GoPHP5: The Revenge

Showing 1-56 of 56 messages
GoPHP5: The Revenge Larry Garfield 10/30/13 5:07 PM
Hi folks.  Because we clearly don't have enough plates in the air[1], I
want to put another idea out there.

Short version:

http://w3techs.com/technologies/details/pl-php/5/all
- We all know that PHP 5.3 is about to lose even security support in the
first half of next year.
- PHP 5.3 is still the most widely used PHP version, with the completely
unsupported 5.2 a strong second.
- 5.4 hasn't even reached 10% yet, and 5.5, which is current stable,
barely registers.

This is particularly troublesome for us in Drupal as we prepare to
release Drupal 8; Drupal is the first 5.3-based version of Drupal, and
there's a strong push from some developers[2] to just bite the bullet
and go to 5.4 (both for new language features and for security reasons).
  Of course, the old chicken-and-egg problem rears its ugly head: Not
enough servers run it, but why should they if we don't require it?

The last time this big of a chicken-and-egg issue existed was around
moving to PHP 5.x at all, which took *for frickin' ever* to supplant PHP
4.  What finally make it happen was the GoPHP 5 project[3], in which a
collection of projects banded together to say "hosts, we're all going to
to 5.2 next year; get with the program or go out of business. kthxbye."
  In short order, 100 projects and 200 hosts signed on and PHP Internals
caught up to announce they were EOLing PHP 4[4].  We, as a community,
assassinated a programming language.

I believe it is time to discuss round 2 of that effort.  I also believe
that it would be good for FIG to play a leading role in such an effort
if possible.  It wouldn't be a PSR, but a collective action by various
projects.  Naturally we cannot require any projects to participate, but
a statement from FIG could encourage many to do so, both members and
non-members.

If nothing else, FIG is the best place to discuss such cooperation for a
number of reasons.  For one, consistent version requirements help
interoperability.  For another, at this point FIG represents pretty much
every project in the PHP world that matters except WordPress, plus many
others[5].

If we follow the previous effort's pattern, the model would be to
collectively agree on a date and announce "any release we have after
this date will be PHP 5.5 and up only. Hosts, you've got until then."
And then have a page where we track projects that have signed on and
hosts that want to advertise that they're offering that version by default.

In the above, note I said 5.5, not 5.4.  If we were to do this, we'd
want to give ample warning.  it's not an end-of-this-year thing; June is
probably the earliest we could target, though of course the intent would
be that hosts move before then.  By June 2014, PHP 5.5 will be just
about a year old and PHP 5.4 will be fading into limited support[6].
PHP's more aggressive release cycle means we need to move a bit faster
to keep up.  Moving forward, it would be great if we could collectively
sync our plans to PHP's release cycle in order to have a more
predictable upgrade cycle for everyone.

The other question is what constitutes a "release".  For Drupal, we'd be
talking about Drupal 8, most likely, although it may be Drupal 9
depending on how quickly we get this release out (and Drupal 9 is years
away)[7].  So far we haven't even discussed 5.5, and 5.4 is a challenge.
  I know others like Symfony have a much longer-term commitment to their
current version on 5.3 that they likely cannot break until their next
major release (Symfony 3 in that case).  Others, like Aura, are already
leaning toward 5.4 because certain things are so much easier.  I'm not
aware of any member projects that are already 5.4, although that just
may be my ignorance.  I am not sure what the best approach is here to
not hamstring current projects and current users but to still send a
strong enough message to hosts to drive a mass-upgrading effort as we
did before.  I welcome input on this front.

Disclaimer: Although I am the Drupal representative to FIG, I have not
pre-cleared this proposal with the other Drupal project leads so cannot
promise that Drupal would be gung-ho on it.  This is a proposal from
Larry Garfield: PHP busybody and community evangelist, not from Larry
Garfield: Drupal rep.

*dons flame retardant suit*

--Larry Garfield

[1] This is sarcasm
[2] https://drupal.org/node/1498574
[3] http://www.garfieldtech.com/blog/go-php-5-go
[4] They never admitted that GoPHP5 was the reason for the sudden change
of heart, but the correlation seems unlikely to be a coincidence.
[5] I will allow everyone to decide in which category they fall; there
are also some major apps not yet represented that we may want to think
about inviting at some point.
[6] If my timing is off, I'm sure Lukas Smith will correct me.
[7] One irony of GoPHP5 is that Drupal was a founding project and one of
the main drivers of the effort; yet our Drupal 6 release slipped late
and just barely came *after* the deadline to require PHP 5.2, yet still
had a PHP 4 requirement.  Fortunately by that point the impact had
already been made and no one really noticed, but I mention it in the
name of transparency.
Re: GoPHP5: The Revenge Zachary King 10/30/13 5:18 PM
Personally I think it sounds like a great idea. Though I don't have a large code base to port over :)

I would love to see hosts moving to current versions faster.

-Zach




--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/52719F5C.9050709%40garfieldtech.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge Drak 10/30/13 5:31 PM
One of the real places to exert pressure are linux distributors. One of the biggest holdups have have been the likes of RHEL and consequently CentOS. Fortunately a lot more hosts are using Debian/Ubuntu but the enterprise linux distros are the real places we need to pressure.

Drak




--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/52719F5C.9050709%40garfieldtech.com.
For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge John Mertic 10/30/13 5:38 PM
I think before we get too deep into this, we should engage the Wordpress guys. Andrew Nacin did a great presentation at php|tek this past spring, which while it was received with a bit of hostility, it really spoke well to folks on the other side of the issue in the "I don't care which version of PHP I'm using as long as things just work".

Just my two cents here, take it for what it's worth...

John


Re: GoPHP5: The Revenge Michael Dowling 10/30/13 5:38 PM
> The last time this big of a chicken-and-egg issue existed was around
moving to PHP 5.x at all, which took *for frickin' ever* to supplant PHP
4.  What finally make it happen was the GoPHP 5 project[3], in which a
collection of projects banded together to say "hosts, we're all going to
to 5.2 next year; get with the program or go out of business. kthxbye."
  In short order, 100 projects and 200 hosts signed on and PHP Internals
caught up to announce they were EOLing PHP 4[4].  We, as a community,
assassinated a programming language.

I'd love for the next version of Guzzle to require, at minimum, PHP 5.4.
Really, with the adoption of 5.4 being so low and new distros  moving to
5.5, I'd rather see projects commit to moving to 5.5 as a next "Go PHP"
type effort. Promoting 5.4 at this point would just put us in this
situation in a year or two from now.


