How? With a shell script and symlinks, my dear Watson. As with all shell scripting this is probably something you can do drunk.

First I need to get a list of all my articles so that I can link to them. Next I take this permalink address, pipe it through sha1sum, take the last 10 characters and make a symlink from the permalink path and presto: short urls

The shell one-liner becomes something like this (note: I've put the echo in there so you can see what happens).

#!/bin/bash
NB_ARCHIVES=/home/miekg/miek.nl/blog/archives
NB_SHORT=/home/miekg/miek.nl/s
for i in $(find $NB_ARCHIVES -type d -wholename "*2???/*/*/*" -print)
do
    HASH=$(echo $i | sha1sum)
    echo ln -sf $i $NB_SHORT/${HASH:30:10}
done

This scripts can be dropped in your plugin directory, so that it runs after each update in nanoblogger.

In your short urls directory (here: miek.nl/s) you can just use ls and grep to find out which short url links to what article.

So for now a simple shell script, which you can find here.

Usage within a template

In your templates you should include the following:

$(/home/miekg/miek.nl/blog/bin/related $NB_EntryID)

the shell script returns the related articles separated with comma's.

Update (200911-08): I've stopped using this, because it made the archive generating fail in mysterious ways...

As you might have seen I have now these nice (opinions may vary) "more..." links in my blog. This is done with some JavaScript and shell foo and in this entry I will explain how I did it.

The setup also works with text browsers and people who disable JavaScript. In the latter case you will get the normal nb behavior where you see the entire article. When using a text browser you also see the entire article but there is also a (non functional) 'more' link.

User side

When writing a blog entry the only thing you need to do is to put the following on a line

- read more -

(technically you need more -'s to make it work, but this will suffice for this example).

Anything you type below that line will become collapsible with a 'more' link. Anything above will be directly visible.

Implementation

First a shell script which reads your entire blog entry and replaces the - read more- with the html and JavaScript to make it work.

#!/bin/bash
EPOCH=$(date +%s)
ART_ID="art_$$_$EPOCH"
HREF_ID="href_$$_$EPOCH"
DATA=$(cat)

# escape galore
repl="<script type=\"text\/javascript\">\n\
<!--\n\
function toggle_visibility(id, anchor) {\n\
   var e = document.getElementById(id);\n\
   var a = document.getElementById(anchor);\n\
   if(e.style.display == 'block') {\n\
      e.style.display = 'none';\n\
      a.innerHTML = \"\&laquo; more ...\";\n\
   } else {\n\
      e.style.display = 'block';\n\
      a.innerHTML = \"\&laquo; less\";\n\
   }\n\
}\n\
\/\/-->\n\
<\/script>\n\
<div class=\"readmore\">\n\
<a class=\"readmore\" style=\"display: none;\" id=\"$HREF_ID\"\n\
href=\"javascript:\/\/\" onclick=\"toggle_visibility(\'$ART_ID\', \'$HREF_ID\');\">\
\&raquo; more ...<\/a>\n\
<\/div>\n\
<div id=\"$ART_ID\">\n\
<script type=\"text\/javascript\">\n\
<!--\n\
e = document.getElementById('$ART_ID');\n\
e.style.display = 'none';\n\
e = document.getElementById('$HREF_ID');\n\
e.style.display = 'block';\n\
\/\/-->\n\
<\/script>\n\
"
if echo "$DATA" | egrep -q -- '-{3,}.?read.more.?-{3,}'; then
echo "$DATA" | sed -r "s/-{3,}.?read.more.?-{3,}/$repl/"
echo "</div> <!-- id=$ART_ID -->";
else
echo "$DATA"
fi

As I'm using stdin twice I am capturing it in the variable $DATA.

The variable $repl is filled and is used by sed the replace - read more-. It was a real pain to correctly quote everything (as you might see).

