You can run Pyblosxom as a CGI script with many web servers. This document covers setting Pyblosxom up as a CGI script.
You need an account on a web server configured to run CGI scripts. It helps to know how to run CGI scripts on that server, too.
Copy the pyblosxom.cgi file from the blog directory (the directory which you created with pyblosxom-cmd create ./blog/) into your CGI directory.
Edit the pyblosxom.cgi file.
The top of the file looks something like this
1 2 3 4 5 6 7 8 9 10 11
#!/usr/bin/env python # -u turns off character translation to allow transmission # of gzip compressed content on Windows and OS/2 #!/path/to/python -u import os, sys # Uncomment this line to add the directory your config.py file is # in to the python path: #sys.path.append("/path/to/directory/")
Make sure the first line points to a valid python interpreter. If you’re using virtualenv, then make sure it points to the python interpreter in the virtual environment.
Uncomment the sys.path.append("/path/to/directory/") line and make sure the path being appended is the directory that your config.py file is in.
Make sure the pyblosxom.cgi file has the correct permissions and ownership for running a CGI script in this directory for the server that you’re using.
Make sure your blog directory has the correct permissions for being read by the process executing your CGI script.
Run your pyblosxom.cgi script by doing:
% ./pyblosxom.cgi test
If that doesn’t work, double-check to make sure you’ve completed the above steps, then check the trouble-shooting section below, then ask for help on IRC or the users mailing list. More details in Project details, contact information, and where to go for help.
If that does work, then try to run the CGI script from your web browser. The url is dependent on where you put the pyblosxom.cgi script and how CGI works on your web server.
We’re going to try to break this down a bit into categories. Bear with us and keep trying things.
If you have problems and have gone through this section to no avail, ask a question on the pyblosxom-users mailing list or ask us on IRC. Details for both of these are on the website.
First, you should check to see if you have the minimum requirements for Pyblosxom on your system. They’re listed in the Requirements section of the Install chapter. If not, then please install them.
If Python is installed on your system, make sure the first line in pyblosxom.cgi points to the correct Python interpreter. By default, pyblosxom.cgi uses env to execute the Python interpreter. In some rare systems, /usr/bin/env doesn’t exist or the system may have odd environment settings. In those cases, you may edit the first line to point to the Python interpreter directly. For example:
Then try running ./pyblosxom.cgi again.
If Python is installed on your system and the first line of pyblosxom.cgi is correct, check for permissions issues. pyblosxom.cgi is a script, so it needs execute permission in order to function. If those aren’t set, then fix that and try running ./pyblosxom.cgi again.
Check the error logs for your web server.
When you try to look at your blog and you get a HTTP 404 error, then you’re using the wrong URL. Here are some questions to ask yourself:
At this point, running ./pyblosxom.cgi at the command prompt should work fine. If you haven’t done that and you’re busy trouble-shooting, go back and review the deployment instructions.
If the problem is with Pyblosxom and not your web server, then you should see a pretty traceback that will help you figure out what the specific problem is.
If the traceback and information doesn’t make any sense to you, ask a question on the pyblosxom-users mailing list or ask us on IRC. Details for both of these are on the website.
If you don’t see a traceback, then you either have a configuration problem with your web server or a configuration problem with Python. The first thing you should do is check your web server’s error logs. For Apache, look for the error.log file in a place like /var/logs/apache/ or /var/logs/httpd/. If you don’t know where your web server’s error logs are, ask your system administrator.
Does the account your web server runs as have execute access to your pyblosxom.cgi script? If your web server does not have the permissions to read and execute your pyblosxom.cgi script, then your blog will not work.
Do you have plugins loaded? If you do, comment out the load_plugins setting in your config.py file so that Pyblosxom isn’t loading any plugins.
py["load_plugins"] = ['plugina', 'pluginb', ...]
would get changed to:
# commenting this out to see if it's a plugin problem # py["load_plugins"] = ['plugina', 'pluginb', ...]
Check to see if the problem persists. Sometimes there are issues with plugins that only show up in certain situations.
