On Mon, Aug 20, 2018, at 1:28 PM, Andy Moreton wrote:
> On Mon 20 Aug 2018, Jostein Kjønigsen wrote:
>> As I've mentioned in another reply on this subject, decoupling
>> cc-mode>> releases from Emacs-releases will make it harder for third-party major->> modes deriving from cc-mode to maintain compatibility when cc-mode
>> introduces breaking changes.
>> Currently we can inspect "cc-mode version" by checking emacs-version,>> and dispatching compatible code based on that.
> 
> Checking version numbers is the wrong approach for feature detection,> and should only be a last resort if checking behaviour is not
> practical.> 
> If an ELPA package results in more frequent updates to cc-mode then
> third party extension authors will simply have to adapt to a slightly> faster release cadence.
> 
>     AndyM
> 

Agreed, but (to my knowledge) Emacs lacks proper introspection
capacities, and there's no way up front to know if  a function requires
3 or 5 parameters until you've called it and your code has crashed.
Compared to that, verifying version numbers against known points of
change seem like a much better strategy. It is also the advice I've
received from other people around in the Emacs-community.
--
Regards
Jostein Kjønigsen

jostein@kjonigsen.net 🍵 jostein@gmail.com
https://jostein.kjonigsen.net