The JavaScript trick to do this collapsing can be easily found on the internet (google: hide/show divs).

interfacing with nanobloggger

You need to edit two templates, entry.htm and permalink.htm, in the first one you enable the 'read more'-feature and the latter you need to get rid of the '- read more -' line.

entry.htm

In entry.htm replace

$NB_EntryBody

with

$(echo "$NB_EntryBody" | ${BLOG_DIR}/bin/readmore)

Adjust this to where you have saved the shell script from above.

permalink.htm

On your permalink page this read more stuff is ridiculous, so there we must get rid of it.

Where it says

$NB_EntryBody

replace that with

$(echo "$NB_EntryBody" | sed -r 's/-{3,}.?read.more.?-{3,}//')

to filter out the line.

Basicly these blog items serve as a personal reminder and documentation. Also the code needs to be refactored a bit and some html tags need to be removed from it.

Files needed

The following files are needed for this comment system:

• comment.php, this is the PHP file the implements everything

• nbadmin, a small shell script that implemented the comment moderation (just a mv of the comment to the correct directory)

• nbnofity, small shell script that notifies you a new comments (to be put in cron)

• com.css, the css code for the comment pages.

Templates

In your templates you will need to require the php script, so a

require_once "/home/miekg/miek.nl/comment.php";

must be there somewhere.

Further more all pages that do something with comments need the following code:

• Get the current comments and put them in a PHP array

$comments = gather("$NB_EntryID");

• Show the current comments as HTML

show($comments);

• Check the current form input and act accordingly

if ($_POST['preview'] == "Preview Comment") {
        preview();
}
if ($_POST['submit'] == "Submit Comment") {
        submit();
}

• Show the comment input form

form("$NB_EntryID");

• Show a "recent comment" section in your side title

recent_comments();

And that should be it.

• there are no user accounts
• has a preview function
• once a comment is submitted you cannot edit it anymore
• all comments are moderated
• a small set of BB tags are allowed
• implemented as one PHP file
• the PHP script has some html in it
• needs a writeable directory in your document root

You need this file: comment.php.

The first few lines allow for some customization:

$NB_TIME = "%Y-%m-%d %T %Z"; # strftime function
$NB_COM_BASE="/home/miekg/miek.nl/blog/comments";
$NB_COM = 1;    # 0 -> disable comments
$NB_COM_CLOSE = 2592000; # number of seconds after which comments are closed
$NB_COM_MAX_SIZE = 2000; # max number of chars in a comment

which have the following meaning:

1. NB_TIME: how to format the time and date of the comment
2. NB_COM_BASE: where to put the comment files
3. NB_COM: emergency comment disable, when 0 all commenting is disabled
4. NB_COM_CLOSE: number of seconds when commenting on an article is disabled
5. NB_COM_MAX_SIZE: maximum number of character per comment.

In NB_COM_BASE all comments are put, each article gets a directory named after its $NB_EntryID. In that directory, two sub dirs are made ok and new. All comments are first put in the new directory, where they wait for moderation. See below for the moderation script.

Further more you need to tweak apache to allow PHP code in .html files or make nb create .php files.

In all files you will need to following code:

<?php
require "/home/miekg/miek.nl/comment.php";
?>

Of course you need to use a different path for your situation.

comment string

In articles you will need something like comments: (5) with comment.php the following is needed (remember the shell is touching this code too, so you'll need to escape the PHP stuff):

<?php
echo commentstr($NB_EntryID, ${ARCHIVES_PATH}$NB_EntryPermalink);
\$comments =  gather($NB_EntryID);
echo "(" . count(\$comments) . ")";
?>

Or whatever you want to put there. The gather() function will retrieve the current comments of this article. It will also create a link to the permalink of the article as I decided that new comments are to be added there.

adding comments

In my permalink template I have the following code: Note: 1) the shell is touching is and 2) my templates are written in m4, where I've made '[' and ']' mean something, hence the double '[[' and ']]'):

<?php
show(\$comments);
if (\$_POST[['preview']] == "Preview Comment") {
    preview();
}
if (\$_POST[['submit']] == "Submit Comment") {
    submit();
}
form("$NB_EntryID");
?>

Well this kinda speaks for itself. A user can preview comments when pressing 'Preview' in the form and for submitting submit is pressed.

