Query: Should ZF2 bump the minimum required PHP version?

classic Classic list List threaded Threaded
27 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Query: Should ZF2 bump the minimum required PHP version?

weierophinney
Administrator
Hey, all --

We're running into a few situations where bumping the minimum required
PHP version for ZF2 would be useful; in fact, there's at least one
situation where staying with 5.3.3 actually prevents progress on a few
issues.

The specific issue we have is that, until 5.3.9, PHP did not allow the
following:

    interface Foo
    {
        public function send();
    }

    interface Bar
    {
        public function send();
    }

    class FooBar implements Foo, Bar
    {
        public function send()
        {
            // do something
        }
    }

Essentially, implementing multiple interfaces that define the same
method, using the same signature.
Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.

Having this would allow us to fix a situation with the way translation
works across components; not having it means we're stuck with some of
those problems.

There are other issues as well: ArrayObject has had a lively history
of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
in the object model as well that have been corrected starting in 5.3.7
and up.

Considering PHP 5.3 has already reached end of life status
(http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
version seems like "a good idea."

My question, then is:

- Should we up the minimum required PHP version for ZF2?
- If your answer was "no", why not?
- If your answer was "yes", what version should become the next
minimum supported PHP version? Why?

NOTE: we are not announcing that we will up the minimum required
version at this time; I'm soliciting feedback so we can make a
decision.

Thanks in advance!

--
Matthew Weier O'Phinney
Project Lead            | [hidden email]
Zend Framework          | http://framework.zend.com/
PGP key: http://framework.zend.com/zf-matthew-pgp-key.asc
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Stewart Lord
My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3
and we have customers on that distro. This would mean we are stuck on
older ZF2 at least for a while. Not a huge deal really, but that's my
vote FWIW :)

I also note that Symfony2 is at 5.3.3.

Stew


On 2013-11-05 11:34 AM, Matthew Weier O'Phinney wrote:

> Hey, all --
>
> We're running into a few situations where bumping the minimum required
> PHP version for ZF2 would be useful; in fact, there's at least one
> situation where staying with 5.3.3 actually prevents progress on a few
> issues.
>
> The specific issue we have is that, until 5.3.9, PHP did not allow the
> following:
>
>      interface Foo
>      {
>          public function send();
>      }
>
>      interface Bar
>      {
>          public function send();
>      }
>
>      class FooBar implements Foo, Bar
>      {
>          public function send()
>          {
>              // do something
>          }
>      }
>
> Essentially, implementing multiple interfaces that define the same
> method, using the same signature.
> Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.
>
> Having this would allow us to fix a situation with the way translation
> works across components; not having it means we're stuck with some of
> those problems.
>
> There are other issues as well: ArrayObject has had a lively history
> of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
> in the object model as well that have been corrected starting in 5.3.7
> and up.
>
> Considering PHP 5.3 has already reached end of life status
> (http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
> version seems like "a good idea."
>
> My question, then is:
>
> - Should we up the minimum required PHP version for ZF2?
> - If your answer was "no", why not?
> - If your answer was "yes", what version should become the next
> minimum supported PHP version? Why?
>
> NOTE: we are not announcing that we will up the minimum required
> version at this time; I'm soliciting feedback so we can make a
> decision.
>
> Thanks in advance!
>

Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Artur Bodera
Voting no.

As much as I'd love to bump it even higher (*ekhem* borked DateTime *ekhem*) it will probably confuse the hell out of user base. We've had this discussion several times on different occasions (on ml, irc, even some stuff on wiki). 

We'll have to live with 5.3.3 until ZF3. With ZF3 we'll have to live with 5.4.X (which will probably be aligned with most popular revision per linux distros) and we'll probably have similar discussions in the future ... and for the sake of consistency, my answer will then also probably be "no".

With love :-)

Art.



