BM Systems

Caps Lock for Xmonad Modifier

2011-09-05T08:31:00.000-07:00

I like to configure Caps Lock as mod4, so that I can use it as my Xmonad modifier key. I noticed that mod4 also seems to act as Super_R, because pressing Caps Lock in my VirtualBox Windows 7 guest machine would bring up the Start menu. Thus, I made mod4 only react to Hyper_L and mapped Caps Lock to Hyper_L in my .Xmodmap:

clear lock

clear mod4

keysym Caps_Lock = Hyper_L

add mod4 = Hyper_L

This works in Slackware 13.37.

Workaround for Flash Suspend Resume Problems

2011-09-01T08:42:00.000-07:00

After resuming my computer after suspend, Flash videos in Chrome play at double the speed without any sound or sometimes don't play at all. Although I don't know how to fix this, I came up with a workaround to kill the flash plugin before suspending. This allows me to simply refresh the page to resume playing the Flash video!

On Slackware 13.37, I created a file called 00-kill-chrome-flash-plugin.hook in the /etc/pm/sleep.d directory with the following contents:

#!/bin/sh

lsof | grep chrome | grep snd | awk '{print $2}' | xargs kill

When you suspend your computer, you should see all the flash videos turn into frownie faces... :-) Of course, you could always terminate the plugin via the "View Background Pages" window.

Thanks to the posters on various Ubuntu and Arch Linux forums for the lsof and grepping for snd ideas.

Antec Solo Front Audio Interference

2010-07-18T11:35:00.000-07:00

DISCLAIMER: I do not know anything about electronics, so the solution I have implemented may be hazardous.

I discovered that the sound coming out of the front headphone jack contained a lot of squealing noise correlated to my machine's activity and mouse movement. A Silent PC Review (SPCR) thread pinpointed the problem to a double earthing of the headphone jack to the case and the motherboard. The difference in potential of these two grounds emerges as noise transmitted through the headphone jack.

In the SPCR thread, it looked like the easiest solution was to have Antec send a replacement front audio header panel. I sent Antec an e-mail, and they swiftly sent me a replacement part which I received about three weeks later. It looked like the promising new header did not have an extra ground wire:

I replaced the old header with the new header and found that none of the interference had gone away. However, by trial and error, I discovered that disconnecting the header's USB cable from the motherboard completely got rid of the noise! I decided to connect only the old header's audio cable to the motherboard's audio socket, and the new header's usb cable to the motherboard's usb port. No noise! Thus, it seemed like the new header's headphone jack was still grounded twice, and the solution was to somehow make the headphone jack independent from the USB ports.

I used an X-Acto knife to painstakingly chip away at the outer shell of the header, until I shockingly discovered that all the internal wires were immortalized in another plastic glob.
PLEASE USE GLOVES when doing this or else you will cut up your fingers. However, I did see signs of an exposed black wire that seemed to connect the USB and headphone jacks.

After hours of meticulous chiseling and painful cutting, I ended up destroying the header anyway, but I discovered what seemed to be the extra wire that had been discussed in the threads:

I started the whole cutting process again with the new header and disappointingly discovered that every mold is different. However, I was still able to identify the cable and cut it after a couple more hours of grueling cutting. It seems it is always a wire attached to the large soldering glob on the headphone jack as opposed to the four pins below the jack. Cutting too much near the three main wires, I had to do some clever case wiring to make sure the wires would not touch each other in the end.

When I cut up the new header, I strategically cut the plastic, so I could put it back together just like brand new:

Thanks to Antec for providing an extra part. I hope they will reexamine this issue and fix this solution across all their cases. I appreciate their courteous and swift support on the issue though.

UL30A Headphones Don't Mute Speakers

2010-05-05T00:09:00.000-07:00

Following these instructions, create /etc/modprobe.conf and add:

options snd-hda-intel model=auto

You can then do the following to avoid restarting after quitting all the applications that use sound:

modprobe -r snd-hda-intel && modprobe snd-hda-intel

Now your headphones and speakers won't be independent on your ASUS UL30a!