The form() function displays the commenting form.

moderation script

After a comment has been left in the new/ directory it needs to be moved to the ok/ dir. I've written nbadmin for this purpose. A typical run looks like this:

> Article: e2009-01-12T19_34_47.txt
 >> Comment found: 1232783070_uvTSOy
<a href="http://miek.nl">Miek Gieben</a>
Test 123
(A)approve (D)discard (N)skip [A]: A
new/1232783070_uvTSOy ok/1232783070_uvTSOy

And with that the comment is okayed.

cron job

Of course you want to be notified of new comments, this is implemented as a cronjob which checks if there are any files in a new/ directory.

The quick and dirty script:

#!/bin/sh
# send email for the comments that are found in my website
COM_BASE="/home/miekg/miek.nl/blog/comments"
BODY=""

for art in $COM_BASE/*.txt; do
for com in $art/new/*; do
    if [ $com = $art/new/* ]; then continue; fi
    BODY="$BODY\n o $com"
done
done

if [ ! -z "$BODY" ]; then
echo "New comments found for miek.nl"
echo ${BODY}
fi

And my cronjob:

% crontab -l
MAILTO=miek@atoom.net
# check if there are new comments
10 8-22/2 * * * /home/miekg/bin/nbnotify

Comment systems for nb

I've been searching for a good comment system that I want to use for nb. It should meet the following criteria:

1. Simple moderation system. With all the spam nowadays every comment made should be moderated.
2. Should not require a database, the filesystem is enough of a database to handle these tasks. This does imply that there will be no user accounts, nor the possibility to edit your comments.
3. Simple interface with nanoblogger.

I've looked at blogkomm, haloscan and cgicomment.

blogkomm

Looks nice, easy to setup, but I could not get the moderation to work. Also I looked like all the comments are kept in a single file called comments.txt, but I prefer the comments to be more linked to the article they belong to.

haloscan

This is a external site which takes care of the comments. So here I'm I running nb because I prefer to keep things simple and then I outsource my comment system to a site which I do not control and which does rely on databases and what not.

cgicomment

Could ... not ... found ... the ... damn ... thing ... anywhere.

Conclusion

I need to write my own system. I have the following in mind:

• Simple PHP based system
• All comments are put in a 'to-be-moderated' directory. A simple tool called nbadmin will control if a comment is let through. You can compare it with listadmin which you can use when you use mailman. If you 'OK' a comment, the comment will be moved to another directory,
• Notification of new comments to 'someone'. This will probably will be implemented as a cron job which scans the 'to-be-moderated' directories.
• Some PHP functions which allow the display of comments and which can be integrated with the rest of nanoblogger.

The implementation will (probably) result in one PHP file called comment.php which needs to be included in your templates.

I'm still in the early stages of this project and I do not yet know if it will work like I want.

This is also something that is done much better in zsh.

Why do want this? SPEED!

Shell scripting is a very easy way to program, but forking all these helper programs takes a lot of time. So one way to speed up your shell program is to load these programs into the shell and making them a builtin.

bash

I needed the bash source for this to work. So download it and compile it. Then in the examples/loadables you have some new builtins, like cat, cut, head and then some.

Making cat a builtin, you need the enable command

$ type cat
cat is /bin/cat

$ enable -f /tmp/bash-3.2/examples/loadables/cat cat
$ cat type
cat is a shell builtin

Same goes for other executables. When updating my site with nanoblogger, using cat and others as builtins made the update much quicker. I shaved off some 20 seconds on a total of 250 seconds; so it may be worth the effort.

zsh

Here everything worked out of the box, but there is currently no cat builtin for zsh (I created one). But there are other interesting builtins.

What you need: man zshmodules and zmodload. With zmodload you can create new builtins.

For example, I made a cat builtin. Next I copy the cat.so to /usr/lib/zsh/4.3.4/zsh

% type cat
cat is /bin/cat

% zmodload zsh/cat
% type cat
cat is a shell builtin

Unload with zmodload -u. Works like a charm.

You can fetch it here.

I'm putting this up for download, so that other people may use this too. gallery.sh If you want to regenerate the slides, you can just remove the html and slides directories.