|
From: | Raif S. Naffah |
Subject: | Re: [Classpathx-crypto] TestOfSerpent |
Date: | Mon, 10 Jun 2002 12:36:49 +1000 |
User-agent: | Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.0.0) Gecko/20020530 |
Casey Marshall wrote:
Raif S. Naffah wrote: [...]since we will be using code from these tools for the test cases, should we look into re-factoring that code to improve code re-usability across the library?It did not even occur to me to use that code; I like this idea, however --- one could envision a point where writing a full-conformance test for a particular algorithm would amount to simply including the appropriate test vectors and pointing the program at them. It seems smarter than keeping large, static String arrays.
absolutely. Nistxxx tools _emit_ the test vectors, but they also include the code to verify the inverse operation of what they generate.
re-factoring that code and re-using your patterns (from TestOfSerpent) would allow:
a. ending up with new tools (or added functionality to the existing ones) that _read_ the conformance test vectors and verify them, this time instead of _writing_ them out.
b. exposing some methods in these tools, or have'em all use common Helper(s) to do their job and making writing TestOf<block-ciphers> as simple as feeding the test-case a collection of files (or a directory where those files would be found).
does this make sense? cheers; rsn
[Prev in Thread] | Current Thread | [Next in Thread] |