Here's the interesting thing about all this; HHVM will always be developed because it's important to Facebook's bottom line and Open Source because it only benefits them to keep it out there and have other people testing it and improving on it.
Now that they're getting rid of direct PHP support, HHVM is only going to get better. This will unlock a whole host of language improvements that HHVM couldn't otherwise make.
HHVM is faster relative to PHP now, and it will only get faster with these changes. Typing is an important part of making JITed code fast and unless PHP ever decides to fully add it, it will never have the potential to catch up. This is important to PHP-based companies as they grow and want to optimize on cost and development efficiency.
Undoubtedly, this split will be painful initially for those of us who are bought into the symbiosis of the HHVM and PHP ecosystem together. How painful it is to split will just be a question of where members of the PHP community want to go (or both). The nice thing is that converting something from PHP to HHVM isn't terribly hard; not anywhere near like converting from PHP to Golang. For HHVM, it's mostly just adding type annotations.
While this is probably still true[1], it's certainly less of a concern now than it was with 5.x. Would (often negligible) performance boosts be enough for someone with a 5.x PHP codebase to choose Hack over PHP 7.x? I can't see that for most cases.
We typically do yearly performance audits with Percona to ensure our databases are optimally configured. We disable the query cache. We set innodb_buffer_pool_size based on a % of total memory (as MySQL will use more than that in total, much for allocating connection structs and things like that). We set innodb_flush_log_at_trx_commit to 0 which is not ideal for data integrity but gives us more performance. In practice, because of Google's Live Migration technology we have never experienced a crash due to hardware on our master nodes (LM will run an emergency migration for any server that has hardware that is detected to be going bad before it becomes a problem) and disks are abstracted away even further. Our main risk is the kernel crashing or MySQL crashing which we've been fortunate enough to not happen on our masters. Spanner provides ACID compliance by default with all the scaleability and performance we get out of it.
Our website was down as well for 16 minutes. My guess is that it was a bad route that was pushed out simultaneously (probably was not the intention). It happened once before, sometimes last year, if I remember correctly. We'll have to wait and see what the definitive cause was though.
Now that they're getting rid of direct PHP support, HHVM is only going to get better. This will unlock a whole host of language improvements that HHVM couldn't otherwise make.
HHVM is faster relative to PHP now, and it will only get faster with these changes. Typing is an important part of making JITed code fast and unless PHP ever decides to fully add it, it will never have the potential to catch up. This is important to PHP-based companies as they grow and want to optimize on cost and development efficiency.
Undoubtedly, this split will be painful initially for those of us who are bought into the symbiosis of the HHVM and PHP ecosystem together. How painful it is to split will just be a question of where members of the PHP community want to go (or both). The nice thing is that converting something from PHP to HHVM isn't terribly hard; not anywhere near like converting from PHP to Golang. For HHVM, it's mostly just adding type annotations.