I'm spent...and an apology
I just spent the last few hours polishing up my site for maximum awesomeness. It doesn’t look much different, you say, but under the hood is a completely different beast. As I mentioned in a previous post, I spent last night converting the backend of this site to byteflow, which is built on django (python based) and running on top of lighttpd.
Previously (read; years ago) I had built a blogging system based on a LAMP stack, specifically Apache 2.0, MySQL and PHP. Slowly, over time I added to and ‘improved’ the software. Time and time again it felt like an exercise in futility. It seemed to be a never ending battle to keep up with the times and stay exposed. Since I started working for ag interactive I’ve fallen in love with Python. Quickly thereafter I started searching for a good, ”developing” blog package written in python. I longed for something I could extend and something that is portable. I long no longer…(?) The django project has provided many facilities to build a great blog (and cms type thing) and I’m glad someone has finally stepped up to take advantage of them. I’m sure I’ll be submitting patches to this project as I discover things to tweak or improve.
To the byteflow crew: Great Job so far!
Now onto some of the details about what I did.
I love lighttpd, all of my recent development has taken advantage of it. It’s configuration is intuitive, it’s feature rich for it’s size, and it’s performance is hard to match. Converting the config for this was relatively simple. This is, of course, coming from a sysadmin who deals with hundreds of web servers daily :). Anywho, here’s the config with paths cleaned up for clarity:
# Main server settings
server.modules = (
"mod_access",
"mod_alias",
"mod_accesslog",
"mod_rewrite",
"mod_redirect",
"mod_status",
"mod_fastcgi",
"mod_compress",
"mod_expire",
)
server.document-root = "/docroot"
server.errorlog = "/var/log/lighttpd/error.log"
index-file.names = ( "index.php", "index.html",
"index.htm", "default.htm" )
accesslog.filename = "/var/log/lighttpd/access.log"
url.access-deny = ( "~", ".inc" )
server.pid-file = "/var/run/lighttpd.pid"
dir-listing.encoding = "utf-8"
server.dir-listing = "enable"
debug.dump-unknown-headers = "enable"
server.username = "www-data"
server.groupname = "www-data"
#status.status-url = "/server-status"
#status.config-url = "/server-config"
# Expires stuff, per yahoo tips :)
$HTTP["url"] =~ "^/" {
expire.url = ( "" => "access 8 hours" )
}
# Set mime-types.
include_shell "/usr/share/lighttpd/create-mime.assign.pl"
# Allow for vhost configs
include_shell "/usr/share/lighttpd/include-conf-enabled.pl"
# Compress settings
compress.cache-dir = "/var/tmp/lighttpd/cache/compress/"
compress.filetype = ("text/plain", "text/html")
# Custom settings
# Set-up some aliases
alias.url = (
"/admin-media" => "/django/contrib/admin/media",
"/images" => "/docroot/images",
"/static" => "/docroot/static"
)
# Startup the fastcgi server, point it at my custom
# fastcgi script.
fastcgi.server = (
"/byteflow.fcgi" => (
"main" => (
"socket" => "/tmp/byteflow.socket",
"bin-path" => "/docroot/byteflow.fcgi"
)
)
)
# Make sure we're passing on these urls
# bare, don't process them with django.
url.rewrite-once = (
"^(/rpc.*)$" => "$1",
"^(/images.*)$" => "$1",
"^(/static.*)$" => "$1",
"^(/admin-media.*)$" => "$1",
# Then push the rest through django
"^(/.*)$" => "/byteflow.fcgi$1"
)
# Legacy url, redirect it to new feedburner.
url.redirect = (
"^/rpc([\/]*)" => "http://feeds.feedburner..truncated..",
)Here is my fast-cgi script, which resides in my docroot.
#!/bin/sh # byteflow.fcgi export PYTHONPATH=$PYTHONPATH:/home/bsmith/Dev export DJANGO_SETTINGS_MODULE=byteflow.settings /home/bsmith/Dev/byteflow/manage.py runfcgi
Compared to a run of the mill apache setup, this is cake walk. I commented the important bits so you can get the gist of what’s going on. If you have a question required more detail, let me know.
After spending a few hours converting my data to the model that django built for byteflow, I then focused on cloning my old theme. As you can see, I didn’t to a bad job! I tried to clean it up a bit and make sure I have more content for the search engines to munch on. Other than that bit of extra work, building the templates for this was pretty simple.
As you may have noticed, I also began using feedburner to serve up my rss feeds. That’s where the apology comes in..If you’re subscribed, most likely you have experienced duplicate posts..like..all of them. So, with that said So Sorry!.
