Ramblings of an Eggplant


My initial experience with Google Inbox

This past week I finally decided to try out Google Inbox. The feature that really drew me to it was the ability to snooze emails. This feature allows you to make an email leave your inbox and come back at some later date and time. This is a really cool feature and a nice way to delay dealing with an email until you need to. In addition to this it is really easy to create filters that add emails to bundles (labels). These bundles can be set to appear in the inbox or not and you have some control over when the bundles appear in the inbox. When a bundle appears in the inbox it shows up as a wide message, once opened you see all of the messages in the bundle. This is a nice way to be able to group messages; you can see your labels in the inbox in a compact fashion. You can also decide which bundles will trigger notifications in the android app.

After using Inbox for about a week, I've decided to go back to using GMail. Here are my reasons:

  1. The keyboard shortcuts in the web interface are lacking.
    • No keyboard shortcut to goto a label/bundle
    • No keyboard shortcut to type in the name of a bundle to move to. There is a shortcut '.' to open the move to menu though.
  2. The bar on the left side showing the bundles don't show how many unread messages are in the bundle
  3. When you choose to have bundles show up in the inbox you can select as the messages arrive, once a day (7:00) and once a week (Monday 7:00). I would really like to be able to at least pick the time for once a day and once a week. It would be nice to be able to pick the day on the once a week.
  4. I like to make sure all messages that I keep are assigned at least one label. The Inbox interface doesn't allow me to see what labels have been applied to a message. This makes me very concerned that I will loose messages by them being archived and not assigned any labels. GMail's search interface is great, but I really like to be able to find my messages by label.

If Google fixes these features I will give Inbox a try again, until then I'm sticking with GMail.


IPv6 on Comcast Residential

Comcast has now opened up their IPv6 service to residential customers. If you have a supported modem from Comcast and a device connected to it that understands IPv6 you can connect. You might ask why would I want to setup IPv6? And that is a good question. One reason is to stay up to date with current networking technology. Another reason is that we're running out of IPv4 addresses and we will eventually need to switch to IPv6. Currently many sites on the Internet are supporting IPv4 and IPv6 to help with the adoption of IPv6. Another reason for IPv6 support is that this setup can give you a subnet of public IPv6 addresses to use in your house. Meaning that you can allow computers on your internal network to be accessible from the outside world. Of course this also means that you could potentially open up your computers to the outside world, so you need to be careful and setup your firewall to keep your internal computers secure unless you want them accessible. This also removes any issues with NAT as IPv6 doesn't have any NAT support.

For my setup I have a compatible modem from Comcast and a Linux computer as my router. My Linux computer is running Ubuntu. These instructions are specific to my setup, but should be able to be used by others running most any Linux distribution.

The first thing you should do is secure your network from IPv6 so that something doesn't get in while you're setting things up. Here is my IPv6 firewall setup, it's very similar to my IPv4 setup, except the port numbers for DHCP are different. Outbound traffic is allowed and inbound traffic is denied. I've also disabled forwarding of traffic, this prevents inbound traffic directly to the internet work. This script needs to be located at "/usr/local/sbin/firewall-ipv6-start" for the radvd script at the end of this post to work properly.




# accept everything by default

 -m state --state RELATED,ESTABLISHED \
 -m comment --comment "allow inbound traffic for established and related connections" \
 -i ${LAN_IFACE} -o ${INET_IFACE} \
 -m comment --comment "allow all Internet bound traffic from the internal network" \
${IPTABLES} -A FORWARD -p ipv6-icmp \
 -m comment --comment "forward any ICMP traffic" \

 -m state --state RELATED,ESTABLISHED \
 -m comment --comment "allow inbound traffic for established and related connections" \

 -i ${LO_IFACE} \
 -m comment --comment "allow any local-only traffic" \

 -p ipv6-icmp \
 -m comment --comment "allow ICMP traffic from anywhere" \

 -p udp -m udp --dport 546 \
 -m comment --comment "Accept DHCP traffic" \

 -p udp -m udp --dport 547 \
 -m comment --comment "Accept DHCP traffic" \

The remainder of this post is based upon this post on using DHCPv6 with prefix delegation.