Here is my aplay -l output just in case there are variations of the UL30a:
~
**** List of PLAYBACK Hardware Devices ****
card 0: Intel [HDA Intel], device 0: ALC269 Analog [ALC269 Analog]
Subdevices: 0/1
Subdevice #0: subdevice #0
card 0: Intel [HDA Intel], device 3: INTEL HDMI 0 [INTEL HDMI 0]
Subdevices: 1/1
Subdevice #0: subdevice #0

Slackware Linux 2.6.33.3-smp #2 SMP Sat May 1 00:26:54

UL30A Three Finger Middle Mouse Click

2010-05-04T23:17:00.000-07:00

It turns out that you can middle click using two or three fingers.

It seemed like there was no way to do this, because there were some bugs filed about incorrectly identifying the touchpad. While the bug correctly states that Linux does not recognize the touchpad, it seems that for now we can function with a sensitive but functional touchpad...

I'm using the Slackware Linux stock 2.6.33.3 SMP hugesmp kernel. Now I can use middle click for pasting from the clipboard and clicking on those links in rxvt-unicode!

UL30A Triple Boot Windows 7 + Linux + ExpressGate Setup

2010-04-23T22:52:00.000-07:00

These are the steps I used to leave Windows 7, ExpressGate, and other recovery modes intact while installing Slackware Linux using GRUB. I installed GRUB onto the root of the Slackware partition and then modified NTLDR to have a Linux menu option:

Used Windows 7's included "Disk Management" tool to shrink the Windows partition and recover about 50 GB.
Changed the "Boot Device Priority" in the BIOS to prioritize "Removable Hard Drive" to load the USB image. Hold F2 down when restarting the machine to enter the BIOS.
Created a Slackware USB boot disc by following the instructions.
Booted into the Slackware setup using the USB boot disc. Launched cfdisk and created 2 additional primary swap (sda3) and root (sda4) partitions. This made the first 1.05 mb partition unusable due to the limit of 4 primary partitions. I was still able to boot ExpressGate and Windows 7, so I don't think it was used for anything.
Used a second USB thumb stick with slackware-current's slackware and extra packages to install the entire Slackware system. Used installpkg --root /mnt to install grub and some additional extra packages.

I decided to use GRUB, because LILO requires you to run lilo every time you change the kernel. In addition, you had to copy the first 512 bytes over to C: every time.
Chrooted into /mnt and performed some mounting trickery in order to run grubconfig:

mount -o bind proc /mnt/proc
chroot /mnt
cat /proc/mounts > /etc/mtab

I installed GRUB to the root of the /dev/sda4 partition in order to keep the NT bootloader intact at the MBR of the drive. I didn't install GRUB to the MBR, because I figured I would not be able to boot into ExpressGate or other special recovery modes if I overwrote the MBR.
Used bcdedit according to these instructions to modify NTLDR to display a Linux option. I copied the first 512 bytes off the root of /dev/sda4 where GRUB was installed and edited the NTLDR menu via bcdedit to show a Linux option.

UL30A Unable to recover the BIOS ROM contents

2010-04-20T23:22:00.001-07:00

The ASUS UL30A's BIOS has an utility called "Easy Flash Utility" to flash the BIOS, so you don't need to keep Windows around just to keep your system up to date.

One caveat is that the BIOS utility can only flash from FAT32 formatted file systems, so while it displayed the BIOS update in the file manager, it complained about being "[u]nable to recover the BIOS rom contents." Thus, I used a USB key which showed up as "D:", selected the uncompressed file, and flashed successfully. The USB key showed up as drive "D:", because I had already removed the auxiliary D: partition and shrunk C: to maximize the room for Linux.

YouTube Fullscreen White Bars Fix

2010-04-11T23:46:00.000-07:00

I use Xinerama to move windows across my ThinkPad T43's LCD display and an external monitor. It seems a lot of Flash movie players are not coded properly to handle Xinerama though and scale the movie wrong leaving huge areas of distracting white space at the top and bottom. Thus, I use xrandr to trick these players into thinking I have two of the same monitors in order to enjoy a properly scaled movie:

xrandr -s 1400x1050 --dpi 96 --output LVDS --mode 1400x1050 --output DVI-0 --same-as LVDS

When I want to return to working with dual monitors, I reissue a slightly different command:

xrandr -s 2680x1050 --dpi 96 --output LVDS --mode 1400x1050 --output DVI-0 --mode 1280x1024 --right-of LVDS

In conclusion, you can use xrandr to get past these improperly coded Flash movie players.

