Hi,
FYI. Given the shift of the 7.0.0 release date by one week off the usual two
weeks release schedule, 7.0.1RC1 is planned for December 10th and to be
tagged on December 8th. This serves two goals
- there are already some bug fixes and it makes full sense them to reach the
world ASAP - the release schedule gets synchronized with PHP 5 line
So starting with PHP 7.0.1 and 5.6.17 releases of both 5 and 7 branches
will be delivered consolidated and on the usual schedule of two weeks with
RCs and final. The end release of 7.0.1 is to be expected on December 24th.
Regards
Anatol
Hi,
FYI. Given the shift of the 7.0.0 release date by one week off the usual two
weeks release schedule, 7.0.1RC1 is planned for December 10th and to be
tagged on December 8th. This serves two goals
- there are already some bug fixes and it makes full sense them to reach the
world ASAP- the release schedule gets synchronized with PHP 5 line
So starting with PHP 7.0.1 and 5.6.17 releases of both 5 and 7 branches
will be delivered consolidated and on the usual schedule of two weeks with
RCs and final. The end release of 7.0.1 is to be expected on December 24th.Regards
Anatol
Hello,
Is December 24th a good date to release ?
Just asking.
Next release will also have a bunch of security patches.
I'm working on 5.5 tree actually to get in security patches, then I'll
merge up to 5.6 and 7.0 few days before release.
Julien.Pauli
Hi,
-----Original Message-----
From: julienpauli@gmail.com [mailto:julienpauli@gmail.com] On Behalf Of
Julien Pauli
Sent: Tuesday, December 8, 2015 11:26 AM
To: Anatol Belski anatol.php@belski.net
Cc: PHP Internals internals@lists.php.net; Ferenc Kovacs tyrael@php.net;
Remi Collet remi@php.net
Subject: Re: PHP 7.0.1 schedulingHi,
FYI. Given the shift of the 7.0.0 release date by one week off the
usual two weeks release schedule, 7.0.1RC1 is planned for December
10th and to be tagged on December 8th. This serves two goals
- there are already some bug fixes and it makes full sense them to
reach the world ASAP- the release schedule gets synchronized with PHP 5 line
So starting with PHP 7.0.1 and 5.6.17 releases of both 5 and 7
branches will be delivered consolidated and on the usual schedule of
two weeks with RCs and final. The end release of 7.0.1 is to be expected on
December 24th.Regards
Anatol
Hello,
Is December 24th a good date to release ?
Just asking.
Next release will also have a bunch of security patches.
I'm working on 5.5 tree actually to get in security patches, then I'll merge up to
5.6 and 7.0 few days before release.
Yeah, particularly our release schedule collides with many holidays this year. For a security release a holiday time is obviously very bad. So we might want to shift it altogether to the after new year? It'd probably give room for security fixes. And yet for 7.0.1RC2 on 24th, if needed. One is firm that 7.0.1 has to go out consolidated with 5.5 and 5.6.
Regards
Anatol
Le 08/12/2015 12:36, Anatol Belski a écrit :
Yeah, particularly our release schedule collides with many holidays this year. For a security release a holiday time is obviously very bad.
+1
Our Release Process state:
" At least one release per month, more at wish "
Perhaps we should consider new release on
" 1st thursday of the month "
(It seems every year, Christmas is end of month ;)
Remi.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Le 08/12/2015 13:04, Remi Collet a écrit :
Perhaps we should consider new release on
" 1st thursday of the month "
(It seems every year, Christmas is end of month ;)
As Thanksgiving and Dec 31th.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iEYEARECAAYFAlZmx+wACgkQYUppBSnxahjuKQCfZ3LChwq0hcWZ6lmUgGan9YsW
u8cAn1HDFBeXDcvBc7+jrXxJmt9z9XHP
=+WW7
-----END PGP SIGNATURE
Le 08/12/2015 12:36, Anatol Belski a écrit :
Yeah, particularly our release schedule collides with many holidays this year. For a security release a holiday time is obviously very bad.
+1
Our Release Process state:
" At least one release per month, more at wish "
Perhaps we should consider new release on
" 1st thursday of the month "
(It seems every year, Christmas is end of month ;)
Yep, end of year has always collided with our release cycle.
Perhaps we should change this latter as proposed by Remi ?
Also, we had another idea.
Why not use RC2 as a chance for some security patches to get tested ?
If we have a RC2 (which is not bad IMO, we have a time slice here at
end of 2015 that enables us to release an RC2) , we then could merge
into this one some well-selected security patches, f.e, some that are
low classified , aka that can't get trigerred remotely by user
payload.
What you think ?
CCing Stas as well, as he got many experience in security.
Julien.P
Le 08/12/2015 12:36, Anatol Belski a écrit :
Yeah, particularly our release schedule collides with many holidays
this year. For a security release a holiday time is obviously very bad.+1
Our Release Process state:
" At least one release per month, more at wish "
Perhaps we should consider new release on
" 1st thursday of the month "
(It seems every year, Christmas is end of month ;)
Yep, end of year has always collided with our release cycle.
Perhaps we should change this latter as proposed by Remi ?
Also, we had another idea.
Why not use RC2 as a chance for some security patches to get tested ?
If we have a RC2 (which is not bad IMO, we have a time slice here at
end of 2015 that enables us to release an RC2) , we then could merge
into this one some well-selected security patches, f.e, some that are
low classified , aka that can't get trigerred remotely by user
payload.What you think ?
CCing Stas as well, as he got many experience in security.
