Re: git: ece1cb52e7ff - main - security/sssd: mark port deprecated

On 2024-02-07 19:00, John Hixson wrote:
> On Tue, Feb 06, 2024 at 11:23:45AM +0100, Michael Osipov wrote:
>> On 2024-02-03 07:25, Gleb Popov wrote:
>>> On Sat, Feb 3, 2024 at 8:53 AM John Hixson <jhixson@freebsd.org> wrote:
>>>>
>>>> The branch main has been updated by jhixson:
>>>>
>>>> URL: https://cgit.FreeBSD.org/ports/commit/?id=ece1cb52e7ffa225bc1c8ed659c5974ef1c0088d
>>>>
>>>> commit ece1cb52e7ffa225bc1c8ed659c5974ef1c0088d
>>>> Author:     John Hixson <jhixson@FreeBSD.org>
>>>> AuthorDate: 2024-02-03 05:51:53 +0000
>>>> Commit:     John Hixson <jhixson@FreeBSD.org>
>>>> CommitDate: 2024-02-03 05:51:53 +0000
>>>>
>>>>       security/sssd: mark port deprecated
>>>>
>>>>       - Very old version of this port no longer supported
>>>>       - sssd-devel has current versions
>>>
>>> Why not just update this port to the devel version?
>>
>> I agree with Gleb. This is unnecessarily confusing and many will likely
>> refuse to use a -devel port in production. Other ports do have regular ports
>> and real -devel ports as well. For me, this violates POLA somewhat.
>>
>> M
> 
> What is the proper route to take here? sssd-devel isn't really a "devel"
> port. It's just the up to date version of sssd. The current sssd port is
> very old (like 5 years or more). I would like to just replace it with
> the new version but don't want to force an update for anyone who may
> rely on it. Should sssd-devel be renamed sssd2 and sssd be renamed sssd1
> or some such? What is this POLA that I am possibly violating by the way?

I see your point it is partially valid to me. I'd expect sssd-devel to 
form off a non-release version, namely a Git hash. As for security/sssd: 
Almost just like you said. Deprecate and introduce security/sssd2. Very 
similar to what security/krb5* or security/openssl* are don't. That 
scheme would resemble. I wouldn't touch security/sssd in terms of 
renaming anymore. When sssd 3 comes out, you can have another parallel 
port for that.
POLA violation for me is that an apparently valid port is replaced with 
a devel one.

Michael