--Michael
Re: GoPHP5: The Revenge Drak 10/30/13 6:03 PM
On 31 October 2013 00:38, John Mertic <jme...@gmail.com> wrote:
I think before we get too deep into this, we should engage the Wordpress guys. Andrew Nacin did a great presentation at php|tek this past spring, which while it was received with a bit of hostility, it really spoke well to folks on the other side of the issue in the "I don't care which version of PHP I'm using as long as things just work".

If Wordpress committed it would be a slam dunk... 
Re: GoPHP5: The Revenge Hari K T 10/30/13 6:11 PM
Great initiative Larry Garfield .

Go 5.5 . Nice number too :) . GoPHP5 , GoPHP55 ;)

Hari K T

You can ring me : +91 9388 75 8821

Skype  : kthari85
Twitter : harikt


--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/CANAnSg3_rwR7pUpzcXb_CPLBkSZg-%2BJDCid8M26UZW-RaXOG6A%40mail.gmail.com.

For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge pmjones 10/30/13 6:13 PM

On Oct 30, 2013, at 7:07 PM, Larry Garfield <la...@garfieldtech.com> wrote:

> If we follow the previous effort's pattern, the model would be to collectively agree on a date and announce "any release we have after this date will be PHP 5.5 and up only. Hosts, you've got until then." And then have a page where we track projects that have signed on and hosts that want to advertise that they're offering that version by default.

I spoke with Larry about this earlier and pledged Aura's support of this.  While I find it hard to get behind "PHP 5.5 and up only" per se, since that implies somehow using a PHP 5.5-only feature in every library to make it non-PHP-5.4 compatible, I can totally get behind "100% tested on PHP 5.4 and 5.5" (or some variation thereof) as well as advertising packages that are 5.5-only.

(Larry, sorry to backtrack on my earlier statement to you; please accept my apologies.)


> Others, like Aura, are already leaning toward 5.4 because certain things are so much easier.

PHP 5.4 is a requirement for all Aura v1 packages (and the framework too as a result).  For exactly the reasons you state, and with heavy heart, we are reworking v2 packages that do not use traits so they are PHP 5.3 compatible to grant them a wider audience.  Yes, the "callable" typehint and the short array syntax are *so* nice.


--
Paul M. Jones
pmjo...@gmail.com
http://paul-m-jones.com


Re: GoPHP5: The Revenge John Mertic 10/30/13 6:38 PM
Hi,

On Oct 30, 2013, at 9:03 PM, Drak <dr...@zikula.org> wrote:

On 31 October 2013 00:38, John Mertic <jme...@gmail.com> wrote:
I think before we get too deep into this, we should engage the Wordpress guys. Andrew Nacin did a great presentation at php|tek this past spring, which while it was received with a bit of hostility, it really spoke well to folks on the other side of the issue in the "I don't care which version of PHP I'm using as long as things just work".

If Wordpress committed it would be a slam dunk...\

I hear what you are saying... but the dialog is a bit larger than this. And frankly, something I'm not seeing as a part of the charter of this group. 

Again, just my two cents here...

John
Re: GoPHP5: The Revenge Ryan McCue 10/30/13 7:56 PM
John Mertic wrote:
> I think before we get too deep into this, we should engage the Wordpress
> guys. Andrew Nacin did a great presentation at php|tek this past spring,
> which while it was received with a bit of hostility, it really spoke
> well to folks on the other side of the issue in the "I don't care which
> version of PHP I'm using as long as things just work".

I'm happy to act as the WordPress liaison for this.

Our big issue with moving beyond 5.2 is usage, which was also the case
for 4. We still have over half [1] of our userbase on 5.2, which is well
over the point where we're ready to drop it. PHP 4 support was dropped
only after we reached ~10% on <5.2 [2].

The reason we've followed the hosts rather than leading is that we
always believe in maximum compatibility (as I believe Nacin spoke about
at php|tek). If we release versions of WP that require newer versions,
we're leaving users behind who will be unable to upgrade, and that's
much worse to us.

In addition, to my knowledge, there's no really compelling features that
make us want 5.3+. Anonymous functions don't work well with our callback
system (since you can't deregister them later), our classes aren't
suddenly going to change to namespacing (huge backwards compatibility
break if so), etc.

As much as I'd love to push for this personally, I don't think WordPress
is going to get behind this.

[1]: http://wordpress.org/about/stats/
[2]: http://wordpress.org/news/2010/07/eol-for-php4-and-mysql4/

--
Ryan McCue
<http://ryanmccue.info/>
Re: GoPHP5: The Revenge Larry Garfield 10/30/13 9:07 PM
On 10/30/2013 08:13 PM, Paul M. Jones wrote:
> On Oct 30, 2013, at 7:07 PM, Larry Garfield <la...@garfieldtech.com> wrote:
>
>> If we follow the previous effort's pattern, the model would be to collectively agree on a date and announce "any release we have after this date will be PHP 5.5 and up only. Hosts, you've got until then." And then have a page where we track projects that have signed on and hosts that want to advertise that they're offering that version by default.
> I spoke with Larry about this earlier and pledged Aura's support of this.  While I find it hard to get behind "PHP 5.5 and up only" per se, since that implies somehow using a PHP 5.5-only feature in every library to make it non-PHP-5.4 compatible, I can totally get behind "100% tested on PHP 5.4 and 5.5" (or some variation thereof) as well as advertising packages that are 5.5-only.
>
> (Larry, sorry to backtrack on my earlier statement to you; please accept my apologies.)

I should clarify what I mean by support here.  As with the previous
GoPHP5 project, I don't see any value in requiring people to use
something in 5.5 just to piss off users of older versions.  Rather, it's
a statement to say "we only support X version and up... and we are not
going to bother checking if something before that works or not."  
Something like, say, Monolog probably wouldn't have much reason to use
the new features in 5.5.  In its case, it would mostly come down to
removing the pre-5.5 lines from its Travis CI configuration and then if
something 5.5-ish looks useful, meh, use it.  If not, meh, don't.  It
may well continue to work on 5.4 for a long while after that, which is
fine, just "unsupported".

If there's nothing specific in 5.5 that Aura would benefit from, that's
OK.  Participation would just mean you don't bother to check if you're
doing something that would not work in 5.4. If you don't actually break
5.4 for another year, *shrug*.

--Larry Garfield
Re: GoPHP5: The Revenge Larry Garfield 10/30/13 9:18 PM
Honestly, I didn't think Wordpress would be interested given how openly
hostile Matt was toward GoPHP5 the first time around:

http://ma.tt/2007/07/on-php/

I totally get the desire to not orphan existing users; however, everyone
taking that attitude is why PHP 4 outlived its usefulness so long and
why PHP 5.3 and even PHP 5.2 are still hanging on. Dislodging fear of
new things requires boldness.  It worked before, for the betterment of
everyone, and I think it will again.

