Older blog entries for bcully (starting at number 11)

ices 0.2.1 is out. Now it runs on Solaris and the BSDs. None of those bugs were My Fault (tm).

My s3kr3t radio station is back on line in its new colo home, and I'm happy. I missed it.

I'm beginning to get pretty depressed about how hard it is to find decent work right now. I'm running out of time, too...

Yay! That !@#$ beam-it story has finally been bumped off the main page. Thank you, new stories.

I've added VVV's account-hook to mutt, working with multiple IMAP accounts is not so bad now.

I should be releasing ices 0.2 (first in the development branch) today. Chock full of new features, including the multiple streams and vorbis reencoding...

I've checked in support for vorbis source files in ices. I'm in time for Ogg/Vorbis 1.0rc1 too :)

Now to get back to mutt hacking...

ices 0.1.0 is out, which means immediately afterwards I merged in the multiple stream code. now to do error-handling. The icecast project is really fun, I'm enjoying working with jack and michael too.

I really have to get a job soon...

Well, I'm now the ices maintainer, so the multistreaming code will probably be official ices code some time next week. Before that I'm bringing the current CVS tree back to health. It compiles and runs again, the only thing left before I tag it and merge in multistreaming is to make it backwards compatible with the last stable release of LAME.

Well, mutt 1.3.19 is out the door, so I took a little break. That means hacking on something else, of course :)

This time it was icecast. I just taught ices how to reencode its stream at multiple bitrates/channels to multiple mountpoints. This is handy for web radio stations (like the one I'm developing in s3kr3t), since you can broadcast the same stream for broadband and dialup users.

I sent the patch to icecast@xiph.org, it's also here (gzipped).

nice day, maybe I'll go for a bike ride.

Mutt now has the last feature I plan to add before 1.4 (you can select your authentication method). For the next couple weeks I expect nothing but bug fixes. And I've closed about 25% of the open bugs in the last few days...

I had a look at computer jobs in Toronto on monster.com, and they are not attractive. I wonder where the good-looking jobs hang out...

I've been on a mutt tear - fixed a zillion little network bugs and worked on concurrent access to the same IMAP folder. Now

• it doesn't seem to crash
• if you change message flags in one session they get picked up automatically in the others

I don't know how many people will notice unless they're playing around, but it looks cool.

I've gotten almost all the features in that I want for 1.4 (still want to make authentication method configurable), after that it's pure bug-hunting. I think.

It looks like the driver API will have to wait. I raised the topic on mutt-dev and it looks like we're going to start the push for 1.4. I'll be doing bugfixes for the next couple of weeks - caught a nice bug with multiple sessions on the same mailbox today. I should work on what happens when you lose your connection now.

I've decided to take the plunge and move the libimap API into a driver model (aka a structure with a bunch of function pointers). I hope this will reduce the rampant #ifdef USE_IMAPs in the code and generally make things less incestuous. The dirty relationship between all the functions makes it hard to work on robustness - too many side effects. Anyway I've gotten just enough code done now that nothing compiles :)

Took an online pretest for a job in Toronto - about twenty really basic questions on C and java. and there I was getting nervous about it.