Dvorak Layout for Mac Japanese Input

2010-03-16T17:30:00.000-07:00

It took me some time to figure how to use the Dvorak layout with Japanese input. What is tricky is that the "Kotoeri Preferences" menu item only appears when you change to "Hiragana" mode. It won't appear if you have "U.S." or some other input selected.

Goto "Operation" in the "Kotoeri Preferences" dialog and choose "Dvorak" from "Keybord Layout for Romaji" to change from "U.S."

recordMyDesktop Broken pipe: Overrun occurred.

2010-03-14T17:07:00.000-07:00

While using recordMyDesktop to record the output from a music application, I encountered lack of sound, crackling noises, and error messages regarding broken pipes overrunning.

To solve the lack of sound in the recordings, I had to make sure the "CAPTURE" channel in alsamixer was enabled with both left and right enabled. At first, I somehow managed to only have "L" highlighted, which resulted in sound only in the left ear. I also reduced the volume of the "CAPTURE" channel to a minimum to avoid clipping. Finally, I made sure that the "MIX" channel also had "L" and "R" highlighted in order to record the sound coming from the music application.

I followed funk's article on recordMyDesktop to resolve the crackling noises and buffer overrun messages. Here is the command I eventually came up with to render a smooth crackle-free recording:

recordmydesktop -x 1 -y 1 -width 1395 -height 1045 --overwrite --no-cursor -s_quality 10 -device default --channels 2 --freq 44100 --buffer-size 65536 --overwrite -o file.ogv

Installing recordMyDesktop was painless using sbopkg and SlackBuilds.org packages. Although I tried gtk-recordMyDesktop, I found it didn't have the audio options I needed above to get rid of the noise issues. Thus, I ended up just using the command line program and uninstalled gtk-recordMyDesktop. In conclusion, recordMyDesktop works well after some tweaking, and maybe it could be better if some of the defaults were changed to make the out of box experience better.

YouTube supports the Ogg Theora videos that recordMyDesktop outputs making it very easy to publish your videos. I was able to publish my music video without any problems.

/dev/dsp: Device or resource busy

2010-03-02T20:51:00.000-08:00

Once in a blue moon, an innocent program like scim-bridge claims the sound device and fails to release it. Use "lsof | grep snd" to list the programs using the sound device and then follow it up with a "kill -9 " to stop those programs. I have no idea why scim-bridge is using sound at all though...

Rxvt-unicode Tips

2010-02-21T12:42:00.000-08:00

I recently switched to rxvt-unicode, so that I could properly see Japanese characters on IRC. Here is what I've tweaked to become comfortable with it:

Clickable Links
If you put this snippet in your .Xdefaults, you can click on links in your terminal using the middle mouse button. Make sure you have enabled Perl when compiling rxvt-unicode though.

URxvt.perl-ext-common: default,matcher
URxvt.urlLauncher: firefox
URxvt.matcher.button: 2
URxvt.matcher.pattern.1: \\bwww\\.[\\w-]\\.[\\w./?&@#-]*[\\w/-]
Daemon Mode
urxvt can be used in a server-client model by starting the urxvtd daemon with urxvtc clients. This supposedly can decrease start up times and save resources. Put this line in your .xinitrc. The -o flag makes urxvtd attach to the X display, so that it will be killed automatically when you kill X.

urxvtd -q -f -o

Start your terminals using urxvtc instead of urxvt now.
Disable ISO 14755 Mode
I use the Ctrl + Shift key combination to switch to Japanese. Unfortunately, shows a strange yellow window about some keyboard picture mode and ISO 14755 mode. It seems you need to recompile urxvt with --disable-iso14755 to disable this.

Android Album Art Covers Solved

2010-01-27T21:42:00.000-08:00

Here is the formula I use on Slackware Linux to organize and tag my music for Android. Some of the variables I mention here may not actually matter, but why mess with a formula that works right?

