Older blog entries for robocoder (starting at number 312)

Now I remember why we didn't move these websites sooner -- some scripts depended on CPAN modules that aren't installed on the servers from our hosting provider. Anyway, other than a few minor hiccups (e.g., hardcoded absolute paths buried in a php script), the sites are up ... and running (for the most part).

Tomorrow: install Ubuntu on the web server (was Red Hat), and set-up a couple of Piwik environments.

I'm in IT administration mode today. It's time to take care of some long deferred tasks, or face putting them on next year's list of New Year Resolutions.

  • Move all remaining public facing websites to externally hosted green web servers.
  • Update the OS on our in-house web server and dedicate it to development and staging.
  • Update DNS entries for what's being moved, shelved (redirect), or monetized.
Tis the season...
20 Sep 2008 (updated 20 Sep 2008 at 12:59 UTC) »

With all the attention on benchmarking the performance of next generation Javascript engines -- Google's V8 vs Mozilla's TraceMonkey vs WebKits's SquirrelFish Extreme -- I'll point out that Wine folks have been steadily making advances on their jscript.dll implementation based on SpiderMonkey. The implication is that we'll see Javascript executing faster in Wine than native on Windows. LOL.

CAPTCHAs as a usability improvement? As a defense against automated attempts to harvest our MLS database (e.g., screen scraping), our MLS now uses reCAPTCHA So instead of being booted off the system (and redo'ing whatever it was you were in the middle of), you'll be prompted to enter the two words shown.

In the upgrade to Ubuntu Hardy, users began complaining when Nautilus no longer preserved a file's timestamp when copied.

After protracted debate about "the right way" and rationalizing "that's how cp works" -- whether right or wrong -- Nautilus developers listened to the users and pushed out a fix.

A new File Management Preference (under the "Behaviors" tab) would have been my preference, but that's just me...

4 May 2008 (updated 4 May 2008 at 06:01 UTC) »

Ran into a weird bug (applies to Apache httpd 1.3.39 and 2.0.40 -- not sure about later 1.x and 2.0.x releases -- but certainly doesn't apply to 2.2.8, where it appears to work correctly) where mod_rewrite exposes the absolute path in the redirect (which then fails).

In a subdirectory off my web root (and root of a virtual host), I have a .htaccess file:


RewriteEngine On
RewriteCond %{HTTP_HOST} .
RewriteCond %{HTTP_HOST} !^anthonpang\.com [NC]
RewriteRule (.*) http://anthonpang.com/$1 [R,L]

The URL (yes, I left off the trailing slash):


http://example.com/anthonpangcom

Mysteriously becomes:


http://example.com//home/robocoder/public_html/anthonpang

I say mysterious, because the behaviour depends on the domain I use in RewriteRule. For example, anthonpang.ca does get rewritten correctly.

That said, it's probably "minor" as I don't expect to publish or link to the subdirectory-based URL when this goes "live".

30 Apr 2008 (updated 1 May 2008 at 15:48 UTC) »

Thanks sneakums. My bad. DomainKeys != DKIM. I'm amending my previous blog comment as follows:

DomainKeys is dead. Long live DKIM (Domain Keys Identified Mail).
Oh goody...it looks like cpanel does have support for setting up DKIM. I'll have to ask my hosting provider if it's available on shared hosting accounts.
24 Apr 2008 (updated 24 Apr 2008 at 02:59 UTC) »

Taking a page from the domain cybersquatter playbook, I registered typos of my name, Anthon -- because the 'h' is silent. But rather than a simple redirect or creating a CNAME, I threw up some single page web sites, ala Wikipedia dab pages: Anton Pang.com and Anton Pang.ca. I also embedded a little analytics tracking code, so I could gauge the utility (or futility, as the case may be) of this exercise. (The self-deprecating reference to typo-squatting is in jest.)

I can't say this was a welcome/needed distraction, given everything else on my plate, but this was "low hanging fruit", so to speak, and I needed to feel like I accomplished something today.

16 Apr 2008 (updated 16 Apr 2008 at 06:59 UTC) »

Just shortly after I put our family tree online, I discover the 4.1.4 update of PhpGedView still packages an outdated (and vulnerable) version of PunBB. (A slightly newer version of PunBB is in the PhpGedView repository head.) Granted this is a contributed module, but this is a nuisance.

In this case, it's easier to apply the changes between 1.2.12 and 1.2.17, to the version of 1.2.12 included with PGV. (The patchfile is 66K in size. In contrast, the diff file between vanilla 1.2.12 and the hacked version included with PGV was over 1M in size, and touched nearly twice as many files.)

10 Apr 2008 (updated 10 Apr 2008 at 14:42 UTC) »

RSS/Atom kept it simple (and for the most part, kept it small). But Holy Hijack! News feeds (like the Big-G's) are pushing the limits by embedding oodles of HTML (table, divs, images, links, ...) ... blurring the boundary between RSS reader and browser. As an example, the Sci/Tech news is 150+K, while the RSS feed is a little over 90K ... for little more than the teaser text on just 10 stories? WTF?!

Obligatory AppEngine blurb: on the first day, the Big-G received registrations from its requisite 10,000 beta developers; other applicants are now being on put on a wait list.

303 older entries...

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!