LScube

As you may already know from the git commits, feng is getting some new features added and some code is being reshaped in order to be easier to understand. Right now we didn't do much with the new lighttpd based configuration beside adding support for listening to an arbitrary number of sockets and other minor changes, before starting to push more interesting stuff we'd like to receive more feedback and test it a bit more.

Big news for felix, after a major refactoring we have completed the working support for creating a TCP bridge between two felix nodes. Now it is possible to transfer packets from the producer to feng under a more reliable transport instead of UDP, trading latency for delivery guaranties

I spent some time hacking a bit our git setup and I eventually revived cia and in the mean time also created a mailing list to hold the commits so people can watch them. I started the process to register our mailing lists to gmane. Hopefully the irc channel will be revived too (please join it).

We are setting up the latest details for the internet governance forum meeting that will be held in Geneva on 26 February 2008 right now.

Recently we had been overly quiet about what we are doing beside the commits in gitweb. I spent lots of time hacking on the lemon based lighttpd syntax, today I eventually switched to lighttpd completely. Once I clean up the remaining details this will be the default syntax and I'll start implementing some of the features you already enjoy in lighttpd in feng.

While feng is undergoing a major rewrite of its configuration system, libnemesi has started to compile under win32 and its simplest example has started to work. Tested under XP and Vista, the latest commit on the git repository has a partially working version of dump_info which is able to send the request to the server and receive the response (while still hanging on thread clean up).

Today we eventually got the vorbis parser in libnemesi in a better shape, it seems to work fine enough to let us hammer some bugs in the MPlayer integration tomorrow we'll try to get aac support in and then decide if the api is fine as is or if we should try to improve it a bit more before a complete release (ridden for sure of more bugs that we'd like to, but it's always like that)

So if you like vorbis or you are interested in trying it over rtp please fetch the latest sources ^^

netembryo is supposed to be a portable abstraction to the network layer, sadly the assumption was a bit wrong at least on freebsd I hopefully fixed the bugs that were preventing netembryo from building on freebsd 7beta1.5, now I'm waiting for some patches to address other weirdness in freebsd 6.2.

In the other news the vorbis-rtp rfc is getting lots of comments that I'm trying to address, so far just clarifications that won't change the behavior.

libnemesi is eventually get in shape for a formal release, more or less.

What it has:

• mpegvideo, h263, h264, mpegaudio, speex parsers more or less in shape
• the api more or less shaped in an usable form (as tested by integrating it on MPlayer)
• tcp and sctp support working fine (tested with mplayer) even on quite shoddy lines
• precise seeking

What's missing:

• aac, vorbis and theora support
• interoperability tests with anything different from feng
• doxy documentation improvements (proofreading mostly)

Lately we had been quite quiet and you couldn't see much here beside the menu restyle done by Dario and some updates about our thesis proposals.

We are still striving to get libnemesi and feng released with all the features we added working properly. Sadly seems it is taking a lot mostly because we are getting to the point that we want to have all the corner case issues addressed. If you want to follow the development and maybe help you may look at the git trees and maybe test the live snapshots.

The summer is almost ended already here, with enough rain to make this late August to look like September…

Alessandro is still fighting with MPlayer and pointed some ugly mistake I made on libnemesi h264 parser… Both code is quite ugly and broken, but at least we can play the first seconds of our videos already (then everything hangs)…