On Tue, Nov 5, 2013 at 8:55 PM, Stewart Lord <[hidden email]> wrote:
My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3 and we have customers on that distro. This would mean we are stuck on older ZF2 at least for a while. Not a huge deal really, but that's my vote FWIW :)

I also note that Symfony2 is at 5.3.3.

Stew



On 2013-11-05 11:34 AM, Matthew Weier O'Phinney wrote:
Hey, all --

We're running into a few situations where bumping the minimum required
PHP version for ZF2 would be useful; in fact, there's at least one
situation where staying with 5.3.3 actually prevents progress on a few
issues.

The specific issue we have is that, until 5.3.9, PHP did not allow the
following:

     interface Foo
     {
         public function send();
     }

     interface Bar
     {
         public function send();
     }

     class FooBar implements Foo, Bar
     {
         public function send()
         {
             // do something
         }
     }

Essentially, implementing multiple interfaces that define the same
method, using the same signature.
Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.

Having this would allow us to fix a situation with the way translation
works across components; not having it means we're stuck with some of
those problems.

There are other issues as well: ArrayObject has had a lively history
of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
in the object model as well that have been corrected starting in 5.3.7
and up.

Considering PHP 5.3 has already reached end of life status
(http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
version seems like "a good idea."

My question, then is:

- Should we up the minimum required PHP version for ZF2?
- If your answer was "no", why not?
- If your answer was "yes", what version should become the next
minimum supported PHP version? Why?

NOTE: we are not announcing that we will up the minimum required
version at this time; I'm soliciting feedback so we can make a
decision.

Thanks in advance!



Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

richard
My vote is for the PHP version the majority of stable / LTS OS's support by default.

Regards


On Tue, Nov 5, 2013 at 8:03 PM, Artur Bodera <[hidden email]> wrote:
Voting no.

As much as I'd love to bump it even higher (*ekhem* borked DateTime *ekhem*) it will probably confuse the hell out of user base. We've had this discussion several times on different occasions (on ml, irc, even some stuff on wiki). 

We'll have to live with 5.3.3 until ZF3. With ZF3 we'll have to live with 5.4.X (which will probably be aligned with most popular revision per linux distros) and we'll probably have similar discussions in the future ... and for the sake of consistency, my answer will then also probably be "no".

With love :-)

Art.



On Tue, Nov 5, 2013 at 8:55 PM, Stewart Lord <[hidden email]> wrote:
My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3 and we have customers on that distro. This would mean we are stuck on older ZF2 at least for a while. Not a huge deal really, but that's my vote FWIW :)

I also note that Symfony2 is at 5.3.3.

Stew



On 2013-11-05 11:34 AM, Matthew Weier O'Phinney wrote:
Hey, all --

We're running into a few situations where bumping the minimum required
PHP version for ZF2 would be useful; in fact, there's at least one
situation where staying with 5.3.3 actually prevents progress on a few
issues.

The specific issue we have is that, until 5.3.9, PHP did not allow the
following:

     interface Foo
     {
         public function send();
     }

     interface Bar
     {
         public function send();
     }

     class FooBar implements Foo, Bar
     {
         public function send()
         {
             // do something
         }
     }

Essentially, implementing multiple interfaces that define the same
method, using the same signature.
Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.

Having this would allow us to fix a situation with the way translation
works across components; not having it means we're stuck with some of
those problems.

There are other issues as well: ArrayObject has had a lively history
of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
in the object model as well that have been corrected starting in 5.3.7
and up.

Considering PHP 5.3 has already reached end of life status
(http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
version seems like "a good idea."

My question, then is:

- Should we up the minimum required PHP version for ZF2?
- If your answer was "no", why not?
- If your answer was "yes", what version should become the next
minimum supported PHP version? Why?

NOTE: we are not announcing that we will up the minimum required
version at this time; I'm soliciting feedback so we can make a
decision.

Thanks in advance!




Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Marco Pivetta
This post has NOT been accepted by the mailing list yet.
In reply to this post by Stewart Lord
I'm actually +1 to bumping the requirements in the next minor version of ZF.
5.3.* is perfectly ok, even for older LTS releases.
I don't see a reason for not upgrading the lesser minor version of PHP, and am actually wondering why it's being freezed blindly like that, given that php-src itself is very (very!) careful about BC breaks.

The bug you presented in your question is also quite annoying to "handle" when testing against 5.3.3.

What I'm wondering is if the features that you want to implement couldn't instead slip into ZF3 - is there any detail on what we're talking about exactly?


On 5 November 2013 20:55, Stewart Lord [via Zend Framework Community] <[hidden email]> wrote:
My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3
and we have customers on that distro. This would mean we are stuck on
older ZF2 at least for a while. Not a huge deal really, but that's my
vote FWIW :)

I also note that Symfony2 is at 5.3.3.

Stew

Symfony is not at 5.3.3. There's bugs that are marked as "won't fix" when dealing with 5.3.3.
 
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Ben Scholzen 'DASPRiD'
In reply to this post by Stewart Lord
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

EvanDotPro
In reply to this post by weierophinney
First, I agree with the RHEL / LTS distro arguments, and can see that having a negative effect on these companies. My gut reaction is to say no, we should respect the version we committed to originally, despite the cost (hey, we can't use traits either... too bad). I can personally confirm that many companies specifically choose Zend Framework because of the reliability on these types of matters (more-so backwards compatibility breaks, but PHP version support plays a large role as well).

That said, due to the ZF policy of backporting security fixes to previous minor version releases, which matches that of the distributions being cited here, I don't actually see bumping the PHP version requirement as a problem. These enterprise users will simply have to be feature-frozen and only receive minor version updates when a security fix is backported, period. Missing out on new features and (non-critical) bugfixes after such a change is a cost they have already accepted as part of their policy to limit themselves strictly to distribution-provided packages.

So, ultimately, I can't say I'd actually be upset either way. All of my clients who are using enterprise / LTS distributions will still be fine to use all of the ZF2 features they've come to depend on up until this point, and we'll still be able to update their ZF2 minor version when/if future security fixes are released. If such a change meant that enterprise / LTS distro users would be unable to get critical security fixes, then I'd strongly be -1 to this, but that is simply not the case here.


--
Evan Coury, ZCE


On Tue, Nov 5, 2013 at 12:34 PM, Matthew Weier O'Phinney <[hidden email]> wrote:
Hey, all --

We're running into a few situations where bumping the minimum required
PHP version for ZF2 would be useful; in fact, there's at least one
situation where staying with 5.3.3 actually prevents progress on a few
issues.

The specific issue we have is that, until 5.3.9, PHP did not allow the
following:

    interface Foo
    {
        public function send();
    }

    interface Bar
    {
        public function send();
    }

    class FooBar implements Foo, Bar
    {
        public function send()
        {
            // do something
        }
    }

Essentially, implementing multiple interfaces that define the same
method, using the same signature.
Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.

Having this would allow us to fix a situation with the way translation
works across components; not having it means we're stuck with some of
those problems.

There are other issues as well: ArrayObject has had a lively history
of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
in the object model as well that have been corrected starting in 5.3.7
and up.

Considering PHP 5.3 has already reached end of life status
(http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
version seems like "a good idea."

My question, then is:

- Should we up the minimum required PHP version for ZF2?
- If your answer was "no", why not?
- If your answer was "yes", what version should become the next
minimum supported PHP version? Why?

NOTE: we are not announcing that we will up the minimum required
version at this time; I'm soliciting feedback so we can make a
decision.

Thanks in advance!