I use easytag 2.1.6 by modifying the SlackBuild (2.1 does not suffice) through sbopkg. Using this bleeding edge version unlocks some vital options (Settings > Preferences > ID3 Tag Settings) that I set for my Android music:

Use ID3v2.3 tags with Unicode. I'm actually not sure if the version matters, but I definitely need Unicode to support Japanese song titles.
Do not write ID3v1.x tags. It seems that Android will use the version 1 tags even if version 2 tags are present.
Automatically convert old IDv2 tag versions. This may be in older versions of EasyTag, but its best to set this to ensure you have a consistent formula for tagging.

Here are the other critical points in my setup:

Strip all pictures from your MP3 files. Although Android claims to read album art attached to MP3s, it does not seem to work reliably for me.
Create a folder per album and drop an AlbumArt.jpg with the cover image. This means you only have one image per album, so if you group music other than by album, this may not suffice for you.

In mysterious ways, Android likes to cache album covers by populating an albumthumbs directory in your SD card with numerical filenames. I have no idea what it is doing, but it is probably tying database values to these filenames. You might have to flush these out if you still don't get album covers to show up properly. However, this method works completely for me, so you shouldn't have this problem...

This works on the Nexus One at least...

Nexus One ADB No Permissions

2010-01-24T17:01:00.000-08:00

I started getting "no permissions" with a bunch of question marks when using adb with the Nexus One. Many sites recommend using root to start the ADB server to get around this problem. Unplugging and plugging in the phone and running lsusb showed that the vendor was no longer 0bb4 but 18d1, so I added a new rule to the 51-android.rules file:

SUBSYSTEM=="usb", ATTR{idVendor}=="0bb4", MODE="0666"
SUBSYSTEM=="usb", ATTR{idVendor}=="18d1", MODE="0666"

I restarted udevd in Slackware with the command /etc/rc.d/rc.udevd restart, and that was it!

Best Skins Ever Nexus One Screen Skin

2010-01-24T12:26:00.000-08:00

I recently applied a BestSkinsEver Nexus One screen skin. The detailed description instructing customers to "[s]imply send [BestSkinsEver] a link where [customers] have shared a short review of the Nexus One Skin...[to get a] ...refund up to $25.00" motivated me to write this review which will actually be just a list of pointers and notes:

I bought two skins, because I figured that I may initially screw up. I did "screw up" a bit, but I believe I applied the skin well enough to leave it on till it gets scratched up.
The instructions say to mix 10 oz of water with 1/4 teaspoon of baby shampoo or dish soap. I used baby shampoo, but I don't think you need 1/4 teaspoon of this concentrated stuff. The solution just gets too bubbly with too much shampoo.
I used an anti-static camera cloth to wipe off the dust from my phone. I'd do this after creating the water/soap solution, because while you are making the solution, dust settles on your phone and you have to clean it again.
Really wait until the solution settles down and the bubbles move toward the edge of the bowl. You don't want soap bubbles and foam to be trapped underneath when applying the skin.
Do not try to stretch the skin even a millimeter with your thumb to position the skin after you have removed the water. You will create bubbles and ripples in your skin this way!
You will probably be unsatisfied with the look of the skin after immediately applying it, but it miraculously improves in the first 24 hours.

While I do not think application of the skin is terribly difficult on hindsight, I think the temptation to get it just right can lead to one's downfall. If I had not tried to stretch the skin just one millimeter to make it perfect, I wouldn't have some ripples and bubbles in the corner. Its not worth making it match perfectly, because you can't even see the skin on the Nexus One after you apply it!

Overall, I cannot even see the screen protector on the phone, unless you focus on the upper left corner where I gallantly tried to stretch it. Some claim that these phones do not need protection, but my mind remains at rest after applying this additional layer since I always lose the phone holders. See you next time when I remove this skin and use the second skin I ordered which I hope is not for a long time!

CyanogenMod 4.2.7.1 and Slackware 13.0 Bluetooth OBEX File Transfer

2009-11-24T18:11:00.000-08:00

CyanogenMod 4.2.7.1 allows OBEX file transfer with an Android phone. All you have to do on the phone end is to enable Bluetooth which is as simple as clicking the Bluetooth icon in the power widget.

Slackware 13.0 comes with the BlueZ stack installed, but its not ready to use out of the box:

1. Stick with the existing /etc/bluetooth/passkeys/default rather than creating a separate file like /etc/bluetooth/passkeys/00:AA:BB:CC:DD:EE, because it seems to cause BlueZ to infinitely loop.

