A distributed RPC framework that allows for consensus based state replication. Or a cryptocurrency for short
>the future of finance
nah. But it's a fun little project
/dpt/ - Daily Programming Thread
Justin Baker
Kevin Reyes
>I honestly don't know what's so hard about
maybe it's not about being hard, but unnecessary? why bloat the kernel with stuff you can do in userspace?
Samuel Ramirez
Anyone of you has ever made a translator? Any tips on how to deal with the strings and replacement?
Brandon Morales
Wrote a scheme script that takes 2 directories and uses openssl to generate a checksum for every file inside of them recursively and checks to make sure that all the files are the same.
Grayson Rivera
>stuff you can do in userspace
Because the userspace AIO implementation is working so good, right?
Andrew Rivera
Doesn't work
Justin Miller
don't know, I don't think anyone uses it; I was referring to the fact that you can easily build a callback abstraction in userspace on top of the readiness-notification mechanism
Joseph Gonzalez
in this company we want it BIG. you hear me? BIG.
No small talks allowed, focus on the project.
Leo Lee
learning bigtalk if true
Matthew Reed
The point is having uniform asynchronous I/O interface for any kind of file descriptor. Epoll currently only works well for sockets. User space tools don't work at all, and even if they did they wouldn't be as well integrated as a kernel mechanism.
We can already epoll nearly anything, even signals with signalfd and it's a vast improvement to software interrupt-based signals handling. People really should start working on a IOCP equivalent that's integrated into epoll.