--
Matthew Weier O'Phinney
Project Lead            | [hidden email]
Zend Framework          | http://framework.zend.com/
PGP key: http://framework.zend.com/zf-matthew-pgp-key.asc

Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Anthony Shireman
In reply to this post by Ben Scholzen 'DASPRiD'
In terms of ZF deployments, it would be great to see this updated again along with some additional information.
http://www.sasaprolic.com/2013/02/list-of-current-php-version-in-major.html

While those are the distros on the list, what is the breakdown for actual ZF deployments? If 20% of deployments are on 5.3.3 and 70% of deployments are on 5.3.9 and the remaining 10% are on 5.3.27, then that would be a compelling case to bump the minimum to hit the 70%. I think determining that breakdown would be very important to making a good decision. If it's a close split where 50% are on 5.3.3 and 40% are on 5.3.9, that would be a tougher call.

With that in mind, here's some data on the PHP versions installed: (at least those systems which can be measured, there's probably some 'hidden' systems which don't show that they're running PHP, but, let's just go with what we have here)
http://w3techs.com/technologies/details/pl-php/5.3/all
http://w3techs.com/technologies/details/pl-php/5.4/all

According to that only 7% of all websites running PHP are using 5.4 (bummer) and of the 41.2% of all PHP websites using 5.3, 21% are specifically using 5.3.3...but that means that 79% of sites running 5.3 are using a version *higher than that*. If we still wanted to ensure that a solid majority of those websites would be fine, then bumping up to 5.3.10 captures the 70% that run at that version or higher. So while 5.3.3 is used on the highest percentage of 5.3 websites, overall it seems that actual deployments are at a version greater than that. I also realize that this is just PHP versions and not specifically ZF deployments which may be an important factor as mentioned above.

I don't think that this data tells us everything that we would want to know, but I think it's interesting to see that lower versions of PHP (with regards to 5.3, go back and look at the data for 5.2...scary) are on the way out. This can allow room to start pushing for bumps in requirements.

Data to chew on.

Regards,



Tony



On Tue, Nov 5, 2013 at 1:25 PM, Ben Scholzen 'DASPRiD' <[hidden email]> wrote:
That's not completely true thought:

http://developerblog.redhat.com/2013/08/01/php-5-4-on-rhel-6-using-rhscl/

On 05.11.2013 20:55, Stewart Lord wrote:
> My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3
> and we have customers on that distro. This would mean we are stuck on
> older ZF2 at least for a while. Not a huge deal really, but that's my
> vote FWIW :)
>
> I also note that Symfony2 is at 5.3.3.
>
> Stew
>
>
> On 2013-11-05 11:34 AM, Matthew Weier O'Phinney wrote:
>> Hey, all --
>>
>> We're running into a few situations where bumping the minimum required
>> PHP version for ZF2 would be useful; in fact, there's at least one
>> situation where staying with 5.3.3 actually prevents progress on a few
>> issues.
>>
>> The specific issue we have is that, until 5.3.9, PHP did not allow the
>> following:
>>
>>      interface Foo
>>      {
>>          public function send();
>>      }
>>
>>      interface Bar
>>      {
>>          public function send();
>>      }
>>
>>      class FooBar implements Foo, Bar
>>      {
>>          public function send()
>>          {
>>              // do something
>>          }
>>      }
>>
>> Essentially, implementing multiple interfaces that define the same
>> method, using the same signature.
>> Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.
>>
>> Having this would allow us to fix a situation with the way translation
>> works across components; not having it means we're stuck with some of
>> those problems.
>>
>> There are other issues as well: ArrayObject has had a lively history
>> of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
>> in the object model as well that have been corrected starting in 5.3.7
>> and up.
>>
>> Considering PHP 5.3 has already reached end of life status
>> (http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
>> version seems like "a good idea."
>>
>> My question, then is:
>>
>> - Should we up the minimum required PHP version for ZF2?
>> - If your answer was "no", why not?
>> - If your answer was "yes", what version should become the next
>> minimum supported PHP version? Why?
>>
>> NOTE: we are not announcing that we will up the minimum required
>> version at this time; I'm soliciting feedback so we can make a
>> decision.
>>
>> Thanks in advance!
>>
>


