Hi,
After doing the last evaluations in the RMs circle before going for 7.0.0
RTM preparations, we came to the conclusion that the current state does not
look reasonable to be packaged as the final release.
The RCs before RC6 was looking acceptable, which gave the reason to announce
the planned RTM date on Nov 12th as final. From the todays perspective, the
most of the issues discovered and fixed since RC6 are still minor by
themselves - 7 crashes (bug #70805 rather critical), and 3 functionality
regressions. However, given the amount and the combination of them
altogether, the state is in our opinion unsuitable for starting the life
cycle of the next major.
Thus, resuming the pre-release cycle seems in our opinion the right thing to
do ATM. Hence this writing to inform the community about the intention and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.
Regards
Anatol
Hi!
Thus, resuming the pre-release cycle seems in our opinion the right thing to
do ATM. Hence this writing to inform the community about the intention and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.
I think this is the right thing to do, we can wait for a bit if it means
better stability.
Do I understand right that the segfault fixes that are now marked as
7.0.1 would be instead released as RC7? And the prospective GA date
(provided RC7 is stable enough) would be now December 10th?
(BTW, by an interesting coincidence, December 10th is the birthday of
Ada Lovelace :).
Stas Malyshev
smalyshev@gmail.com
Hi Stas,
-----Original Message-----
From: Stanislav Malyshev [mailto:smalyshev@gmail.com]
Sent: Sunday, November 8, 2015 8:46 PM
To: Anatol Belski anatol.php@belski.net; internals@lists.php.net
Cc: 'Kalle Sommer Nielsen' kalle@php.net; 'Ferenc Kovacs'
tyrael@php.net
Subject: Re: [PHP-DEV] PHP 7 RTM dateHi!
Thus, resuming the pre-release cycle seems in our opinion the right
thing to do ATM. Hence this writing to inform the community about the
intention and to gather the opinions. PHP 7 is still going down the
home stretch and is very close to the finish, the next RC can be well
the last. Though we'd better ensuring the quality as many times as
needed before, instead of casually disservicing the consumers.I think this is the right thing to do, we can wait for a bit if it means
better
stability.Do I understand right that the segfault fixes that are now marked as
7.0.1 would be instead released as RC7? And the prospective GA date
(provided
RC7 is stable enough) would be now December 10th?
(BTW, by an interesting coincidence, December 10th is the birthday of Ada
Lovelace :).
Exactly, the relevant patches are going into RC7 which is then out on 12th
instead of GA. The next closest release slot were then Nov 26th, either for
GA or for RC8 (but better not try to predict). If it stands good, we might
have no reason to wait until the Mrs. Lovelace birthday party :)
Regards
Anatol
Exactly, the relevant patches are going into RC7 which is then out on 12th
instead of GA. The next closest release slot were then Nov 26th, either for
GA or for RC8 (but better not try to predict). If it stands good, we might
have no reason to wait until the Mrs. Lovelace birthday party :)
We should try really hard to get it out on the 26th. Once we get into
December people start to disappear. Nov.26 is American Thanksgiving
though, but I don't think that should interfere.
-Rasmus
If nothing else, it will give us an excuse to avoid the Black Friday
madness ("Sorry hon, I need to update the server!").
Scott Arciszewski
Chief Development Officer
Paragon Initiative Enterprises https://paragonie.com
Exactly, the relevant patches are going into RC7 which is then out on
12th
instead of GA. The next closest release slot were then Nov 26th, either
for
GA or for RC8 (but better not try to predict). If it stands good, we
might
have no reason to wait until the Mrs. Lovelace birthday party :)We should try really hard to get it out on the 26th. Once we get into
December people start to disappear. Nov.26 is American Thanksgiving
though, but I don't think that should interfere.-Rasmus
Hi Rasmus,
-----Original Message-----
From: Rasmus Lerdorf [mailto:rasmus@lerdorf.com]
Sent: Monday, November 9, 2015 2:36 AM
To: Anatol Belski anatol.php@belski.net; 'Stanislav Malyshev'
smalyshev@gmail.com; internals@lists.php.net
Cc: 'Kalle Sommer Nielsen' kalle@php.net; 'Ferenc Kovacs'
tyrael@php.net
Subject: Re: [PHP-DEV] PHP 7 RTM dateExactly, the relevant patches are going into RC7 which is then out on
12th instead of GA. The next closest release slot were then Nov 26th,
either for GA or for RC8 (but better not try to predict). If it stands
good, we might have no reason to wait until the Mrs. Lovelace birthday
party :)We should try really hard to get it out on the 26th. Once we get into
December
people start to disappear. Nov.26 is American Thanksgiving though, but I
don't
think that should interfere.
