Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:97411
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain blastwave.org from 209.17.115.50 cause and error)
To: internals@lists.php.net
References: <cb4383a0-e8ce-160d-d754-4f359ac71d2f@gmx.de>
 <CAJW__o3pTZLh4Zop7FCFyEU1kqLgw2ze2Zg7ZTNo-9WMHNxa4w@mail.gmail.com>
 <ab56a62b-f521-449c-c6a5-169ad859b1e6@blastwave.org>
 <AF8CFFFD-DE51-4771-BDC7-BFD26614367C@heroku.com>
Message-ID: <99ca9c55-e56d-8fc3-5245-baa4d7e27e9e@blastwave.org>
Date: Wed, 14 Dec 2016 15:36:57 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Thunderbird/45.5.1
MIME-Version: 1.0
In-Reply-To: <AF8CFFFD-DE51-4771-BDC7-BFD26614367C@heroku.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] PHP 5.6 end of active support
From: dclarke@blastwave.org (Dennis Clarke)

On 12/14/2016 01:29 PM, David Zuelke wrote:
> On 14.12.2016, at 16:15, Dennis Clarke <dclarke@blastwave.org> wrote:
>>
>> On 12/14/2016 06:35 AM, Kalle Sommer Nielsen wrote:
>>> Hi
>>>
>>> On Dec 14, 2016 12:23, "Christoph M. Becker" <cmbecker69@gmx.de> wrote:
>>>>
>>>> Hi!
>>>>
>>>> The end of active support for PHP 5.6 is documented to be on December,
>>>> 31th[1].  Does that mean that there'll be no further release with
>>>> "normal" bug fixes (but only security fixes)?
>>>
>>> Yes, 5.6 was extended to compensate for 5->7 adoption afair from an rfc
>>>
>>
>> This is entirely too soon. At the present moment neither of the PHP 7.0
>> releases will compile clean on a strict POSIX environment. At all. The
>> version 5.6.x tree is perfectly stable and works out of the box without
>> an endless compile nightmare whereas 7.0.14 and 7.1.0 won't even
>> compile. I guess I need to file more bug reports
>
> Yes, you do. PHP 7.0.0 was released a year ago.
>
>> and push through this
>> or the 5.6.x version will be dropped with no valid replacement that
>
> 5.6 will receive security fixes for another 24 months.
>
>> works in a strict environment.  Perhaps the gcc compiler is an absolute
>> requirement and if that is true then the code isn't acceptable to any
>> other compiler regardless if it is C99 compliant or otherwise.
>
> It appears to be, from looking at https://bugs.php.net/bug.php?id=67229, but that affects PHP 5 as well, not just 7.
>

Actually looking at that bug we see the problem again. This is a
Solaris 10 system, which is a very strict POSIX system, but the
compiler being used is gcc and the option --std=gnu99 makes reference
to a "defacto standard" that exists no where.  So this is exactly the
problem I see in many many places, code gets releases that compiles
with gcc on some systems but will not pass even a basic compile on a
strict system with a very tightly conforming compiler such as the
Oracle Studio 12.5 tool set. Some code is amazingly clean like libgmp
and we also have PHP 5.6.x which compiles with some extensions but the
7.x codebase won't compile. At all.  Yet. Very close however.[1]

So I need to dig into this as I am sure whatever things I can uncover
and perhaps patch will be of benefit across all systems everywhere. I
think that is what ( now I am just being pedantic ) standards are all
about. Total portability.

Dennis


[1] discussion happened last month in "[PHP-DEV] C89 vs. C99" thread