I don't know enough about Wordpress's architecture to say specifically
what might be beneficial, but I'd be very surprised if there was nothing
in two significant releases of PHP that Wordpress would find useful
without breaking an API. :-)  Honestly, from the Wordpress front silence
or a non-committal "good luck guys" would be a significant step up from
2007.  I didn't really expect more than that.

--Larry Garfield
Re: GoPHP5: The Revenge Ryan McCue 10/30/13 9:29 PM
Larry Garfield wrote:
> Honestly, I didn't think Wordpress would be interested given how openly
> hostile Matt was toward GoPHP5 the first time around:
>
> http://ma.tt/2007/07/on-php/

It's worth emphasising that those were Matt's personal views, not
necessarily the views of the whole core team. The core team is also
significantly bigger now than it was 6 years ago, and Matt's less
involved in the day-to-day decisions like this.

> I totally get the desire to not orphan existing users; however, everyone
> taking that attitude is why PHP 4 outlived its usefulness so long and
> why PHP 5.3 and even PHP 5.2 are still hanging on. Dislodging fear of
> new things requires boldness.  It worked before, for the betterment of
> everyone, and I think it will again.

Personally I totally agree, but I also agree with WordPress' policy.
I've talked to some of the host representatives I know to try and push
this forward, and that's where the discussion needs to happen for us,
not by unilaterally declaring that we're dropping support.

> I don't know enough about Wordpress's architecture to say specifically
> what might be beneficial, but I'd be very surprised if there was nothing
> in two significant releases of PHP that Wordpress would find useful
> without breaking an API. :-)

The biggest feature to us would probably be the performance
improvements. Apart from that, I'm not sure anything would truly benefit
us *right now*. Having said that, we also said this with regards to PHP
5, and there are some features we're definitely using there.
(Reflection, for example, is something that we're looking at using in
the REST API.)

> Honestly, from the Wordpress front silence
> or a non-committal "good luck guys" would be a significant step up from
> 2007.  I didn't really expect more than that.

I can't offer you an official opinion of the core team, but... good luck
guys. :)

However, I'll be here participating, and I'll start reaching out to
hosts right now to get a hold of what we can start doing there. If
anyone has any specific hosts to target, let me know, otherwise I'll
just start from the top.

Cheers,
Re: GoPHP5: The Revenge Larry Garfield 10/30/13 10:00 PM
On 10/30/2013 11:29 PM, Ryan McCue wrote:
> Larry Garfield wrote:
>> Honestly, I didn't think Wordpress would be interested given how openly
>> hostile Matt was toward GoPHP5 the first time around:
>>
>> http://ma.tt/2007/07/on-php/
> It's worth emphasising that those were Matt's personal views, not
> necessarily the views of the whole core team. The core team is also
> significantly bigger now than it was 6 years ago, and Matt's less
> involved in the day-to-day decisions like this.

Good to know.  I don't know much (read: anything) about Wordpress
internal politics in this regard.

>> I don't know enough about Wordpress's architecture to say specifically
>> what might be beneficial, but I'd be very surprised if there was nothing
>> in two significant releases of PHP that Wordpress would find useful
>> without breaking an API. :-)
> The biggest feature to us would probably be the performance
> improvements. Apart from that, I'm not sure anything would truly benefit
> us *right now*. Having said that, we also said this with regards to PHP
> 5, and there are some features we're definitely using there.
> (Reflection, for example, is something that we're looking at using in
> the REST API.)

Out of curiosity, what does Wordpress.com run?  (I suspect that accounts
for a huge percentage of the Wordpress install base, if not a majority;
I have no data for that suspicion, of course.)

>
>> Honestly, from the Wordpress front silence
>> or a non-committal "good luck guys" would be a significant step up from
>> 2007.  I didn't really expect more than that.
> I can't offer you an official opinion of the core team, but... good luck
> guys. :)

Thanks. :-)

> However, I'll be here participating, and I'll start reaching out to
> hosts right now to get a hold of what we can start doing there. If
> anyone has any specific hosts to target, let me know, otherwise I'll
> just start from the top.
>
> Cheers,

That would be awesome.  If we go through with it, we'll want to keep a
running list of latest-version-friendly hosts as before, in addition to
projects.  To the extent possible it should be a collective action, not
a "hosts you suck".  Free advertising for hosts that are on board is an
important carrot.

--Larry Garfield
Re: GoPHP5: The Revenge Evert Pot 10/30/13 10:05 PM
> I should clarify what I mean by support here.  As with the previous GoPHP5 project, I don't see any value in requiring people to use something in 5.5 just to piss off users of older versions.  Rather, it's a statement to say "we only support X version and up... and we are not going to bother checking if something before that works or not."  Something like, say, Monolog probably wouldn't have much reason to use the new features in 5.5.  In its case, it would mostly come down to removing the pre-5.5 lines from its Travis CI configuration and then if something 5.5-ish looks useful, meh, use it.  If not, meh, don't.  It may well continue to work on 5.4 for a long while after that, which is fine, just "unsupported".
>
> If there's nothing specific in 5.5 that Aura would benefit from, that's OK.  Participation would just mean you don't bother to check if you're doing something that would not work in 5.4. If you don't actually break 5.4 for another year, *shrug*.

Speaking from SabreDAV. The next major version will require PHP 5.4. This is much to the dismay of some users, but they will have to get over it.
For an application like SabreDAV, the most common use-case has an extremely high likelihood of requiring a dedicated environment, and it's my opinion that the environment should adapt to the application, and not the other way round.

If there's a large enough group supporting GoPHP5.5, I'd also happily push PHP 5.5 for the version right after that. Can't wait to start using generators.

My only concern is that I think hhvm will become more and more relevant; I may instead start using 'current hhvm' as a baseline.

Evert
Re: GoPHP5: The Revenge Ryan McCue 10/30/13 10:06 PM
Larry Garfield wrote:
> Out of curiosity, what does Wordpress.com run?  (I suspect that accounts
> for a huge percentage of the Wordpress install base, if not a majority;
> I have no data for that suspicion, of course.)

WordPress.com runs 5.4 according to my sources, but WordPress.com also
a) doesn't affect any of the statistics as shown on WordPress.org (they
handle their own upgrading, so there's no need to check for updates),
and b) would count as a single install if it did appear.