Absolutely, that is why the planned date was the 12th - to have it before
one starts to thinks about X-mas. So we're still striving hard to reach this
goal.
Regards
anatol
Exactly, the relevant patches are going into RC7 which is then out on 12th
instead of GA. The next closest release slot were then Nov 26th, either for
GA or for RC8 (but better not try to predict). If it stands good, we might
have no reason to wait until the Mrs. Lovelace birthday party :)We should try really hard to get it out on the 26th. Once we get into
December people start to disappear. Nov.26 is American Thanksgiving
though, but I don't think that should interfere.
I think thanksgiving is not a great time to releasing something so significant. Most folks in the US are on time off and you won’t get as much interest.
If we do need to postpone I’d aim for Nov 30th or Dec 1st then. My experience is that as long as you’re in the first two weeks of December you still get people. After the 14th you slow down….
I know we are open-source and release when it’s ready but I think for such a huge deal it is worth waiting three more working days.
Btw, do we want to do something special on php.net http://php.net/? Front page? Not the sexiest site as-is so maybe gives us an excuse to spice up the front page?
Just my 2 cents.
Andi
Exactly, the relevant patches are going into RC7 which is then out on 12th
instead of GA. The next closest release slot were then Nov 26th, either for
GA or for RC8 (but better not try to predict). If it stands good, we might
have no reason to wait until the Mrs. Lovelace birthday party :)We should try really hard to get it out on the 26th. Once we get into
December people start to disappear. Nov.26 is American Thanksgiving
though, but I don't think that should interfere.I think thanksgiving is not a great time to releasing something so significant. Most folks in the US are on time off and you won’t get as much interest.
If we do need to postpone I’d aim for Nov 30th or Dec 1st then. My experience is that as long as you’re in the first two weeks of December you still get people. After the 14th you slow down….
I know we are open-source and release when it’s ready but I think for such a huge deal it is worth waiting three more working days.
Btw, do we want to do something special on php.net http://php.net/? Front page? Not the sexiest site as-is so maybe gives us an excuse to spice up the front page?
Just my 2 cents.
Andi
Hey Andi.
Why not, but we should release on a thursday, as our RM planning is
focused on this day. That would lead us to Dec 3rd.
Exceptions can happen, though it would be better to fit our
2-weeks-centered-on-thursdays cycle.
Julien.Pauli
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)
At least at Heroku we have a blackout policy around Thanksgiving and Christmas for platform changes, and I'd imagine other hosting platforms do as well.
So maybe either Tuesday, or even better, Thursday, since that's release day anyway.
Exactly, the relevant patches are going into RC7 which is then out on 12th
instead of GA. The next closest release slot were then Nov 26th, either for
GA or for RC8 (but better not try to predict). If it stands good, we might
have no reason to wait until the Mrs. Lovelace birthday party :)We should try really hard to get it out on the 26th. Once we get into
December people start to disappear. Nov.26 is American Thanksgiving
though, but I don't think that should interfere.I think thanksgiving is not a great time to releasing something so significant. Most folks in the US are on time off and you won’t get as much interest.
If we do need to postpone I’d aim for Nov 30th or Dec 1st then. My experience is that as long as you’re in the first two weeks of December you still get people. After the 14th you slow down….
I know we are open-source and release when it’s ready but I think for such a huge deal it is worth waiting three more working days.
Btw, do we want to do something special on php.net http://php.net/? Front page? Not the sexiest site as-is so maybe gives us an excuse to spice up the front page?
Just my 2 cents.
Andi
Hi!
