WordPress upgrade

In the past 6 years, this blog hasn’t changed that much. I added a plugin here and there, integrated google buzz posts, etc., but the core wordpress install remained the same. After procrastinating the upgrade several years, I thought, mmm.., spammers don’t bother anymore to hack an ancient wordpress install! So I left it at that. :)

Unfortunately, lately, spammers have become less lazy, so my blog still got targetted again. Today I finally made the switch to WordPress 3.3.1 and although I didn’t expect wordpress to support the giant version leap, it did!

I was tempted to leave the design just the way it was (you know, good old kubrick), and leave it at just a back-end upgrade. But then I thought it would be nice to have a descent display on mobile devices, which comes for free with the newer themes of course. Also my beloved antispam plugin Spam Karma 2 which was abandoned a few years ago (Thanks anyway Dave!), was in for replacement.
So I decided to switch to the fancy twentyeleven theme and tweak the header (about the only thing I care to tweak :)). The tweak consists of making the header smaller and move it up:

#branding img {
height: 180px;
margin-bottom: -7px;
margin-top: -100px;
width: 100%;

The best way to apply this tweak is by creating a child theme.

Btw, the header image is based on a photo of one of the layers present in an LCD display (thanks to sonic840).

Using widgets I was able to recreate the same look as in the old theme (having the categories, archive, etc).
I was hesitating to keep with wordpress, but I’m glad I did (for a blog at least):

  • Upgrade went well from an ancient version
  • Killer feature: install/upgrade new plugins from the web interface, it just works!
  • I love the fullscreen authoring mode

Google to disable site scraping feed service

For ages, I had been requesting a site monitoring feed for sites that don’t have an rss feed. It’s not so difficult, keep track of changes and post the diff in an RSS feed (kind of ;) ). In the beginning of this year, google finally implemented that! I didn’t expect it.

Anyway, short joy: google has announced it will discontinue this service. :(

Here is a tip for people like me having hundreds of feed to find back those special feeds more easily: go to manage subscriptions and filter on “webfeed”. All those site scraping feeds start with “http://www.google.com/reader/public/atom/webfeed/”. Happy page2rss migration. :)

Best. Plugin. Ever.

I was trying to watch a vid on a website which was quicktime only. Since I’ve had some trouble a few days ago to play a wmv file on a windows media only site, I decided to give the firefox mplayer plugin a try. I had heard of it before but never found a compelling reason to install it. Until now. It just plays _everything_! I can’t believe how I could live so long without it. :) Or maybe because 99% of the web’s vids are flash? ;) (which works, of course)

Here is the supported formats list:
Window Media: wmv, avi, asf, wav and asx
QuickTime: mov and smil
MPEG Video and Audio: mpeg and mp3
Ogg Vorbis: ogg
AutoDesk FLI: fli and flc
Vivo: vivo
Real Player: ram and rm

Securing mysql

Everyone knows, first thing you have to do after installing mysql, is setting its root password. I used to do this using the mysqladmin tool like this:

/usr/bin/mysqladmin -u root password 'new-password'

For starters, you should not forget to do the same command for your other host names (-h option), but also there are still some other tasks like disabling anonymous access and removing the test databases. Today, having another mysql install, I noticed that the install script mentions this neat tool packaged with mysql to do just all that. :)

Just run /usr/bin/mysql_secure_installation and you’re done!

Freenode raising the bar for routers

This week I noticed a strange problem with my irc client.
First day (Tuesday), it would fail to connect. Since my login to the campus network expired, I had to relogon, so I thought the problem was simply related to that and that the irc client wouldn’t recover until an application restart.
Second day, it still didn’t connect, which I thought was strange, but I didn’t have time to chat anyway, so I neglected.
Today, I frowned and decided to take a look :)

First of all, I noticed it wasn’t really a connect problem, it couldn’t even resolve the hostname. Heh? When I ran “nslookup irc.freenode.org” it spit out:

;; Truncated, retrying in TCP mode.
;; Connection to for irc.freenode.org failed: connection refused.

I went to the freenode site, to see if there were any news items reporting problems or outage. There didn’t seem to be a news section. So I assumed: no news == good news. :)

Then, I noticed the actual hostname for irc is really chat.freenode.net instead of irc.freenode.org. “chat” or “irc” in front works both. Indeed, that did work. So why did irc.freenode.org suddenly stop working? Was it because of a recent patch (although I couldn’t remember applying any in the last week)? Was it because of my firewall? Both answers are negative.

The friendly chaps at #freenode kindly tipped me off that the dns record had probably become too large to be transmitted through UDP and that TCP dns requests seemed blocked. Servers get added to the network on a regular basis, so that would explain why it suddenly stopped working. It became obvious that my dusty old router would be the culprit (quick tip: never blame your linux kernel first ;) ). After all, according to contemporary standards, one could declare mine “ancient”.

Quite recently I already encountered some problems with it. After an upgrade of the campus network, I kept having strange packet loss when playing multiplayer games for more than 2 minutes. I had never updated my firmware, so I thought I might try that. Although the latest firmware release dated back to 2004, it did help! mattie=1, router=0.

After I pulled the router out of the chain, the dns problem was solved indeed! Unfortunately, updating firmware is no longer an option, so I guess the router and I are even now. ;)

On #freenode, I was told I wasn’t the first one to report this problem, so I guess a lot of people will need to update their firmwares, buy new routers, or simply use a more local dns name. ;)

New google reader – render problem in opera

Whether you like the new look or not, it seems there is an issue with long feed names being wrapped at the icon position which wastes valuable space (read my problem description and support request here). The day after, I already received a fix from an Opera employee :) very simple, just use the following user css file:

.scroll-tree .name {
padding-left: 0 !important;
position: relative !important;
left: 17px !important;

If you are unfamiliar with user css: right-click on the page and edit the site preferences. In the display tab, point the “my style sheet” to the relevant user css file. Make sure “my style sheet” usage is enabled in Preferences->content->style options->presentation modes->author mode->my style sheet (also more easy accessible through manage modes under the style mode button on the extended view bar if you use that).

Opera 10: it’s unbelievable

Opera 10 seems to bring ground-breaking news for people using Opera Mail. It seems they finally got hold of the resources needed to implement the 2 most asked features ever! Behold:

  • Rich text email composition
  • Remove mail from server after X days

Can you believe it??!! After all these years of denial? I for one surely didn’t see that one coming. :) Read the full changelog here.