Vote up your favourite bash commands. I love the linux community.

simple note logging in bash. I could not resist tweaking the file name creation bit as follows...

function n {
NAMESTRING="$*"
NAMESTRING=${NAMESTRING// /_}
gedit ~/Dropbox/notes/$NAMESTRING.txt
}

This works, but I commented out the mkdir command for a directory where new images are being added.

Good tutorial on using lftp as a batch client in particular how to mirror a local directory to a remote directory.

Ask permission to use the redesigns. Nice CSS and semantic mark up.

Perl scrip works. I had to take out all the narrow table gunge and add in a $description field into the section that prints the html. I recast the thing as a definition list.Produces an html formatted fragment ideal for generating an inc file for use with Cugley's sed script.

I've already got an xslt processor installed!

XML tools for transforming xml documents

I'm thinking of a script that will wget the rss feed for this page on pinboard and pull it down as a text file, then using Cugley's sed filter script to render it in html and then have it uploaded as part of the web site...

Debian: gnome-terminal sessions need to be set so that they use commands like a log-in shell. Then you can add folders to your path in the .bash_profile and they will appear in a gnome-terminal session run under X. Last post is the key!

Applications | System Tools | Configuration Editor | Apps | Gnome Terminal | Profiles | Default | check the login_shell option.

Editors and version control

another one that uses make. Used in Webconvert. Could use the make bit to automate the indexes.

getting closer

some more stuff about html from bash prompt

How to add a header and footer...

The Terminal Window That Never Forgets

Add a few arguments to ls in this script and we could be away with automatic indexing. The -c argument works and produces a list of files in order of the modification time. Use the *.html wildcard to limit indexing to html files. Chop out the html page head and body tags so we get a fragment of clean html for use with Cugley's sed script. We are away.

Building html lists from a directory listing. On Debian, ls defaults to three column lists, so to get a list ordered by modification time so the most recent is first, you would use ls -1 -c.

A series of annotated examples including reading a text file into an array one line per element.

Python implementation of something very close to what I'm after. He's using make to glue a series of programs together.

LFTP uses the modified date to decide which files to upload to a server. Can use a script file. Would need to ensure that old pages were not 'touched' by the local publishing script. Wordpress and similar no good. MT might be ok if you didn't 'republish all pages' too often. LFTP is in the Debian repositories. rsync needs ssh, can't work over ftp.

A simple development package including apache and mysql with phpmysqladmin. I'll try getting a local word press working with this and then see if I can set the mod rewrite rules that wordpress needs to make 'sensible' urls. Then I might be publishing bodmas.org as static pages.

EDIT: installs fine, easy to install wordpress script to /opt/lampp/htdocs, easy to set up a local wordpress database using phpmysqladmin, easy to export from real wordpress blog, harder to get http paths right due to paths being hard coded into the wordpress database. I'll crack that one.

My particular server uses a version 4 MySQL and Xampp uses a version 5 one, so need to use a legacy install and latin1 encoding. A few errors due to symbols found but corrected. Need to find a way of increasing the 2Mb upload limit!

Edit2: modrewrite just works.

How to use wordpress to publish to a web site that does not allow scripting / have a database / or just because you are tired of updating the thing when they gte hacked again.

Edit: I've reached step 5 now. The local version of word press does not run especially fast, so the wget stage could take a bit of time...

How to use the linux / mac os wget command to save an entire blog or web site. It works, and the main drawback is that the links between pages in the site are 'absolute local'. I'm going to try using a multi file search and replace to make them all relative links. The other glitch is that the cascading style sheet is not saved (but my own site uses only a couple of style sheets for the whole thing, so I can save those manually). I was impressed that all the images and linked pdfs and hot potatoes quzzes came down as well. Back up your blog now! Via daringfireball and pinboard. EDIT I was wrong about the local absolute links, they are actually relative links and it fetches the style sheets. Magic.

The comments are better than the article. I'm going to try the grep piped to sed method. I want to change an absolute local web address to fully relative so I can burn a pile of linked web pages to a cd. See next bookmark...