> That would be awesome.  If we go through with it, we'll want to keep a
> running list of latest-version-friendly hosts as before, in addition to
> projects.  To the extent possible it should be a collective action, not
> a "hosts you suck".  Free advertising for hosts that are on board is an
> important carrot.

Agreed; anything to encourage hosts is good in my book. Plus, it means
we can give the stick to the hosts that don't ;)
Re: GoPHP5: The Revenge William DURAND 10/31/13 1:45 AM
I like the idea! However, most of my libs are PHP 5.3 compliant, and I don't feel the need to bump the minimum PHP version as I don't rely on new features... Should we just bump the minium requirement to *invite* people to upgrade their PHP setup?

Note that Propel2 requires at least PHP 5.4. Propel1 is still PHP 5.2 compliant... That's why we decided to choose PHP 5.4 as minimum requirement earlier last year, we can't bump the minimum version afterward.

William

--
William Durand | http://www.williamdurand.fr


--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/5271E550.5020005%40rotorised.com.


Re: GoPHP5: The Revenge Drak 10/31/13 2:39 AM
On 31 October 2013 02:56, Ryan McCue <li...@rotorised.com> wrote:
The reason we've followed the hosts rather than leading is that we
always believe in maximum compatibility (as I believe Nacin spoke about
at php|tek). If we release versions of WP that require newer versions,
we're leaving users behind who will be unable to upgrade, and that's
much worse to us.

But in the case of PHP 5.3, it was already the case a long time ago that there was sufficient competition in the market to allow anyone to switch. It's not like they could not upgrade, the options were always there. It's also becoming very common for hosts to support multiple versions of PHP so you can just switch in the hosting control panel.
 
In addition, to my knowledge, there's no really compelling features that
make us want 5.3+. Anonymous functions don't work well with our callback
system (since you can't deregister them later), our classes aren't
suddenly going to change to namespacing (huge backwards compatibility
break if so), etc.

Just in terms of speed and performance, PHP 5.3 makes sense. It's incredibly more efficient than 5.2. PHP 5.4 and 5.5 are even more efficient.
There should be all the incentive.
 
As much as I'd love to push for this personally, I don't think WordPress
is going to get behind this.

One thing Wordpress could do is to liaise with the hosting companies they act as affiliate for and encourage them to upgrade to 5.5, or at least add it as an option to their control panels. The likes of Bluehost etc positively thrive on their Wordpress customers. That would go a LONG way to helping the movement. They already recommend these hosts, and then they could recommend officially to use PHP 5.5 for all the speed and memory improvements and say "these hosts already support it". It would be a positive feedback loop for them.

Regards,

Drak

[1]: http://wordpress.org/about/stats/
[2]: http://wordpress.org/news/2010/07/eol-for-php4-and-mysql4/

--
Ryan McCue
<http://ryanmccue.info/>

--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/5271C6C5.5050907%40rotorised.com.


Re: GoPHP5: The Revenge Alexander Makarov 10/31/13 3:05 AM
I support this initiative and will help whatever I can for it to achieve its goal.

Yii2 decided to use 5.4 recently. We did a research and it seems not enough servers is not entirely true: https://docs.google.com/spreadsheet/ccc?key=0Agccij5yGAXadFZBcVV4U0NaUWwyNko4MTBNcHI1T1E
So the initiative should work well if done properly.

Would be great to get phpbb, vbulletin, IPB, Joomla, Drupal and Wordpress on board. Currenlty Joomla even doesn't work throwing errors on 5.4. That's one of the reasons shared hosting companies aren't upgrading.
The fact is forcing these to require 5.5 will be nearly impossible to achieve.

Ryan McCue, do you think it's possible to put a large visible "Wordpress works much better with PHP 5.5" on the website?
I think same "flexible" approach will work for products that have outdated userbase.
Re: GoPHP5: The Revenge Ryan McCue 10/31/13 3:11 AM
Drak wrote:
> But in the case of PHP 5.3, it was already the case a long time ago that
> there was sufficient competition in the market to allow anyone to
> switch. It's not like they could not upgrade, the options were always
> there. It's also becoming very common for hosts to support multiple
> versions of PHP so you can just switch in the hosting control panel.

What matters to us is what they're on. It might be easy to switch, but
it's not seamless, and they might have parts of their site that break.
Allowing the hosts to handle that with their customers is a better
experience for everyone involved.

> Just in terms of speed and performance, PHP 5.3 makes sense. It's
> incredibly more efficient than 5.2. PHP 5.4 and 5.5 are even more efficient.
> There should be all the incentive.

I thought I mentioned this in my original email; apologies. I agree, but
it's not something we *need*. If it still works on 5.2, there's no need
to require something newer.

> One thing Wordpress could do is to liaise with the hosting companies
> they act as affiliate for and encourage them to upgrade to 5.5, or at
> least add it as an option to their control panels. The likes of Bluehost
> etc positively thrive on their Wordpress customers. That would go a LONG
> way to helping the movement. They already recommend these hosts, and
> then they could recommend officially to use PHP 5.5 for all the speed
> and memory improvements and say "these hosts already support it". It
> would be a positive feedback loop for them.

That is exactly what's happening. Both myself and other core developers
are already working with hosts and plan to continue to do so. In fact,
Bluehost is pushing all of their 5.2 customers to 5.4 in the coming
weeks [1].

Alexander Makarov wrote:
> *Ryan McCue*, do you think it's possible to put a large visible
> "Wordpress works much better with PHP 5.5" on the website?
> I think same "flexible" approach will work for products that have
> outdated userbase.

It's one of the things we can look in to, but it's not really a
priority. Most people are at the mercy of their hosts with regards to
this, so pushing for it there is the preferred course of action.

[1]:
http://lists.automattic.com/pipermail/wp-forums/2013-October/010243.html
Re: GoPHP5: The Revenge Jordi Boggiano 10/31/13 3:23 AM
Yes please. I am tired of 5.3, but I also know Composer will probably
be the last one being able to switch, so I'm not holding my breath. Do
it though!

Cheers

--
Jordi Boggiano
@seldaek - http://nelm.io/jordi
Re: GoPHP5: The Revenge Benjamin Eberlei 10/31/13 6:05 AM
Going for 5.5 is not a credible threat for hosters, as no library actually uses 5.5 features.



--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/CA%2BGmM%3Dxaqf2nN-H0Wjhe7DznM9T1%3DAU8U517NX%3D%2B0aUS6yQzaw%40mail.gmail.com.


Re: GoPHP5: The Revenge Beau Simensen 10/31/13 8:58 AM
I've considered moving Sculpin to 5.4 or 5.5. The biggest reasons for me are access to newer Markdown libraries (one promising one[1] supports 5.4+) and continued support from React[2] (0.4+ only supports 5.4+). The biggest things holding me back is fear of not getting enough adoption.