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible
e-commerce site, or b) unable to roll out PHP 7 because their
customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and
Christmas for platform changes, and I'd imagine other hosting
platforms do as well.So maybe either Tuesday, or even better, Thursday, since that's
release day anyway.
I don't think anybody would be deploying PHP 7 in production right after
the release. Moreover, I don't think anybody would be (or should be, for
that matter) deploying anything important into production pretty much
starting late November and until January, for the reasons described
above and also because too many people would either be gone or filling
in for people that are gone and don't want to stir any trouble. So I
think what matters is not the operational part but just announcing it
and making people aware of it (i.e. would they read the announcement or
would they have it buried in the post-holiday "omg, I've got 9000 emails
while I was out" pile. From that point, I agree that Nov 26 release is
less optimal.
--
Stas Malyshev
smalyshev@gmail.com
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and Christmas for platform changes, and I'd imagine other hosting platforms do as well.
So maybe either Tuesday, or even better, Thursday, since that's release day anyway.
Let's be realistic. This is not a critical security fix for production
servers. This is a .0.0 release. For developers this marks the point
where they need to start making upgrade plans if they haven't already. I
realize for Heroku and the few similar cloud computing services out
there it might be slightly different in that you want to be quick to
provide the first golden PHP 7 to your customers.
My only concern with the timing is whether we are adding undue
inconvenience to our RM team and all the other folks on our side that
are affected by a release.
-Rasmus
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and Christmas for platform changes, and I'd imagine other hosting platforms do as well.
So maybe either Tuesday, or even better, Thursday, since that's release day anyway.
Let's be realistic. This is not a critical security fix for production
servers. This is a .0.0 release. For developers this marks the point
where they need to start making upgrade plans if they haven't already. I
realize for Heroku and the few similar cloud computing services out
there it might be slightly different in that you want to be quick to
provide the first golden PHP 7 to your customers.My only concern with the timing is whether we are adding undue
inconvenience to our RM team and all the other folks on our side that
are affected by a release.
Anybody moving their current stable infrastructure to PHP7 without fully
testing deserves everything they get? PHP7 should be 'rolled out' as a
secondary option for savy users who understand the implications since
the code base WILL need to be reworked before it becomes a stable
replacement for the current PhP5.x infrastructure. Actually a cloud
service getting something out as a taster for PHP7 before the holiday
could only be helpful since it would give more eyes on real world
migration problems in much the same way that PHP5 roll-out was probably
helped by RC's being available on some hosting services back then. ARE
any cloud services providing that facility today with PHP7 RC's?
--
Lester Caine - G8HFL
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk
2015-11-10 10:26 GMT+01:00 Lester Caine lester@lsces.co.uk:
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible
e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and
Christmas for platform changes, and I'd imagine other hosting platforms do
as well.So maybe either Tuesday, or even better, Thursday, since that's
release day anyway.
Let's be realistic. This is not a critical security fix for production
servers. This is a .0.0 release. For developers this marks the point
where they need to start making upgrade plans if they haven't already. I
realize for Heroku and the few similar cloud computing services out
there it might be slightly different in that you want to be quick to
provide the first golden PHP 7 to your customers.My only concern with the timing is whether we are adding undue
inconvenience to our RM team and all the other folks on our side that
are affected by a release.Anybody moving their current stable infrastructure to PHP7 without fully
testing deserves everything they get? PHP7 should be 'rolled out' as a
secondary option for savy users who understand the implications since
the code base WILL need to be reworked before it becomes a stable
replacement for the current PhP5.x infrastructure. Actually a cloud
service getting something out as a taster for PHP7 before the holiday
could only be helpful since it would give more eyes on real world
migration problems in much the same way that PHP5 roll-out was probably
helped by RC's being available on some hosting services back then. ARE
any cloud services providing that facility today with PHP7 RC's?
