Older blog entries for will (starting at number 9)

Well today I managed to dodge sleepiness to get through an exam. *sigh*

I've been charged with getting the XFree86 4.2.0 packages for FreeBSD in line for the upcoming 5.0-DP1, for which the final package build is supposed to start in about 4 hours from now. Sigh. Just a few more issues to take care of and it'd be all ready. Now if only KDE would build. Silly binutils... 8)

Speaking of KDE.. I've also been charged with getting 3.0 packages ready for FreeBSD for tomorrow's release. I finally fixed a few more bugs in the build system and it's now doing kdelibs... properly, I think. We'll see how far it gets after this round.

I've finally updated my entry on Advogato at Leimy's badgering. Things have changed quite a lot since I last posted here, so I'll just let the summary speak for itself. Suffice it to say there's way too much to go over so I'm not gonna bother, and will just post more stuff here for now.

I'm still Vice President at the Purdue IEEE Computer Society, which hosts my KDE/FreeBSD Project homepage and its mailing lists.

Well, it's been a few days since my last diary entry, so I thought I'd give it another shot. We had a Purdue University IEEE Computer Society meeting in Purdue University Materials Science and Electrical Engineering building room B012 on Thursday where a few guys (and one woman :-) from IBM's S/390 Server group came and gave a presentation on their new server series. It was pretty interesting: their "Z Series" servers are supposed to have "zero downtime" (which is why it's called "Z Series"). They showed a typical IBM server of today with 12 CPUs being slightly redesigned to do real-time intimate server upgrades to the OS, BIOS, microcode, and other things. It was pretty cool, and I got free pizza out of the deal too.

Oh, that reminds me. I'm now one of the Vice Presidents of the Computer Society at Purdue University as well. Woot. ;)

I did some more work on OpenPackages' make(1) and sent a diff to Marc Espie. Still waiting for a reply. :)

I had a very bad week last week mostly because most nights I couldn't get to bed until 3 or 4am, and when you consider that on two mornings I had 7:30am classes... well, they both got skipped. And because I usually went to work right after these classes.. well, I skipped work too. :(

So I ended up sleeping a good portion of my weekend. I went to bed around 4am on Saturday morning and didn't get up until 1pm, and a few hours later took a short nap. Then I didn't get back to sleep until 6am today. :-(

Woke up at 3pm, did laundry & homework, and hacked some more. Then I went and bought 14" meat pizza, 7up, hawaiian punch, and 3/4 lb of sour candy. I think I've already been suffering from "Freshman 15" for a few months now, so... I really need to get back in shape. :-)

Later, freaks. Go wank off at amihotornot.com. :D

Whee.. another day goes by. I went to an interview with a company that designs and builds aviation equipment as well as a host of other cool stuff (like GPS's and embedded systems). I think it went well. :)

Tried to work some more on OpenPackages (OP for short) make(1). What I'm trying to do now involves porting the ODE make modifiers that NetBSD put into their make(1) implementation. Marc Espie did a significant rewrite of OpenBSD's make(1) that involves huge amounts of code breakup and simplification. He also did a lot of work towards making make(1) more standards-compliant. Now I'll be specific about what exactly I'm trying to implement in this make(1): a concept coined by Marc called "personalities". Basically, what we would do is this: if we encounter differences in traditional BSD make's behavior regarding certain things (for OpenPackages' purpose, variable modifiers), we will interpret it differently based on whether a certain personality switch was turned on or not. This allows us great flexibility in what kind of behaviors OP make(1) can emulate. It also comes fairly cheap in terms of lines of code and binary data.

Other things this "personality" idea could be used for are GNU make compatibility bits. For OP's purpose, this would allow fewer pkgsrc bits to need GNU make in order to build. The logic behind this is that a significant number of Makefiles use only small bits of code that are specific to GNU make.

By the way, it would be really nice if some more people could start participating in OpenPackages discussions, because the idea has a lot of merit to it. There just needs to be more eyes looking and more hands typing comments. :-)

Oh yes, and more coders would be nice. :>

That should do it for today, I'm tired and I have to get up early. sigh.

