Le 06/08/2014 14:36, Zeev Suraski a écrit :
I opened the voting on the phpng RFC:
https://wiki.php.net/rfc/phpng#vote
Voting ends on Thursday, August 14th.
After speaking with other members of AFUP (French user-group) about this
RFC, here's a short summary of our thoughts:
The idea of (greatly: 20-30%) improving performances is really
interesting, be it for performances themselves or for the reputation of
the language.
That's even more true if this gets us closer to JIT in the future.
Having to update all extensions is not "perfect", but is not really a
problem either: developers of extensions know this is to be expected for
each major version of PHP -- and, after the branch is merged to master,
they will still have something like two years to update their extensions.
We are a bit sad this has been developed for quite some time in secret
and the community has learned about it only after most of the work had
been done: it doesn't really fit with the way we think about "open-source".
Same thing about the fact several evolutions have all been worked on in
the same branch, the same RFC and the same vote, instead of each being
independent from the others.
We feel there might have been too much "public" communication about
"phpng" and its performance improvements, before the decision of merging
it was (or not) taken.
Because of this, at the time of the vote, we kind of feel like we don't
have much of a choice: if the branch is not merged, PHP will look quite
ridiculous, as "phpng" is already expected by many to be used as a a
first step for PHP 7.
Judging from some discussions here, we are afraid the quality and
maintainability of PHP's source code might suffer a bit from this merge.
For us, a major version is the right time (it's going to be used for
something like 10 years) to work on those. (still, easy to say, when
only a few of us contribute code to PHP).
PHP 7 will be a "major" version.
It must not be released earlier than planned only to publish some
performance improvements as soon as possible (even if those are great).
It is important to also think about features, reworks and other major
improvements: with one major versions every 10 years, it's pretty much
now or never!
We would prefer PHP 7 to be released as a great version in two years,
rather than seeing it too early in one year.
In the end, there have been two opposite opinions on our UG's
mailing-list -- and both are well represented, with strong posts on both
"yes merge" and "no don't merge":
* Some of us are OK to merge right now and improve things later,
* While others would prefer to improve things now and merge only after.
In any case, we all agree there is still a lot to be done before
reaching PHP 7 ;-)
--
Pascal MARTIN
http://blog.pascal-martin.fr
@pascal_martin
--
PHP Internals - PHP Runtime Development Mailing List
To unsubscribe, visit: http://www.php.net/unsub.php