Currently, the target audience for Sculpin is aimed more at more technical people so there is a higher chance that they will either have or be able to add at least PHP 5.4 to their systems pretty easily. On the other hand, I myself have been more than happy to stick with stock 5.3.x on both my Mac and on my Ubuntu LTS servers and virtual machines... you may disagree with my choices but I mention it only to point out that until a few months ago I wasn't really setup to actively work on PHP 5.4 projects. I only made the jump when I was forced to use PHP 5.4 in order to do more extensive work in Stack and decided to go "all in" on PHP 5.5 and ensured that my mac and Linux hosts were all capable of PHP 5.5.

I asked what people thought of the idea. The feedback I've received was generally positive, though there were some caution advised. I think that when I get to the point that I really want to migrate to a new Markdown library I'll have to make the leap.

Re: GoPHP5: The Revenge Don Gilbert 10/31/13 9:19 AM
Ok, which one of you already purchased GoPHP55.org? http://d.pr/i/RyjI

I'm for it personally, but the current LTS version of the Joomla CMS (2.5) has a PHP 5.2 minimum requirement, and it's supported until the end of 2014.

The next planned LTS version of the Joomla CMS (3.5) has a PHP 5.3 minimum requirement, and it will be supported until at least the end of 2016. :(

The CMS does, however, recommend PHP 5.4+ for the server environment, so there's that - http://www.joomla.org/technical-requirements.html

The Joomla Framework is already targeting 5.3.10+ and there's talk about actively supporting 5.4+ only. Although there are no plans to specifically drop support for 5.3, we'll just stop actively checking lower versions, as was mentioned earlier in the thread.

All in all, I think it's a good thing. We can probably push for Joomla 4.x to have a 5.5 minimum requirement, and the first version for that is scheduled to be released Q1 2015.


Re: GoPHP5: The Revenge Larry Garfield 10/31/13 9:59 AM
On 10/31/13 11:19 AM, Donald Gilbert wrote:
> Ok, which one of you already purchased GoPHP55.org? http://d.pr/i/RyjI

Wasn't me.  I picked up gophp.info, so that it's not version-dependent
and we can use it for tracking/coordinating versioning long-term. :-)

> The CMS does, however, recommend PHP 5.4+ for the server environment, so
> there's that - http://www.joomla.org/technical-requirements.html


> The Joomla Framework is already targeting 5.3.10+ and there's talk about
> actively supporting 5.4+ only. Although there are no plans to
> specifically drop support for 5.3, we'll just stop actively checking
> lower versions, as was mentioned earlier in the thread.

That is essentially where Drupal is right now, although if we make that
shift there will be several traits introduced within a week.

> All in all, I think it's a good thing. We can probably push for Joomla
> 4.x to have a 5.5 minimum requirement, and the first version for that is
> scheduled to be released Q1 2015.

Given the different landscape, perhaps a coordinated effort to document
and publicize aggressive upgrade schedules rather than one globally
unified upgrade schedule would be sufficient?  That wouldn't have the
big-splash effect of the original GoPHP5, but could have enough of an
effect that we could all move to PHP 5.4 sooner, at the very least.

--Larry Garfield
Re: GoPHP5: The Revenge Larry Garfield 10/31/13 10:00 AM
On 10/31/13 5:23 AM, Jordi Boggiano wrote:
> Yes please. I am tired of 5.3, but I also know Composer will probably
> be the last one being able to switch, so I'm not holding my breath. Do
> it though!
>
> Cheers

Which reminds me... is it possible to get any data from Composer about
current version requirements for Composer-distributed packages?

--Larry Garfield
Re: GoPHP5: The Revenge Zachary King 10/31/13 10:18 AM
Given the different landscape, perhaps a coordinated effort to document
and publicize aggressive upgrade schedules rather than one globally
unified upgrade schedule would be sufficient?  That wouldn't have the
big-splash effect of the original GoPHP5, but could have enough of an
effect that we could all move to PHP 5.4 sooner, at the very least.

Maybe something like the Python 3 Wall of Superpowers/Shame? https://python3wos.appspot.com/
Re: GoPHP5: The Revenge Jordi Boggiano 10/31/13 10:34 AM
On Thu, Oct 31, 2013 at 6:00 PM, Larry Garfield <la...@garfieldtech.com> wrote:
> Which reminds me... is it possible to get any data from Composer about
> current version requirements for Composer-distributed packages?

I only looked at dev-master versions (that excludes svn/hg packages,
but it's like 99% of packages), we have 17593 of these, and only 12882
specify a requirement on php at all.

Out of those 12882 versions, over 80% require 5.3+, and 5.4 is at 10%.
You can find details here: http://dump.seld.be/phpversions.html

It's not so surprising given the whole Composer ecosystem is mostly
"post-5.3", but it's time to move this forward :)

Cheers

--
Jordi Boggiano
@seldaek - http://nelm.io/jordi
GoPHP5: The Revenge Andrew Embler 11/3/13 10:06 AM
We are (finally) moving concrete5 to require php 5.3 with our next major release. I believe you're going to see active 5.2 installations fall off a cliff over the next 6 months, and that going 5.4.x will be feasible even for larger projects sometime in the middle of next year.
Re: GoPHP5: The Revenge Phil Sturgeon 11/3/13 6:26 PM

PyroCMS will also be releasing 2.3.0-alpha1 shortly which ups requirements to PHP 5.3.

GoPHP5: The Revenge Taylor Otwell 11/3/13 7:13 PM
Laravel will be moving to 5.4 for our May 2014 release.
Re: GoPHP5: The Revenge Andreas Möller 11/3/13 10:16 PM


> I'd love for the next version of Guzzle to require, at minimum, PHP 5.4.
> Really, with the adoption of 5.4 being so low and new distros  moving to
> 5.5, I'd rather see projects commit to moving to 5.5 as a next "Go PHP"
> type effort. Promoting 5.4 at this point would just put us in this
> situation in a year or two from now.

Couldn't agree more!


Best regards,

Andreas
Re: GoPHP5: The Revenge FGM 11/3/13 10:58 PM
Although I'm not the one who should talk about it, I noticed that Silex has been adding lots of traits in recent versions, thus not forcing a 5.4 requirement, but strongly encouraging it by making 5.4 applications contain less code thanks to traits.


2013/11/4 Andreas Möller <a...@localheinz.com>

--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/0C9D82B2-EF25-45C8-AD15-3382FC803190%40localheinz.com.

