Mongrel2 v1.6 Released

Welll, to be honest, life sort of pushed Mongrel2 to the side temporarily. The book, teaching classes, and the new job took over until they stabilized. But, it's back and people were using it in production already so this means more progress to come.

Re: Mongrel2 v1.6 Released

#5

Post by zedshaw » Thu, May 12, 2011, 6:52 PM UTC

evangineer wrote: post #2

DevOps folks are going to love the upcoming "config server" protocol that Zed is proposing to do for v1.8.

That's where I'm getting that feature idea from. I think the control port, config anywhere, fast dynamic restarts, and filters will make it the bee's-knees for hosting.

Re: Mongrel2 v1.6 Released

#6

Post by peterb » Thu, May 12, 2011, 6:53 PM UTC

This is a project to watch. Language agnostic web server with a 0MQ back-end to automagically take advantage of multi-threading/multi-core on modern hardware. Nice.

Plus he is running unit tests against code samples in the manual. Brilliant!

Re: Mongrel2 v1.6 Released

#7

Post by fictorial » Thu, May 12, 2011, 7:06 PM UTC

TNetStrings seems quite similar to Bencode less the size prefix.

http://tnetstrings.org

http://enwp.org/Bencode

(sample implementation) https://github.com/bittorrent/bencode/blob/master/bencode/__...

Re: Mongrel2 v1.6 Released

#8

Post by FooBarWidget » Thu, May 12, 2011, 7:07 PM UTC

I'm wondering why TNetstring types come after the data. This means you have to buffer the entire data before you can parse it. Is this intentional?

Re: Mongrel2 v1.6 Released

#9

Post by zedshaw » Thu, May 12, 2011, 7:26 PM UTC

peterb wrote: post #6

This is a project to watch. Language agnostic web server with a 0MQ back-end to automagically take advantage of multi-threading/multi-core on modern hardware. Nice. Plus he is running unit tests against code samples in the manual. Brilliant!

Here's the coverage report:

http://mongrel2.org/static/coverage/

This release had a drop of about 17% from the last run, but I'll bring it back up in the next one.

Re: Mongrel2 v1.6 Released

#10

Post by amattn » Thu, May 12, 2011, 7:27 PM UTC

FooBarWidget wrote: post #8

I'm wondering why TNetstring types come after the data. This means you have to buffer the entire data before you can parse it. Is this intentional?

It's for backwards compatibility with regular old netstrings. It works very well in practice. Implementations are dead-simple and super fast in just about every language.