Random Ramblings

What Hurts

Thu, 12 Nov 2009 05:32:42 GMT

What hurts
is not
the remembering.

The remembering
is a joyous,
momentous occasion.

What hurts
is the realization
that I had forgotten.

あふれる

Sat, 05 Sep 2009 01:24:19 GMT

Don't say I never gave you nothing.

Ophelia was the sweetheart to a nation overnight

Fri, 30 Jan 2009 08:55:44 GMT

I've had time to review more of the FUDCon 11 videos, specifically the KVM, Moksha, and Desktop videos. The KVM and Moksha videos pretty much only had me saying "neat", not that I thought they were boring, just that I saw nothing to comment further on; I think both are great ideas and I'm sure that both are only at the very beginning of their potential. This then leaves the Desktop video. But I'd like to digress before I give my thoughts about it.

I first heard of Linux through a classmate, via RHL 5.1. I dabbled with it through 7.2, then drifted away. I'm not sure what caused the drift. I think it was games. Anyway, I used Windows up until Windows 2000, when I decided to take a look at this new-fangled thing called "Fedora". I set up my system as dual-boot between Windows 2000 and Fedora Core 2. I didn't boot it into FC2 very much, simply because I already had Windows 2000 set up how I liked it, and to be completely honest, Windows 2000 wasn't terribly bad. In fact, I found that it suited my needs as a desktop well. The only thing I missed was multiple workspaces.

But blah blah blah, enough of that. How does that relate to the Desktop video? Long story short, I saw the GNOME 3.0 demo, and I don't get it. How does any of that help me get things done? Why should I worry about managing the number of workspaces I have? I'm not going to say anything about the launcher application selection since I'm sure it will go through one or more revisions before they come up with a sane scheme.

"Okay big talker, how do you think it should be done?" Well, we'll get to that in a second. First I want to show you an experiment that I'm undertaking, working within the confines of the existing technology:

As you can see, it's very similar to the GNOME default with a few significant changes:

I've told nautilus to not conrol the desktop
I've added a transparent panel to the right side of the screen
I've moved the launchers from the top panel to the top of the right panel
I've added the trash and drive mounter applets to the bottom of the right panel

This mostly works, except for a few issues with (auto)mounting, and with the clock applet not playing nice with the right panel.

What I envision though, goes even further:

A number of changes are in place here. The visible ones:

No more desktop

No more files, media, or icons on the desktop. Use a file manager window to get at them. Media and trash are now stuck to the right side of the display, extending up.

No more workspaces

Instead of a number of individual workspaces, there is only one solid desktop which is much larger than the display. The display need not snap to any boundary short of the actual edge of the desktop. A modifier is used in conjunction with the pointer in order to slide the display around the desktop.

Only one panel

Only one panel, period, along the top stuck to the top-right corner, resizeable on the left.

Detached menu and launchers, in their own space

Programs dragged from the menu can be dropped on the left hand side, which gives them their own launcher. Much like a panel, but transparent parts allow anything underneath them to be manipulated.

Of course, there are also a number of invisible differences as well:

No Minimize or Maximize functionality

Since there are no workspaces, neither option makes sense. Instead Lower (push this window to the bottom Z-level) and Fit (fill visible display) functionality will be used.

Applets on the panel will stick to a side

You will be able to rearrange applets as you like, but they will gravitate towards either the left or the right. If you need them separated from other applets, then... use a separator.

Moving the pointer past the edge of the display will not slide it

In this scheme the sides of the display are important, so you can't push the display just by moving to the edge.

So there you have it. I'm sure there's a few things I missed in there. As always, comments and criticism are welcome.