Re: GoPHP5: The Revenge Phil Sturgeon 11/4/13 6:32 AM

I'd strongly urge packages like Guzzle to maintain a PHP 5.3 version (security updates in a 1.x branch while the 2.x branch cracks on with new stuff) as some CMS' like PyroCMS will need to loiter on PHP 5.3 for another 6 months at least - depending on the specifics of their community.

We'll push it to PHP 5.4 again for 3.0 or 3.1, but 2.3 is only just getting to PHP 5.3 and going any faster will seriously screw us on sales of the Pro version.

Other purely open-source CMS' can just jump straight to 5.4 and 5.5 whenever no doubt. 
Re: GoPHP5: The Revenge Alexander Makarov 11/6/13 2:36 PM
If everyone agree it's a good idea we can start by creating a repo under https://github.com/php-fig and starting work on the PHP 5.5 promo website.

My idea is that it should contain colorful banners/guidelines of 2 types:

1. We're requiring 5.5. We're cutting edge.
2. We're recommending 5.5. Our project is compatible and works much better with it.

Wording and design should be standartised so it will be recognizable and therefore get more attention.

Also ther should be a list of notable projects w/ links.

On Thursday, October 31, 2013 4:07:56 AM UTC+4, Larry Garfield wrote:
Hi folks.  Because we clearly don't have enough plates in the air[1], I
want to put another idea out there.

Short version:

http://w3techs.com/technologies/details/pl-php/5/all
- We all know that PHP 5.3 is about to lose even security support in the
first half of next year.
- PHP 5.3 is still the most widely used PHP version, with the completely
unsupported 5.2 a strong second.
- 5.4 hasn't even reached 10% yet, and 5.5, which is current stable,
barely registers.

This is particularly troublesome for us in Drupal as we prepare to
release Drupal 8; Drupal is the first 5.3-based version of Drupal, and
there's a strong push from some developers[2] to just bite the bullet
and go to 5.4 (both for new language features and for security reasons).
  Of course, the old chicken-and-egg problem rears its ugly head: Not
enough servers run it, but why should they if we don't require it?

The last time this big of a chicken-and-egg issue existed was around
moving to PHP 5.x at all, which took *for frickin' ever* to supplant PHP
4.  What finally make it happen was the GoPHP 5 project[3], in which a
collection of projects banded together to say "hosts, we're all going to
to 5.2 next year; get with the program or go out of business. kthxbye."
  In short order, 100 projects and 200 hosts signed on and PHP Internals
caught up to announce they were EOLing PHP 4[4].  We, as a community,
assassinated a programming language.

I believe it is time to discuss round 2 of that effort.  I also believe
that it would be good for FIG to play a leading role in such an effort
if possible.  It wouldn't be a PSR, but a collective action by various
projects.  Naturally we cannot require any projects to participate, but
a statement from FIG could encourage many to do so, both members and
non-members.

If nothing else, FIG is the best place to discuss such cooperation for a
number of reasons.  For one, consistent version requirements help
interoperability.  For another, at this point FIG represents pretty much
every project in the PHP world that matters except WordPress, plus many
others[5].

If we follow the previous effort's pattern, the model would be to
collectively agree on a date and announce "any release we have after
this date will be PHP 5.5 and up only. Hosts, you've got until then."
And then have a page where we track projects that have signed on and
hosts that want to advertise that they're offering that version by default.

In the above, note I said 5.5, not 5.4.  If we were to do this, we'd
want to give ample warning.  it's not an end-of-this-year thing; June is
probably the earliest we could target, though of course the intent would
be that hosts move before then.  By June 2014, PHP 5.5 will be just
about a year old and PHP 5.4 will be fading into limited support[6].
PHP's more aggressive release cycle means we need to move a bit faster
to keep up.  Moving forward, it would be great if we could collectively
sync our plans to PHP's release cycle in order to have a more
predictable upgrade cycle for everyone.

The other question is what constitutes a "release".  For Drupal, we'd be
talking about Drupal 8, most likely, although it may be Drupal 9
depending on how quickly we get this release out (and Drupal 9 is years
away)[7].  So far we haven't even discussed 5.5, and 5.4 is a challenge.
  I know others like Symfony have a much longer-term commitment to their
current version on 5.3 that they likely cannot break until their next
major release (Symfony 3 in that case).  Others, like Aura, are already
leaning toward 5.4 because certain things are so much easier.  I'm not
aware of any member projects that are already 5.4, although that just
may be my ignorance.  I am not sure what the best approach is here to
not hamstring current projects and current users but to still send a
strong enough message to hosts to drive a mass-upgrading effort as we
did before.  I welcome input on this front.

Disclaimer: Although I am the Drupal representative to FIG, I have not
pre-cleared this proposal with the other Drupal project leads so cannot
promise that Drupal would be gung-ho on it.  This is a proposal from
Larry Garfield: PHP busybody and community evangelist, not from Larry
Garfield: Drupal rep.

*dons flame retardant suit*

--Larry Garfield

[1] This is sarcasm
[2] https://drupal.org/node/1498574
[3] http://www.garfieldtech.com/blog/go-php-5-go
[4] They never admitted that GoPHP5 was the reason for the sudden change
of heart, but the correlation seems unlikely to be a coincidence.
[5] I will allow everyone to decide in which category they fall; there
are also some major apps not yet represented that we may want to think
about inviting at some point.
[6] If my timing is off, I'm sure Lukas Smith will correct me.
[7] One irony of GoPHP5 is that Drupal was a founding project and one of
the main drivers of the effort; yet our Drupal 6 release slipped late
and just barely came *after* the deadline to require PHP 5.2, yet still
had a PHP 4 requirement.  Fortunately by that point the impact had
already been made and no one really noticed, but I mention it in the
name of transparency.
Re: GoPHP5: The Revenge Jamie Hannaford 11/10/13 3:06 PM
I think this is a great idea. Whenever I come across a Github project which requires 5.4+ outright, I'm immediately interested. Although it's crucial to keep the life support plugged in for older versions, having a more gung-ho approach is important.

Could FIG also provide upgrade guides? Like an official doc that can help people not familiar with Linux repos / manually compiling the source themselves...

Jamie
Re: GoPHP5: The Revenge Phil Sturgeon 11/11/13 9:59 AM


On Sunday, 10 November 2013 18:06:42 UTC-5, Jamie Hannaford wrote:
Could FIG also provide upgrade guides? Like an official doc that can help people not familiar with Linux repos / manually compiling the source themselves...

