Change the zipimport implementation to accept files containing

arbitrary bytes before the actual zip compatible archive.  Zipfiles
containing comments at the end of the file are still not supported.

Add a testcase to test_zipimport, and update NEWS.

This closes sf #775637 and sf #669036.

Lib/test/test_zipimport.py

@@ -49,7 +49,7 @@ class UncompressedZipImportTestCase(ImportHooksBaseTestCase):
         zipimport._zip_directory_cache.clear()
         ImportHooksBaseTestCase.setUp(self)
 
-    def doTest(self, expected_ext, files, *modules):
+    def doTest(self, expected_ext, files, *modules, **kw):
         z = ZipFile(TEMP_ZIP, "w")
         try:
             for name, (mtime, data) in files.items():
@@ -57,6 +57,19 @@ class UncompressedZipImportTestCase(ImportHooksBaseTestCase):
                 zinfo.compress_type = self.compression
                 z.writestr(zinfo, data)
             z.close()
+
+            stuff = kw.get("stuff", None)
+            if stuff is not None:
+                # Prepend 'stuff' to the start of the zipfile
+                f = open(TEMP_ZIP, "rb")
+                data = f.read()
+                f.close()
+
+                f = open(TEMP_ZIP, "wb")
+                f.write(stuff)
+                f.write(data)
+                f.close()
+
             sys.path.insert(0, TEMP_ZIP)
 
             mod = __import__(".".join(modules), globals(), locals(),
@@ -181,6 +194,12 @@ class UncompressedZipImportTestCase(ImportHooksBaseTestCase):
                  "some.data": (NOW, "some data")}
         self.doTest(pyc_ext, files, TESTMOD)
 
+    def testImport_WithStuff(self):
+        # try importing from a zipfile which contains additional
+        # stuff at the beginning of the file
+        files = {TESTMOD + ".py": (NOW, test_src)}
+        self.doTest(".py", files, TESTMOD,
+                    stuff="Some Stuff"*31)
 
 class CompressedZipImportTestCase(UncompressedZipImportTestCase):
     compression = ZIP_DEFLATED

Misc/NEWS

@@ -12,6 +12,10 @@ What's New in Python 2.3 release candidate 2?
 Core and builtins
 -----------------
 
+- It is now possible to import from zipfiles containing additional
+  data bytes before the zip compatible archive. Zipfiles containing a
+  comment at the end are still unsupported.
+
 Extension modules
 -----------------
 

Modules/zipimport.c

@@ -655,11 +655,12 @@ read_directory(char *archive)
 	PyObject *files = NULL;
 	FILE *fp;
 	long compress, crc, data_size, file_size, file_offset, date, time;
-	long header_offset, name_size, header_size;
+	long header_offset, name_size, header_size, header_position;
 	long i, l, length, count;
 	char path[MAXPATHLEN + 5];
 	char name[MAXPATHLEN + 5];
 	char *p, endof_central_dir[22];
+	long arc_offset; /* offset from beginning of file to start of zip-archive */
 
 	if (strlen(archive) > MAXPATHLEN) {
 		PyErr_SetString(PyExc_OverflowError,
@@ -675,6 +676,7 @@ read_directory(char *archive)
 		return NULL;
 	}
 	fseek(fp, -22, SEEK_END);
+	header_position = ftell(fp);
 	if (fread(endof_central_dir, 1, 22, fp) != 22) {
 		fclose(fp);
 		PyErr_Format(ZipImportError, "can't read Zip file: "
@@ -689,7 +691,10 @@ read_directory(char *archive)
 		return NULL;
 	}
 
+	header_size = get_long((unsigned char *)endof_central_dir + 12);
 	header_offset = get_long((unsigned char *)endof_central_dir + 16);
+	arc_offset = header_position - header_offset - header_size;
+	header_offset += arc_offset;
 
 	files = PyDict_New();
 	if (files == NULL)
@@ -720,7 +725,7 @@ read_directory(char *archive)
 		   PyMarshal_ReadShortFromFile(fp) +
 		   PyMarshal_ReadShortFromFile(fp);
 		fseek(fp, header_offset + 42, 0);
-		file_offset = PyMarshal_ReadLongFromFile(fp);
+		file_offset = PyMarshal_ReadLongFromFile(fp) + arc_offset;
 		if (name_size > MAXPATHLEN)
 			name_size = MAXPATHLEN;