Try changing the renderer for your blog to the debug renderer. You can do this by setting the renderer property in your config.py file to debug. For example:
py["renderer"] = "debug"
That will show a lot more detail about your configuration, what the web server passes Pyblosxom in environment variables, and other data about your blog that might help you figure out what your problem is.
If that doesn’t help, ask a question on the pyblosxom-users mailing list or ask us on IRC. Details for both of these are on the website.
We encourage you not to try any of this until you’ve gotten a blog up and running.
This section covers additional advanced things you can do to your blog that will make it nicer. However, they’re not necessary and they’re advanced and we consider these things to be very much a “you’re on your own” kind of issue.
If you ever have problems with Pyblosxom and you ask us questions on the pyblosxom-users or pyblosxom-devel mailing lists, make sure you explicitly state what things you’ve done from this chapter. It’ll go a long way in helping us to help you.
In the default installation, the Pyblosxom script is named pyblosxom.cgi.
For a typical user on an Apache installation with user folders turned on, Pyblosxom URLs could look like this:
http://example.com/~joe/cgi-bin/pyblosxom.cgi http://example.com/~joe/cgi-bin/pyblosxom.cgi/an_entry.html http://example.com/~joe/cgi-bin/pyblosxom.cgi/dev/another_entry.html
That gets pretty long and it’s not very good looking. For example, telling the URL to your mother or best friend over the phone would be challenging. It would be nice if we could shorten and simplify it.
So, we have some options:
Both methods are described here in more detail.
There’s no reason that pyblosxom.cgi has to be named pyblosxom.cgi. Let’s try changing it blog. Now our example URLs look like this:
http://example.com/~joe/cgi-bin/blog http://example.com/~joe/cgi-bin/blog/an_entry.html http://example.com/~joe/cgi-bin/blog/category1/another_entry.html
That’s better looking in the example. In your specific circumstances, that may be all you need.
You might have to change the base_url property in your config.py file to match the new URL.
The base_url value should NOT have a trailing slash.
If you’re running on Apache, you might have to tell Apache that this is a CGI script even if it doesn’t have a .cgi at the end of it. If you can use .htaccess files to override Apache settings, you might be able to do something like this:
# this allows execution of CGI scripts in this directory Options ExecCGI # if the user doesn't specify a file, then instead of doing the # regular directory listing, we look at "blog" (which is our # pyblosxom.cgi script renamed) DirectoryIndex blog # this tells Apache that even though "blog" doesn't end in .cgi, # it is in fact a CGI script and should be treated as such <Files blog> ForceType application/cgi-script SetHandler cgi-script </Files>
You may need to stop and restart Apache for your Apache changes to take effect.
Apache has a module for URL rewriting which allows you to convert incoming URLs to other URLs that can be handled internally. You can do URL rewriting based on all sorts of things. See the Apache manual for more details.
In our case, we want all incoming URLs pointing to blog to get rewritten to cgi-bin/pyblosxom.cgi so they can be handled by Pyblosxom. Then all our URLs will look like this:
http://example.com/~joe/blog http://example.com/~joe/blog/an_entry.html http://example.com/~joe/blog/category1/another_entry.html
To do this, we create an .htaccess file (it has to be named exactly that) in our public_html directory (or wherever it is that /~joe/ points to). In that file we have the following code:
RewriteEngine on RewriteRule ^blog?(.*)$ /~joe/cgi-bin/pyblosxom.cgi$1 [last]
The first line turns on the Apache mod_rewrite engine so that it will rewrite URLs.
The second line has four parts. The first part denotes the line as a RewriteRule. The second part states the regular expression that matches the part of the URL that we want to rewrite. The third part denotes what we’re rewriting the URL to. The fourth part states that after this rule is applied, no future rewrite rules should be applied.
If you do URL rewriting, you may have to set the base_url property in your config.py accordingly. In the above example, the base_url would be http://example.com/~joe/blog with no trailing slash.
For more information on URL re-writing, see the mode_rewrite chapter in the Apache documentation for the version that you’re using.