On Tuesday, 03 February 2015 at 13:53, Kevin J. McCarthy wrote: > David Champion wrote: > > * On 03 Feb 2015, Kevin J. McCarthy wrote: > > > > > > > > see > > > > http://en.wikipedia.org/wiki/Data_Encryption_Standard#Security_and_cryptanalysis > > > > > > > > I'm attaching a patch that instead creates a default of "aes256". > > > > > > I have no problem with this, but let's just poll the list first. Does > > > anyone have a problem with setting the default smime algorithm to > > > aes256? > > > > In principle no. The main consideration with this kind of change > > is whether the new default is available to the currently supported > > installed base. (E.g. if we "support" fedora 12 and fedora 12's openssl > > doesn't have aes256, then there's a problem.) I don't think that there > > are any actual issues here, but that's the underlying question I think. > > > > So +0, +1 if it's certain that aes256 is pretty much universal at this > > point. > > Hi David, > > Thanks for your input! I did do a little searching around. > > In the openssl changelog for 0.9.7 [31 Dec 2002]: > https://www.openssl.org/news/changelog.html > it mentions Rijndael and aes a few times. > > I also found a ticket from 2/2004: > https://rt.openssl.org/Ticket/Display.html?id=834&user=guest&pass=guest > that indicates the -aes256 flag was already supported back then. > > I can't say for certain that it's universal at this point though, as I > haven't really paid much attention on openssl up to this point! :-) > > So unless someone can proclaim that with certainty, I'll just push the > des3 default in a few days. At least the other options are documented > and people can adjust as they see fit.
I think I'd prefer to go the other way, and default to the stronger cipher. On those systems where it's unavailable the default can be overridden, but the user will have her eyes open.