--
Ben Scholzen 'DASPRiD'
Community Review Team Member | [hidden email]
Zend Framework               | http://www.dasprids.de

Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

EvanDotPro

On Tue, Nov 5, 2013 at 3:23 PM, Anthony Shireman <[hidden email]> wrote:
In terms of ZF deployments, it would be great to see this updated again along with some additional information.
http://www.sasaprolic.com/2013/02/list-of-current-php-version-in-major.html

While those are the distros on the list, what is the breakdown for actual ZF deployments? If 20% of deployments are on 5.3.3 and 70% of deployments are on 5.3.9 and the remaining 10% are on 5.3.27, then that would be a compelling case to bump the minimum to hit the 70%. I think determining that breakdown would be very important to making a good decision. If it's a close split where 50% are on 5.3.3 and 40% are on 5.3.9, that would be a tougher call.

With that in mind, here's some data on the PHP versions installed: (at least those systems which can be measured, there's probably some 'hidden' systems which don't show that they're running PHP, but, let's just go with what we have here)
http://w3techs.com/technologies/details/pl-php/5.3/all
http://w3techs.com/technologies/details/pl-php/5.4/all

According to that only 7% of all websites running PHP are using 5.4 (bummer) and of the 41.2% of all PHP websites using 5.3, 21% are specifically using 5.3.3...but that means that 79% of sites running 5.3 are using a version *higher than that*. If we still wanted to ensure that a solid majority of those websites would be fine, then bumping up to 5.3.10 captures the 70% that run at that version or higher. So while 5.3.3 is used on the highest percentage of 5.3 websites, overall it seems that actual deployments are at a version greater than that. I also realize that this is just PHP versions and not specifically ZF deployments which may be an important factor as mentioned above.

I don't think that this data tells us everything that we would want to know, but I think it's interesting to see that lower versions of PHP (with regards to 5.3, go back and look at the data for 5.2...scary) are on the way out. This can allow room to start pushing for bumps in requirements.

I agree that this data does not tell us everything, so I just want to follow up with a disclaimer when looking at data like this in our current context:

As tempting as it can be to cite these statistics, you need to keep in mind that even a huge install-base that we'd indubitably want to continue to support in all ways possible (let's even say multiple thousands of large, enterprise ZF2 installations) would likely not even register as a fraction of one percent on any of these statistics — thus, the fact that 79% of PHP 5.3 installs are > 5.3.3 is most likely largely irrelevant to our decision in this matter. (Warning: Do not try to apply this same logic when talking about IE 6.)

--
Evan Coury
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

weierophinney
Administrator
In reply to this post by Artur Bodera
On Tue, Nov 5, 2013 at 2:03 PM, Artur Bodera <[hidden email]> wrote:

> Voting no.
>
> As much as I'd love to bump it even higher (*ekhem* borked DateTime *ekhem*)
> it will probably confuse the hell out of user base. We've had this
> discussion several times on different occasions (on ml, irc, even some stuff
> on wiki).
>
> We'll have to live with 5.3.3 until ZF3. With ZF3 we'll have to live with
> 5.4.X (which will probably be aligned with most popular revision per linux
> distros) and we'll probably have similar discussions in the future ... and
> for the sake of consistency, my answer will then also probably be "no".

Actually, you're mistaken.

We bumped the minimum required version of ZF1 to 5.2.4 (from 5.1.6)
for 1.7. We bumped it again to 5.2.10 or 5.2.12 for either the 1.11 or
1.12 release.

In other words, we HAVE allowed bumps to the PHP version as long as
they happen with a minor release or greater of ZF. That is what I'm
proposing here - a bump with the 2.3.0 release of ZF.