Gonna put some wheels in motion, get ready 'cause we're coming through

Thu, 22 Jan 2009 10:41:04 GMT

I just finished watching the XO/RPM FUDCon video, and here are my thoughts, in no particular order:

Packaging activities as rpms is a lost cause; there are almost none important enough to require root capabilities to install. Package known-good versions of Browse and Terminal with Sugar, and move on.
Activities are exactly like DVCS repos:

The master is the versions of the activity from upstream. 1.0, 1.1, 2.0, etc. All versions would be contained within it.
User-modified versions are branches, each based on a specific version; Kim's is a UI hack based on 1.0, while Miguel's adds additional transports and is based off of 1.1.
Branches are separate from the trunk; both the trunk and each of the branches can be moved around independently, and the branches depend on the trunk in order to function properly. It will be the job of rainbow or whatnot to reintegrate the activity so that it can be run.
The icon on the Activity view is the normal icon for the activity, and represents the latest version of the master. It can be right-clicked to either rebase the existing activity to a user branch (or the master as the case may be), or add another icon representing the user branch. User branches would be indicated with the homunculus of the creator as an emblem.

Activities can still be stored in a central place, perhaps mediated by an "Activity Manager" daemon or process.

Comments and criticism welcome.

Happy new year-number!

Thu, 01 Jan 2009 05:04:37 GMT

Every second is a new year; a chance for new opportunities and new successes.

And just so that this post doesn't become completely touchy-feely, here you go.

One direction, only way to go...

Fri, 05 Dec 2008 01:18:55 GMT

Whew! I finally got Python 2.6 into Rawhide. I'd like to thank the following people for their invaluable help:

Jesse Keating
Toshio Kuratomi
Alex Lancaster
Panu Matilainen
Rex Dieter

And of course, every Fedora Contributor that maintains or uses a package in Fedora that uses Python in one fashion or another, since your kind assistance will be needed in order to complete the integration of this new version.

I want you stronger, I want you harder

Tue, 25 Nov 2008 00:49:22 GMT

Fedora 10 on the XO

blah blah normal Fedora 10 on the XO on a USB drive. Here's how:

Find a machine that can boot USB devices and that you can install Fedora with.
Get a drive big enough. In my case it's a 40GB.
Install Fedora 10 i386 normally, with these options:

Separate /boot as the first primary partition
Install grub to /sdX1, not the MBR

Boot off the USB drive.
Install the i586 kernel. Use yum shell if you need to remove the i686 kernel at the same time:

# yum shell
> remove kernel.i686
> install kernel.i586
> run

Put this xorg.conf in /etc/X11.
Create the /boot/security directory and put your developer key in it. (optional)
Create the /boot/boot directory and put this olpc.fth in it.
Open /boot/grub/grub.conf and /boot/boot/olpc.fth and do the following:

Replace $args at the bottom of olpc.fth with the i586 kernel arguments from grub.conf.
Replace vmlinuz0 and initrd0.img with the filenames of the kernel and the initrd respectively.

You may be able to skip the kernel-swapping step by passing i586 at the install boot prompt. I did not try this.

Python 2.6

Preparations are now underway to get Python 2.6 into Fedora. Watch the skies in the next few days.

Change of Plans

Thu, 06 Nov 2008 16:56:06 GMT

I had originally agreed to give a presentation about packaging a part of the Fedora Classroom series. As time passed I realized 2 things:

There isn't enough bandwidth in IRC to deal with the intricacies of editing/completing spec files, and I don't believe making our gobby server open to the public would be a good idea.
I've already written a very goo (if incomplete) article that covers most of packaging.

So instead of talking about packaging, I'm going to be talking about packages. How to identify them, how to use them, and so on. The talk will be a little more disjoint due to the running around that needs to be done, but it's better than me just typing the article out in IRC.

So you'll still have to read the Building Packages Guide if you want to learn about packaging (and I urge you to read it), but I'll be more than happy to answer any questions you have about it during or after the session, with the exceptions of "why is it still a draft" and "when will you finish it" (because, and whenever :P ).

How to Destroy Microsoft in Three (Mostly) Easy Steps

Sat, 18 Oct 2008 22:05:45 GMT

Step the First: Write an Application for Windows

Actually, applications. Lots of them. Use as many varied core Windows technologies as you can. COM, ActiveX, DirectX, you name it.

