[bind10-dev] Coding guidelines and tests

[Michal 'vorner' Vaner]

Hello

On Wed, Oct 12, 2011 at 05:34:26PM +0200, Tomasz Mrugalski wrote:
> Tests should be as broad as possible. For example in tests for DHCPv6
> option that conveys IPv6 addresses, I check:
> - :: (any address, a valid value that must be supported),
> - ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff (to see if there are no
> integer overflows),
> - ff02::face:b00c (to see that multicast addresses are handled ok)
> - and some other addresses to verify that tested code does not have
> first 32 bits hardcoded to 2001:db8.

If you have a reason to test the range, test ::, and so on specifically, it's
OK.

As I understand the guidelines, they suggest the range of addresses to maintain
some consistency and ensure the address is not stolen from someone (which might
make him unhappy or something). Anyway, they are for cases when any one address
is as good as other, so if I'm trying to create a test A record, I put
"192.0.2.1" into the data, because it doesn't matter, what is inside. It's the
same we use "example.org" domain within our tests, just to be little consistent.
If there was a reason to think it would break in case of "root.cz" or whatever
address and not other, then it's much stronger reason to use it than the reason
of guidelines.

With regards

-- 
How many Lisp programmers does it take to change a light bulb?
(((H)mmm,) (I'm ((not) sure, better))) (find (out))...

Michal 'vorner' Vaner
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <https://lists.isc.org/pipermail/bind10-dev/attachments/20111014/8ce2a880/attachment.bin>