summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorXiang Zhang <angwerzx@126.com>2017-05-09 03:38:32 (GMT)
committerGitHub <noreply@github.com>2017-05-09 03:38:32 (GMT)
commit9da408d15bdef624a5632182cb4edf98001fa82f (patch)
tree55707952ef7b86c4c333122add2bc27101c27ef0
parentfa5abac1e6cd74979557d5a6f960a55f40a10b0e (diff)
downloadcpython-9da408d15bdef624a5632182cb4edf98001fa82f.zip
cpython-9da408d15bdef624a5632182cb4edf98001fa82f.tar.gz
cpython-9da408d15bdef624a5632182cb4edf98001fa82f.tar.bz2
bpo-29990: Fix range checking in GB18030 decoder (#1495)
When decoding a 4-byte GB18030 sequence, the first and third byte cannot exceed 0xFE.
-rw-r--r--Lib/test/test_codecencodings_cn.py6
-rw-r--r--Misc/NEWS2
-rw-r--r--Modules/cjkcodecs/_codecs_cn.c4
3 files changed, 11 insertions, 1 deletions
diff --git a/Lib/test/test_codecencodings_cn.py b/Lib/test/test_codecencodings_cn.py
index 3bdf7d0..c8a410c 100644
--- a/Lib/test/test_codecencodings_cn.py
+++ b/Lib/test/test_codecencodings_cn.py
@@ -48,6 +48,12 @@ class Test_GB18030(multibytecodec_support.TestBase, unittest.TestCase):
(b"abc\x84\x32\x80\x80def", "replace", 'abc\ufffd2\ufffd\ufffddef'),
(b"abc\x81\x30\x81\x30def", "strict", 'abc\x80def'),
(b"abc\x86\x30\x81\x30def", "replace", 'abc\ufffd0\ufffd0def'),
+ # issue29990
+ (b"\xff\x30\x81\x30", "strict", None),
+ (b"\x81\x30\xff\x30", "strict", None),
+ (b"abc\x81\x39\xff\x39\xc1\xc4", "replace", "abc\ufffd\x39\ufffd\x39\u804a"),
+ (b"abc\xab\x36\xff\x30def", "replace", 'abc\ufffd\x36\ufffd\x30def'),
+ (b"abc\xbf\x38\xff\x32\xc1\xc4", "ignore", "abc\x38\x32\u804a"),
)
has_iso10646 = True
diff --git a/Misc/NEWS b/Misc/NEWS
index 5e5ce59..b934888 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -317,6 +317,8 @@ Extension Modules
Library
-------
+- bpo-29990: Fix range checking in GB18030 decoder. Original patch by Ma Lin.
+
- bpo-29979: rewrite cgi.parse_multipart, reusing the FieldStorage class and
making its results consistent with those of FieldStorage for
multipart/form-data requests. Patch by Pierre Quentel.
diff --git a/Modules/cjkcodecs/_codecs_cn.c b/Modules/cjkcodecs/_codecs_cn.c
index 1a070f2..bda175c 100644
--- a/Modules/cjkcodecs/_codecs_cn.c
+++ b/Modules/cjkcodecs/_codecs_cn.c
@@ -279,7 +279,9 @@ DECODER(gb18030)
REQUIRE_INBUF(4);
c3 = INBYTE3;
c4 = INBYTE4;
- if (c < 0x81 || c3 < 0x81 || c4 < 0x30 || c4 > 0x39)
+ if (c < 0x81 || c > 0xFE ||
+ c3 < 0x81 || c3 > 0xFE ||
+ c4 < 0x30 || c4 > 0x39)
return 1;
c -= 0x81; c2 -= 0x30;
c3 -= 0x81; c4 -= 0x30;