News for hackers. Stuff that matters.
Re:RustgreSQL by iagooar on Monday January 09, 2017 @01:15AM (#13354393)
I would suggest that the author creates a clone of a "simpler" DB like SQLite or Redis. Those have, as far as I understand, good C codebases, so understanding the system should be much easier.

This way you could also say: hey! I have successfully ported a DB written in C in the past, and it has been an extremely positive experience, how about trying the same with Postgres?

Re:RustgreSQL by nathancahill on Monday January 09, 2017 @01:17AM (#13354406)
SQLite for sure. The test suite is top tier.
Re:RustgreSQL by dllthomas on Monday January 09, 2017 @01:27AM (#13354451)
Hopefully you could even reuse it!
Re:RustgreSQL by rosser on Monday January 09, 2017 @02:00AM (#13354567)
If you wanted to write a drop-in SQLite replacement in Rust, using the original project's (incredible) test suite would be on the mandatory side, wouldn't it?
Re:RustgreSQL by dllthomas on Monday January 09, 2017 @04:00PM (#13359750)
To a degree. Certainly the end-to-end tests should be something you can expect to reuse. Other tests may be hit or miss. In the case of SQLite, where there's >100x as much test code as application code, there's a large range between "can use some of it" and "can use most of it" - as mathw gets at I think we should be able to expect the latter here.
Re:RustgreSQL by mathw on Monday January 09, 2017 @04:10AM (#13355043)
Definitely! And Rust is capable of that kind of thing (producing a drop-in replacement for a C library) so there'd be no excuses either.
Re:RustgreSQL by riffraff on Monday January 09, 2017 @03:33AM (#13354881)
redis-in-rust : https://github.com/seppo0010/rsedis

it seems many people re-implement redis (I recall at least also Go and Scala).

It looks like it's the new "backend language todo list project".