Phonon To Break Git History

Phonon is currently preparing to move its repositories from Gitorious to Git.KDE.org. Unfortunately we noticed that our current main Phonon repository suffers from some metadata problems.

Due to this unfortunate situation we will have to rewrite the affected part of the Git history to comply with modern Git version’s strictness. This bares certain implications if you happen to have a clone of this repository, since it will essentially break your clone’s references to the original repository.

Therefore we decided to give you a 2 weeks grace period, so you can take preliminary measures.

Some time after December 20, history will be rewritten.

If you have changes for Phonon, please make sure to get them into the repository before that date.

Please also pass this information on to any interested parties as to prevent confusion.

I’ll post more information on the move to git.kde.org once the time has come. The Gitorious repository will also be rewritten and remain read-only once the move is done.

Parley gone Fluffy

Beware that the following video might be too fluffy for some people.

[Theora Version]

How to swarm some code

As if a “how does one make such a code swarm?” comment was not to be anticipated….

First of all you’ll need codeswarm in order to use it you’ll also need sun-java6-jdk (or whatever your distro might call it) and ant. (on Kubuntu I recommend you install without recommends though)

Now it gets tricky. You’ll need to get a log from your version control system and convert that using the script in convert_logs. If you are using bzr you are boned since the script does not parse the default log output of bzr, neither is a plugin available for bzr that would output something the script can handle (or directly to the XML format as produced by the converter script)… if anyone wants to create one… 🙂
So, if you are using bzr it’s probably best to convert to something supported (beware bzr fast-export|git fast-import is broken if you have tags that include :, such as used to mark an epoch in debian packaging). I used hg (aka Mercurial) since it got a halfway decent converter plugin that converts from about everything to hg (also you get additional goody magic, see below).

Now just take a look at the data directory and tweak stuff to your needs. Execute run.sh and you should get a live swarm in a window.
If you set

TakeSnapshots=true

in the config you will get loads of png files, those you can abuse to render a video (with mencoder for example).

—

If you want it a bit more automized…http://rakjar.de/shared_codeswarm/readme.html
With that fancy script you bascially just need one config fie with loads of hg convert lines, then run script and you’ll end up with a ready to go rendered avi file.
The config might look like this:

`–> cat simple.config

banshee: hg convert /home/me/src/git/banshee banshee

amarok: hg convert /home/me/src/git/amarok amarok

It executes the given command and hopes to end up with a directory named like the entry, then gets an hg log from that, parses it, combines it with the other entries in the config to end up with a shared codeswarm. Then it just executes the codeswarm app with an autogenerated config, mencodes the resulting pngs and plays that back in mplayer.

I hope you understand that I don’t give a step to step tutorial, the documentation is all there and….one remark: Since codeswarm is javaware and community_codeswarm is pyware you are going to have to do fiddling to get it work properly 😛

Have fun swarming 🙂