The next thing we need to do is get an address from Comcast along with a prefix (subnet) to hand out to the computers on the internal network. Comcast doesn't appear to be using router advertisements for IPv6, so we'll need to use DHCP over IPv6. For this I could use the ISC DHCP server that I'm using for IPv4, but it doesn't support prefix delegation which I need to give the other computers in my house IPv6 addresses. For this I installed wide dhcp client. Ubuntu includes this in the package wide-dhcpv6-client. A side advantage to using a different DHCP client for IPv6 is that you can turn it off to disable IPv6 support without messing with your IPv4 network. Once you install the client edit /etc/wide-dhcpv6/dhcp6c.conf to look like this. You will need to modify the interface used and possibly the sla-len. I found the sla-len by trial and error. You won't get a prefix if the value is incorrect.

interface eth0 { # external facing interface (WAN)
 send ia-na 1;
 send ia-pd 1;

 request domain-name-servers;
 request domain-name;

 script "/etc/wide-dhcpv6/dhcp6c-script";

id-assoc pd 1 {
 prefix-interface eth1 { #internal facing interface (LAN)
 sla-id 0; # subnet. Combined with ia-pd to configure the subnet for this interface.
 ifid 1; # IP address "postfix". if not set it will use EUI-64 address of the 
         # interface. Combined with SLA-ID'd prefix to create full IP address of interface.
 sla-len 0; # prefix bits assigned. Take the prefix size you're assigned
            # (something like /48 or /56) and subtract it from 64. 
            # In my case I was assigned a /64, thus the value is 0

id-assoc na 1 {
 # id-assoc for external interface

When you start the wide DHCP client and all is happy you will find that your external interface has an address. In my case it starts with 2001:558:6014. See the output of "ip addr show dev eth0" changing the interface as appropriate. Below is the output for my system with the IP addresses masked out.

2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
 link/ether 00:d0:b7:3f:4d:18 brd ff:ff:ff:ff:ff:ff
 inet XXX.XXX.XXX.XXX/22 brd scope global eth0
 valid_lft forever preferred_lft forever
 inet6 2001:558:6014:XXXX:XXXX:XXXX:XXXX:XXXX/128 scope global 
 valid_lft forever preferred_lft forever
 inet6 fe80::XXXX:XXXX:XXXX:XXXX/64 scope link 
 valid_lft forever preferred_lft forever

Once this is setup you can ping IPv6 addresses from your router. You can test this with "ping6 google.com".

Now to allow your local network talk to the Internet via IPv6 you'll need to allow forwarding and then assign them IPv6 addresses.

First we'll tell the kernel to allow forwarding by modifying adding the file 70-ipv6-routing.conf to /etc/sysctl.d. Note that net.ipv6.conf.all.accept_ra is set to 2. Any other value will not work due to how the router advertisements are handled.

# only set this on the external interface, otherwise we don't get a
# default route for IPv6



Once you change these values you will need to reboot or use the sysctl utility to set them immediately.

Now to hand out IPv6 addresses to the rest of the network. This will be done by setting up radvd. The package 'radvd' on Unbuntu contains this daemon. Once installed you can setup /etc/radvd.conf for the prefix that Comcast gave you. However when your IP address changes you'll need to update the file. So instead I have created a script that can be run from wide dhcp client. Put the following in "/usr/local/sbin/update-ipv6-setup.sh" and add a call to this script from the end of /etc/wide-dhcpv6/dhcp6c-script. You'll need to change the interface in this script to be your internal interface.


debug() { ! "${log_debug-false}" || log "DEBUG: $*" >&2; }
log() { printf '%s\n' "$*"; }
warn() { log "WARNING: $*" >&2; }
error() { log "ERROR: $*" >&2; }
fatal() { error "$*"; exit 1; }
try() { "$@" || fatal "'$@' failed"; }

mydir=$(cd "$(dirname "$0")" && pwd -L) || fatal "Unable to determine script dir

prefix=$(ip -6 addr show dev eth1 scope global \
 | grep inet6 \
 | awk '{print $2}') \
 || fatal "Unable to get prefix"

cat > /etc/radvd.conf.new <<EOF

interface eth1
 AdvSendAdvert on;
 AdvIntervalOpt on;
 MinRtrAdvInterval 60;
 MaxRtrAdvInterval 300;
 AdvLinkMTU 1280;
 AdvOtherConfigFlag on;
 AdvHomeAgentFlag off;
 prefix ${prefix}
 AdvOnLink on;
 AdvAutonomous on;
 AdvRouterAddr on;


diff /etc/radvd.conf.new /etc/radvd.conf > /dev/null
if [ $? -ne 0]; then 
 # only move if there are differences
  try mv -f /etc/radvd.conf.new /etc/radvd.conf
  try service radvd restart
try /usr/local/sbin/firewall-ipv6-start

Now you have IPv6 setup on your router and your home network. I found that Linux, Windows and Mac automatically recognize the IPv6 router advertisements and grab addresses and setup routes appropriately.



