During testing, I discovered that the odd-ball task failed to run, and found the following error in the system log:

sudo: sorry, you must have a tty to run sudo

I traced this error to a line trying to invoke a perl command as a user called dynamic:

sudo -u dynamic /usr/bin/perl run-periodic-tasks --load 5 --randomly

A simple Google search turned up an obvious solution to the error: use visudo to disable sudo’s tty requirement, allowing sudo to be invoked from any shell lacking a tty (including cron). This would have solved my problem, but it just felt wrong, dirty, and most troublingly insecure.

One reason why sudo ships with the requiretty option enabled by default is, among other reasons, to prevent remote users from exposing the root password over SSH. Disabling this security precaution for a simple maintenance task already running as root seemed totally unnecessary, not to mention irresponsible. Moreover, client‘s script didn’t even need a tty.

Thankfully, there’s a better way: use su --session-command and send the whole job to the background.

su --session-command="/usr/bin/perl run-periodic-tasks --load 5 --randomly" dynamic &

This line launches a new, non-login shell (typically bash) as the other user in a separate, background process and runs the command you passed using the shell’s -c option. Sending the command to the background (using &) continues execution of the rest of the cron job.

A process listing would look like this:

root     28109     1  0 17:10 ?        00:00:00 su --session-command=/usr/bin/perl run-periodic-tasks --load 5 --randomly dynamic
dynamic  28110 28109  0 17:10 ?        00:00:00 bash -c /usr/bin/perl run-periodic-tasks --load 5 --randomly

Note the parent process (PID 28109) is owned by root but the actual perl process (PID 28110) is being run as dynamic.

This in-script solution that replaces sudo -u user cmd with su --session-command=cmd user seems much better than relying on a change in sudo‘s default (and more secure) configuration to me.

I spent about an hour looking into this this afternoon and it turns out that much of this information is readily available on the filesystem in the Apple WikiServer’s data store. I whipped up the following shell script to extract this information in CSV format, exactly as requested.

I’m posting this script here in case someone else wants similar “export a list of WikiServer pages to a comma-separated values (CSV) file” functionality but isn’t sure how to go about getting it. To use this, just edit the line that reads http://my-server.example.com/groups/wiki/ so that it refers to the wiki base URI of your own server.

Update: The latest version of this script is now available at its Github-hosted repository. You should probably use that instead of the script below.

#!/bin/sh -
#
# Script to extract data from an Apple WikiServer's data store by querying the
# filesystem itself. Creates a 'wikipages.csv' file that's readable by any
# spreadsheeting application, such as Numbers.app or Microsoft Excel.app.
#
# USAGE:   To use this script, change to the WikiServer's pages directory, then
#          just run this script. A file named wikipages.csv will be created in
#          your current directory. For instance:
#
#              cd /Library/Collaboration/Groups/mygroup/wiki  # dir to work in
#              wikipages2csv.sh                               # run the script
#              cp wikipages.csv ~/Desktop                     # save output
#
# WARNING: Since the WikiServer's files are only accessible as root, this script
#          must be run as root to function. Additionally, this is not extremely
#          well tested, so use at your own risk.
#
# Author:  Meitar Moscovitz
# Date:    Mon Sep 22 15:03:54 EST 2008

##### CONFIGURE HERE ########

# The prefix to append to generated links. NO SPACES!
WS_URI_PREFIX=http://my-server.example.com/groups/wiki/

##### END CONFIGURATION #####
# DO NOT EDIT PAST THIS LINE
#############################

WS_CSV_OUTFILE=wikipages.csv
WS_PAGE_IDS_FILE=`mktemp ws-ids.tmp.XXXXXX`

function extractPlistValueByKey () {
    head -n \
      $(expr 1 + `grep -n "<key>$1</key>" page.plist | cut -d ':' -f 1`) page.plist | \
        tail -n 1 | cut -d '>' -f 2 | cut -d '<' -f 1
}

function linkifyWikiServerTitle () {
    echo $1 | sed -e 's/ /_/g' -e 's/&amp;/_/g' -e 's/&gt;/_/g' -e 's/&lt;/_/g' -e 's/\?//g'
}

function formatISO8601date () {
    echo $1 | sed -e 's/T/ /' -e 's/Z$//'
}

function csvQuote () {
    echo $1 | grep -q ',' >/dev/null
    if [ $? -eq 0 ]; then
        echo '"'$1'"'
    else
        echo $1
    fi
}

ls -d [^w]*.page | \
  sed -e 's/^\([a-f0-9][a-f0-9][a-f0-9][a-f0-9][a-f0-9]\)\.page$/\1/' > $WS_PAGE_IDS_FILE

echo "Title,ID,Date Created,Last Modified,URI" > $WS_CSV_OUTFILE
while read id; do
    cd $id.page
    title=$(extractPlistValueByKey title)
    created_date="$(formatISO8601date $(extractPlistValueByKey createdDate))"
    modified_date="$(formatISO8601date $(extractPlistValueByKey modifiedDate))"
    link=$WS_URI_PREFIX"$id"/`linkifyWikiServerTitle "$title"`.html
    cd ..
    echo `csvQuote "$title"`,$id,$created_date,$modified_date,`csvQuote "$link"` >> $WS_CSV_OUTFILE
done < $WS_PAGE_IDS_FILE
rm $WS_PAGE_IDS_FILE

For those new to the Wiki Server, this introduction to the Apple WikiServer for web developers may be of interest.

If you’re not afraid of the Terminal, this can be avoided. Put the following lines in your ~/.profile (or similar):

alias rmds='find . -name ".DS_Store" -type f -print0 | xargs -0 rm'

What this does is creates a new command that you can use (rmds) which recursively finds and deletes any regular file named “.DS_Store” starting from the current directory. Thus, running this command in the folder you are about to create an archive out of will clean it first, and will prevent unnecessary confusion on the part of your archive file recipient.

Alternatively, another way to do this is to use the command-line zip program and an (admittedly more complicated) pipeline to remove the .DS_Store files after they have been added to the archive. To do that, use this series of commands:

zip -d ZIPfile.zip `unzip -l ZIPfile.zip | grep .DS_Store | awk '{print $4}'`

where, naturally, ZIPfile.zip is the ZIP archive you want to remove the .DS_Store files from. Creating an alias out of that command (and making it work for paths that contain spaces) is left as an exercise for the reader. ;)

As an aside, the alias, find and xargs commands are incredibly useful in their own right and can be used to do a lot of pretty amazing things. As always, man command will give you the nitty gritty.

Also as an aside, you can stop the Finder from creating .DS_Store files entirely when browsing network volumes (like Windows shares) with another command, documented in Apple’s Knowledge Base.

for i in *; do
    v=`grep 'version = "' "$i/current/$i/"*.info |
      cut -d ':' -f 2 |
        sed -e 's/^version = "/v/' -e 's/"$//'`
    svn cp "$i/current" "$i/$v"
done;

It’s a bit buggy for some modules that have multiple .info files, but I’m sure a few more pipeline stages can fix that. (Which, because I’m done with this at the moment, I will leave as an exercise to the reader.)

Chalk this one up as another testament to the power of shell scripting and how it can help every developer get their job done faster.