Should we have a TechComm-driven meeting about this ASAP?
Like others, I don't expect that there will be disagreement about the way
to go, but there is a lot to discuss about what needs to be done,
resourcing, etc.
It would be nice to have Ori around for it, to pick his brains about any
undocumented or little-known knowledge about the HHVM migration that could
bite us when migrating to PHP 7.x if we don't know about it.
On Tue, Sep 19, 2017 at 9:07 AM, Moritz Muehlenhoff <
mmuhlenhoff(a)wikimedia.org> wrote:
On Tue, Sep 19, 2017 at 10:13:47AM +1000, Tim Starling
wrote:
On 19/09/17 06:58, Max Semenik wrote:
> Today, the HHVM developers made an announcement[1] that they have
plans of
ceasing
to maintain 100% PHP7 compatibility and concentrating on Hack
instead.
The HHVM team did tell us privately that they were planning on
changing their strategy, basically as you describe it above. The
surprising things for me in this announcement were:
* The plan to also drop PHP 5 compatibility, on a short timeline (1
year).
* Rather than "drifting away" from PHP,
their top priority plans
include removing core language features like references and destructors.
While this does not mean that we need to take an
action immediately,
eventually we will have to decide something.
Actually, I think a year is a pretty short time for ops to switch to
PHP 7. I think we need to decide on this pretty much immediately.
The next step would be the upgrade of the mw* fleet to Debian stretch
while still using HHVM 3.18 (to minimise disruption since we've stabilised
3.18 in it's current build). That work is tracked at T174431. 3.18 will
be supported by upstream for at least another six months (and if the
migration drags
further I can roll custom 3.18 security backports from later LTS releases)
Debian stretch ships PHP7, so that'd be a good stepstone to migrate
back to Zend.
Cheers,
Moritz
_______________________________________________
Wikitech-l mailing list
Wikitech-l(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l