Filed under: Tech Tips Comments Off

Skitch for easy sharing of drawings

I use a Mac at work and was recently introduced to Skitch in a training class. This application is really handy if you need to share images or screen shots with others. Skitch allows you to easily take a screen shot of what you're doing and then annotate it to state what is important in the image. You can also easily do simple drawings and then share them. Once you have an image that you like you can just drag it to your email program, or if you sign up for an account with Skitch you can have the image uploaded there and accessible from their website.

Filed under: Mac, Tech Tips Comments Off

Finally able to use my iPod Touch without Windows or Mac

I've finally made it. I can now use my iPod Touch just with Linux. Apple adding over the air updates really helped, but I still couldn't sync my music. I also found that pulling large videos off via Dropbox didn't work as the app would time out. I had been trying to get gtkpod to work, but kept running into a problem with the database checksum on the music database. Recently I found Phone Drive and this has solved my problem. This app lets me copy images and videos out of my camera roll into the app and then the app will bring up a web server and an ftp server on the device. I can then browse the files from my web browser and download the files to my desktop. I can also put files up this way. Phone Drive also has a built in music player that will play any directory as a playlist and supports users adding their own playlists. So I wrote a little script (below) that will generate a playlist for all music and one for each artist on my computer and then I can upload those and all of my music via ftp. I did have a little trouble with some ftp programs that want to open multiple connections. In the end I used ncftp as it doesn't try to open multiple connections and it has the ability to upload directories recursively. I did find one oddity that I needed to turn off the auto conversion of ascii files otherwise playlist files got their line endings changed and Phone Drive wouldn't recognize them.

So my current list of apps that I regularly use is this:

  • Calendar syncing with Google, either via Active Sync or via caldav (this allows the colors to sync)
  • Contacts syncing with Google as an Exchange account
  • GMail app for mail from Google as I prefer the way Google does conversations
  • Appigo Todo syncing with Toodledo
  • PlainText for notes
  • Dropbox for keeping family pictures and moving images between my device and my computer
  • MiniKeePass on my device and KeePassX syncing through Dropbox
  • Podcaster for audio and video podcasts
  • ReadItLater for offline access to web pages
debug() { ! "${log_debug-false}" || log "DEBUG: $*" >&2; }
log() { printf '%s\n' "$*"; }
warn() { log "WARNING: $*" >&2; }
error() { log "ERROR: $*" >&2; }
fatal() { error "$*"; exit 1; }
mydir=$(cd "$(dirname "$0")" && pwd -L) || fatal "Unable to determine script directory"
cd "${mydir}/mp3"
# generate artist playlists
find . -maxdepth 1 -type d -print0 | while read -d $'\0' dir
 if [ "${short_dir}" != "." ]; then
 log "Processing ${short_dir}"
 printf "" > "${playlist}"
 find "${short_dir}" -type f -name '*.mp3' -printf "/music/%p\n" >> "${playlist}"
 mv "${playlist}" "${dir}"
# generate all music playlist
log "Generating all music playlist"
find . -type f -name '*.mp3' \
 -fprintf "${mydir}/mp3/aaa_all_music-new.m3u" "/music/%P\n"


Filed under: iPod Touch Comments Off

Appigo Todo now supports Dropbox

On my iPod Touch I'm using Appigo's Todo application. I had been syncing the data to Toodledo so that I had a backup in case something happened to my itouch. Now Appigo's Todo app supports Dropbox. The advantage here is that the place that I'm syncing to, dropbox, supports all of the features of the Todo app. This is important for recovery. Toodledo doesn't support hierarchical todo items. However when syncing with dropbox it's in Appigo's own format so all features are supported. And it turns out the native format is an sqlite database, so I can pull all of the items out of it in case Appigo goes away!

2/20/2012 update. I've gone back to Toodledo.com for syncing as the Dropbox sync for the todo app was just too slow. The Toodledo sync is faster, but it's not a smart as it keeps trying to sync when I'm offline and then popping up an error dialog. I've contacted Appigo about this, but haven't seen a fix yet.

