BIND 10 #1954: Implement perfdhcp command parser

[BIND 10 Development]

#1954: Implement perfdhcp command parser
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  marcin
  stephen                            |                Status:  reviewing
                       Type:         |             Milestone:  Sprint-
  enhancement                        |  DHCP-20120528
                   Priority:         |            Resolution:
  medium                             |             Sensitive:  0
                  Component:         |           Sub-Project:  DHCP
  perfdhcp                           |  Estimated Difficulty:  16
                   Keywords:         |           Total Hours:  42
            Defect Severity:  N/A    |
Feature Depending on Ticket:         |
        Add Hours to Ticket:  6      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by marcin):

 Most of the suggestions implemented and pushed with commit number:

 3f32b68d6f2d0f02af6cebac046a604751ad4dcc

 Q: ''initRandomRange(): "errmsg" does not need to be "std::string" - it
 could be declared as "const char*" (which avoids the need to call c_str()
 before using it).''

 A: It could be, but I want to be consistent with other functions where I
 declare string arguments as const std::string&


 Q: ''initRandomRange(): why is "randomRange" (which should be random_range
 - see note on variable names) declared as "long long"? The "opt" argument
 is cast to "long long", then that value is cast to uint32_t. Why not cast
 to uint32_t directly?''

 A: I want to verify that random range is not negative. If I cast to
 uint32_t I loose information that user specified negative value and I
 approve it. Lexical_cast does not throw exception if I cast negative value
 (string) to unsigned. Or it does?

 Q: ''IpVersion: need a test that an exception is thrown if both -4 and -6
 are given together.''

 A: There is such a test in TEST_F(CommandOptionsTest, IpVersion)

 Q: ''IpVersion: the comment "negative test cases" - these checks are
 checking that invalid combinations of flags are caught.''

 A: By negative test cases I exactly mean ''invalid combinations of flags
 are caught''. Isn't it what negative tests are about?

 Q: ''RandomRange: according to usage(), the "-R" switch specifies the
 number of different clients. It does not indicate that anything is random.
 Is the use of the word "random" justified?''

 A: Random range value is a range of values that we use to replace last
 four octets of default DUID value in outgoing packages to simulate
 different clients. Eventually, different clients have different random
 values so this value maps to number of clients. In order to avoid
 confusion I renamed some getter methods, members etc to reflect that it is
 number of clients. In the future maybe we could change this in
 documentation too?

 Q: ''Base: need to check invalid values passed as arguments to "mac=" and
 "duid=" are caught.''

 A: There are test cases like this in TEST_F(CommandOptionsTest, Base)

 Q: ''DropTime: question: reading usage() suggests that DropTime should be
 a double value. Yet DropTime is stored as as array of two doubles: why?''

 A: Comments in the code updated

-- 
Ticket URL: <http://bind10.isc.org/ticket/1954#comment:10>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development