[openssl-project] A proposal for an updated OpenSSL version scheme (v2)

[Matt Caswell]

On 21/09/18 13:52, Richard Levitte wrote:
> Note that this is for the text form, which is separate from our
> numeric encoding (something that isn't covered by semver at all).
> That is the only place where I propose to have an epoch, and it's for
> one purpose only, that the value of that number should be greater in
> the next version number.  
I see no reason for the epoch value to be greater. Equal is sufficient
(i.e. setting it to 1 is fine). 0 would not be fine.

There are two requirements for OPENSSL_VERSION_NUMBER afaict:

1) The encoded number should obey the same precedence rules as semantic
versioning, so that equality and inequality tests between version
numbers work as you might expect.

2) We don't break things when compared to how this version number is
used in practice today


Matt