...more recent posts
I complain here a lot about running a mail server. There's a new package out based on Qmail (which is what I presently run, but in a slightly out of date manner,) called Qmail Toaster Plus. I'm very excited about it. It's just the basic Qmail package, but with a whole bunch of patches and add-ons selected and bundled together for you. Or, in other words, it looks like it will take a lot of pain out of installing, and more importantly, staying up to date, as it's now available as a yum repository. Probably the coolest tool that QTP includes is qtp-newmodel which will upgrade a Qmail Toaster installation automatically, but it does it by building the new installation in a sandbox - without touching your current installation. At that point you can do some testing to make sure everything works before you take your still running older installation down and replace it with the new one. That just feels so much safer to me. Why doesn't everything work like that?
Great script for analyzing MySQL my.cnf values: tuning-primer.sh
cd /usr/local/sbin wget http://www.day32.com/MySQL/tuning-primer.sh chmod 700 tuning-primer.sh tuning-primer.sh
Google's new App Engine was the big story yesterday. Developers can now build web services and run them on Google's computers. Amazon's combinded offerings of S3, EC2, and SimpleDB is probably the closest competitor. But here is an intriguing theory about their plans which even goes beyond what Amazon is doing:
App Engine comes with its own development setup that runs off your computer (available for Windows, OSX, and Linux). You develop the application on your computer, run it, test it, add features, and then upload it to Google's computers. My question is this: What's stopping Google from turning the local development code into a full desktop-based runtime for web applications?In this version of things Google is taking on not just Amazon, but Adobe AIR and Microsoft's Silverlight.
Not sure about the security benefits, but I love this idea of adding a hash= attribute to any html element that can take a href= or src= attribute. The hash= value is "a base 32 encoding of the SHA of the object that would be retrieved." This would be good for caching, as the article suggests, but I'm more interested in the reverse - it would help with the problem of old objects already in the cache being used instead of updated objects on the server. In this case the hash values wouldn't match so it would force the browser to fetch the updated object.
My friend recently bought a MacBook. It really comes pretty set up right out of the box. There are just a couple things I would do for a new machine, and here they are
I just finished a project which basically amounted to fixing a blog based website that had been originally built by one person, and then much later "fixed" by someone else. Except here "fixing" was equivalent to "badly breaking". Nothing worked anymore. MySQL errors all over the place if you even tried to sign into the admin back end. Anyway, it's a strange hell to dive into someone else's code. I certainly wouldn't wish mine on anyone. But this was much worse than I could have imagined. Obviously it was originally done by someone with a little HTML knowledge and almost no PHP. And then "fixed" by someone who knew even less. Still, it worked out in the end and even made me look pretty good in the eyes of the guy who brought me in to do the work.
But the reason I'm posting this is because the site used TinyMCE to give the textarea posting boxes more full fledged word processing ability. You get a control strip across the top of the box and you can click to make bold, italic, insert links, make lists, etc... I've seen it before, and from reading it seems like TinyMCE is the best solution at the moment.
But, uh, wow. It's just really bad I think. Especially the way it was implemented on this site: with absolutely every option turned on. It's ridiculous. It's ugly. And it just seems to fight you every step of the way. Is it so hard to learn that a couple HTML tags? Well, yeah, I guess some people just don't want to do that. So as if in answer to my struggles this morning I found Markitup!, a much lighter weight solution. This looks really nice. It doesn't hide the HTML from you, but it does give you a bare minimum of tastefully designed short cut buttons that will insert the tags so at least you don't have to remember the exact syntax. I'll definitely be using this one in the future. Nice job.
Google on Monday said it has a plan to have American consumers from Manhattan to rural North Dakota surfing the Web on handheld gadgets at gigabits-per-second speeds by the 2009 holiday season.
Epic battle with qmail today. I believe I am victorious.
From the links you hope to never need department: recovering from an rm -rf on an ext3 file system. Very interesting as this is widely reported to be impossible - most notably by one of the ext3 developers!
Youtube.com has released a javascript player API for controlling its embedded movie player on a webpage. You can even customize the player with your own chrome, user controls, custom playlists, and menus. Very nice. When do the mandatory ads before every video begin?