p14nd4's blog is the best blog ever.

After nearly five years of using my trusty (Palm OS-based) Treo 700p (preceded by over three years with a Palm OS-based Treo 300, preceded by over a year with a Treo 180, etc…), I finally bit the bullet and moved to Android. I actually did pick up a used Samsung Moment (landscape QWERTY slider) off a friend a couple summers ago, but never activated it; pretty much all of my reference point for Android comes from that device, though (running a custom rom, rooted, Android 2.1). I wasn’t especially thrilled with the Moment for a few reasons, which convinced me it was alright to stick with my Treo for a while longer, and keep my eye on the market. A few months ago, there were some leaks about an upcoming portrait QWERTY device from Motorola on Sprint that was supposed to be pretty fancy, and even back then, I had pretty much made up my mind: this would be my next phone.

Details in subsequent leaks changed a little bit, and the rumored release date came and went without additional information, but at long last, the Motorola Admiral was released at the end of October. I was hoping to find some other reviews online right away, but they didn’t come immediately. A few people on forums ordered the devices promptly and answered a few questions, and someone posted a Youtube overview, but there wasn’t exactly a thorough review out. Nevertheless, I decided to pull the trigger, and ordered mine (requiring upgrading my plan to Everything Data) this past Wednesday. I was fairly impressed with Sprint’s customer service that got my plan changed and corporate discount applied over online chat with virtually no pain, and then I ordered the phone, and they had it to my door around noon the next day. Nice work, guys!

Read more »

I recently purchased a Super Micro X7SPA-HF-D525 motherboard, and was excited to try out the IP-KVM feature. While I was able to log into the web interface fine, launching the console redirection java webapplet partially proceeded, but eventually failed for me with an error no iKVM64 in java.library.path. I was able to experiment a bit and developed a patch (below) for the launch.jnlp file served for the iKVM applet that enabled this to run on my system (i.e. downloaded the launch.jnlp, patched it, then ran with javaws launch.jnlp). When run without the patch, Java fails to download liblinux_x86_64.jar for me, because it only tries the following two URLs, which don’t contain the relevant file: https://192.168.0.207:443/liblinux_x86_64.jar?version-id=1.0.3 and https://192.168.0.207:443/liblinux_x86_64.jar (while the actual path is something like /liblinux_x86_64__V1.0.3.jar.pack.gz).

Read more »

I’ll make this brief, but wanted to document my fix to potentially save someone else some aggravation in the future.

I’ve been working on a c++ project utilizing Google protobuf, and at one point ran into an error during application startup (static initialization):

libprotobuf ERROR google/protobuf/descriptor_database.cc:314] Invalid file descriptor data passed to EncodedDescriptorDatabase::Add().
libprotobuf FATAL google/protobuf/descriptor.cc:862] CHECK failed: generated_database_->Add(encoded_file_descriptor, size):
Abort (core dumped)

Read more »

I recently developed a new application on my local machine (in Visual Studio/C++ 2008), and then moved it into our source/revision control system at work, IBM/Rational ClearCase (cc). The new app uses some third-party libraries for which we only have DLLs (no source), so I checked those into ClearCase, too. Unfortunately, the first time I made any changes to the version in cc and tried to step through the app in the debugger, I was immediately presented with an unfriendly error and app termination:

Debugger:: An unhandled non-continuable exception was thrown during process load

along with some return code (0xc0000022).

Read more »

As mentioned in my previous post, I’ve been playing with WordPress, vBulletin, vBridge, and some other stuff as part of my work for my other site, TechwareLabs. One of the other plugins I utilize on that site is WP Super Cache, which I discovered was an easier way to accomplish basic static caching instead of trying to jump into the deep end of learning mod_cache, memcache, and other sorts of fancy things for an actual elegant solution.

Unfortunately, these are complicated systems, and vbridge in particular seems slightly less polished than some more widely utilized components (though, in fairness, its user base is considerably smaller than most wordpress plugins).