That is more of a job for PHP The Right Way than PHP-FIG.
Re: GoPHP5: The Revenge Jamie Hannaford 11/11/13 11:06 AM
Ah okay. Yeah, I wasn't sure whether this was in FIG's remit - PHPTRW looks like a really cool alternative.
Re: GoPHP5: The Revenge Josh Hall-Bachner 11/11/13 2:22 PM
On Wednesday, October 30, 2013 5:07:56 PM UTC-7, Larry Garfield wrote:

PHP's more aggressive release cycle means we need to move a bit faster
to keep up.

Apologies in advance for taking this and complicating it even more, but... I feel like the above is the bigger-picture issue here. Hosts are used to ignoring new PHP versions and lingering in long-outdated versions of the language, but the PHP project has adopted a release strategy that runs headlong in exactly the opposite direction. It's all well and good to push for people to move to 5.5 now, but what happens in another three years when 5.8 is out and 5.5 is EOLed? The approach for core language releases that's being used now probably calls for a different approach from distros and hosting providers than we've seen historically.
Re: GoPHP5: The Revenge Don Gilbert 11/11/13 2:48 PM
I had a great talk with Andrew Nacin from WordPress this weekend at the Joomla World Conference. (He came Saturday morning since Matt was keynoting on Sunday)

We reviewed a bunch of stats and talked about a lot of things, one of which was the GoPHP55 concept, and he said it's not for WordPress at this point due to usage stats (which I hope he'll be able to stop by and clarify.)
Re: GoPHP5: The Revenge Ryan McCue 11/11/13 4:06 PM
Nacin has been watching this thread, so fingers crossed he can drop in. :)

(See my previous post and http://wordpress.org/about/stats/ for the
usage stats)

Also, word back from hosts on my end:

* BlueHost is rolling out 5.4 en-masse over this month
* HostGator is working on data centre migrations for the next month or
two, but plans to try and hit 5.4 early next year
* GoDaddy should be getting 5.4 this week or so, then moving on to 5.5.
After that, the plan is to retire MySQL 4, then PHP 5.2; no timeframes
on those.
* DreamHost still have 5.2, but they're informing their customers that
they should upgrade to 5.3+. (I believe old versions of Joomla that
don't work on 5.3+ are a concern for them.)
* 1and1 have 5.2 as an option, but default to 5.4. No word on rolling
out past that, at least that I've heard.

There are a couple of other hosts I've only heard murmurs from, and I'm
trying to contact them further.

--
Ryan McCue
<http://ryanmccue.info/>
Re: GoPHP5: The Revenge Ian Thomas 11/12/13 12:38 PM
Although Dreamhost support PHP 5.2, they recommed PHP 5.3 and support 5.4.

They also support older versions of PHP if you've got domains that are still running on them (I had one running PHP4 until recently).

Regards,
Ian Thomas
Re: GoPHP5: The Revenge William DURAND 1/23/14 7:49 AM
Obviously, the world is not ready for such a plan: https://github.com/willdurand/StackNegotiation/pull/4.

My idea was to set the minimum PHP version to 5.4, not because I am using 5.4 features, but rather to *invite* people to upgrade to 5.4 and above. Is that bad?

William

Le mardi 12 novembre 2013 21:38:58 UTC+1, Ian Thomas a écrit :
Re: GoPHP5: The Revenge Marco Pivetta 1/23/14 7:50 AM

On 23 January 2014 16:49, William DURAND <william...@gmail.com> wrote:
My idea was to set the minimum PHP version to 5.4, not because I am using 5.4 features, but rather to *invite* people to upgrade to 5.4 and above. Is that bad?

Nope, it's a good thing :-)

As already discussed with other folks in the past, if people stick with older versions "because stability", they will get the "stability" of no new features ;-)
Re: GoPHP5: The Revenge Hari K T 1/23/14 8:05 AM
Hey William DURAND,

I noticed the repo issue. I feel you are wrong in it. If you are not using any of the 5.4 features then there is no need of forcing people.

What is the gain you are seeing ?

You will loose most of the people who are looking for a 5.3 one :)

My 2 cents.

Hari K T

You can ring me : +91 9388 75 8821

Skype  : kthari85
Twitter : harikt


--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/CADyq6s%2BCZf18R5j5YyUXotjr49tLrWpFBYMr-sGryV%2BzhaoCPQ%40mail.gmail.com.

For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge William DURAND 1/23/14 8:17 AM
As I said in the GitHub comments, pretty much every new 5.4/5.5 features are available in 5.3, either as a library (password_compat for the Password Hashing API), using a similar syntax (array() vs []), or even "hacks".

Thinking that way won't change anything, and we will keep 5.3 stuff forever. According to the PHP 5.3 EOL RFC, in March, 5.3 will stop receiving security patches.

Here is the gain, preventing security issues, and preparing the future.

William
Re: GoPHP5: The Revenge Sebastian Krebs 1/23/14 8:19 AM



2014/1/23 Hari K T <ktha...@gmail.com>

Hey William DURAND,

I noticed the repo issue. I feel you are wrong in it. If you are not using any of the 5.4 features then there is no need of forcing people.

Nobody was ever forced to use this package, so nobody was ever forced to update their PHP setup. That's exactly, what annoyed me: Some _wants_ to use the package, but werent willing (or whatever) to update their environment. Let aside the reasons, why they can't upgrade, but I don't think the right solution is to _downgrade_ the package because of this. I understand people, when they blame the maintainer, that he breaks their setup, when he changes the dependencies, but that didn't happened here.

Interestingly I've never seen such a discussion about Desktop-Software. Or about Games and GPUs. ;)
 

What is the gain you are seeing ?

The _possibility_ to use 5.4-features. Ever tried to upgrade such a dependency _after_ release? Once you decide for a version you are stuck with it for a while, because else you will probably break existing setups. I've seen this in PR-comments, when Symfony2-bundles dropped their 2.1-support (and that isn't even an interpreter)

 

You will loose most of the people who are looking for a 5.3 one :)

You can't loose, what you never had: This package was never 5.3-compatible.
 

My 2 cents.



On Thu, Jan 23, 2014 at 9:20 PM, Marco Pivetta <ocra...@gmail.com> wrote:

On 23 January 2014 16:49, William DURAND <william...@gmail.com> wrote:
My idea was to set the minimum PHP version to 5.4, not because I am using 5.4 features, but rather to *invite* people to upgrade to 5.4 and above. Is that bad?

Nope, it's a good thing :-)

As already discussed with other folks in the past, if people stick with older versions "because stability", they will get the "stability" of no new features ;-)

--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/CADyq6s%2BCZf18R5j5YyUXotjr49tLrWpFBYMr-sGryV%2BzhaoCPQ%40mail.gmail.com.

