Related
Run Ubuntu 10.10 on Nexus One
--------------------------------
Background:
In June 2010, XDA zedomax got Ubuntu 9.04 running on Nexus One, tutorial is available at this thread:
http://forum.xda-developers.com/showthread.php?t=718952
I continued studying the tutorial and worked out a way to make Ubuntu 9.04 running better, smoother and cause less error, here's my post:
http://forum.xda-developers.com/showthread.php?t=815532
Recently HTC_Linux team has released version 0.3 of their Ubuntu image, compare to the original guide provided by zedomax, the v0.3 release is based on the latest Ubuntu 10.10. So I thought about making this tutorial to guide you to run Ubuntu 10.10 on your Nexus One.
--------------------------------
What benefits will you get from running Ubuntu 10.10 on Nexus One (at least):
1. Fully functional Linux terminal, with all standard Linux command line utilities
2. Fully functional desktop browser Firefox 3.6.13
3. Ability to compile and run native Linux applications
4. Bunch of graphical Internet and utility applications as shown in the screenshots
--------------------------------
Credit:
Thanks to the great XDA-developers community.
Thanks to Zedomax for providing the original tutorial of running Ubuntu on Nexus One.
Thanks to HTC_Linux team for creating the optimised Ubuntu 10.10 image file.
--------------------------------
Now the tutorial starts.
Preparation:
1. A micro SD card at least 4GB in capacity, 8GB/16GB is recommended
2. (Optional) If you're using Froyo App2SD, move all your applications which are on SD card back to Internal Storage, or you may also uninstall all applications that are on SD card. If you fail to do so, you won't be able to enter Ubuntu. However I've discovered a way that you won't need to move or uninstall apps on SD card, will discuss later.
3. Basic Linux shell knowledge
Then:
1. Get Ubuntu 10.10 from this post (file name: hd2-ubuntu_0.3.zip): http://forum.xda-developers.com/showthread.php?t=889433
2. Download attached scripts.zip
3. On your computer, make a directory on a partition that has more than 3.5GB available space, name the directory “ubuntu” (case-sensitive!)
4. Extract three scripts (b, e, u) from scripts.zip to the directory ubuntu
5. Once you finish download hd2-ubuntu_0.3.zip, extract linux/rootfs.ext2 to the directory ubuntu
6. In your ubuntu directory, rename rootfs.ext2 into ubuntu.img (case-sensitive!)
7. Now verify that you have the following files in your computer's ubuntu directory: b, e, u, ubuntu.img
8. Copy the whole ubuntu directory onto your SD card, put the SD card back to your phone
9. Power up your Nexus One
Now we have some script work:
1. Run a terminal application from your phone (e.g. Android Terminal, Better Terminal Emulator)
2. type “cd /sdcard/ubuntu” (enter)
3. type “sh u” (enter)
4. If you see “Type 'b' to boot Ubuntu”, then you're almost there!
5. Type “b” (enter)
6. If you see “[email protected]:/#” (like shown in screenshot) Congratulations! Now you're running Ubuntu 10.10 on your Nexus One!
Whenever you want to enter Ubuntu again, simply type “b” (enter) in your Android Terminal.
To properly exit from Ubuntu, you must type “exit” from Ubuntu terminal.
If your Android terminal application gets killed by Android system (i.e. you lost Ubuntu terminal), you need to go back to Android terminal, type “e” (enter), that will properly unmount Ubuntu.
Very often that you can get “loop device busy” error when exiting from Ubuntu, in that case, type “e” (enter) multiple times (it's harmless to run “e” script multiple times, don't worry). If still you get “loop device busy” error, don't worry, ignore it then.
--------------------------------
(Optional) If you would prefer working with GUI, now let's do some extra script work:
1. Install “android-vnc-viewer” from Market (freeware)
2. Enter Ubuntu terminal
3. Make sure that you have network connection (Wi-fi recommended)
4. type “apt-get update” (enter), and wait till it finishes
5. type “apt-get install tightvncserver” (enter) and answer “y” to the question
6. After installed tightvncserver, type “cd /bin” (enter)
7. Type the following commands (all are case-sensitive!):
“cat > x” (enter)
“rm -r -f /tmp” (enter)
“mkdir /tmp” (enter)
“chmod 777 /tmp” (enter)
“export USER=root” (enter)
“export XKL_XMODMAP_DISABLE=1” (enter)
“vncserver -geometry 1024x800” (enter)
(enter)
(press and hold trackball and click D on keyboard)
“chmod 777 x” (enter)
8. The above procedure makes a script named “x”, whenever you want to start GUI, you need to type “x” (enter).
Notice that the first time (only the first time) you run “x” script you'll be prompt to enter a password for VNC connection, then you'll be asked whether to create a “view-only” password, since it is not necessary, answer “n” to the question.
9. Start “android-vnc-viewer”, create a new connection with the following settings:
Nickname: (anything you like to identify the connection)
Password: the VNC connection password that you created when ran the “x”script
Address: localhost
Port: 5901
Color format: 24-bit color (4 bpp)
Click “Connect” button, and in about a minute, you'll see your Ubuntu 10.10 desktop!