Step the Second: Gain Popularity

Get people to like and use your applications. Lots of people. People with money. People with power.

Step the Third: Keep an Eye on New Windows Technologies

And then promptly ignore them. .NET? Who cares. WPF? Passing fad. Force Microsoft to spend billions of dollars to keep your apps and their 20-year old API calls working. The building pile of legacy in their codebase will compound and eventually overwhelm them.

I'm getting played by you, déjà vu...

Sun, 28 Sep 2008 08:27:16 GMT

Yes, I'm a lazy git, but here it is.

Crystal, blister, it's all over now...

Mon, 28 Jul 2008 17:21:28 GMT

jdahlin has officially fired the PyGTK salvo for Python 3000, so I took the liberty of updating the package to beta 2. Enjoy.

Everyone Loves Web Interfaces

Thu, 12 Jun 2008 01:07:04 GMT

Version 0.1 of fas-planet is now available. It needs the patches from FAS tickets 57 and 59 in order to work. Feel free to tear up the code and tell me what's wrong. I know I'm missing a lot of i18n, and the error-handling is sub-par, but I think I got most of it right.

And for the non-techies, here's an action shot:

The planet awakens...

Fri, 06 Jun 2008 02:02:41 GMT

You're absolutely right Alex; it sucks. It sucks, it sucks, it sucks. Now, having said that, it is better (for some people, at least) than the scheme we had before, which is to say... no scheme at all.

In fact, I would go so far as to say that what we had was an antischeme. Sure, all you had to do was send your feed and hackergotchi URLs to someone and they would add it to their text file. While trivial for non-technical people to use, it has a couple of flaws:

It requires a person on the other side to do some work. If they're busy or on vacation, you're not getting added to the Planet.
It requires said person to maintain a text file containing a rather large number of entries. The tedium is astounding.

Now, you could throw more people at the task, but that has diminishing returns beyond approximately 3 people (No, I have no scientific proof for that "fact". Read some Brooks if you want more details about my reasoning.).

But the simple truth is that we already have this handy web app that we use to handle our accounts. Why not let it handle the necessary data, while still allowing a self-serve scheme? Let it collect the data, and store it, and process it, and fold, spindle, and mutilate it. After all, that's what machines are good at (heck, that's the only thing they're good at). FAS2 already has a plugin system (although I don't believe that it's been battle-tested), so now we just need someone to actually sit down and write it.

Needed, actually. I've decided to sit down and write it myself. I can't guarantee that it will appear in a timely fashion seeing as to how much I have to learn/figure out, but, it'll be there. And hopefully far less sucky.

A long-overdue update

Fri, 30 May 2008 10:07:58 GMT

Get it while it's hot!

Blah blah packaging bugs blah blah.

A preview release of my own

Sun, 20 Apr 2008 03:24:08 GMT

Enjoy.

As always, packaging bug reports are welcome.

O_o What? ... Oh fine...

Thu, 10 Apr 2008 18:02:45 GMT

[ignacio@ignacio ~]$ history | awk '{a[$2]++ } END{for(i in a){print a[i] " " i}}'|sort -rn|head
149 cd
99 ls
68 git
62 rpm
58 man
54 less
42 sudo
37 make
35 df
32 find

[ignacio@localhost ~]$ history | awk '{a[$2]++ } END{for(i in a){print a[i] " " i}}'|sort -rn|head
135 ls
90 rpmbuild
90 cd
82 sudo
59 less
57 vi
50 make
45 rpm
42 mock
33 rm

No, I still haven't named my laptop yet.

Security-enhanced Python

Sun, 16 Mar 2008 04:58:10 GMT

Time to put this down, now that I've had some time to think it over.

Imagine names in Python treated as files in SELinux. Yeah, that's SEPython.

More details as I actually get some of it done.

