry_pathname(entry) == NULL) {
+#endif
 		archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
 		    "Can't record entry in tar file without pathname");
 		return (ARCHIVE_FAILED);
diff --git a/libarchive/test/test_ustar_filename_encoding.c b/libarchive/test/test_ustar_filename_encoding.c
index cc62453f1..1242bd1d3 100644
--- a/libarchive/test/test_ustar_filename_encoding.c
+++ b/libarchive/test/test_ustar_filename_encoding.c
@@ -390,3 +390,105 @@ DEFINE_TEST(test_ustar_filename_encoding_CP932_UTF8)
 	assertEqualMem(buff, "\xE8\xA1\xA8.txt", 7);
 }
 
+DEFINE_TEST(test_ustar_filename_encoding_UTF16_win)
+{
+#if !defined(_WIN32) || defined(__CYGWIN__)
+	skipping("This test is meant to verify unicode string handling"
+		" on Windows with UTF-16 names");
+	return;
+#else
+	struct archive *a;
+	struct archive_entry *entry;
+	char buff[4096];
+	size_t used;
+
+	/*
+	 * Don't call setlocale because we're verifying that the '_w' functions
+	 * work as expected when 'hdrcharset' is UTF-8
+	 */
+
+	/* Part 1: file */
+	a = archive_write_new();
+	assertEqualInt(ARCHIVE_OK, archive_write_set_format_ustar(a));
+	if (archive_write_set_options(a, "hdrcharset=UTF-8") != ARCHIVE_OK) {
+		skipping("This system cannot convert character-set"
+		    " from UTF-16 to UTF-8.");
+		archive_write_free(a);
+		return;
+	}
+	assertEqualInt(ARCHIVE_OK,
+	    archive_write_open_memory(a, buff, sizeof(buff), &used));
+
+	entry = archive_entry_new2(a);
+	/* Set the filename using a UTF-16 string */
+	archive_entry_copy_pathname_w(entry, L"\u8868.txt");
+	archive_entry_set_filetype(entry, AE_IFREG);
+	archive_entry_set_size(entry, 0);
+	assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
+	archive_entry_free(entry);
+	assertEqualInt(ARCHIVE_OK, archive_write_free(a));
+
+	/* Check UTF-8 version. */
+	assertEqualMem(buff, "\xE8\xA1\xA8.txt", 7);
+
+	/* Part 2: directory */
+	a = archive_write_new();
+	assertEqualInt(ARCHIVE_OK, archive_write_set_format_ustar(a));
+	assertEqualInt(ARCHIVE_OK, archive_write_set_options(a, "hdrcharset=UTF-8"));
+	assertEqualInt(ARCHIVE_OK,
+	    archive_write_open_memory(a, buff, sizeof(buff), &used));
+
+	entry = archive_entry_new2(a);
+	/* Set the directory name using a UTF-16 string */
+	/* NOTE: Explicitly not adding trailing slash to test that code path */
+	archive_entry_copy_pathname_w(entry, L"\u8868");
+	archive_entry_set_filetype(entry, AE_IFDIR);
+	archive_entry_set_size(entry, 0);
+	assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
+	archive_entry_free(entry);
+	assertEqualInt(ARCHIVE_OK, archive_write_free(a));
+
+	/* Check UTF-8 version. */
+	assertEqualMem(buff, "\xE8\xA1\xA8/", 4);
+
+	/* Part 3: symlink */
+	a = archive_write_new();
+	assertEqualInt(ARCHIVE_OK, archive_write_set_format_ustar(a));
+	assertEqualInt(ARCHIVE_OK, archive_write_set_options(a, "hdrcharset=UTF-8"));
+	assertEqualInt(ARCHIVE_OK,
+	    archive_write_open_memory(a, buff, sizeof(buff), &used));
+
+	entry = archive_entry_new2(a);
+	/* Set the symlink target using a UTF-16 string */
+	archive_entry_set_pathname(entry, "link.txt");
+	archive_entry_copy_symlink_w(entry, L"\u8868.txt");
+	archive_entry_set_filetype(entry, AE_IFLNK);
+	archive_entry_set_symlink_type(entry, AE_SYMLINK_TYPE_FILE);
+	archive_entry_set_size(entry, 0);
+	assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
+	archive_entry_free(entry);
+	assertEqualInt(ARCHIVE_OK, archive_write_free(a));
+
+	/* Check UTF-8 version. */
+	assertEqualMem(buff + 157, "\xE8\xA1\xA8.txt", 7);
+
+	/* Part 4: hardlink */
+	a = archive_write_new();
+	assertEqualInt(ARCHIVE_OK, archive_write_set_format_ustar(a));
+	assertEqualInt(ARCHIVE_OK, archive_write_set_options(a, "hdrcharset=UTF-8"));
+	assertEqualInt(ARCHIVE_OK,
+	    archive_write_open_memory(a, buff, sizeof(buff), &used));
+
+	entry = archive_entry_new2(a);
+	/* Set the symlink target using a UTF-16 string */
+	archive_entry_set_pathname(entry, "link.txt");
+	archive_entry_copy_hardlink_w(entry, L"\u8868.txt");
+	archive_entry_set_size(entry, 0);
+	assertEqualInt(ARCHIVE_OK, archive_write_header(a, entry));
+	archive_entry_free(entry);
+	assertEqualInt(ARCHIVE_OK, archive_write_free(a));
+
+	/* Check UTF-8 version. */
+	assertEqualMem(buff + 157, "\xE8\xA1\xA8.txt", 7);
+#endif
+}