@@ -1,5 +1,18 @@

 2009-05-29  Michal Ludvig  <michal@logix.cz>

+	* S3/Utils.py: New function replace_nonprintables()

+	* s3cmd: Filter local filenames through the above function

+	  to avoid problems with uploaded filenames containing invalid 

+	  XML entities, eg  

+	* S3/S3.py: Warn if a non-printables char is passed to

+	  urlencode_string() - they should have been replaced earlier 

+	  in the processing.

+	* run-tests.py, TODO, NEWS: Updated.

+	* testsuite/crappy-file-name.tar.gz: Tarball with a crappy-named

+	  file. Untar for the testsuite.

+

+2009-05-29  Michal Ludvig  <michal@logix.cz>

+

 	* testsuite/blahBlah/*: Added files needed for run-tests.py

 2009-05-28  Michal Ludvig  <michal@logix.cz>

@@ -8,6 +8,8 @@ s3cmd 1.0.0

 * Added --recursive support for [cp] and [mv], including

   multiple-source arguments, --include/--exclude, --dry-run, etc.

 * Added --exclude/--include and --dry-run for [del], [setacl].

+* Neutralise characters that are invalid in XML to avoid ExpatErrors.

+  http://boodebr.org/main/python/all-about-python-and-unicode

 s3cmd 0.9.9   -   2009-02-17

 ===========

@@ -344,19 +344,19 @@ class S3(object):

 					# systems.

 					#           [hope that sounds reassuring ;-)]

 			o = ord(c)

-			if (o <= 32 or		# Space and below

+			if (o < 0x20 or o == 0x7f):

+				error(u"Non-printable character 0x%02x in: %s" % (o, string))

+				error(u"Please report it to s3tools-bugs@lists.sourceforge.net")

+				encoded += replace_nonprintables(c)

+			elif (o == 0x20 or	# Space and below

 			    o == 0x22 or	# "

 			    o == 0x23 or	# #

-			    o == 0x25 or	# %

+			    o == 0x25 or	# % (escape character)

 			    o == 0x26 or	# &

 			    o == 0x2B or	# + (or it would become <space>)

 			    o == 0x3C or	# <

 			    o == 0x3E or	# >

 			    o == 0x3F or	# ?

-			    o == 0x5B or	# [

-			    o == 0x5C or	# \

-			    o == 0x5D or	# ]

-			    o == 0x5E or	# ^

 			    o == 0x60 or	# `

 			    o >= 123):   	# { and above, including >= 128 for UTF-8

 				encoded += "%%%02X" % o

@@ -256,6 +256,29 @@ def unicodise_safe(string, encoding = None):

 	return unicodise(deunicodise(string, encoding), encoding).replace(u'\ufffd', '?')

+def replace_nonprintables(string):

+	"""

+	replace_nonprintables(string)

+

+	Replaces all non-printable characters 'ch' in 'string'

+	where ord(ch) <= 26 with ^@, ^A, ... ^Z

+	"""

+	new_string = ""

+	modified = 0

+	for c in string:

+		o = ord(c)

+		if (o <= 31):

+			new_string += "^" + chr(ord('@') + o)

+			modified += 1

+		elif (o == 127):

+			new_string += "^?"

+			modified += 1

+		else:

+			new_string += c

+	if modified:

+		warning("%d non-printable characters replaced in: %s" % (modified, new_string))

+	return new_string

+

 def sign_string(string_to_sign):

 	#debug("string_to_sign: %s" % string_to_sign)

 	signature = base64.encodestring(hmac.new(Config.Config().secret_key, string_to_sign, sha1).digest()).strip()

@@ -12,8 +12,6 @@ TODO list for s3cmd project

   - Option --mime-type should set mime type with 'cp' and 'mv'. 

     If possible --guess-mime-type should do as well.

   - Skip files that disapper during upload. Now it fails.

-  - Neutralise characters that are invalid in XML to avoid ExpatErrors.

-    http://boodebr.org/main/python/all-about-python-and-unicode

 - For 1.0.0

   - Add 'geturl' command, both Unicode and urlencoded output.

@@ -223,6 +223,8 @@ test_s3cmd("Buckets list", ["ls"],

 ## ====== Sync to S3

 test_s3cmd("Sync to S3", ['sync', 'testsuite/', 's3://s3cmd-autotest-1/xyz/', '--exclude', '.svn/*', '--exclude', '*.png', '--no-encrypt', '--exclude-from', 'testsuite/exclude.encodings' ],

+	must_find = [ "WARNING: 32 non-printable characters replaced in: crappy-file-name/non-printables ^A^B^C^D^E^F^G^H^I^J^K^L^M^N^O^P^Q^R^S^T^U^V^W^X^Y^Z^[^\^]^^^_^? +-[\]^<>%%\"'#{}`&?.end",

+	              "stored as 's3://s3cmd-autotest-1/xyz/crappy-file-name/non-printables ^A^B^C^D^E^F^G^H^I^J^K^L^M^N^O^P^Q^R^S^T^U^V^W^X^Y^Z^[^\^]^^^_^? +-[\\]^<>%%\"'#{}`&?.end'" ],

 	must_not_find_re = [ "\.svn/", "\.png$" ])

 if have_encoding:

@@ -638,6 +638,8 @@ def _get_filelist_local(local_uri):

 				## for now skip over

 				continue

 			relative_file = unicodise(os.path.join(rel_root, f))

+			if not cfg.verbatim:

+				relative_file = replace_nonprintables(relative_file)

 			if relative_file.startswith('./'):

 				relative_file = relative_file[2:]

 			sr = os.stat_result(os.lstat(full_name))

new file mode 100644

Binary files /dev/null and b/testsuite/crappy-file-name.tar.gz differ