My Landlord Must Die

Wed, 27 Feb 2008 01:01:09 GMT

So, I'm moving. Or rather, I'm supposed to be.

There aren't many suitable apartments around here at the price I'm looking for. I thought I had found a few decent choices, but they need a reference from my current landlord. The same landlord that only shows up once or twice a week. But I digress. I filled out all the information as far as I knew it, including my landlord's phone number. All good.

Fast forward a few days.

They've all attempted to verify my reference information. And lo and behold, the person at the phone number isn't my landlord. It's someone who's never heard of me, and is certainly not renting out an apartment at my location. Oh joy.

He showed up tonight, and I asked him what his phone number was. His response? "I don't have a phone." Of course, I know that he's LYING since I doubt he's talking to himself upstairs right now. And I can't just go "Hey F*CKFACE, what's your F*CKING PHONE NUMBER?!", for as much has he may deserve it.

So now I'm going to pay him half the rent twice a month until I get out of here. If he dares to refuse those terms then I'm just going to remind him of the apartments I was SO close to having, but lost because of his idiocy.

git Madness!

Tue, 05 Feb 2008 18:05:34 GMT

This one's for you, Mike:

git pull
git branch -r live
git rebase master live
git push

And it further (only slightly-related) news, Fedora 9 "Sulphur" Alpha is out.

Drinking in L.A.

Mon, 04 Feb 2008 08:03:23 GMT

I finally got around to doing an update to my ~~Bran Van~~ Python 3000 package. It should now be mostly 64-bit clean.

C extensions will need some changing before they'll build. I'll let GeoIP-Python-1.2.1-py3000.patch from the python3000-GeoIP SRPM do the talking:

diff -up GeoIP-Python-1.2.1/py_GeoIP.c.py3000 GeoIP-Python-1.2.1/py_GeoIP.c
--- GeoIP-Python-1.2.1/py_GeoIP.c.py3000	2008-02-04 02:21:54.000000000 -0500
+++ GeoIP-Python-1.2.1/py_GeoIP.c	2008-02-04 02:22:27.000000000 -0500
@@ -22,7 +22,7 @@
 #include "GeoIP.h"
 #include "GeoIPCity.h"
 
-staticforward PyTypeObject GeoIP_GeoIPType;
+static PyTypeObject GeoIP_GeoIPType;
 
 typedef struct {
   PyObject_HEAD;
@@ -297,11 +297,16 @@ static PyMethodDef GeoIP_Class_methods[]
   {NULL, NULL, 0, NULL}
 };
 
-DL_EXPORT(void)
+#ifndef PyMODINIT_FUNC	/* declarations for DLL import/export */
+#define PyMODINIT_FUNC void
+#endif
+PyMODINIT_FUNC
 initGeoIP(void) 
 {
   PyObject *m, *d, *tmp;
-  GeoIP_GeoIPType.ob_type = &PyType_Type;
+  GeoIP_GeoIPType.tp_new = PyType_GenericNew;
+  if (PyType_Ready(&GeoIP_GeoIPType) < 0)
+    return;
 
   m = Py_InitModule("GeoIP", GeoIP_Class_methods);
   d = PyModule_GetDict(m);
@@ -328,11 +333,11 @@ initGeoIP(void) 
   PyDict_SetItemString(d, "country_names", cname);
   PyDict_SetItemString(d, "country_continents", ccont);
 
-  tmp = PyInt_FromLong(0);
+  tmp = PyLong_FromLong(0);
   PyDict_SetItemString(d, "GEOIP_STANDARD", tmp);
   Py_DECREF(tmp);
 
-  tmp = PyInt_FromLong(1);
+  tmp = PyLong_FromLong(1);
   PyDict_SetItemString(d, "GEOIP_MEMORY_CACHE", tmp);
   Py_DECREF(tmp);
 }

Fontastic!

Sun, 13 Jan 2008 09:00:27 GMT

Wheeeee!