Lester,
With the advent of a wider variety of (container based) hosting solutions
(i.e. Heroku (https://devcenter.heroku.com/changelog-items/679) and
dokku-esque providers) and i.e. Docker providing the latest RC's via
'official' images (https://hub.docker.com/_/php) And ofcourse the vagrant
boxprovided by Rasmus (https://github.com/rlerdorf/php7dev) have allowed a
far easier upgrade path for developers trying out new stuff. This is a
vastly different landscape than back when PHP5 came to be. (or even 5.3 for
that matter) I as an app developer (and thus PHP user) now have alot more
options to try out different versions of PHP (even RC's) with faw lower
barrier of entry.
Kudo's to the PHP development team / RM's for bringing this delay up before
hand and not silently postponing the release. As a developer watching /
lurking on this list, it's great to see descisions made in such open manner!
Looking forward to 7!
--
Regards
Robin Speekenbrink
With the advent of a wider variety of (container based) hosting solutions
(i.e. Heroku (https://devcenter.heroku.com/changelog-items/679) and
dokku-esque providers) and i.e. Docker providing the latest RC's via
'official' images (https://hub.docker.com/_/php) And ofcourse the vagrant
boxprovided by Rasmus (https://github.com/rlerdorf/php7dev) have allowed a
far easier upgrade path for developers trying out new stuff. This is a
vastly different landscape than back when PHP5 came to be. (or even 5.3 for
that matter) I as an app developer (and thus PHP user) now have alot more
options to try out different versions of PHP (even RC's) with faw lower
barrier of entry.
Certainly today it is a lot easier to maintain a range of versions which
may actually be something of a noose also since in many cases there is
less incentive to replace PHP5.2/5.3 services. However the same noose
also provides a much easier way for hosting providers to offer multiple
versions in parallel making testing and transferring a hell of a lot
easier today? And PHP7 just dovetails in on top of that easily, so we
don't end up with the legacy problems of hosts 'upgrading the
infrastructure' and everything failing.
Now we just need to convince people that the days of NOT maintaining the
current stable system while also playing with the next new 'improved'
version are dead? There should be no reason today that multiple systems
can't exist side by side and one only switches off the older one once
the new build is stable and accepted?
--
Lester Caine - G8HFL
Contact - http://lsces.co.uk/wiki/?page=contact
L.S.Caine Electronic Services - http://lsces.co.uk
EnquirySolve - http://enquirysolve.com/
Model Engineers Digital Workshop - http://medw.co.uk
Rainbow Digital Media - http://rainbowdigitalmedia.co.uk
November 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and Christmas for platform changes, and I'd imagine other hosting platforms do as well.
So maybe either Tuesday, or even better, Thursday, since that's release day anyway.
Let's be realistic. This is not a critical security fix for production
servers. This is a .0.0 release. For developers this marks the point
where they need to start making upgrade plans if they haven't already. I
realize for Heroku and the few similar cloud computing services out
there it might be slightly different in that you want to be quick to
provide the first golden PHP 7 to your customers.My only concern with the timing is whether we are adding undue
inconvenience to our RM team and all the other folks on our side that
are affected by a release.Anybody moving their current stable infrastructure to PHP7 without fully
testing deserves everything they get? PHP7 should be 'rolled out' as a
secondary option for savy users who understand the implications since
the code base WILL need to be reworked before it becomes a stable
replacement for the current PhP5.x infrastructure. Actually a cloud
service getting something out as a taster for PHP7 before the holiday
could only be helpful since it would give more eyes on real world
migration problems in much the same way that PHP5 roll-out was probably
helped by RC's being available on some hosting services back then. ARE
any cloud services providing that facility today with PHP7 RC's?
Sure: http://phpversions.info/paas-hosting/
The default PHP version in the absence of any explicit declaration on Heroku will remain 5.6.latest for a while, but if you have a composer.json with a bunch of requirements for package "php" like ">=5.3.3", and not a single one with an upper bound, then on the day we make PHP 7.0.0 available, your next deploy will get PHP 7, because your composer.json says that's fine (unless you also require extensions that are not PHP 7 ready; it all gets resolved correctly).
Of course, I'll roll out a warning soon to remind users of this, but many will miss it because they tend to never read what's on their screens, so the more stable 7.0.0 is, the better!