2. /etc/dbus-1/system.d/bluetooth.conf needs to be modified to have a line for the Passkey Agent:

After this you can install obexfs from SlackBuilds.org which works with the stock Slackware 13.0 fuse package to mount your phone as a file system. Unfortunately, the phone's root directory ends up as /sdcard/bluetooth which I wish was configurable. Overall, its easier to mount the phone as a hard drive via USB than bother with setting up Bluetooth.

SeaMonkey Patch Leads to Eclipse Instability

2009-11-22T22:02:00.000-08:00

When I tried to start working on some Android projects again, I couldn't work in Eclipse before crashing with a Java segmentation fault regarding libxpmcore. I reverted the seamonkey package to the original Slackware 13 one, and I haven't experienced anymore crashing. In the snafu, I also lost all auto completion but reenabled it by navigating to Window > Preferences > Java > Editor >Content Assist > Advanced > Restore Defaults.

More Suspend/Resume Problems with Flash Player 10 + FireFox

2009-11-22T12:17:00.000-08:00

I have the Flash Player 10 installed via the SlackBuilds.org package, but I couldn't get the sound to play on YouTube after resuming from suspend. I've always had to restart FireFox to get sound back. It seems the sound system just needs to be poked by some other program though. If I start playing music in moc, I can then hear sound in Firefox. This isn't so bad as having to restart FireFox.

Stock S13 Huge Kernel 2.6.29.6-smp
Flash Player 10 flash-player-plugin-10.0_r32-i386-1_SBo

Thinkpad T43 Resume Freezing Fix

2009-11-22T12:05:00.000-08:00

Although I fixed the blank screen issue, I encountered periodic freezing upon resuming my laptop. After blacklisting the acpi_cpufreq module, I haven't had this problem in weeks now, and I used to crash upon resuming every other day. I created a new file called config in the existing /etc/pm/config.d directory and put in:

SUSPEND_MODULES="acpi_cpufreq"

I hope this fixes the issue. I like the new suspend and resume work in the new kernel, because it seems to take under 2 seconds to suspend or resume!

Successfully Resurrecting the PS3 from the YLOD

2009-11-15T23:06:00.000-08:00

While trying to make Snake crawl into an air duct, my PS3 blacked out and began emitting the dreaded YLOD (Yellow Light of Death). Initially, I searched for RLOD (Red Light of Death), because the light eventually stays red. Red lights seem to occur when the PS3 overheats, so I figured I could just wait a couple hours. Alas, I still got the puzzling sequence of green, yellow, and red lights accompanied by a couple beeps. I then discovered that was actually the YLOD symptom that I had scoffed at and assumed would never happen to me.

To fix the YLOD, I combined the techniques found in Gilksy's videos and DJ Whetzel's videos. I used a level to make sure that no parts like capacitors or resistors would slide once the heat gun started melting the board. I also preheated the board on the low setting of the heat gun as DJ Whetzel advised. Since it was very cold outside and observing that Gilksy applied only 15 seconds to each of the affected sections versus how long DJ warmed up the CPU and GPU, I decided to apply about 30 seconds to each of the sections. After applying heat to the board, the system board became a little warped which frightened me, but it eventually became flat again. Besides that though, I was able to follow the instructions in both of those videos to fix my PS3 without any issues.

Finally, I fretted over the fact that I had not carefully marked where each screw had went. However, I was able to figure out where each screw eventually went. If it is hard to screw it in, then don't continue, because that's probably not the right screw. Try to use the coarse threaded screws when screwing in from metal to plastic. Use the machine screws near the end to fasten the power supply and the Blu-ray drive back to the chassis. The two very tiny stubby screws are for the hard drive basket. The blue screw is for the hard drive, and the torx screw is the final screw that fastens the PS3's lid. If you follow these rules, you shouldn't have any screws left over. :-)

Maybe this is just one of the levels in MGS where I am supposed to fix the system from enemy hackers :-D

MediaPlayer SeekTo and the Sprint Hero AppWidget Bugs

2009-11-08T23:18:00.000-08:00

While the rapid updates to the Android platform have brought long awaited features and needed improvements, it seems that there have been some damaging regressions that have inconvenienced both developers and users. Here is the latest update on the situation at BM Software:

I recently took down "Ongoing Music Adapter Pro" from the Android market, because Android 1.6 ships with a MediaPlayer regression that no longer lets you seek in a stream. Thus, it is no longer possible to resume streams as the application had promised. While this may seem like an error with "Ongoing Music Adapter Pro," the underlying OpenCore system regressed. By the suggestion of a helpful user, I filed a bug which has been not only acknowledged but being addressed! The plan is to wait until 2.1 (or whenever it is fixed) to refresh the "Ongoing Music Adapter" suite of applications.