Next up: Convincing someone to add a checkbox that reads "Only show glyphs found in this font".

Waking up the Neighbors

Thu, 20 Dec 2007 17:54:57 GMT

It seems that my connection having gone up and down like a yoyo last week has had good consequences. I think someone at my ISP noticed the problem and sent a technician out to the CO to figure out what was going on. Anyways, long story short, my connection not only got unf*cked-up, but I'm now getting speeds closer to what I'm paying for (but not getting due to distance issues). So a huge thank you to whoever at Sympatico noticed and fixed this.

Acer TravelMate 5720-6410

Tue, 11 Dec 2007 08:23:26 GMT

I decided to pick up an Acer TravelMate 5720-6410 and install Fedora 8 x86_64 on it. Here's a summarized report on it:

Things that work:

Video

The Intel X3100 works fine, DRI included.

CPU

LM, VMX, speed-stepping, dual-core, all good.

Networking

Both wired (tg3) and wireless (iwl3945) work, excepting NM bugs.

Things that mostly work:

Audio

Audio playback and recording work, but only via the built-in speakers and mic. The front audio jacks can't seem to be selected as input or output.

Keyboard

Most keys work. A few keystrokes (as well as the front-panel killswitches) give unknown keycodes. hal-info doesn't have all the keystrokes or models yet obviously. Ctrl_L+Shift_L doesn't always seem to "register". Changing the backlight brightness via keyboard doesn't show the brightness gauge on the display.

Webcam

Works with the UVC driver, but you'll want a small gstreamer patch for, say, cheese.

Backlight Control

gnome-power-manager doesn't seem to be controlling the backlight completely properly. I don't know what's going wrong, but it's irritating.

Media Reader

I only have SD to test with. No SDHC, MMC, MS, MS Pro, or xD.

HVM

Virtualized OSes boot, but suffer from various strange bugs (panics during DHCP, etc.).

Things that don't work:

Fingerprint Scanner

Oh well. I'm not terribly worried, but I wouldn't mind seeing it supported.

Front-panel Indicators

Neither the Bluetooth nor wireless indicators light up, regardless of Bluetooth or wireless status.

Bluetooth Killswitch

Okay, maybe this shouldn't be working per se seeing as how this submodel doesn't have built-in Bluetooth, but it would be nice if it did.

Things I didn't test/haven't tested yet:

DVD Burner

Just haven't got around to it yet.

External video

I tried neither VGA Out nor S-Video Out.

FireWire

I don't have any FW peripherals to try out.

IrDA

I have an IrDA device around here somewhere...

PSTN Modem

Mowhat?

Anyways, if anyone has any specific tests they want me to run on this machine then just let me know.

And now for the gory details:

00:00.0 Host bridge: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub (rev 03)
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)
00:1a.0 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Contoller #4 (rev 03)
00:1a.1 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Controller #5 (rev 03)
00:1a.7 USB Controller: Intel Corporation 82801H (ICH8 Family) USB2 EHCI Controller #2 (rev 03)
00:1b.0 Audio device: Intel Corporation 82801H (ICH8 Family) HD Audio Controller (rev 03)
00:1c.0 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI Express Port 1 (rev 03)
00:1c.1 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI Express Port 2 (rev 03)
00:1c.2 PCI bridge: Intel Corporation 82801H (ICH8 Family) PCI Express Port 3 (rev 03)
00:1d.0 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Controller #1 (rev 03)
00:1d.1 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Controller #2 (rev 03)
00:1d.2 USB Controller: Intel Corporation 82801H (ICH8 Family) USB UHCI Controller #3 (rev 03)
00:1d.7 USB Controller: Intel Corporation 82801H (ICH8 Family) USB2 EHCI Controller #1 (rev 03)
00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev f3)
00:1f.0 ISA bridge: Intel Corporation 82801HEM (ICH8M) LPC Interface Controller (rev 03)
00:1f.1 IDE interface: Intel Corporation 82801HBM/HEM (ICH8M/ICH8M-E) IDE Controller (rev 03)
00:1f.2 SATA controller: Intel Corporation 82801HBM/HEM (ICH8M/ICH8M-E) SATA AHCI Controller (rev 03)
00:1f.3 SMBus: Intel Corporation 82801H (ICH8 Family) SMBus Controller (rev 03)
02:00.0 Ethernet controller: Broadcom Corporation NetLink BCM5787M Gigabit Ethernet PCI Express (rev 02)
04:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02)
0f:06.0 CardBus bridge: Texas Instruments PCIxx12 Cardbus Controller
0f:06.1 FireWire (IEEE 1394): Texas Instruments PCIxx12 OHCI Compliant IEEE 1394 Host Controller
0f:06.2 Mass storage controller: Texas Instruments 5-in-1 Multimedia Card Reader (SD/MMC/MS/MS PRO/xD)
0f:06.3 Generic system peripheral [0805]: Texas Instruments PCIxx12 SDA Standard Compliant SD Host Controller

