[bind10-dev] [svn] commit: r1108 - in /branches/each-nsec3/src/lib/dns/cpp: Makefile.am base32.cc base32.h sha1.cc sha1.h tests/Makefile.am tests/base32_unittest.cc tests/sha1_unittest.cc

[Francis Dupont]

> The reason I used the word "initial" in my commit message is that
> I expect both sha1 and base32 to be replaced with something else.

=> my concern is about sha1 only because crypto is special.

> In year 2 I reckon we'll probably switch to crypto++ or something.

=> I prefer the word "plug" to "switch". The idea is to implement
a crypto following a signature from BIND 10 with what one'd like
(crypto++ or something).
Note in case of NSEC3 this (defining a signature) should be easy
as there is one function with a few known parameters, and there is
no clear benefit from partial applications.

Francis Dupont <fdupont at isc.org>

PS: I use the implementation and signature concepts from programming
language theory. Please answer if you don't understand me (as C++
is far from a clean design from this point of view).