Julien.P
we are just iterating by two weeks which makes the monthly release not
monthly but every 28 days, which is fine but that also means that sometimes
the release dates are not always at the same date on each month (stating
the obvious).
I think that there is no better alternative, we just have to handle the
exceptional cases.
As we seem to have security fixes (and unfortunatelly at least one of those
were directly pushed by laruance so now it is public information) maybe we
should release it one week earlier (cutting the RC period one week shorter)
instead of delaying it until january?
Hi Ferenc,
-----Original Message-----
From: tyra3l@gmail.com [mailto:tyra3l@gmail.com] On Behalf Of Ferenc
Kovacs
Sent: Tuesday, December 8, 2015 2:41 PM
To: Julien Pauli jpauli@php.net
Cc: Remi Collet remi@php.net; PHP Internals internals@lists.php.net;
Anatol Belsky ab@php.net; Stanislav Malyshev stas@php.net
Subject: Re: [PHP-DEV] Re: PHP 7.0.1 schedulingLe 08/12/2015 12:36, Anatol Belski a écrit :
Yeah, particularly our release schedule collides with many holidays
this year. For a security release a holiday time is obviously very bad.+1
Our Release Process state:
" At least one release per month, more at wish "
Perhaps we should consider new release on
" 1st thursday of the month "
(It seems every year, Christmas is end of month ;)
Yep, end of year has always collided with our release cycle.
Perhaps we should change this latter as proposed by Remi ?
Also, we had another idea.
Why not use RC2 as a chance for some security patches to get tested ?
If we have a RC2 (which is not bad IMO, we have a time slice here at
end of 2015 that enables us to release an RC2) , we then could merge
into this one some well-selected security patches, f.e, some that are
low classified , aka that can't get trigerred remotely by user
payload.What you think ?
CCing Stas as well, as he got many experience in security.
Julien.P
we are just iterating by two weeks which makes the monthly release not
monthly but every 28 days, which is fine but that also means that sometimes the
release dates are not always at the same date on each month (stating the
obvious).
I think that there is no better alternative, we just have to handle the exceptional
cases.
As we seem to have security fixes (and unfortunatelly at least one of those were
directly pushed by laruance so now it is public information) maybe we should
release it one week earlier (cutting the RC period one week shorter) instead of
delaying it until january?
Yeah, that works for 7.0 alone in any case. If it's about 7.0.1 only, it were
- 7.0.1RC1 on 10th December
- 7.0.1 final on 17th December
- 7.0.2RC1 on 24th or 31st
- 7.0.2 final next year together with 5.5.31 and 5.6.17
Probably it could make sense, 7.0.1 in any case on 17th then - what was disclosed was only one patch and relevant for 7.0 only. 17th is also holiday free, hopefully :) If PHP 5 branches could not be so far, we should indeed get 7.0.1 out this year, and we would sync 5 and 7 release dates starting with 7.0.2 and next year. How does it sound?
Regards
Anatol
Hi!
Why not use RC2 as a chance for some security patches to get tested ? If we have a RC2 (which is not bad IMO, we have a time slice here at end of 2015 that enables us to release an RC2) , we then could merge into this one some well-selected security patches, f.e, some that are *low* classified , aka that can't get trigerred remotely by user payload.
Generally, if we publish security patch we should release ASAP, as
having public unpatched security issue is bad. Fortunately, the issue
disclosed is not that bad, I'd even doubt it is security issue per se,
as I see no real way to trigger it without code access (it may be I miss
something though).
Accelerating the schedule is OK with me (as is keeping the current one,
too) but please do not forget to merge pending patches (or please tell
me when you want me to merge them).
Thanks,
Stas Malyshev
smalyshev@gmail.com
On Tue, Dec 15, 2015 at 12:55 AM, Stanislav Malyshev smalyshev@gmail.com
wrote:
Hi!
Why not use RC2 as a chance for some security patches to get tested ? If we have a RC2 (which is not bad IMO, we have a time slice here at end of 2015 that enables us to release an RC2) , we then could merge into this one some well-selected security patches, f.e, some that are *low* classified , aka that can't get trigerred remotely by user payload.
Generally, if we publish security patch we should release ASAP, as
having public unpatched security issue is bad. Fortunately, the issue
disclosed is not that bad, I'd even doubt it is security issue per se,
as I see no real way to trigger it without code access (it may be I miss
something though).Accelerating the schedule is OK with me (as is keeping the current one,
too) but please do not forget to merge pending patches (or please tell
me when you want me to merge them).Thanks,
Stas Malyshev
smalyshev@gmail.com
so Anatol already tagged 7.0.1, I'm waiting with the 5.6.17 tag for Julien
to tag 5.5.31.
Julien what is the status there?
Ferenc Kovacs in php.internals (Wed, 16 Dec 2015 15:57:14 +0100):
so Anatol already tagged 7.0.1, I'm waiting with the 5.6.17 tag for Julien
to tag 5.5.31.
Julien what is the status there?
Any progress? 5.6.17 was initually is planned for Dec 17th ...
Jan
Ferenc Kovacs in php.internals (Wed, 16 Dec 2015 15:57:14 +0100):
so Anatol already tagged 7.0.1, I'm waiting with the 5.6.17 tag for Julien
to tag 5.5.31.
Julien what is the status there?Any progress? 5.6.17 was initually is planned for Dec 17th ...
Jan
--
hi,
as Julien wasn't available that time we delayed the 5.5 and 5.6 release for
early January.
--
Ferenc Kovács
@Tyr43l - http://tyrael.hu