Closed (won't fix)
Project:
Drupal core
Version:
8.0.x-dev
Component:
base system
Priority:
Normal
Category:
Plan
Assigned:
Unassigned
Reporter:
Created:
30 Oct 2015 at 21:29 UTC
Updated:
16 Jul 2018 at 19:34 UTC
Jump to comment: Most recent
Comments
Comment #2
chx commentedComment #3
catchWe could/should actively recommend PHP 7 from 8.1 onwards due to the performance gains. This has much more practical reasons for a specific recommendation than #2524432: Suggest PHP 5.6 as the recommended version does for 5.6.
Actually dropping 5.5 and/or 5.6 support is more drastic. I think it's reasonable for us to drop support for PHP versions when PHP itself does, not sure about before that, really needs to be a compelling reason to do so.
http://php.net/supported-versions.php
That would mean PHP 5.5 drops in 2015/2016 and PHP 5.6 drops in 2016/17 depending on whether you take the beginning or end of security-only support as the prompt to drop it.
So I would definitely think about dropping 5.5 support in 2016.
Dropping 5.6 support would require, for a start, that Drupal 7 runs smoothly on PHP 7, including contrib, no idea if that's the case or not. However it's a reasonable expectation, and one we shouldn't break, that you can run Drupal 7 and Drupal 8 on the same host without issues.
Comment #4
chx commenteda) By the time 8.2 drops if we keep 6 months twice, 5.6 is in security fixes only. b) if we want to do this we need to announce it now so that people can code/port with the right expectations. I think so many ppl will go PHP7 ASAP because of the real huge performance boost that if a contrib is not PHP7 compatible, bug reports will come in for that and so even if we don't hard require PHP7 core and contrib will be PHP7 compatible from day one.
If we don't do this then by 8.4 we are still stuck on PHP 5 and there's no more PHP 5 support any more.
And I think we could be trailblazers without taking a huge risk -- if by Drupal 8.1.0 the PHP 7 hosting support landscape is bleak then we announce a postpone to 8.3.0. Easier to postpone then than not announcing now. It's a very unique moment in time, when a lot of outside attention will be on Drupal.
Comment #5
chx commentedAlso ... if we go with this then we could perhaps get the PHP 7 people to put it into their announcement on the 12th.
Comment #6
chx commentedComment #7
stefan.r commented@tim.plunkett mentioned we'd be breaking semantic versioning by requiring PHP 7 in a minor release. In Drupal 8 core we wrote some code that was broken in PHP 7, so when contrib and custom modules do the same and their code stops working with Drupal 8.2 we have a BC break.
I wonder how far out we think Drupal 9 might be. If it's 3-4+ years from now, it may make sense to break semver and drop PHP 5 support in a minor release (whether that's 8.2, 8.3, 8.4, ...), and plan for that now so we can give advance notice prior to releasing 8.0.0.
I agree it's a reasonable expectation that actively maintained contrib modules will have fixed any PHP7 issues by the time PHP 5.6 is EOL. Any PHP 5.4 issues in D7 contrib were also fixed by the time 5.4 security support ended.
Comment #8
dawehnerIt would be great to quote that. Yes I could understand that you should be able to at least follow the security releases with version. For that usecase we could start dropping support for 5.5 after the first LTS.
I would not really worry too much about PHP 7 compatibility, such things can be patched quickly, but yeah I would watch out how hosting providers make the move to PHP 7. I could imagine that the additional performance motivates a lot of them to offer it at least.
Comment #9
chx commented> In Drupal 8 core we wrote some code that was broken in PHP 7
And caught it because bots now provide PHP 7 support but also we caught bugs in PHP 7 itself so the situation will be different (and better) for everything coming after.
Ever since I came up with the idea of cross-marketing D8 and PHP 7 together the more sure I am this is the right thing to do. Push together for a quicker transition, raise awareness for Drupal, win all around.
Comment #10
catchIf this issue title was 'recommend PHP 7' I'd be +1.
However it's not, and I think we should be more explicit and retitle it to be 'Drop PHP 5.6 support for Drupal 8.2' - and I think that is impractical.
I do think we can recommend PHP 7, and I personally would like to drop PHP 5.5 support (at least explicit support via DrupalCI testing) fairly soon after PHP itself does.
I opened #2607222: [policy, no patch] Default to PHP 7 for Drupal core patch testing as a concrete step towards this we can do in the next couple of weeks - would prefer to focus on things like that than what would primarily be a marketing move that annoys people actually trying to use Drupal (which is what would happen if we prevented 8.2 from actually running on PHP 5.6).
Comment #11
chx commentedComment #12
catchWell we should have open issues around dropping php 5.5 and 5.6 support. And issues on recommending php7.
Comment #13
effulgentsia commentedFWIW, I disagree with PHP dropping support as being the relevant trigger. Historically, various OS distributions continue security support of PHP past when PHP does. For example, Ubuntu 14.04 might continue providing secure builds of PHP 5.5 until 2019. So for me, dropping PHP 5.5 and/or PHP 5.6 during the 8.x lifetime should only happen when there's a very compelling reason, even after their official EOL. I'd certainly be fine with making Drupal 9 require PHP 7 though.
Comment #14
wim leershttps://groups.drupal.org/node/518200