Title: MultiMarkdown-CMS
Author: Fletcher T. Penney
Date: January 31, 2010
Tags: MultiMarkdown, web, server
NOTE: MultiMarkdown CMS is now updated to work with MMD 3.0. If you need a version that is compatible with MMD 2.0, you can choose an older version from the downloads page. MMD 3.0 must be installed separately.
Also, please note that MMD-CMS has been modified to output HTML that should be compatible with HTML5, in ancticipation of these changes being made in MMD as well.
MultiMarkdown CMS is basically a collection of utilities that allows you to run a web site without any other software. I have tried a variety of packages when hosting my own web site. In the end, however, I was basically using those programs to put a wrapper around my content, which was always managed by MultiMarkdown. Additionally, none of those programs really fit into my workflow, and the performance was often not what I would like.
So, instead, I decided to run my web site with a minimal amount of software. In fact, it could be run with MultiMarkdown alone, but I would miss out on some of the interactive features.
The basic premise is:
-
Create a plain text file (in MMD syntax) for each page of the site
-
use
mmd2web.pl
to convert to HTML, including some markup to format the pages for the web site -
upload the HTML file (and optionally the txt file) to your web server, and you're off
The newly formatted HTML file includes SSI code that causes Apache to add certain templates to the page. This allows you to standardize the appearance of each page (e.g. headers, footers, sidebar, etc).
Additionally, CGI scripts are included to manage archives, Atom feeds, tags, searching, OpenID authentication, comments, and more.
When building the system, I started off with a collection of essentially static web pages. I then added cgi scripts only where I thought they were absolutely necessary to give my site the interactive features I desired. If you're someone who likes lots of widgets and gadgets on their site --- I won't be programming them. However, because of the simple philosophy behind this approach, the pages should be compatible with just about anything you want to add that doesn't require a backend database.
However, keep in mind that my goal is simplicity. I don't want lots of plugins, extensions, modules, etc. I don't want lots of different themes that I can choose from. If you want these things, then another package might be for you. But if you want a simple method of creating a web site that focuses on content, this may be for you.
This package certainly requires more tinkering than a prebuilt solution. If you're not familiar with .htaccess files, configuring apache, changing file permissions, and examining error logs then you might run into some roadblocks. However, that could also be considered an opportunity to track down solutions!
You can download the package:
http://github.com/fletcher/MultiMarkdown-CMS
You can also use git to clone the source:
git clone git://github.com/fletcher/MultiMarkdown-CMS.git
If you know how to use git, I recommend that approach for reasons that will become clear later.
Once you download and unpack the software, place the directory where your web server software expects it, and rename it as desired.
Please note: if you have trouble installing the software, do a web search to figure out what to do. I'm happy to help troubleshoot my software, but I don't want to get into trying to troubleshoot everyone's web server setup....
For example, I place mine in /Users/fletcher/Sites/mmd-static
.
You should now be able to go to something like:
http://127.0.0.1/~fletcher/mmd_static/index.html
And see something --- it will likely give you some error messages, however.
MultiMarkdown CMS is fairly easy to install with Apache. You are only required to modify your Apache configuration files so that:
AllowOverride none
becomes
AllowOverride All
On a mac, for example, you would modify /etc/apache2/users/username.conf
.
There are two versions of the software --- the regular, and the advanced.
The regular version of the software is designed to be easier to install. It can go in any folder/URL in your web host (i.e. it does not have to be at the root). You do have to configure the .htaccess file so that the software knows where it is located so that links are coded properly. When using this version, when you want to link to your home page, for example you would link to "index.html", NOT "/index.html".
This flexibility comes at a cost --- links within a page, e.g. "#footnote1",
will not work with the regular version of MMD CMS. This is a limitation of the
<base href>
feature.
If you want to allow internal links, then you need to use the advanced version of MMD CMS. This version does need to be installed so that it lives at a top-level URL. This can be accomplished by putting it at the webserver root folder (wherever that is for your particular server), or by using a virtual host. I can't support everyone in figuring out how to do that, but I have included what I did to get it working on my particular machine.
For Apache 2 on a Mac, do the following (all others will need to seek help elsewhere):
-
as an admin, go to
/etc/apache2/extra
-
sudo pico httpd-vhosts.conf
(or similar) -
comment out the existing virtualhost sections (add a "#" at the beginning of the line)
-
then, add something like the following at the end of the file:
<VirtualHost *:80> DocumentRoot "/Users/fletcher/Sites/mmd_static" ServerName mmd.local </VirtualHost>
-
then add the following to
/etc/hosts
:127.0.0.1 mmd.local
-
finally,
sudo pico /etc/apache2/httpd.conf
and uncomment the following line:Include /private/etc/apache2/extra/httpd-vhosts.conf
If you restart Web Sharing in the control panel, you should now be able to
access your site by pointing your browser to http://mmd.local/
. It should
look much better than the first time, but the default CSS is ugly. There are
two sample articles, and one sample tag, to get you started.
You can then modify the .htaccess
file just like the regular version.
You can add content anywhere within the MultiMarkdown CMS directory. You can
create folders, and bury your content away. Keep in mind that the archives
feature only locates pages in a /YYYY/MM/
folder, e.g. the sample posts in
/2010/01/
. Other folders are not treated as "blog posts". My site is an
example of this structure.
To process the files into HTML, you need to have a working installation of MultiMarkdown.
Once you add the text files, you need to run mmd2web.pl
to create an html
file. This uses the xhtml-static-site.xslt
file to add the appropriate
templates to your documents. You will also need to change the permissions on
your .html
files so that the execute bit is enabled (required by the
XBitHack
feature in order to enable SSI without screwing everything else up).
