Package: release.debian.org
Severity: normal
Tags: wheezy
User: release.debian....@packages.debian.org
Usertags: pu
This package has two bugs that can cause significant issues. Both result
in a failure of a DNS lookup; one with a timeout that should not occur and the
other with a traceback.
The fixes are trivial in complexity/risk so I would like to fix wheezy.
Debdiff attached.
diff -u py3dns-3.0.2/debian/changelog py3dns-3.0.2/debian/changelog
--- py3dns-3.0.2/debian/changelog
+++ py3dns-3.0.2/debian/changelog
@@ -1,3 +1,16 @@
+py3dns (3.0.2-1+deb7u2) wheezy; urgency=medium
+
+ * Backport change to fix timeouts associated with only one of several
+ available nameservers being unavailable (Closes: #718579):
+ - Only raise timeout error after trying all available servers
+ - Stop lookups once an answer is gotten
+ * Backport fix from 3.1/trunk to correctly deal with source port already in
+ use errors (LP: #1318264)
+ - Note that this bug was originally reported to the upstream project by a
+ Debian user
+
+ -- Scott Kitterman <sc...@kitterman.com> Sat, 10 May 2014 20:11:01 -0400
+
py3dns (3.0.2-1+deb7u1) wheezy; urgency=low
* Revert AAAA query result type change and return raw bytes instead of
only in patch2:
unchanged:
--- py3dns-3.0.2.orig/DNS/Base.py
+++ py3dns-3.0.2/DNS/Base.py
@@ -193,7 +193,7 @@
break
except socket.error as msg:
# Error 98, 'Address already in use'
- if msg[0] != 98: raise
+ if msg.errno != 98: raise
def conn(self):
self.getSource()
@@ -282,6 +282,11 @@
# servers worked:
first_socket_error = first_socket_error or e
continue
+ except TimeoutError as t:
+ first_socket_error = first_socket_error or t
+ continue
+ if self.response:
+ break
if not self.response and first_socket_error:
raise first_socket_error
@@ -319,6 +324,11 @@
except socket.error as e:
first_socket_error = first_socket_error or e
continue
+ except TimeoutError as t:
+ first_socket_error = first_socket_error or t
+ continue
+ if self.response:
+ break
if not self.response and first_socket_error:
raise first_socket_error
