Google Sitemaps needs a list of URLs to optimize crawling. Usually, this is no problem, since Google supplies a script you can run on your server to build that list.

But that fails, if the content of your site is not stored in HTML-files, but in TXT-files like DokuWiki does. So here is what I did to build that list of URLs:

We do all of this in the directory where DokuWiki stores the pages:

cd /home/hdocs/beta.linuxbasics.org/data
find ./ -iname "*.txt"

give us

./wiki/syntax.txt
./wiki/dokuwiki.txt
./wiki/playground.txt
./start.txt
./tutorials/pre/start.txt
... 

which is the URL except that:

• the first '.' need to be replaced with the base-URL “http://LinuxBasics.org”
• .txt needs to be .html

The editor 'sed' can help us with those replacements. It is the source of Perl's s/-command, so if you know Perl, this will be familiar: <code bash> sed -e 's#^./#http://LinuxBasics.org/#g ; s/.txt$g' </code>

• “sed -e” will execute the command given to the standard-input and print the result to the standard-output.
• “s#^.#http://LinuxBasics.org#g” will replace the dot at the beginning of the line ('^') with the base-URL.

This uses '#' as a delimiter instead of '/'. Why? Because it looks much better then the version with slashes: “s/^./http:\/\/LinuxBasics.org/g”

• “s/.txt$/.html/g” will replace the ”.txt” at the end of the line ('$') with ”” which removes ”.txt”.
• ”;” seperates the two commands.

find ./ -iname "*.txt" | sed -e 's#^./#http://LinuxBasics.org/#g ; s/.txt$//g'

gives us what we want:

http://LinuxBasics.org/wiki/syntax
http://LinuxBasics.org/wiki/dokuwiki
http://LinuxBasics.org/wiki/playground
http://LinuxBasics.org/start
http://LinuxBasics.org/tutorials/pre/start
http://LinuxBasics.org/tutorials/pre/md5sum