BIND 10 #185: review: more tests and bug fixes for xfrin
BIND 10 Development
do-not-reply at isc.org
Fri May 28 23:15:49 UTC 2010
#185: review: more tests and bug fixes for xfrin
--------------------+-------------------------------------------------------
Reporter: jinmei | Owner: zhanglikun
Type: defect | Status: reviewing
Priority: major | Milestone: 04. 2nd Incremental Release: Early Adopters
Component: xfrin | Resolution:
Keywords: | Sensitive: 0
--------------------+-------------------------------------------------------
Comment(by jinmei):
Replying to [comment:3 zhanglikun]:
> I have did some review on your change, yeah, looks quite well. I did
some minor change in r1948, and fixed failed unit tests in r1961. The
problem is there are some log message after running unittest, but anyway
it can be avoided once we use logsys in future.
Likun, thanks for the review.
r1948 looks correct, good catch.
We cannot adopt r1961, however:
- "3.3.3" will be rejected in BSD's getaddrinfo() implementation. This
is because their implementation doesn't fully conform to the RFC (by
applying stricter checks), but aside from that, what we should do in this
context is to provide a "perfectly invalid" address string and let an
exception be raised. So I use "3.3.3.3.3" (five 3's) and revised the test
so that an exception will be detected.
- we shouldn't allow 65536. Clearly it's an invalid number for a UDP/TCP
port. Apparently Linux (glic's) implementation is buggy and incorrectly
accepts such out-of-range numeric port numbers. I've added an additional
(and redundant for non Linux implementations) validation ot work around
this bug.
The changeset is r1968. Please check.
Assuming these are okay, is this patch ready to be merged to trunk? Or is
this your intermediate report?
--
Ticket URL: <https://bind10.isc.org/ticket/185#comment:4>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list