Bus 002 Device 001: ID 0000:0000  
Bus 007 Device 001: ID 0000:0000  
Bus 006 Device 001: ID 0000:0000  
Bus 005 Device 001: ID 0000:0000  
Bus 001 Device 004: ID 064e:a101 Suyin Corp. 
Bus 001 Device 001: ID 0000:0000  
Bus 004 Device 001: ID 0000:0000  
Bus 003 Device 003: ID 08ff:1600 AuthenTec, Inc. 
Bus 003 Device 001: ID 0000:0000

Living on Video

Tue, 20 Nov 2007 15:40:54 GMT

The state of TV playback in Fedora with an ivtv-based card is abysmal. There's quite a few apps for it, but they all have problems of some sort:

tvtime runs, but doesn't have MPEG decoding support and can't do anything with the tuner
VLC can view, but can't tune
mplayer needs all sorts of magic command-line options, and can't change the tuner (and gmplayer is no help either)
xawtv is all sorts of fail (but at least v4lctl is useful)

Not one to be easily daunted, I went and looked at other options:

zapping works, but is buggier than a zero-star motel
kdetv seems resistant to having people actually use it
I couldn't for the life of me figure out how to get fftv to work

So at the moment I'm using a combination of VLC, zapping, and v4lctl to watch TV on my machine. Time to go find out what work elisa needs in order to get TV playback working...

Legacy media Fails again

Tue, 13 Nov 2007 15:46:36 GMT

Every Monday evening from 8pm to 11pm Eastern I watch Chuck, Heroes, and Eureka. I normally watch them religiously because, well... that's the kind of person I am. However, last night I was exhausted and so fell asleep part way through Heroes. I figured that wouldn't be a problem since you can watch Heroes episodes online and the station that shows Eureka here doesn't have a bug up their ass about only showing it once a week (Okay, only thrice. But that's still better than once).

This morning I decided to watch Heroes, so I went to their site. Pick the episode... pick the scene...

"We're sorry, but the clip you[sic] selected isn't available from your location. Please select another clip."

Oh fun. That's okay, I figured I'd just go to the website that they display during the airing of the show.

So I get there, and find the appropriate page. It says "looking up player . . ." which is... strange. I give it a few minutes just in case it really *is* loading the player and I'm just being impatient.

A few minutes later, still nothing. At this point I know something's wrong. I start up Firebug and reload the page.

2 Errors

sysReqContainer is not defined
sysReqContainer is not defined

Long story short, the JavaScript on the page will only load the player if you're running Windows or OS X, and the site designer forgot to add the box that contains the message that comes up if you're running anything else.

At this point I had 3 options:

1) Wait until the end of the season and buy the DVD set
2) Wait several years for reruns
3) Download an illicit copy of the show

Can you guess which 2 options I didn't take? Thanks, NBC and Global Television, for turning me into a fugitive. Jackasses.