BIND 10 #1957: Implement perfdhcp Interface Manager
BIND 10 Development
do-not-reply at isc.org
Mon Jun 25 12:39:43 UTC 2012
#1957: Implement perfdhcp Interface Manager
-------------------------------------+-------------------------------------
Reporter: | Owner: marcin
stephen | Status: reviewing
Type: | Milestone: Sprint-
enhancement | DHCP-20120703
Priority: | Resolution:
medium | Sensitive: 0
Component: | Sub-Project: DHCP
perfdhcp | Estimated Difficulty: 24
Keywords: | Total Hours: 25
Defect Severity: N/A |
Feature Depending on Ticket: |
Add Hours to Ticket: 25 |
Internal?: 0 |
-------------------------------------+-------------------------------------
Changes (by stephen):
* owner: UnAssigned => marcin
Comment:
Reviewed commit 0e1996af4d06ab528f76e13a3ce7c4c29b12d046.
Some minor changes to spacing around operators ands header comments have
been made and pushed.
'''src/lib/dhcp/iface_mgr.{h,cc}'''[[BR]]
openSocketFromAddress()/openSocketFromIface(): the description says that
the method will return the socket descriptor if the operation was
successful. However, if neither the interface nor address is found, the
methods will return a value of zero. Two options:
a) Any failure - including not finding the interface or not finding the
address - should result in an exception.
b) If a return value is required in the case of not finding an interface
or address, I would suggest a value of -1, as 0 is a valid descriptor
number.
(In both cases, the variable "sock" can be declared at the point of first
use. In case (b), the return status can be hard-coded - there is no need
to use "sock".)
getLocalAddress(): as remote_endpoint is not being shared outside this
method, a shared_ptr is not needed - a scoped_ptr eliminates the overhead
of reference counting.
'''src/lib/dhcp/tests/iface_mgr_unittest.cc'''[[BR]]
As a general rule, I suggest that the pointer to !NakedIfaceMgr be
encapsulated in a scoped_ptr and the "delete" at the end of the method
removed. Although OK now, if the code is subsequenly edited to include an
ASSERT_XXX test, a memory leak could ensue.
socketsFromIface: the comment says that the v4 sicket ''(sic)'' be bound
to the 10548 port, yet the code binds it to 10547.
Suggest that the "magic" port numbers 10547 and 10548 be coded as symbolic
constants.
--
Ticket URL: <http://bind10.isc.org/ticket/1957#comment:5>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list