The "Sticky Memo Widget" line of applications also faced problems on the Android 1.5 based Sprint Hero phones. Users reported that their memos were being lost on reboot. In the end, this turned out to be a problem in the tweaked HTC version of Android 1.5. Yet another helpful user suggested I band with fellow developers to narrow down the problem. I was able to track down the issue, and I have received a response from HTC that they both acknowledge and are fixing this as we speak! I've decided to keep these applications in the market, because this is a problem specific to the Sprint Hero...

Thanks to the understanding users who gave me suggestions on how to tackle the problems. Hope you continue to have fun with your phone! I am! :-)

Thinkpad T43 Slackware 13.0 Radeon X300 Display Resume Fix

2009-10-08T12:40:00.000-07:00

Recently my computer started to show a blank screen after resuming from suspend. To diagnose the situation, I started playing a MP3 before suspending. It continued to play after resuming, so I concluded that somehow the screen wasn't being properly initalized.

The current solution that seems to work is:

pm-suspend --quirk-s3-bios

My system specifications are:

IBM Thinkpad T43
Slackware 13.0 - Using stock kernel of vmlinuz-huge-2.6.29.6-smp
ATI Radeon X300

Sticky Memo Widget Lite 1.7

2009-06-15T22:30:00.000-07:00

I am happy to release "Sticky Memo Widget Lite" 1.7 which has the ability to control what widgets appear in the widget list. I initially released "Sticky Memo Widget Lite" with 16 sizes appearing in the widget list. I thought that freedom to choose would trump the seldom usage of the widget list. However, I have learned that it is pretty annoying and would be lying if I didn't admit it annoyed me too. In 1.7, you can now limit the sizes that appear in the list by starting the bundled "Sticky Memo Widget Sizes" application and checking the sizes you want. One caveat is that you need to restart your phone after making changes, because Android needs to reload the application package. Have fun! :-)

Sticky Memo Widget Lite 1.2

2009-06-02T01:23:00.001-07:00

Now I really regret touching the widget sizes :-) I have put them back the way they were to avoid the confusion. Have fun!

Sticky Memo Widget Lite 1.1

2009-06-01T23:14:00.000-07:00

Sticky Memo Widget Lite is a widget for Android 1.5 that lets you place sticky notes on the home screen. You edit the notes by tapping on them, and you view them just by looking at them on the home screen!

In my obsession to make the smallest and fastest applications, I usually sacrifice on appearance, and some probably feel I have done that again by not using a real life sticky note picture here. I tried to save some space by using Android's Drawable infrastructure to generate a semi-translucent yellow gradient to depict a sticky note. Despite not being exactly like a sticky note, I really like how it turned out.

I rushed to bring out this application, and I now regret offering all the possible 16 widget sizes in the list. Although I had intended to allow maximum versatility, I may have ended up just creating more confusion. In this version, I prefixed what I think would be the less popular widget sizes with a "Z" to put them at the end of the list. Removing the widget sizes altogether would have made a backwards incompatible change, so this was the trade off I made.

Nevertheless, I am pleased with the result. I cherish both Astrid and AK Notepad, but this simple application will suffice for my needs. I hope you enjoy the application too!

Ongoing Music Adapter Pro Version 1.1

2009-04-19T18:48:00.001-07:00

This version introduces a behavior change to combat connectivity problems. If you are listening to a link when you lose connectivity, the Android music player will just stop playing, and playback cannot be resumed without opening a different link and then resuming the prior link. Now you can just go to the Ongoing Music Adapter Pro, select the link that lost connectivity, and it will restart the player and resume where you left of.

In the prior version the Ongoing Music Adapter Pro would notice that the same file is already playing and would merely transfer you to the player. Now it will always stop and then start the player before resuming playback.