Control: reopen -1 ! Thanks for pushing this forward, Cyril. Sorry for not communicating my intent to make a pu or acting on it sooner. On Wed, Oct 02, 2013 at 12:17:24PM +0200, Cyril Brulebois wrote: > Only impacted file is subversion/bindings/swig/core.i, fix is different > from upstream's (which isn't in the version in unstable anyway), Upstream's fix is in unstable, although the incorrect “fix” uploaded in 1.7.9+nmu1/1.6.17dfsg-4+deb7u2 is there as well. Updated pu has upstream's backported fix and removes the other attempted fix. Cheers, -- James GPG Key: 4096R/331BA3DB 2011-12-05 James McCoy <jamessan@debian.org>
diff -u subversion-1.6.17dfsg/debian/changelog subversion-1.6.17dfsg/debian/changelog
--- subversion-1.6.17dfsg/debian/changelog
+++ subversion-1.6.17dfsg/debian/changelog
@@ -1,3 +1,12 @@
+subversion (1.6.17dfsg-4+deb7u4) wheezy; urgency=low
+
+ * Non-maintainer upload.
+ * patches/python-swig205: Backport upstream patch to fix Python bindings
+ when built against swig 2.0.5+. (Closes: #683188)
+ * Remove patches/chunksize-integer.patch
+
+ -- James McCoy <jamessan@debian.org> Wed, 02 Oct 2013 21:40:37 -0400
+
subversion (1.6.17dfsg-4+deb7u3) wheezy-security; urgency=high
* Non-maintainer upload by the Security Team.
diff -u subversion-1.6.17dfsg/debian/patches/series subversion-1.6.17dfsg/debian/patches/series
--- subversion-1.6.17dfsg/debian/patches/series
+++ subversion-1.6.17dfsg/debian/patches/series
@@ -37,6 +37,6 @@
-chunksize-integer.patch
cve-2013-1845
cve-2013-1846
cve-2013-1849
CVE-2013-1968.patch
CVE-2013-2112.patch
+python-swig205
reverted:
--- subversion-1.6.17dfsg/debian/patches/chunksize-integer.patch
+++ subversion-1.6.17dfsg.orig/debian/patches/chunksize-integer.patch
@@ -1,17 +0,0 @@
-Author: W. Martin Borgert <debacle@debian.org>
-
-Origin: vendor, http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=683188#78
-Bug-Debian: http://bugs.debian.org/683188
-Forwarded: not-needed
-Last-Update: 2016-04-16
---- subversion-1.7.9.orig/subversion/bindings/swig/python/svn/core.py
-+++ subversion-1.7.9/subversion/bindings/swig/python/svn/core.py
-@@ -145,7 +145,7 @@
- # read the rest of the stream
- chunks = [ ]
- while 1:
-- data = svn_stream_read(self._stream, SVN_STREAM_CHUNK_SIZE)
-+ data = svn_stream_read(self._stream, int(SVN_STREAM_CHUNK_SIZE))
- if not data:
- break
- chunks.append(data)
only in patch2:
unchanged:
--- subversion-1.6.17dfsg.orig/debian/patches/python-swig205
+++ subversion-1.6.17dfsg/debian/patches/python-swig205
@@ -0,0 +1,28 @@
+Fix python bindings for swig 2.0.5. Upstream r1351117.
+Somehow swig 2.0.5 produces a long integer instead of an integer in this
+situation - I'm not entirely clear on specifics. But tolerate both as
+inputs to svn_stream_read().
+
+
+--- a/subversion/bindings/swig/core.i
++++ b/subversion/bindings/swig/core.i
+@@ -337,12 +337,17 @@
+ */
+ #ifdef SWIGPYTHON
+ %typemap(in) (char *buffer, apr_size_t *len) ($*2_type temp) {
+- if (!PyInt_Check($input)) {
++ if (PyLong_Check($input)) {
++ temp = PyLong_AsLong($input);
++ }
++ else if (PyInt_Check($input)) {
++ temp = PyInt_AsLong($input);
++ }
++ else {
+ PyErr_SetString(PyExc_TypeError,
+ "expecting an integer for the buffer size");
+ SWIG_fail;
+ }
+- temp = PyInt_AsLong($input);
+ if (temp < 0) {
+ PyErr_SetString(PyExc_ValueError,
+ "buffer size must be a positive integer");
Attachment:
signature.asc
Description: Digital signature