On 26 November 2014 at 08:49, Ferenc Kovacs <tyr...@gmail.com> wrote: >> That's a rather extreme reaction to trying to patch string operations that >> real-world frameworks use to handle crypto secrets, don't you think? >> > and there are at least that much, but probably lot more usages in the > wild(see https://github.com/search?l=php&q=bin2hex&type=Code&utf8=%E2%9C%93 > for example) where there is nothing to do with security so there is no gain > for being constant time, but those users would get the performance > degradation.
Agreed. I've never really thought of those functions as being cryptographically significant. > I think it would be better to introduce constant time alternatives for > functions like this instead of trying to replace them and require everybody > to pay the performance price. +1 on this too: if we added a new function for constant time string comparison (which is, IMO, the far more common crypto case requiring this sort of work) rather than modifying strcmp(), I don't see why any of these should be different. I do wonder if these need to be in core, or if a PECL extension would be more appropriate. Adam -- PHP Internals - PHP Runtime Development Mailing List To unsubscribe, visit: http://www.php.net/unsub.php