> On Tue, Nov 5, 2013 at 8:55 PM, Stewart Lord <[hidden email]> wrote:
>>
>> My vote is 'no' because (as far as I know) RHEL 6 is currently at 5.3.3
>> and we have customers on that distro. This would mean we are stuck on older
>> ZF2 at least for a while. Not a huge deal really, but that's my vote FWIW :)
>>
>> I also note that Symfony2 is at 5.3.3.
>>
>> Stew
>>
>>
>>
>> On 2013-11-05 11:34 AM, Matthew Weier O'Phinney wrote:
>>>
>>> Hey, all --
>>>
>>> We're running into a few situations where bumping the minimum required
>>> PHP version for ZF2 would be useful; in fact, there's at least one
>>> situation where staying with 5.3.3 actually prevents progress on a few
>>> issues.
>>>
>>> The specific issue we have is that, until 5.3.9, PHP did not allow the
>>> following:
>>>
>>>      interface Foo
>>>      {
>>>          public function send();
>>>      }
>>>
>>>      interface Bar
>>>      {
>>>          public function send();
>>>      }
>>>
>>>      class FooBar implements Foo, Bar
>>>      {
>>>          public function send()
>>>          {
>>>              // do something
>>>          }
>>>      }
>>>
>>> Essentially, implementing multiple interfaces that define the same
>>> method, using the same signature.
>>> Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.
>>>
>>> Having this would allow us to fix a situation with the way translation
>>> works across components; not having it means we're stuck with some of
>>> those problems.
>>>
>>> There are other issues as well: ArrayObject has had a lively history
>>> of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
>>> in the object model as well that have been corrected starting in 5.3.7
>>> and up.
>>>
>>> Considering PHP 5.3 has already reached end of life status
>>> (http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
>>> version seems like "a good idea."
>>>
>>> My question, then is:
>>>
>>> - Should we up the minimum required PHP version for ZF2?
>>> - If your answer was "no", why not?
>>> - If your answer was "yes", what version should become the next
>>> minimum supported PHP version? Why?
>>>
>>> NOTE: we are not announcing that we will up the minimum required
>>> version at this time; I'm soliciting feedback so we can make a
>>> decision.
>>>
>>> Thanks in advance!
>>>
>>
>



--
Matthew Weier O'Phinney
Project Lead            | [hidden email]
Zend Framework          | http://framework.zend.com/
PGP key: http://framework.zend.com/zf-matthew-pgp-key.asc
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Artur Bodera
In reply to this post by EvanDotPro

On Tue, Nov 5, 2013 at 10:55 PM, Evan Coury <[hidden email]> wrote:
These enterprise users will simply have to be feature-frozen and only receive minor version updates when a security fix is backported, period. Missing out on new features and (non-critical) bugfixes after such a change is a cost they have already accepted as part of their policy to limit themselves strictly to distribution-provided packages.

I think it's the gist of the problem. 

There's also quite lot of assumptions, which basically boils down to: "if they use old PHP versions, it's their problem and they must know and acknowledge all downsides, accept all the risks".  Well, it's actually a logical conclusion but the question is: how much do they rely on frameworks such as ZF2 to "shim away" their problems ?

If we went this way, on our side, this means very careful PR merging process so backports do receive as many of the upgrades as possible. Not sure how much pain that would be actually :-| @mwop ?

will still be fine to use all of the ZF2 features they've come to depend on up until this point

Yes, but those features will slowly diminish with time, as only security stuff will be backported. So basically our decision here means: "no more feature upgrades and bug fixes for you from now on unless you upgrade PHP".

Reply | Threaded
Open this post in threaded view
|

Re: [fw-general] Re: [zf-contributors] Query: Should ZF2 bump the minimum required PHP version?

Marco Pivetta

On 6 November 2013 00:22, Artur Bodera <[hidden email]> wrote:
On Tue, Nov 5, 2013 at 10:55 PM, Evan Coury <[hidden email]> wrote:

> These enterprise users will simply have to be feature-frozen and only
> receive minor version updates when a security fix is backported, period.
> Missing out on new features and (non-critical) bugfixes *after* such a
> change is a cost they have already accepted as part of their policy to
> limit themselves strictly to distribution-provided packages.
>

I think it's the gist of the problem.