Filed under: iPod Touch No Comments

Passwords for iTouch and iPhone

Until recently I've been using MyKeePass to store passwords on my iTouch. This allows me to use KeePassX on my desktop. What I'd really like though is Dropbox integration. Well recently on Hak5 they said that there was a free app for IOS that understood KeePass password databases and MyKeePass isn't free. So I went looking and found MiniKeePass and discovered that it's open-source and free. On top of that it integrates nicely with Dropbox! You still need to manually sync the database, but it's really easy if you keep it in Dropbox. When you make changes on your desktop, you just open the Dropbox application on your iTouch and open the database and it imports right into MiniKeePass, then delete the old database from MiniKeePass. When you make changes on your iTouch, you can export it to Dropbox and overwrite the one there so that your desktop sees the changes.

Filed under: iPod Touch No Comments

Backing up cloud data

This year I replaced my Palm Treo with an iPod Touch. To make sure that I could access all of my data on my iPod as well as on my Linux desktop (and anywhere else for that matter) I moved all of my data to the cloud. This is great for accessibility, however I still want backups in case something happens to the cloud provider. Remember that all data needs to be backed up or you run the risk of losing it.

The 3 cloud providers that I am currently using and want backups for are Google Contacts, Google Calendar and Toodledo. My backups are done from my Linux desktop, so I'm using unix tools to do the work. I'm sure people can find replacements under Windows as well.

I wrote a script for each job (based upon various solutions that I found on the net). The data for each cloud service is backed up into a directory named by the date and I choose to keep 30 days worth of backups, so that I can go back and pull information that I may have accidentally changed or deleted.

First up is Google Calendar.

#!/bin/bash -e
date_dir=`date +%Y%m%d`
mkdir -p ${base}
find ${root} -maxdepth 1 -type d -mtime +30 -exec rm -rf \{\} \;
CURL="curl -s -S"
${CURL} -o ${base}/cal.xml "PRIVATE_URL_FROM_GOOGLE"

Here I setup some variables to determine the name of the directory based upon date. I then delete anything older than 30 days. The real key here is the last line. This tells curl to download the calendar and save it locally as "cal.xml". The PRIVATE_URL_FROM_GOOGLE is found by going to calendar settings for the calendar in Google. At the bottom is the private address with 3 links for XML, ICAL, and HTML. I choose to download the XML version, but you can pick any that you want to work with. Since this script has your private calendar URL, you should protect it from others as they can change your calendar with this URL.

Backing up Google Contacts is a little harder because you have to login to get the contacts. I found a script to do this searching Google. I found a script written in python that dumps the contacts in Google's XML format. I don't know the best way to load back from this format, but I'm sure I could figure it out if needed.

#!/usr/bin/env python

import gdata.contacts.service

gd_client = gdata.contacts.service.ContactsService()
gd_client.ClientLogin('user@gmail.com', 'password')

query = gdata.contacts.service.ContactsQuery()
query.max_results = 1000 # change for max contacts returned

feed = gd_client.GetContactsFeed(query.ToUri())
print feed

Then I wrote a wrapper script like the one for Google Calendar that puts the data where I want it.

#!/bin/bash -e
date_dir=`date +%Y%m%d`
mkdir -p ${base}
find ${root} -maxdepth 1 -type d -mtime +30 -exec rm -rf \{\} \;
export base
${HOME}/bin/export-gcontact.py > ${base}/contacts.xml

I'm sure I could have done this in python too, but I wanted to keep all of my scripts as much the same as possible.

The last cloud service I needed a backup of was Toodledo. This one uses curl like the one for Google Calendar. Except that it does the login too. This script has curl post to the login page, save the cookies, then visit the XML export page. Since the cookie is saved, we don't get prompted for a login again. I tried to do this trick on the Google Contacts page as well, but it doesn't work because Google embeds a random hash in the login page as a hidden form field that needs to be passed on for the login to work. Anyway here's my Toodledo backup script:

#!/bin/bash -e
root=${HOME}/backup/pim/tododate_dir=`date +%Y%m%d`base=${root}/${date_dir}mkdir -p ${base}
cd ${base}
curl \  
  --output /dev/null  \ 
  --form 'email=YOUR_EMAIL' \
  --form 'pass=YOUR_PASSWORD' \ 
  --form 'remember=1' \ 
  --cookie-jar ${root}/cookies.txt \ 
  http://www.toodledo.com/signin.php \ 
  --output toodledo.xml \ 

Note that this script has your username and password in it, so you'll want to set the permissions on the script to keep others out.

To backup my email from gmail, I've found offlineimap to be a great application and easy to setup.