You'd be surprised how many of our users have composer.json templates from weirdo frameworks with "minimum-stability": "dev", and they don't even notice that they're getting PHP 7 RCs because of that.
David
Sure: http://phpversions.info/paas-hosting/
The default PHP version in the absence of any explicit declaration on Heroku will remain 5.6.latest for a while, but if you have a composer.json with a bunch of requirements for package "php" like ">=5.3.3", and not a single one with an upper bound, then on the day we make PHP 7.0.0 available, your next deploy will get PHP 7, because your composer.json says that's fine (unless you also require extensions that are not PHP 7 ready; it all gets resolved correctly).
Of course, I'll roll out a warning soon to remind users of this, but many will miss it because they tend to never read what's on their screens, so the more stable 7.0.0 is, the better!
You'd be surprised how many of our users have composer.json templates from weirdo frameworks with "minimum-stability": "dev", and they don't even notice that they're getting PHP 7 RCs because of that.
Of course I don’t have any influence in what Heroku does but I would recommend not to deploy 7.0.0 on release day.
Actually part of the value we provide our customers with Zend Server is to deliver hardened, tested versions which is why we tend to wait a bit before we deliver the latest and greatest versions. Our customers depend on us delivering a version they can absolutely count on.
We’ve had PHP releases in the past which have been lemons and thanks to our waiting policy we saved our customers major outages. I agree with Rasmus that a .0.0 release will have a higher risk of having issues once the masses test their applications with it. The reality is that many people don’t bother testing their apps with RCs and wait for a GA version is out.
So while I understand Heroku likes to be ahead of the curve my friendly recommendation would be not to make it available too quickly. Give it at least a couple of weeks to ensure nothing major blows up…
Andi
-----Original Message-----
From: Rasmus Lerdorf [mailto:rasmus@lerdorf.com]
Sent: Tuesday, November 10, 2015 1:50 AM
To: David Zuelke dz@heroku.com; Andi Gutmans andi@zend.com
Cc: Anatol Belski anatol.php@belski.net; Stanislav Malyshev
smalyshev@gmail.com; internals@lists.php.net; Kalle Sommer Nielsen
kalle@php.net; Ferenc Kovacs tyrael@php.net
Subject: Re: [PHP-DEV] PHP 7 RTM dateNovember 30 is Cyber Monday, where people are either
a) focused on maxing out their credit cards on every possible
e-commerce site, or
b) unable to roll out PHP 7 because their customers are busy with a)At least at Heroku we have a blackout policy around Thanksgiving and
Christmas for platform changes, and I'd imagine other hosting platforms do
as
well.So maybe either Tuesday, or even better, Thursday, since that's release
day
anyway.Let's be realistic. This is not a critical security fix for production
servers. This is a
.0.0 release. For developers this marks the point where they need to start
making upgrade plans if they haven't already. I realize for Heroku and the
few
similar cloud computing services out there it might be slightly different
in that
you want to be quick to provide the first golden PHP 7 to your customers.My only concern with the timing is whether we are adding undue
inconvenience
to our RM team and all the other folks on our side that are affected by a
release.
Probably targeting 26th were indeed inconvenient for the colleges in USA who
test and build. Though as far as I was asking around, it were still doable.
But if there'll be a decision we can't hold our 2-weeks cycle, IMHO Julien's
suggestion about 3rd December were feasible (as exception for 7.0.0 only).
Ferenc, Kalle and me believe also, that we will see more on the factual
matter after having RC7 outdoors for some time.
Regards
Anatol
Hey Andi,
Andi Gutmans wrote:
Btw, do we want to do something special on php.net http://php.net/? Front page? Not the sexiest site as-is so maybe gives us an excuse to spice up the front page?
7 is the first new major version in more than a decade, it'd be a shame
if we didn't capitalise it!
A big banner across the top of the site announcing PHP 7 (perhaps
visible on manual pages too?) and some party streamer-type effects (not
animated ;) might be good.
(Tangential: It'd be really cool if we used some media other than plain
text for once, for such a release. What if there was a photo montage of
all PHP 7's developers? Of course, I don't want any efforts to take away
time from making sure 7.0.0 works :)
Thanks!
