3 ms·
There is no testing of whether a disabled feature actually gets disabled. The only openssl test cases that exist are explicitly to test that a feature works, no
by peterwwillis 12y ago
There is no testing of whether a disabled feature actually gets disabled. The only openssl test cases that exist are explicitly to test that a feature works, not that it doesn't work.
The 'test' directory of openssl is just a big pile of giant C programs and shell scripts which make up the test cases, and a lot of the features and options are piled together, and there's no documentation. There isn't even an indication of which test case fails when one does fail. I don't see how anyone could get an idea from this of whether a feature works as expected or not, or how many regressions you might have between releases or after new features. As is typical of openssl, it's a big mess.
- justcommenting 12y agoagreed--as phk put it in his operation orchestra talk (link in previous post) regarding openssl: 1)API is a nightmare, 2)documentation is deficient and misleading, and 3) defaults are deceptive kudos again to the akamai team for reporting the bug.