There's also quite lot of assumptions, which basically boils down to: "if
they use old PHP versions, it's their problem and they must know and
acknowledge all downsides, accept all the risks".  Well, it's actually a
logical conclusion but the question is: how much do they rely on frameworks
such as ZF2 to "shim away" their problems ?

I don't think ZF, SF or whatever framework you pick should become "the jQuery of PHP". A framework is not an excuse to use older PHP versions without worrying.
 

If we went this way, on our side, this means very careful PR merging
process so backports do receive as many of the upgrades as possible. Not
sure how much pain that would be actually :-| @mwop ?

Yes, that would be very painful. Instead of that solution, I' rather suggest keeping 5.3.3 in ZF2 and delaying whatever is causing the failures for ZF3 (where the bump is possible).
 

> will still be fine to use all of the ZF2 features they've come to depend
on up until this point

Yes, but those features will slowly diminish with time, as only security
stuff will be backported. So basically our decision here means: "no more
feature upgrades and bug fixes for you from now on unless you upgrade PHP".

That's kind-of the point of upgrading generally...
 


--
[hidden email]
<a href="tel:%2B48%20695%20600%20936" value="+48695600936">+48 695 600 936
http://thinkscape.pro


Marco Pivetta

http://twitter.com/Ocramius     

http://ocramius.github.com/
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Anthony Shireman
In reply to this post by EvanDotPro


On Tue, Nov 5, 2013 at 2:42 PM, Evan Coury <[hidden email]> wrote:
<snip/>
I agree that this data does not tell us everything, so I just want to follow up with a disclaimer when looking at data like this in our current context:

As tempting as it can be to cite these statistics, you need to keep in mind that even a huge install-base that we'd indubitably want to continue to support in all ways possible (let's even say multiple thousands of large, enterprise ZF2 installations) would likely not even register as a fraction of one percent on any of these statistics — thus, the fact that 79% of PHP 5.3 installs are > 5.3.3 is most likely largely irrelevant to our decision in this matter. (Warning: Do not try to apply this same logic when talking about IE 6.)

--
Evan Coury

I did mention at the outset that it would be good to get this kind of data specifically for ZF deployments. But, unless there's a hidden reservoir of such data out there this is all I've got for the time being.

I also accept that this data only applies to PHP and not ZF2 as a specific case, and I agree completely that continuing support as much as possible is desirable. But if you think about it like IE6 then it all becomes... sorry, it was too tempting to not throw that in there.

So if there are thousands of large deployments, but millions of small ones that are pushing for new features and bug fixes, who "wins" in that kind of a situation? You mentioned that there is a sort of assumed risk/cost when you accept only distro packages, and as Artur also said that is the decision we're trying to make.

    -How do we balance the needs of the users who cannot/or will not upgrade, with those users who will continue to want fixes in the features they use even if it means updating outside of their distro?-

I think it would even be beneficial to turn it into a policy of sorts that says, "Every 6 months, a ZF release will likely (but may not) have a small PHP version bump. Sooner only if major issues are discovered." Then look at what the smallest PHP bump is that will fix some of the issues that have come up. I'd prefer a more data driven approach if such data can be found. Much past that we're only working on guesses...unless you have some secret information you're sworn not to divulge.