--
Andrea Faulds
http://ajf.me/
I think this is the right thing to do, we can wait for a bit if it means
better stability.
+1
I love that The PHP Development Team is not looking to rush a release and
while I, like many others wish to see PHP 7.0 I would rather have a stable
and solid release over the idea of a rushed release. So I commend the
choice of releasing another RC (RC7) coming up in the next few days
There appears to be some concern though surrounding the holidays and that
is understandable. I would like to recommend that if all matters are not
resolved by November 26 (the new suggested date), that perhaps it may be
better to wait until the start of 2016 after the holidays. This would give
plenty for the development to proceed without being rushed to release and
perhaps relieve any holiday stress.
In any case, thank you for your time, effort, and hard work and
dedication. It is very much appreciated.
On Sun, Nov 8, 2015 at 2:45 PM, Stanislav Malyshev smalyshev@gmail.com
wrote:
Hi!
Thus, resuming the pre-release cycle seems in our opinion the right
thing to
do ATM. Hence this writing to inform the community about the intention
and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.I think this is the right thing to do, we can wait for a bit if it means
better stability.Do I understand right that the segfault fixes that are now marked as
7.0.1 would be instead released as RC7? And the prospective GA date
(provided RC7 is stable enough) would be now December 10th?
(BTW, by an interesting coincidence, December 10th is the birthday of
Ada Lovelace :).Stas Malyshev
smalyshev@gmail.com
Hi,
After doing the last evaluations in the RMs circle before going for 7.0.0
RTM preparations, we came to the conclusion that the current state does not
look reasonable to be packaged as the final release.The RCs before RC6 was looking acceptable, which gave the reason to announce
the planned RTM date on Nov 12th as final. From the todays perspective, the
most of the issues discovered and fixed since RC6 are still minor by
themselves - 7 crashes (bug #70805 rather critical), and 3 functionality
regressions. However, given the amount and the combination of them
altogether, the state is in our opinion unsuitable for starting the life
cycle of the next major.Thus, resuming the pre-release cycle seems in our opinion the right thing to
do ATM. Hence this writing to inform the community about the intention and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.
This is the right decision. The whole point of the RC cycle is to
identify crash bugs like 70805. It would have been great if we had
caught it earlier, but we still caught and fixed it before the release
and hopefully we can still get the final release out in November.
-Rasmus
-----Original Message-----
From: Rasmus Lerdorf [mailto:rasmus@lerdorf.com]
Sent: Sunday, November 08, 2015 9:49 PM
To: Anatol Belski; internals@lists.php.net
Cc: 'Kalle Sommer Nielsen'; 'Ferenc Kovacs'
Subject: Re: [PHP-DEV] PHP 7 RTM dateHi,
After doing the last evaluations in the RMs circle before going for
7.0.0 RTM preparations, we came to the conclusion that the current
state does not look reasonable to be packaged as the final release.The RCs before RC6 was looking acceptable, which gave the reason to
announce the planned RTM date on Nov 12th as final. From the todays
perspective, the most of the issues discovered and fixed since RC6 are
still minor by themselves - 7 crashes (bug #70805 rather critical),
and 3 functionality regressions. However, given the amount and the
combination of them altogether, the state is in our opinion unsuitable
for starting the life cycle of the next major.Thus, resuming the pre-release cycle seems in our opinion the right
thing to do ATM. Hence this writing to inform the community about the
intention and to gather the opinions. PHP 7 is still going down the
home stretch and is very close to the finish, the next RC can be well
the last. Though we'd better ensuring the quality as many times as
needed before, instead of casually disservicing the consumers.This is the right decision. The whole point of the RC cycle is to
identify crash
bugs like 70805. It would have been great if we had caught it earlier,
but we
still caught and fixed it before the release and hopefully we can still
get the
final release out in November.
Thumbs up from me as well (incl. trying to get it out the door in November
if possible).
