Re: [edk2-devel] API breakages and their implications. Was: [PATCH 1/1] MdeModulePkg: UART Dynamic clock freq Support

["Leif Lindholm" <leif@nuviainc.com>]

On Tue, Mar 31, 2020 at 15:23:24 +0200, Laszlo Ersek wrote:
> On 03/31/20 11:22, Leif Lindholm wrote:
> > "This breaks many platforms" is a good argument for seeing if a
> > solution can be found that does not break (as) many platforms. It is
> > not an argument for duplicating drivers when the change needed for
> > those platforms is trivial.
> > 
> > These days, Linux tends to deal with API breakages (and other things)
> > using a semantic patch tool called Coccinelle[1]. It would not be
> > unreasonable, and indeed it would be helpful to us on the non-x86 side
> > if something similar was adopted (and semantic patches mandated) for
> > API breakages in EDK2.
> > 
> > [1] http://coccinelle.lip6.fr/sp.php
> 
> Two comments:
> 
> (1) One of the reasons why I would like to keep all platforms in a
> single tree is to deal with API changes like this.

Agreed.

> That way, someone
> proposing an API change would at least have the chance to fix up all the
> consumer sites. Of course it would require diligent review from the
> other pkg maintainers, but it could be implemented without any temprary
> breakage in the git history even.

And a daily CI job could spot breakages and send out alerts to
platform owners.

It would also provide more incentive for actually upstreaming platform
ports.

> (2) Specifically about this problem. The vendor GUID approach is not a
> bad one. How about the following alternative:

I have no strong comment on your alternative. It seems perfectly
feasible, and I agree there is precedent. Thanks for providing it.

I will let the MdeModulePkg maintainers specify their preference, or
provide other alternative solutions.

Best Regards,

Leif