My vote is yes and for the minimum version to be upped to 5.3.10. Unless there's enough support to jump to 5.5.6... ≖_≖


Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

David Muir-2
In reply to this post by weierophinney
On 06/11/13 06:34, Matthew Weier O'Phinney wrote:

> Hey, all --
>
> We're running into a few situations where bumping the minimum required
> PHP version for ZF2 would be useful; in fact, there's at least one
> situation where staying with 5.3.3 actually prevents progress on a few
> issues.
>
> The specific issue we have is that, until 5.3.9, PHP did not allow the
> following:
>
>      interface Foo
>      {
>          public function send();
>      }
>
>      interface Bar
>      {
>          public function send();
>      }
>
>      class FooBar implements Foo, Bar
>      {
>          public function send()
>          {
>              // do something
>          }
>      }
>
> Essentially, implementing multiple interfaces that define the same
> method, using the same signature.
> Prior to 5.3.9, this raises an E_FATAL. From 5.3.9 forward, it works.
>
> Having this would allow us to fix a situation with the way translation
> works across components; not having it means we're stuck with some of
> those problems.
>
> There are other issues as well: ArrayObject has had a lively history
> of malfunctioning with 5.3 and 5.4, and there are  some odd behaviors
> in the object model as well that have been corrected starting in 5.3.7
> and up.
>
> Considering PHP 5.3 has already reached end of life status
> (http://php.net/archive/2013.php#id2013-07-11-1), upping the minimum
> version seems like "a good idea."
>

It's not quite EOL yet. There's still security fixes until July 2014.
That said, I don't mind the minimum version being bumped to 5.3.9.

Cheers,
David
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Lars Kneschke-2
Am 06.11.2013 23:46:17, schrieb David Muir:
Hello!

It's not quite EOL yet. There's still security fixes until July 2014.
That said, I don't mind the minimum version being bumped to 5.3.9.

Me too. 

--
​Lars Kneschke
Productmanager Tine 2.0
Metaways Infosystems GmbH
Pickhuben 2-4, 20457 Hamburg

Tel: +49 (0)40 31 70 31 - 521
Fax: +49 (0)40 31 70 31 - 921
Mobile: +49 (0)175 930 4324
Web: www.metaways.de

Metaways Infosystems GmbH - Sitz: D-22967 Tremsbüttel
Handelsregister: Amtsgericht Lübeck HRB 4508 AH
Geschäftsführung: Hermann Thaele, Lüder-H. Thaele

Sent with love from the new Tine 2.0 email client ...
Please visit http://www.tine20.com
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Stefano Torresi
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

EvanDotPro
Hello,

On Thu, Nov 7, 2013 at 1:57 AM, Stefano Torresi <[hidden email]> wrote:
FWIW I think a minor requirement bump shouldn't be that tragic. If one
is ok with using an older PHP version, he will be ok using an older
framework version. Afterall 2.2.* should be mature enough to satisfy
their requirements.

That said, backporting fixes is pretty much mandatory, and the
concerns raised about this are legit.
What would have to be backported? Security only or bug fixes in general?

This can easily be argued both ways, but I would say that for the sake of consistency with enterprise / LTS packaging standards and the sanity of the ZF team, it should be security fixes only.

--
Evan Coury
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Stefano Torresi
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

Oliver Koenig
This post has NOT been accepted by the mailing list yet.
In reply to this post by weierophinney
How come that in ZF 2.2.5 are several (at least 7 or so) trait declarations?
Being a newbe ZF2 user, I thought that the PHP requirements were raised to 5.4 now, and that I need a hosting platform that supports it, and asked the owner of the project that I'm working on to order a dedicated cloud server for it, because the shared hosting only has 5.3 on it. In hindsight I find it a bit embarrassing.

Greetings, Oliver
jah
Reply | Threaded
Open this post in threaded view
|

Re: Query: Should ZF2 bump the minimum required PHP version?

jah
This post has NOT been accepted by the mailing list yet.
This post was updated on .
In reply to this post by weierophinney
- Should we up the minimum required PHP version for ZF2?
- If your answer was "no", why not?
- If your answer was "yes", what version should become the next minimum supported PHP version? Why?

Yes for ZF2.3 at PHP5.3.9 but continue support for ZF2.2.x at PHP5.3.3
12