I might have to change editors now…

  • August 31, 2006
  • Jasper Huijsmans

The biggest gripe I’ve had with emacs — alright, besides the finger-breaking key combo’s ;-) — is that it is just plain ugly. It just doesn’t fit with my beautiful gtk desktop with anti-aliased truetype fonts. Gvim got this right a long time ago.

But, no longer does this have to be true:

Emacs-gtk with xft support.

I already knew about the gtk interface, but I never heard of the xft support. I got the instructions from EmacsWiki:

$ cvs -z3 -d:pserver:anonymous@cvs.savannah.gnu.org:/cvsroot/emacs co -remacs-unicode-2 emacs
$ cd emacs
$ ./configure --with-gtk --enable-font-backend --with-xft
$ make bootstrap && make && sudo make install
$ emacs --enable-font-backend --font "Bitstream Vera Sans Mono-10"

nvidia drivers

  • August 28, 2006
  • Brian Tarricone

I just noticed that some people are working on reverse-engineering nvidia's graphics chipsets. Sweet. Maybe we can someday get a decent 3D- (and 2D-) accelerated open source video driver.

Maynard Ferguson Passes

  • August 25, 2006
  • Brian Tarricone

Just read that Maynard Ferguson passed away yesterday. I wasn't a huge fan of his playing style, but nonetheless I have to respect his skill. I heard him play live sometime in HS, and I was quite impressed. Shame...

More 4.6 panel thoughts

  • August 24, 2006
  • Jasper Huijsmans

In this blog entry Aaron Seigo talks about Plasma, the panel/desktop framework for KDE4. The ideas about data engines and data visualizations are really interesting.

I have been thinking about this for Xfce as well: it would be really cool if we could provide plugin writers with data sources and display widgets to easily write status monitors for the panel.

The idea I had is that the display widgets could be in-process, so it would be an internal plugin, but the data to display would be obtained and processed by an external process, through a DBUS protocol or whatever. This has the shared advantage of in-process widgets, which are a lot easier to handle (no more XEMBED, yay), and out-of-process data handling (no GUI blocking or crashing of the panel).

Of course we really should get Xfce 4.4 out before even considering the possibilities for 4.6…

Maybe something to look at for Xfce 4.6?

  • August 22, 2006
  • Jasper Huijsmans

Here’s a post by Ryan Lortie, who has been working on a (possible) new API for GNOME applets. The work was done as part of the Google Summer of Code and it looks very interesting.

It would be kinda cool if we could support the same API for Xfce 4.6. Most of the GNOME dependencies seem to have been dropped… or maybe, by the time we release 4.6, we simply switch to using the GNOME panel, who knows?

Or maybe I should be nice to our plugin writers and not change the panel API every major release :-)

Cognitive Dissonance

  • August 17, 2006
  • Brian Tarricone

While I generally consider myself a liberal (socially at least, even if I'm generally financially conservative), I enjoy reading about the opposition. Unfortunately, I get the impression that the opposing viewpoints will just never find compromise.

Take this article, for example. Summary: "liberals do not appreciate the seriousness of the terrorism threat". And maybe we don't. Sure, 9/11 was a terrible occurrence. But if you consider lives lost to terrorist attacks in the United States over the past 50 years (or whatever), it's pretty insignificant. Somewhat ironically, we might find that the lives lost in our own retaliation for terrorist attacks (e.g., Afghanistan) might rival or even eclipse that number. (I'm not saying that's the case; I've done no research on this topic.)

This article seems to miss the point on several of its arguments, however. I don't view Iran as a terrorist threat. Sure, their government has a nuclear weapons program. But that has very little to do with terrorism, unless, of course, Iran plans to arm terrorist groups, an accusation for which I believe there is no evidence. Iraq also has nothing to do with terrorism. Saying you are against the war in Iraq has nothing to do with your stance on fighting terrorism.

So all that's left is the attempt to make fun of the liberal mindset on preserving due process and civil liberties:

Liberals: Stop the illegal wiretapping and torture! Conservatives: Hah! Stupid bastards, we just prevented a terrorist attack based on what you call illegal wiretapping and torture!

Maybe liberals are just idealists? Maybe some liberals are willing to allow a couple attacks to succeed if the only way to stop them involves an unacceptable erosion of rights? It's hard to draw the line. What if I were a passenger on one of the planes earmarked for destruction? Could I honestly say I'd rather die than to be preemptively saved through means I find reprehensible? I'm not sure I could. Would I, as a potential policy-maker, be willing to make that decision for others? I don't know.

I'm often torn on the point of emotional detachment. On one hand, I feel like policy decisions should be made without regard to anything but the cold hard facts. Is A, B, and C an acceptable consequence of doing D and E, based solely on numbers and an objectively measured effect? In this case, are lost lives worth the preservation of what I consider essential liberties: reasonable expectations of privacy and due process? For me, the answer is a resounding yes.

But on the other hand, there's always a human element. We are emotional creatures. An objective, emotionless response to a problem may seem neat and tidy from a logical perspective, but ignoring how people (including yourself) feel about any particular decision doesn't make sense from a human perspective. People will have an opinion -- often a stong emotionally-driven opinion -- on any decision that has the potential to affect lives. Even if I can say that yes, I'd be willing to sacrifice lives for an ideal, that's too abstract. What if one of those lives turns out to be someone I care about? But does that even matter? Greater good and all that?

I'm not sure where else I'm trying to go with this. Maybe nowhere.

They’ve Already Won

  • August 15, 2006
  • Brian Tarricone

I've said it before, and I'll say it again: the terrorists have already won. Apparently now it's a terrorist act to buy a large quantity of cell phones. Oh, and it helps if your ethnic background is Middle-Eastern as well. Such bullshit.

Drunk and Sober

  • August 7, 2006
  • Brian Tarricone

The BBC has a couple amusing articles on the five stages of drunkenness, followed by the five stages of sobering up. Funny stuff.

Misc Stuff

  • August 4, 2006
  • Brian Tarricone

I'm currently working on a new, more minimalist design for my website. It's not finished yet, but you can preview it here.

I actually did a little hacking on Xfmedia earlier this week. Unfortunately I find myself forgetting fine details of the direction I wanted to take with the reorg, so I feel like it'll be a while before I really jump into it hardcore again. On the other hand, I'm having second thoughts about parts of that document, so maybe it's a good thing that I'm re-evaluating and possibly redesigning some portions.

I attempted to update Xfce Bugzilla to 2.22.0 yesterday, but running checksetup.pl failed with some weird error about the parameter 'languages' not existing. I messed with it for a little while, but ended up downgrading back to the latest stable in the 2.20 series.

I'm also working on a Xfce debugging guide for people with crashes or unexplained behavior who want to help out. I'm seriously getting tired of answering the same "how do I get a backtrace?" questions. I'm also tired of people posting straces in bug reports. While an strace might be the easiest thing to get, and it may look impressive in the amount of output it generates, in my experience it's useless in fixing most crash bugs. A stack dump at the crash from the debugger is really the best way to go. Bonus points if it's done on a debug build with symbols unstripped and no compiler optimisations. Double bonus points if the supporting libraries have debugging information too.

Anyway, this guide isn't finished yet, but I'd appreciate some feedback. Already I see it's gotten way too wordy and long. Probably some parts of it should be broken down into concrete steps instead of paragraphs, and maybe it should be separated out into two separate guides, one being more of a FAQ to help figure out what part of Xfce crashed, and the other more of a generic how-to on gdb. Not sure yet. As I said, comments (either here or via email) are appreciated.