For more options, visit https://groups.google.com/groups/opt_out.

--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/CAESZFtJGcSyDjEu%3DLvyB2dcOyWVORVkXFY71aEs9U9yu%3DEW8tQ%40mail.gmail.com.

For more options, visit https://groups.google.com/groups/opt_out.



--
github.com/KingCrunch
Re: GoPHP5: The Revenge Marc Seiler 1/23/14 8:25 AM
Just my 2 cents and I'm just a lurker here but a long time PHP developer. PHP has been plagued with to much thinking about backwards compatibility (thats why we still have register_globals :P). I say push forward and bump to 5.4. I mean really the EOL of 5.3 is just around the corner its about time for people to upgrade if they want newer features. Don't slow progress the progress of a project to satisfy backwards compatibility for too long...

Thank you,
    Marc Seiler
    863-354-0252



Re: GoPHP5: The Revenge Evert Pot 1/23/14 8:58 AM

On 14-23-1, 11:25 AM, Marc Seiler wrote:
> Just my 2 cents and I'm just a lurker here but a long time PHP
> developer. PHP has been plagued with to much thinking about backwards
> compatibility (thats why we still have register_globals :P).

Funny that you use that example, because we actually no longer have
register_globals in php.

Evert
Re: GoPHP5: The Revenge Marc Seiler 1/23/14 9:09 AM
Yes it was removed in 5.4 :P BUT it was found to be such a bad idea that the default changed to OFF in 4.2 way back in 2002 so lets try not to follow that precedence :D

Thank you,
    Marc Seiler
    863-354-0252




--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/52E14A51.9030308%40gmail.com.

For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge Hari K T 1/23/14 4:55 PM
Hey,
 
What is the gain you are seeing ?

The _possibility_ to use 5.4-features.

In aura, we have always tried to push people to 5.4. But some of the packages can be 5.3 compatible for it only uses short array. What happened is people started forking and keeping them. And many of them asking for the same.

So for v2, we decided to push 5.3 if it can work on 5.3 ( if we don't really need all the traits and stuffs like that which are in 5.4 )
 
Ever tried to upgrade such a dependency _after_ release? Once you decide for a version you are stuck with it for a while, because else you will probably break existing setups. I've seen this in PR-comments, when Symfony2-bundles dropped their 2.1-support (and that isn't even an interpreter)

I haven't used any symfony bundles. So I don't know. And in composer doesn't exactly pointing to the version works ?

I have some where read, symfony is following http://semver.org/ .
Re: GoPHP5: The Revenge Pádraic Brady 1/24/14 4:30 AM
Hi,

On 24 January 2014 00:55, Hari K T <ktha...@gmail.com> wrote:
Hey,
 
What is the gain you are seeing ?

The _possibility_ to use 5.4-features.

In aura, we have always tried to push people to 5.4. But some of the packages can be 5.3 compatible for it only uses short array. What happened is people started forking and keeping them. And many of them asking for the same.

So for v2, we decided to push 5.3 if it can work on 5.3 ( if we don't really need all the traits and stuffs like that which are in 5.4 )

There's two factors here that we all, as programmers, are going to be looking at:

1. We'd like to have the possibility of using new features.
2. PHP 5.3 will very shortly be a dead PHP version.

I get that users, hosts and even distributions may cling to it, but dead is dead. It will be slower, have less features, be a security headache (unless custom patched obviously), and rule users out of employing new PHP 5.4 library versions. Where exactly is the upside?

PHP 5.3 is a corpse.

This also reminds me of https://github.com/padraic/mockery/issues/237 (and the network of resulting PRs and calls on Dave’s time). While I do appreciate the HHVM team taking time to fix HHVM, I get a wee bit uncomfortable when it also involves changing Mockery itself (using its developers' free time). Is Facebook benchmarking HHVM’s progress in reaching PHP parity, or the projects’ progress in reaching parity with HHVM?

 

In any case, my point is that I’m worried about polluting PHP. With PHP 5.6 alpha1, we now have 5 different PHPs to test on. Next there will be 6. Then 7. What is the acceptable gap between the PHP users demand and the PHP that actually exists for us to use? 8?

 

So, PHP 5.3 needs to die. We’re not doing users any favour in constantly enabling their bad habits and we’re certainly not doing ourselves any either.


My long winded 2c :P

Paddy

--
Pádraic Brady

http://blog.astrumfutura.com
http://www.survivethedeepend.com
Zend Framework Community Review Team
Zend Framework PHP-FIG Representative
Re: GoPHP5: The Revenge Phil Sturgeon 1/24/14 7:24 AM
I follow the "3 minor versions" rule. I currently still support PHP 5.3 for PyroCMS and by the time PHP 5.6 is out we'll be onto PHP 5.4. 

Same for the League of Extraordinary Packages. Not much point me making packages that can't be used in projects like PyroCMS. :)

This next month or two will see a few of those packages making a new major version support PHP 5.4+ only though. My only concern for projects I am involved in is alienating a large percentage of the community. We can take the tough love approach, but that has always had more downsides than softly softly (for me). YMMV.
Re: GoPHP5: The Revenge William DURAND 1/24/14 7:39 AM
I don't think it helps the userland to keep supporting dead PHP versions.

--
William Durand | http://www.williamdurand.fr


--
You received this message because you are subscribed to the Google Groups "PHP Framework Interoperability Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to php-fig+u...@googlegroups.com.
To post to this group, send email to php...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/php-fig/a32e1f7c-7374-423b-8e27-b0c224850e5a%40googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

Re: GoPHP5: The Revenge Phil Sturgeon 1/24/14 12:28 PM


On Friday, 24 January 2014 10:39:39 UTC-5, William DURAND wrote:
I don't think it helps the userland to keep supporting dead PHP versions.

--
William Durand | http://www.williamdurand.fr

It's a business decision, not a "whats best for PHP call", which is what any business has to do. 

I'm just saying I understand the hesitation for some, such as Aura. PyroCMS 2.3 comes out soon, which will be 5.3. PyroCMS 3.0 will be PHP 5.4, as it will be a few months out from now.

An extra few months will make a lot of difference, so anyone dropping a major version then - or after - should certainly be considering it. 
Re: GoPHP5: The Revenge Adrian Cardenas 1/25/14 7:25 AM
I can't speak officially for ServerGrove, but we have plans for offering support for 5.4 & 5.5 on shared hosting by, probably, mid-year. We already do for VPSs, which will offer 5.6 probably shortly after it's stable.


PHP's more aggressive release cycle means we need to move a bit faster
More topics »