Read more »

I’m the webmaster (and co-founder) of TechwareLabs, which gives me opportunities to play with some technology, software, and code that I otherwise likely wouldn’t have the motivation (or occasion) to pursue on my own. (This isn’t the main point of the post, but one of the coolest in recent history was using Xdebug profiling in PHP with KCachegrind to debug a performance problem on the live site. Flipped the profiler on for a few seconds, sent a few requests to the server, disabled the profiler, and loaded the output in KCachegrind to immediately get an awesome graphical overview of all the calls made to render the page, and how long they took. I don’t get to use fancy software like that at my day job, so this was very cool to use, and even cooler that it actually helped solve my problem so quickly.)

Anyway, back to the main point of this post! We’re currently working on an integration between the site backend (WordPress) and our forum software (vBulletin). I found a slick plugin for WordPress that’s supposed to accomplish most of this: Complete WordPress/Vbulletin Bridge and a guide explaining that vBulletin’s own utility, ImpEx, already has the ability to import users from WordPress. (Side note: ImpEx is available only to current vBulletin license holders, and can be downloaded from the same page where you can download vBulletin, after logging into the members area on their site. This wasn’t immediately apparent to me from reading that guide.) Woo! I can follow instructions!

Things went pretty smoothly, other than the hiccup I had trying to use the WordPress or WordPress CMS importers in vBulletin ImpEx. I eventually realized I had to click some start over link to get ImpEx to clear my session before I could continue with the BBpress importer, as the guide mentioned above recommended. Things seemed to be working pretty well in my testing, aside from a few pesky PHP errors about headers already being sent on some pages in the WordPress admin control panel, and a few slightly more concerning vBulletin-generated emails about insufficient database connections (yet to be resolved).

The bad news came when someone actually tried to publish a live article, and discovered that the Add media button for uploading images wasn’t working, and was instead presenting an error:

Are you sure you want to do this?

Please try again.

I poked around in media-upload.php, and first noticed that $_GET[‘inline’] wasn’t being defined, so I thought this was the problem. Not so fast! I guess the staff’s been using the NextGEN Gallery Plugin for WordPress, so this does something fancy for the media upload links, so that wasn’t the problem. My google-fu was apparently a little weak, since although I found other people reporting the issue immediately, I had to read through about twenty pages of the plugin thread before I found a solution (which, in retrospect, I wish I would’ve thought to check, myself, earlier).

User ‘skariko’ posted a [possible?] solution for folks having a problem with the WordPress flash-based image uploader, and while this wasn’t exactly what I needed, it was the Eureka! moment I needed to solve it, myself with a few keystrokes of modification. He suggested modifying the following line in vbbridge.php:
[code lang=”php”]#if (basename($_SERVER[‘SCRIPT_NAME’]) == ‘upload.php’) { return; }[/code]
to
[code lang=”php”]if (basename($_SERVER[‘SCRIPT_NAME’]) == ‘async-upload.php’) { return; }[/code]
I’m guessing that async-upload.php is the file that the flash uploader posts to, and for whatever reason needs to be unaffected by the vbbridge code. Anyway, I wasn’t even getting that far, so I just changed it to
[code lang=”php”]if (basename($_SERVER[‘SCRIPT_NAME’]) == ‘media-upload.php’) { return; }[/code]
and everything magically worked!

It turns out that someone else posted this exact same solution on another page of the thread, but I wanted to share it here, too, so maybe someone else in the future can have better luck googling for a solution (or in case I forget how I fixed it, and have to fix it again :-p).

Hahaha. So I’ll try to keep this brief, but I just wanted to post the problem I recently found with VMware Server (1.0.4, 1.0.5) on linux (perhaps others), and what I eventually found to be the cause and workaround.

Problem

Running VMware server on linux, I edit virtual machine settings, open the CD-ROM device, select Use ISO image, and Browse. I navigate to my smbmount (actually a cifs mount line in my /etc/fstab), and VMware hangs for about three or four minutes before displaying the contents of the directory. The following lines are written to /var/log/syslog:

Read more »

This is a break from my regular vein of [not] posting, but I’m livid enough that I need to write this. I’ve had SprintPCS cell phone service since 2003, and have been generally pleased with the service. It’s all subjective, of course, but I haven’t had real issues with dropped calls, and their data services (1xrtt, evdo) have been quite good to me. I log in to their web site regularly to check my account, pay bills etc, and it’s worked perfectly for the past several years. Apparently that was unsatisfactory for Sprint.

