Should we have 5.27.12?

We had more changes than we would like during this period of freeze[1].
I'm thinking maybe it's best if we release another development version.

Thoughts?

[1] We would like zero and we had quite a few more than zero.
0
xsawyerx
5/1/2018 9:08:48 AM
perl.perl5.porters 47343 articles. 0 followers. Follow

3 Replies
42 Views

Similar Articles

[PageSpeed] 32

On 05/01/2018 05:08 AM, Sawyer X wrote:
> We had more changes than we would like during this period of freeze[1].
> I'm thinking maybe it's best if we release another development version.
> 
> Thoughts?
> 
> [1] We would like zero and we had quite a few more than zero.
> 

I'm not opposed to this, but I think the following items should be taken 
into consideration:

* What data, from which humans, do we expect to get from a 5.27.12 
release that we don't have already -- or that we would not get from our 
RC releases?

I ask because, while in our smoke-tests we have a somewhat systematic 
way of collecting data on specific commits, we don't have the equivalent 
for either monthly dev releases or RC releases, both of which require 
testing of tarballs.

* What specific code changes since 5.27.11 do we feel are most in need 
of a 5.27.12 release?

* To what extent are the code changes since 5.27.11 platform-specific? 
What is our plan to gather test data about those platforms with a 
5.27.12 release?

* If we're going to do this, I recommend *not* waiting until May 20, as 
that would push the 5.28.0 release into June.

Thank you very much.
Jim Keenan
0
jkeenan
5/1/2018 12:14:16 PM
On Tue, May 01, 2018 at 12:08:48PM +0300, Sawyer X wrote:
> We had more changes than we would like during this period of freeze[1].
> I'm thinking maybe it's best if we release another development version.
> 
> Thoughts?
> 
> [1] We would like zero and we had quite a few more than zero.

I'm sanguine with 5.28.0-RC1 being the testbed for those extra commits.

-- 
In economics, the exam questions are the same every year.
They just change the answers.
0
davem
5/1/2018 3:45:46 PM

On 05/01/2018 03:14 PM, James E Keenan wrote:
> On 05/01/2018 05:08 AM, Sawyer X wrote:
>> We had more changes than we would like during this period of freeze[1].
>> I'm thinking maybe it's best if we release another development version.
>>
>> Thoughts?
>>
>> [1] We would like zero and we had quite a few more than zero.
>>
>
> I'm not opposed to this, but I think the following items should be
> taken into consideration:
>
> * What data, from which humans, do we expect to get from a 5.27.12
> release that we don't have already -- or that we would not get from
> our RC releases?
>
> I ask because, while in our smoke-tests we have a somewhat systematic
> way of collecting data on specific commits, we don't have the
> equivalent for either monthly dev releases or RC releases, both of
> which require testing of tarballs.

I can't think of any.

>
> * What specific code changes since 5.27.11 do we feel are most in need
> of a 5.27.12 release?

We had a few last minute fixes so I was somewhat concerned.

>
> * To what extent are the code changes since 5.27.11 platform-specific?
> What is our plan to gather test data about those platforms with a
> 5.27.12 release?

Nothing out of the ordinary.

>
> * If we're going to do this, I recommend *not* waiting until May 20,
> as that would push the 5.28.0 release into June.

If we were to release 5.27.12, 5.28 would be in June. Considering Dave
is not concerned (and no one expressed support for this), we'll continue
with RCs. I wanted to leave room open for thoughts by others.
0
xsawyerx
5/9/2018 7:07:42 PM
Reply: