Front page | perl.perl5.porters |
Postings from May 2017
Re: Perl 5.26.0-RC1 is now available!
Thread Previous
|
Thread Next
From:
Sawyer X
Date:
May 22, 2017 09:34
Subject:
Re: Perl 5.26.0-RC1 is now available!
Message ID:
5fc9bf4f-247f-3292-a81c-92446aed86b6@gmail.com
On 05/21/2017 06:08 AM, Karl Williamson wrote:
> On 05/20/2017 09:37 AM, Karl Williamson wrote:
>> On 05/20/2017 03:37 AM, Sawyer X wrote:
>>>
>>>
>>> On 05/19/2017 10:24 PM, Karl Williamson wrote:
>>>> On 05/18/2017 09:31 PM, Karl Williamson wrote:
>>>>> On 05/17/2017 11:17 PM, Karl Williamson wrote:
>>>>>> On 05/17/2017 08:54 AM, Ævar Arnfjörð Bjarmason wrote:
>>>>>>> On Wed, May 17, 2017 at 4:32 PM, Karl Williamson
>>>>>>> <public@khwilliamson.com> wrote:
>>>>>>>> On 05/17/2017 07:16 AM, Dave Mitchell wrote:
>>>>>>>>>
>>>>>>>>> On Tue, May 16, 2017 at 04:13:33PM +0200, Sawyer X wrote:
>>>>>>>>>>>
>>>>>>>>>>> Are you in a position to be able to coalesce soon? If not I
>>>>>>>>>>> could
>>>>>>>>>>> have a
>>>>>>>>>>> go.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> Also, worth noting: I *have* coalesced them, just with some
>>>>>>>>>> mistakes.
>>>>>>>>>> Feel free to add it yourself to perldelta directly.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Yeah sorry, I realised that later - I must have been looking
>>>>>>>>> at the
>>>>>>>>> wrong
>>>>>>>>> branch when I saw an emptyish perldelta.
>>>>>>>>>
>>>>>>>>>> I'll do it today/tomorrow. I have additional cleanups to do. We
>>>>>>>>>> will
>>>>>>>>>> have another RC soon, so I think by then we'll have ironed out
>>>>>>>>>> everything.
>>>>>>>>>>
>>>>>>>>>> If you find anything else, please let me know. I'll go over it
>>>>>>>>>> again and
>>>>>>>>>> again, hopefully cleaning it up properly.
>>>>>>>>>>
>>>>>>>>>> Would you like me to merge your text?
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> I'd suggest that first, Ævar merges his branch as-is (any
>>>>>>>>> bikeshedding
>>>>>>>>> of it can be accommodated with further commits to blead rather
>>>>>>>>> than
>>>>>>>>> updating a separate branch).
>>>>>>>>>
>>>>>>>>> Then I'll merge in my @INC text,
>>>>>>>>>
>>>>>>>>> Then people can tweak anything as they see fit.
>>>>>>>>>
>>>>>>>>> ?
>>>>>>>>>
>>>>>>>>
>>>>>>>> +1
>>>>>>>
>>>>>>> I've just done that. Yves had a patch on top which I squashed into
>>>>>>> mine and changed him to be the author (since the text was all
>>>>>>> his at
>>>>>>> that point), sans a typo in a commit message and a minor whitespace
>>>>>>> fix what I pushed was what I noted above.
>>>>>>>
>>>>>>> Thanks everyone for the feedback!
>>>>>>>
>>>>>>
>>>>>> I have made some updates to perldelta, but have run out of time
>>>>>> tonight.
>>>>>> There's a lot more work needed than I realized.
>>>>>>
>>>>>> The commit messages indicate things I hope others will look at.
>>>>>>
>>>>>
>>>>>
>>>>> I pushed a couple of nits fixes. I'm only about 10% of the way
>>>>> through the file.
>>>>>
>>>>
>>>> I've pushed a few more changes, stopping when my eyes started to glaze
>>>> over reading the thing.
>>>>
>>>> I believe the section on documentation changes needs extensive
>>>> changes. The stuff in there is mostly useless. "Removed redundand dSP
>>>> from an example" may be true, but its not really useful. And
>>>> specifying these would lead a reader to think that the section really
>>>> is extremely detailed and accurate, whereas, what is in it is kind of
>>>> random.
>>>
>>> I think this has grown from a people documenting what was important
>>> on a
>>> particular dev release. When merged together for 5.26.0, the changes
>>> are
>>> now competing on the whole, and render themselves relatively useless.
>>>
>>>> The rewrite of the first part of perlre and removal of the final part
>>>> were not mentioned at all, and are far more important than most of
>>>> what is in there.
>>>
>>> +1
>>>
>>>> If we are going to have this section, the various pod files should be
>>>> examined for commits that occurred in this release. I would go to the
>>>> history links like
>>>>
>>>> https://perl5.git.perl.org/perl.git/history/HEAD:/pod/perlre.pod
>>>
>>> I would like to release RC2 soon. (As in, as soon as possible.) How
>>> long
>>> of a process might this be?
>>>
>>
>> I wasn't volunteering to do this. I did want to point out that the
>> delta contains a lot of not important info, while omitting important
>> stuff.
>
> I ended up doing it, and committing it. There was an easier way to
> find the differences than I first thought.
Thank you for doing this. Can you share what you used to find the
differences?
>>
>> I think the delta needs a bunch of work, though the early sections
>> are now good enough IMO. Someone, e.g., the pumpking should make
>> sure that the enhancements section is properly ordered so the most
>> important things are first.
>
> Isn't there some code that needs to be added before final release?
> Can RC2 be made without the final perldelta?
There was one thing: A patch by Jim Keenan, which Dave pushed. I think
we're ready for RC2. While the code is tested, we can continue reviewing
perldelta. I know I have to go over *all* of it again.
>
> I think that there is still a bunch of sloppiness in perldelta which
> would not reflect well on us as a project with our act together if it
> is released without being cleaned up.
>
> I added a section "Removed Deprecations" I think it probably should
> be renamed "Undeprecations" and added to the template.
I'm not sure "undeprecations" is a better term, to be honest.
> utf8_hop_safe() does not belong where it currently is, under "Core
> Enhancements". There are lots more functions and macros added in this
> release for the same "_safe" purpose, and they are listed later, under
> "Internal Changes". I'm thinking there should be a section like
> "Internal API Enhancements" or "XS API Enhancements"
That would be useful. We can also put such changes at the end of the
"Internal Changes." It's difficult to determine what *should* or
*shouldn't* go under that category since it's broad. Perhaps that's the
value of it.
>
> I'm willing to continue to slog through this document, but there's
> only so much I can do per day.
I'm going to try releasing RC2 today to at least give the code changes a
smoking chance and slog through the document again.
Thread Previous
|
Thread Next