Read more »

Introduction

As some of you may know, I bought a refurbished IBM ThinkPad T41 (2373-5U2, iirc) during fall 2005, and have been dual booting it with Windows XP Professional and Ubuntu Linux. The hard drive is only 40 GB, which I partitioned as:

• 26.5 GB Windows (NTFS)
• 5 GB Ubuntu / (ext3)
• 1 GB swap
• 4 GB /home (ext3)
• 3.5 GB unusable/recovery partition

Windows and Visual Studio 2005 Professional managed to eat up about 8 GB, plus another 1.2 GB for the installers I left on the hard drive, and miscellaneous other crap put the Windows partition near its capacity, and I had nearly 1 GB free on the Ubuntu / partition, but when it came time to upgrade to Feisty, I was informed that I needed over 1.1 GiB free. Even after some housecleaning, I still came up short, so I decided to wipe / and perform a fresh Feisty (Ubuntu 7.04) install. I figured I was probably about due anyway, since I’d been dist-upgrading since at least Breezy. The installation went fine (I did it during a meeting here at work), but when I went to enable “Desktop Effects” (compiz), I got some odd results. Actually, immediately after enabling it, the right section of the screen was displaying odd banded/corrupt artifacts (leftovers, I presume, from switching video modes, etc). I disabled the effects for the time being, and went about my life.

Fixing Compiz Display

As I noted above, I had problems getting Compiz working from Ubuntu 7.04 Feisty on my ThinkPad T41 with an ATI Mobility Radeon 7500 (01:00.0 VGA compatible controller: ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500]). One or two Ubuntu version ago, I had some OpenGL performance issues with xmoto (a game), which I kind of resolved by changing my xorg.conf file to explicitly use the ‘radeon’ driver instead of ‘ati’ and forcing AGPMode 2 (and maybe something else, I don’t remember for sure). I hoped that would solve my current problem, but alas, it did not. Perhaps resulting from my inability to accurately describe the results I was seeing, I was unable to immediately find any quick fixes to my problem. I guess that’s why it’s still beta, huh?

Update! I reinstalled Feisty (fresh) on my new laptop hard drive, and spent some time isolating the line that fixed the display rendering. It wasn’t the original line I thought, but rather:

Option "AGPSize" "32"

Fixing Compiz Performance

After getting compiz working, I was still getting weird performance hits, where I could hear my hard drive click, whirr, and see CPU spikes during really any display updates (even scrolling a man page in a gnome terminal window would spike the CPU). I thought this was somehow related to the new IDE device handling by the SCSI subsystem, and possibly not having the drive using DMA. I followed that hunch for a while, since it seemed that other people were indeed reporting issues of that nature in Feisty. The information I could glean from dmesg suggested that the drive was in fact in DMA mode, but during the course of the investigation, I noticed some questionable SMART values, so I decided that maybe my hard drive was just dying, and bought a replacement. However, I wasn’t entirely convinced that was the problem, so I kept poking around. I tried many combinations of options, but eventually found a winner. Again, I’m not 100% sure which line was the actual solution, but I think it may have been the following line in my Device section of xorg.conf:

Option "RenderAccel" "true"
Along the way I also modified my dric (/etc/dric or ~/.dric) to enable hyperz as follows (alternatively accomplished with the package driconf):

dric file »

<driconf>
    <device screen="0" driver="radeon">
        <application name="Default">
            <option name="force_s3tc_enable" value="false" />
            <option name="no_rast" value="false" />
            <option name="fthrottle_mode" value="2" />
            <option name="tcl_mode" value="1" />
            <option name="texture_depth" value="0" />
            <option name="def_max_anisotropy" value="1.0" />
            <option name="no_neg_lod_bias" value="false" />
            <option name="texture_units" value="3" />
            <option name="dither_mode" value="0" />
            <option name="hyperz" value="true" />
            <option name="round_mode" value="0" />
            <option name="color_reduction" value="1" />
            <option name="vblank_mode" value="1" />
            <option name="allow_large_textures" value="2" />
        </application>
    </device>
</driconf>

Full xorg.conf Devicesection:

Full xorg.conf Device section »