Zeev
-----Original Message-----
From: Zeev Suraski [mailto:zeev@zend.com]
Sent: Sunday, November 8, 2015 8:57 PM
To: Rasmus Lerdorf rasmus@lerdorf.com; Anatol Belski
anatol.php@belski.net; internals@lists.php.net
Cc: Kalle Sommer Nielsen kalle@php.net; Ferenc Kovacs tyrael@php.net
Subject: RE: [PHP-DEV] PHP 7 RTM date-----Original Message-----
From: Rasmus Lerdorf [mailto:rasmus@lerdorf.com]
Sent: Sunday, November 08, 2015 9:49 PM
To: Anatol Belski; internals@lists.php.net
Cc: 'Kalle Sommer Nielsen'; 'Ferenc Kovacs'
Subject: Re: [PHP-DEV] PHP 7 RTM dateHi,
After doing the last evaluations in the RMs circle before going for
7.0.0 RTM preparations, we came to the conclusion that the current
state does not look reasonable to be packaged as the final release.The RCs before RC6 was looking acceptable, which gave the reason to
announce the planned RTM date on Nov 12th as final. From the todays
perspective, the most of the issues discovered and fixed since RC6
are still minor by themselves - 7 crashes (bug #70805 rather
critical), and 3 functionality regressions. However, given the
amount and the combination of them altogether, the state is in our
opinion unsuitable for starting the life cycle of the next major.Thus, resuming the pre-release cycle seems in our opinion the right
thing to do ATM. Hence this writing to inform the community about
the intention and to gather the opinions. PHP 7 is still going down
the home stretch and is very close to the finish, the next RC can be
well the last. Though we'd better ensuring the quality as many times
as needed before, instead of casually disservicing the consumers.This is the right decision. The whole point of the RC cycle is to
identify crash
bugs like 70805. It would have been great if we had caught it earlier,
but we
still caught and fixed it before the release and hopefully we can
still
get the
final release out in November.Thumbs up from me as well (incl. trying to get it out the door in November if
possible).
Thanks for the opinions and support. That means we're having the RC7 this week and look forward to the final. Back to the work then :)
Regards
Anatol
As much as I had been looking forward to it soon, haste does makes waste.
Making a more responsible choice for the users is commendable and outweighs
the inconvenience a mere two week delay.
+1 from me.
Scott Arciszewski
Chief Development Officer
Paragon Initiative Enterprises https://paragonie.com
Hi,
After doing the last evaluations in the RMs circle before going for 7.0.0
RTM preparations, we came to the conclusion that the current state does
not
look reasonable to be packaged as the final release.The RCs before RC6 was looking acceptable, which gave the reason to
announce
the planned RTM date on Nov 12th as final. From the todays perspective,
the
most of the issues discovered and fixed since RC6 are still minor by
themselves - 7 crashes (bug #70805 rather critical), and 3 functionality
regressions. However, given the amount and the combination of them
altogether, the state is in our opinion unsuitable for starting the life
cycle of the next major.Thus, resuming the pre-release cycle seems in our opinion the right
thing to
do ATM. Hence this writing to inform the community about the intention
and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.This is the right decision. The whole point of the RC cycle is to
identify crash bugs like 70805. It would have been great if we had
caught it earlier, but we still caught and fixed it before the release
and hopefully we can still get the final release out in November.-Rasmus
Hi,
After doing the last evaluations in the RMs circle before going for 7.0.0
RTM preparations, we came to the conclusion that the current state does not
look reasonable to be packaged as the final release.The RCs before RC6 was looking acceptable, which gave the reason to announce
the planned RTM date on Nov 12th as final. From the todays perspective, the
most of the issues discovered and fixed since RC6 are still minor by
themselves - 7 crashes (bug #70805 rather critical), and 3 functionality
regressions. However, given the amount and the combination of them
altogether, the state is in our opinion unsuitable for starting the life
cycle of the next major.Thus, resuming the pre-release cycle seems in our opinion the right thing to
do ATM. Hence this writing to inform the community about the intention and
to gather the opinions. PHP 7 is still going down the home stretch and is
very close to the finish, the next RC can be well the last. Though we'd
better ensuring the quality as many times as needed before, instead of
casually disservicing the consumers.Regards
Anatol
OCI8 still has a mem leak or two that our developer is looking at, so a delay would
be good.
Chris