Passwords on the iPod Touch

On my Palm I used Keyring to keep track of my passwords. This has a nice interface in jpilot to access it on my desktop. When it came to my iPod Touch I needed to find something else to use. I did some searching and found that there was an open source password program called KeePass that had applications for many platforms (including Windows, Linux, Mac and IOS).  It turns out that there is an old version of the database (1.x) and a new version (2.x). There isn't an application for the new format on Linux yet, so I choose to use the old database format and the desktop application KeePassX. There are two applications for IOS listed on the KeePass website: MyKeePass and iKeepass. I chose  MyKeePass because it supports both the old and new formats and seems to have been updated more recently.

MyKeePass can store multiple databases and those databases can be loaded from a public website, or from it's own web server. You can tell the application you want to import a database file and it starts up a mini web server that just has an upload page and tells you the URL to visit. If your computer is on the same network you can visit this URL and then upload the database and it's on your device. This is pretty nice. It would even better if MyKeePass could read the database file from Dropbox, but that functionality doesn't exist yet. You can have MyKeePass reference a public URL and that could be a public file in Dropbox, but that seems a little to open for me. I'd prefer it to be a private folder on Dropbox that I store my database in.

I found an XSLT to convert the output of the keyring XML exporter to the 2.x format of KeePass, but not to the 1.x format. So I hacked on the 2.x XSLT some and created my version for 1.x. This is a fork of the original that I hope will get pulled into the main.



Offline web pages on iPod Touch

One of the things that I really liked about my Palm was using Plucker to view web pages when I'm not connected to the Internet. I've found a few apps for the iPod Touch that do this and here are the ones I've tried and what I think about them.

I read all of my news using Google Reader and so I wanted an app that would sync with it and keep the articles offline so that I could read them later. I tried Byline and MobileRSS. Both are good RSS readers, however I've found that I prefer MobileRSS. The syncing of all messages and ability to scan through all messages easily is great. Also MobileRSS allows one to send web pages to multiple other services, twitter, email, safari, Instapaper, Read It Later, etc. I'm currently using the free version and it works nicely. I'll probably upgrade to the pro version to get rid of the ads eventually.

For generic web pages I started out using iWebSaver favelet. This favelet takes a given web page and sends it through the iWebSaver favelet site and produces a data page that can be stored on your IOS device. This data page is then opened in Safari when you want to view it. The advantage to this is that you have the whole page as is downloaded onto your device. The problem is that it takes a long time to load and doesn't remember where you are in the page.

Next I tried Instapaper. This has worked very well for most pages. I've been very happy with the scaled down version of web pages. However I found that my Bible study had a problem with rendering. I couldn't see the days. I'm guessing this is some CSS magic that gets stripped out by Instapaper as extra visuals.

I then gave Read It Later a try. At first I thought that it wasn't going to help because the stripped down page looked the same as Instapaper and my Bible study still didn't work. However then I noticed a button at the bottom of the page with a 'T' on it. It turns out that this toggles the page from the stripped down text version to the full page. With a little reading through the options I found that if I disable auto picking of the "right" format to download I can have Read It Later always download both the text and full versions of the page. Then when I'm offline I can toggle between the two. This allows me to view most pages in text and then pages like my Bible study as the full page.


Palm Contacts to Google

I have already been using Google to store some of my contacts for syncing using Zindus. So it seemed pretty natural to use Google to store all of my contacts.  Using Google means that my information is available on any platform. Google provides a nice sync interface for the iPod touch that works well in an offline mode, so I was happy with it.

Originally I started writing an application in python to convert my Palm contacts to Google. As I got close to having this done I discovered that the Birthday and Anniversary fields couldn't get populated via the python API from Google (it may work now, but didn't work then). So I started looking for other options.

It turns out that Google allows you to export your contacts in a CSV format and can import in that CSV format as well. So I entered a few contacts with all of the fields that I wanted and then did an export to see what I got. The CSV file was pretty standard, but had lots of extra columns. So I tried trimming it down and discovered that I actually need to have most of the columns, even if they're blank, otherwise Google wouldn't recognize some of the other non-blank columns. This seems kind of odd, but my guess is that Google is looking for some set of column names to determine if it's a CSV file from Outlook or one from Google.

So now I needed to get my contacts into the Google CSV format. I started by exporting from jpilot and then trying to reorganize the data. This got old real fast. So I wrote another application in C++ using the pilot link library that would output my CSV files. I've put this application up on github so that anyone else can use it as well.

After running this application I had a nice CSV file that I was able to import into Google and all of my contacts were now available to my iTouch. I did end up hacking the IM conversion as I didn't have many contacts with IM information in my Palm. Otherwise the conversion went pretty well.