Section "Device"
        Identifier      "ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500]"
        Driver          "radeon"
        BusID           "PCI:1:0:0"
#       Option          "DisableGLXRootClipping"        "True"
        Option          "DRI"                           "true"
#       Option          "ColorTiling"                   "on" #already the default
        Option          "EnablePageFlip"                "true"
#       Option          "AccelMethod"                   "EXA" #prevents mine from loading
        Option          "XAANoOffscreenPixmaps"
        Option          "AGPMode"                       "4"
        Option          "AGPFastWrite"                  "on"
        Option          "RenderAccel"                   "true"
#       Option          "UseInternalAGPGART"            "no"
#       Option          "EnableDepthMoves"              "true"
#       Option          "GARTSize"                      "64"
        Option          "AGPSize"                       "32"
#or "64"
#       Option          "backingstore"                  "on"
#driconf -> Enable HyperZ for big improvement
#power saving:
        Option          "DynamicClocks"                 "on"
EndSection

After leaving my front page (visually) the same for about two years now (you can check it out for yourself at archive.org), I recently felt somewhat compelled (obligated) to actually do something useful after receiving repeated praises from Josh about my site (though, to be honest, I’m not sure what makes my site particularly any better than his … whatever). The idea actually struck me around 4:30 AM, right after I had lain down for bed; I knew that if I didn’t get up immediately to get the idea down on paper, it would be forever forgotten by the time I awoke. While that same situation plagues me with relative frequency, I almost never actually get up to ensure the idea’s survival.

I missed my first (11:15 AM) class the next morning, but was on time to my 12:45 lecture, where I quickly went to work coding up the idea. The idea is by no means particularly complex, just efficient and elegant. With about six lines of xhtml constituting five objects, and maybe three css id’s, the layout worked perfectly in Opera and Firefox (both boasting W3C standards compliance), and the code was fully compliant to the xhtml 1.0 strict standard. Unfortunately, Internet Explorer was another story. (Duh.) Here was the original xhtml code:

<div id="title">
<h1>p14nd4.com</h1>
<a href="/resources/" title="resources">/resources</a>
<a href="/gallery2/" title="gallery">/gallery</a>
<a href="/blog/" title="blog">/blog</a>
</div>

And the CSS basically consisted of:

div#title { position: absolute; top: 40px; left: 40px; color/border stuff; }
div#title a { display: block; position: relative; bottom: 7px; right: 0px; float: right; margin-right: 15px; color/border stuff; }
div#title a:hover { color stuff; }

Internet Explorer didn’t render the <div id="title"> the correct height, aligned the h1 to the top instead of middle, and didn’t have the links in the correct place. I was able to fix the link placement with moderate ease, by enclosing them all inside another nested <div>, which had position: absolute; right: 0px; bottom: -9px; and then just did float: right; margin-right: 15px; for the <a> tags. I believe this fixed the main div’s vertical size rendering issue as well, in IE. Getting the <h1> to vertical-align: middle; within the main div was another story, though. I probably worked on that issue alone for another hour or more. I tried various combinations of nested elements and other css attributes without success, until I tried nesting the h1 inside a div, with a border enabled. I stumbled across this solution because I start putting borders on all my objects to debug where certain boxes are ending, beginning, hitting each other, etc. Regardless, I figured that just another nested div was the solution, so I immediately went to remove the testing borders, so the site would be suitable for public display (having half a dozen dashed red borders running through the header wasn’t particularly attractive). To my great confusion and horror, though, the h1 regressed to vertical-align: top after removing the borders. Sure enough, adding the border back onto the div re-fixed its alignment. I tried just setting a top border, so there wouldn’t be much visual evidence of this hack, but IE wouldn’t behave with just a top border; it needed all four. Eventually I conceded that it just wasn’t going to look quite right in IE, and moved on for the evening.

Today I finally got around to moving the new header to the main page, but again grew displeased with how it looked in IE. Another idea struck me, though, which managed to elude me yesterday evening. If you were to look at the code, you would again see an additional div containing only the <h1> element, with a style="border: 1px solid black; … which would normally result in an ugly visible border along the bottom of the header… I remedied this situation by overriding the border shorthand property immediately after declaring it, and eliminating the left, right, and bottom borders. In full, the code for that div tag reads as <div style="border: 1px solid black; border-bottom: 0; border-right: 0; border-left: 0;">. Don’t ask me why this happens to fix IE’s rendering, but I’ll live with it, I suppose.