[CFT/RFC]: refactor bsd.prog.mk to understand multiple programs instead of a singular program

Chris Rees utisoft at gmail.com
Sat Oct 27 14:34:52 UTC 2012


On 27 October 2012 15:32, Bryan Drewery <bryan at shatow.net> wrote:
> On 10/27/2012 8:23 AM, Chris Rees wrote:
>> [trim CC list a little to stop people regretting replying to this thread]
>>
>> On 27 October 2012 10:15, Chris Rees <utisoft at gmail.com> wrote:
>>>
>>> On 27 Oct 2012 00:35, "Simon J. Gerraty" <sjg at juniper.net> wrote:
>>>>
>>>>
>>>> On Fri, 26 Oct 2012 22:02:00 +0100, Chris Rees writes:
>>>>> In that case we have a switch time on the order of years, not weeks; 8.3
>>>>> is
>>>>> supported until May '14, and unless we get a :tl etc MFC into 8, even
>>>>> longer.  All this time the ports tree must work with pmake.
>>>>
>>>> I'm pretty sure I was told it is already in 8 and 7
>>>
>>> Not in 8.3 at least:
>>>
>>> svnweb.freebsd.org/base/releng/8.3/usr.bin/make/var.c?view=log
>>>
>>>>> I don't want to discourage you or belittle your excellent work here, but
>>>>> Marcel made me very nervous with his comment on the process being "a few
>>>>> weeks".
>>>>
>>>> That was based on discussions at the last devsummit.
>>>
>>> These discussions need backing up with a real roadmap, including detail on
>>> exactly what 8.3 and 7.4 users will have to do to ensure that the ports tree
>>> still works.
>>>
>>> I don't see where these considerations have been made.
>>
>> OK, so how about this.
>>
>> We (ab)use the security update mechanism to merge the pmake changes
>> (:tl and :tu) into releng/7.4 and releng/8.3 (possibly the earlier
>> releng branches such as 7.3, 8.2, 9.0).  We could then send out a
>> message on ports-announce, giving a few weeks' notice that the change
>> to bsd.port.mk is going through and that users need the latest
>> 'security' patches.
>
> This "weeks" is making a assumptions that users 1. reads ports@ or 2.
> Update to security/errata patches in a timely manner or 3. Read UPDATING

Quite.  This should be at least a few months, otherwise we're making
unreasonable requests of our users, and yet again annoy them by
breaking older versions-- this time with no real benefit for
end-users.

Chris


More information about the freebsd-hackers mailing list