Show HN: C++11 port of Postgres | Hacker News Reader

Show HN: C++11 port of Postgres

github.com

4 pointsjarulraj11 comments

Several questions: * Is this a one-man effort (for now at least?) or is this 'official' * Why refactor the increment operator? (s ++ to s = s + 1) * Is clang supported (or libc++ for example?) * Do you have a roadmap for planned refactoring?

Interesting project, I wonder about the effort which is required. Even slow incremental involves breaking interfaces.

* It is a proof-of-concept implementation to show that porting Postgres over to C++ is feasible with minimal changes. It is not official.

* We need to refactor the increment operator because the variables are of enumeration type.

* We have not yet tested compilation with clang. But, I presume that minimal changes would be required to support compilation with clang. PRs are welcome!

* Not really. Our hope is that the PostgreSQL core team would consider gradually migrating over to C++. I think that this refactoring can be done without breaking external interfaces.

Thanks for the positive feedback.

> Our hope is that the PostgreSQL core team would consider gradually migrating over to C++.

How many are you? I see only commits from you. And did you speak with any of the pgsql core team yet? Might be worth investigating, whether they even consider it... I don't see the point, if they don't.

Good luck with they project, might look into it some time soon!

There is on-going conversation on the pg-hackers mailing list. Thanks for the positive feedback.

It's dissapointing it is not in Javascript.

Whoa, why would you do that ?

Just kidding.

But it would work in the browser!

I'm ignorant, what's the benefit of it?

Certain features of the C++ language and its library should help simplify coding, improve code reuse, and avoid bugs.

Here's a blog post that provides more context: https://petereisentraut.blogspot.com/2013/05/moving-to-c.htm...

The problem is that the current set of portability targets (https://www.postgresql.org/docs/devel/static/supported-platf...) doesn't support a lot of common ground WRT c++ support. Otherwise I (and some others involved in the project) would have started to gently push in that direction.

I see, thanks for sharing this information.

Reply on news.ycombinator.com