Well, I guess I'm back after almost a year of absence from advogato. I don't really know why I remembered my password after being gone that long. But I did...

I guess an update is in order. The 'addport' script I was talking about back in April has been developed to maturity and has many features over its predecessor. For one thing, you can autogenerate a commit message, which saves a *lot* of time when adding new ports to the tree.

I'm still an active FreeBSD developer (and more than I ever was before getting a commit bit). I've mentored six new ports team members. Some of them have been quite active lately (Go James, Mario, Trevor! :-).

I graduated from high school with a cum laude diploma. Parents & most of my family were very proud; open house, etc. I earned the Eagle Scout award in July 2000, and they decided to turn the open house into a reception for that. Ooh boy. Life is fun. ;)

I was accepted to Purdue University (West Lafayette, IN) and Rochester Institute of Technology (Rochester, NY). I chose Purdue. I'm still there; currently taking Calculus III, Electricity & Optics, Psychology, Communications, and C Programming for Engineers (don't laugh :-). I'm pursuing a B.S. in Computer Engineering, and might be pursuing cooperative education. Co-Op seems like a great idea: you work every other semester at some large company on stuff related to your degree, and take classes on off semesters. Co-Op folks get paid pretty well, and job security is good (plus experience). So I like it. I might do it.

Getting back to my Cyberlife (tm).. these days I've been working on make(1) in the FreeBSD tree. I recently fixed a number of bugs in the tree, and merged them into the 4.x-STABLE and 3.x-STABLE trees.

I'm also working on OpenPackages, a project that intends to replace the ports/pkgsrc systems in use by FreeBSD, NetBSD, and OpenBSD. We also have some Darwin people interested in using our project, as well as folks who use other operating systems (e.g. Linux, BeOS, Solaris, HP-UX, and so on). I think our first focus will be to unify Free/Net/Open pkgsrc systems (we decided to call it pkgsrc from now on, to distinguish from architecture ports and such), and later we'll provide easy methods for adopting to other operating systems. And of course it'll all be under the BSD license (a topic I and others have debated somewhat on op-tech@openpackages.org - check out the archives).

Anyways, that's enough for now. Take care people.

I have just imported my FreeBSD port for SDLQuake, the SDL version of Quake, into the FreeBSD ports tree.

The initial reaction for me seemed to be "How the heck does this thing compile and link so fast, and how the heck are these graphics so beautifully and smoothly rendered??!!"

So I played for awhile. I was filled with nostalgia, remembering where the secrets were that I found three years ago. And I had a lot of fun!

addport was used for the importing of the sdlquake port. addport is fairly well-done now, and can probably be used actively soon.

Ok, it seems our little addport script is working. mharo added the misc/boxes port with it, and I just (a few hours ago) added ports such as textproc/xlhtml and www/kwebsearch. From the looks of things, all three were successful, with very minor (and rather invisible) problems.

We may be looking at a substitute for easy-import for FreeBSD committers, very soon now!

Although, after we post these scripts, it would be much appreciated if (somehow) the scripts might be useful for other projects, if the people who use it from that project send us email suggesting useful modifications.

Now to add more ports to the tree... :-)

--will

mharo and I have gotten some more work done now. We've got a perl script that adds an entry to CVSROOT/modules and another perl script that adds the port to its category Makefile.

I'm currently working on a remote script 'addport' which will check ports as needed and do the cvs commits as well as execute the above-named scripts to add the ports to the ports tree.

Hopefully, with any luck, we'll be done with this and have a fairly polished addport suite by next friday.

The next job after that will be creating a suite of tools to assist in making new ports. I wonder how we'll decide to do that in the end...

The last few days have been a little busy. I spoke with mharo about a particular script. We're working on improving the easy-import script that is currently available on freefall.FreeBSD.org. Actually, we're working on a replacement. It is hoped that by next week FreeBSD developers will be able to import new ports remotely. We will be using a series of perl scripts to accomplish this. More to come later.

I'm hoping to help Michael Haro <mharo@FreeBSD.org> write some scripts to help make port maintaining, committing, writing, etc. easier.. anyone got ideas? :-)

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!