Now you just need to customize your templates and CSS files to make everything look pretty....
Included with MultiMarkdown-CMS are several cgi scripts that add additional features:
-
pages in the
/YYYY/MM/
hierarchy are treated as "blog posts"- organized by date -
an archive of blog posts is available at
/archives/
, you can also go to any year or year/month and see a list of posts in that month -
visitors can leave comments on any blog post (by default comments are not accepted on other pages)
-
visitors are authenticated via OpenID for leaving comments
-
two Atom feeds are generated automatically - one for blog posts, and one for comments
-
/sitemap.xml
generates a Google compatible sitemap of the content of your site -
by default, the home page lists the last 15 blog posts
-
by default,
index.html
pages include a list of the other pages within that folder -
a built-in search function is available (though setting up a Custom Search Engine isn't a bad idea either)
-
a list of similar pages can be added to each page in the site
-
several tag related features are available automatically ---
/tags/foo
searches for posts taggedfoo
and/tagmap/
shows a list of current tags with links to those pages -
default CSS allows you to switch from a right-sided sidebar, to a left-sided sidebar with a simple keyword switch in
header.html
---rightmenu
vsleftmenu
(thanks to Matthew James Taylor) -
valid HTML 5 & valid CSS, with the exception of unsupported
<meta>
elements if you have metadata
These features are all demonstrated by the default configuration - feel free to modify the scripts as needed for your own requirements.
The primary metadata used to for this software consists of three tags:
- Title --- Self-explanatory
- Date --- either
01/31/2010 15:59:56
, or01/31/2010
- Tags --- comma separated list of tags (case sensitive)
There are a couple of configuration changes that should be done before going live with this software:
-
the root
.htaccess
file should be changed to reflect your desired time zone -
four files in
/cgi/
contain the phrase "enter your password here" --- this should be changed to a private passphrase for better security when handling OpenID authentication
Obviously, I recommend improving the CSS design, and tweaking the various templates to construct the look of your site.
One of the things I really wanted to be able to do was to manage my site on my laptop, experiment with it to be sure it works after any changes are made, and only then to upload my changes to the live server. I have tried to design the cgi scripts so that they configure themselves as much as possible, so that the same code can be run locally, or from a production server.
My workflow is that I make changes in /Users/fletcher/Sites/mmd_static/
as
described above. That directory was created as a git clone of the
MultiMarkdown-CMS software(as above).
I then add a remote repository to the git repo:
git remote add live ssh://user@my.live.host/path/to/public
where the username and host are what I would normally use to ssh into my web
provider. /path/to/public
is the path to the directory that holds my web
site.
This allows me to use git to upload a copy of my local site to my web host when I am satisfied with the changes. Additionally, by storing my web site in a git repository, I can undo changes when I make a mistake, and I can also go back to any previous version of my site.
One trick, though, is that you have to configure the remote repository to
update the remote copies of all files after every git push
. To do this, ssh
into your account, and go to the public directory. (You may want to back
everything up at this point....) Then do the following to create the remote
git repository:
git init
cd .git/hooks
pico post-update
and then put the following in the file:
#!/bin/sh
#
cd /path/to/public
/usr/bin/env -i /usr/local/bin/git reset --hard
chmod g+s cgi/accept_comment.cgi
What this does is force git on the remote server to reset the remote directory to match the state of the repository itself. Then I change the permissions on accept_comment.cgi so that it is permission to write to the comments files.
You can, of course, add your own commands to this file as well.
If you use the same approach, whenever you want to upload your changes, simply type:
git push live master
Your changes will be uploaded, then the remote repository will be reset so it's up to date, and the permissions will be fixed on accept_comment.cgi.
Remember, since you have a read-only connection to my github repository, your
changes and private files cannot be uploaded. You can still use git pull origin master
to update to the latest version of software I have released.
This should not overwrite any changes you have made, since git is pretty good
at avoiding "collisions". I've only been using this "triple repository"
approach for a little while, but so far it's been convenient in allowing me to
keep private and public files separate.
NOTE: This set-up works for me. I don't promise it will work for you, or that it won't mess something up. I am not a git expert. I simply patched this together with a lot of help from Google. I can't really offer any support if it's not working for you.
This software would not be possible without work previously done for many other projects:
- by Brad Fitzpatrick
- http://search.cpan.org/~mart/Net-OpenID-Consumer/
Used for OpenID authentication.
- by Benjamin Trott
- http://search.cpan.org/dist/Crypt-DH/
Crypt-DH is used to enable the OpenID authentication.
- by Aristotle Pagaltzis
- http://search.cpan.org/~aristotle/XML-Atom-SimpleFeed/
Used to generate the atom feeds.
- by Fletcher T. Penney
- http://fletcherpenney.net/
Used to find similar pages, and perform searches of site content.
- by Benjamin Franz and Jim Richardson
- http://search.cpan.org/dist/Lingua-Stem/
Used by my VectorMap searching software to improve results
- by Fletcher T. Penney
- http://fletcherpenney.net/mmd/
MultiMarkdown is my update to John Gruber's Markdown syntax, and built using modifications to peg-markdown by John MacFarlane.
- Email is "de-obfuscated" when run through an XSLT, and it appears there is no (straightforward) way around this. Not that this does not affect email addresses that are placed in your HTML include statements --- in a sidebar, for example.
Copyright (C) 2010-2011 Fletcher T. Penney <fletcher@fletcherpenney.net>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the
Free Software Foundation, Inc.
59 Temple Place, Suite 330
Boston, MA 02111-1307 USA