10. We still need to create one final script called “s”, now go back to your Ubuntu terminal and type:
“cd /bin” (enter)
“cat > s” (enter)
“export USER=root” (enter)
“vncserver -kill :1” (enter)
“vncserver -kill :2” (enter)
“sudo reboot” (enter)
“sudo halt” (enter)
(enter)
(press and hold trackball and click “D” on keyboard)
“chmod 777 s” (enter)
11. If you have GUI running, you * MUST * type “s” (enter) to stop VNC server before you exit from Ubuntu terminal!
In case that your Android kills your Android Terminal application and you still have VNC server running, do the following steps:
a. Start Android Terminal
b. type “b” (enter) to bring back Ubuntu terminal
c. type “s” (enter) to stop VNC server
d. type “exit” (enter) to exit from Ubuntu terminal
Fail to do so you will have problem with VNC connection next time you use it.
--------------------------------
What are not working in Ubuntu:
1. Sound
2. Some applications may not run (I guess, due to Android memory management mechanism)
--------------------------------
A little trick to access SD card from your Ubuntu:
1. Install “SwiFTP” from Market (freeware)
2. Set up FTP server using SwiFTP
3. In Ubuntu, “Places → Connect to Server” then:
Service type: FTP (with login)
Server: localhost
Port: (port number specified in SwiFTP, default is 2121)
Folder: /
User Name: (user name specified in SwiFTP)
Then press “Connect” and enter your password specified in SwiFTP. Now feel free to access your phone's file system!
--------------------------------
How to bring up Ubuntu if you have applications installed on SD card using Froyo App2SD:
1. Make sure that your Android Terminal application is installed in Internal Storage
2. Put a shortcut to Android Terminal application on your home screen
3. Make sure you have got ubuntu directory and all its contents on SD card
4. Reboot your phone (stay sharp, these following steps require perfect timing!)
5. Once you enter lock screen, Immediately unlock it.
6. Enter Android Terminal and type “b” as quick as possible, do not press Enter yet!
7. Pay close attention to your notification bar, as soon as the “SD: Checking for errors” notification disappears (the little SD icon disappears), Immediately press Enter.
8. Now you should be in Ubuntu terminal, wait 1-2 minutes before you do anything else.
9. After 1-2 minutes, you may do whatever you want.
With some applications installed on SD card, you must do the procedure above to bring up Ubuntu terminal each time you reboot your phone.
--------------------------------
That's it! Thank you for reading this tutorial. Please feel free to make any comment or ask any question!
Awesome. I'll be trying this out as soon as I get a larger SD card. It would be great if you could upload a .zip with all of the scripts, downloads, etc. already placed in one folder; that would make it a lot easier.
Thx U for this Tutorial....it was really helpful...Can u fix the adobe flash thing...i know that sound will not work but just wondering if it can be installed.
yumcax said:
Awesome. I'll be trying this out as soon as I get a larger SD card. It would be great if you could upload a .zip with all of the scripts, downloads, etc. already placed in one folder; that would make it a lot easier.
Click to expand...
Click to collapse
Actually I thought about the same but here my uploading speed is 15kb/s maximum, so I'm not sure if I'm gonna upload a 600mb file...
reubenjack said:
Thx U for this Tutorial....it was really helpful...Can u fix the adobe flash thing...i know that sound will not work but just wondering if it can be installed.
Click to expand...
Click to collapse
Adobe Flash has no support for ARM architecture.. sorry, there's no way to install Adobe Flash plugin for Firefox on Ubuntu ARM.
OMG Im abt to die. Transfering the folder to my sd card takes 20 min. I really nid to get a better memory card (higher speed). SIGH..
Works nice on my N1 running MIUI 12.24..
BTW I have a question, if I wipe my phone(data,cache).. Would I need to perform any steps then?
Keshav_CoolDude said:
Works nice on my N1 running MIUI 12.24..
BTW I have a question, if I wipe my phone(data,cache).. Would I need to perform any steps then?
Click to expand...
Click to collapse
If you re-installed a rom (i.e. /system partition gets rewritten)
You'll need to execute the following commands on Android Terminal:
cd /sdcard/ubuntu
sh u
then type "b" to enter Ubuntu terminal
Happy New Year!
Hey,
In the OP you mention that some apps will not run because of the memory management? Is there a size threshold that this is start to be seen at or is it totally random on which things run/don't?
Thanks
houzuoguo said:
If you re-installed a rom (i.e. /system partition gets rewritten)
You'll need to execute the following commands on Android Terminal:
cd /sdcard/ubuntu
sh u
then type "b" to enter Ubuntu terminal
Happy New Year!
Click to expand...
Click to collapse
Okay.. Thanks for the guide..
Happy New Year..
Chopes said:
Hey,
In the OP you mention that some apps will not run because of the memory management? Is there a size threshold that this is start to be seen at or is it totally random on which things run/don't?
Thanks
Click to expand...
Click to collapse
I'm not very familiar with Linux kernel so sorry I can't give an accurate explanation
According to my observation, many applications in the Ubuntu 10.10 can successfully start (program title appears in task bar) but immediately they turn off by themselves.
And it seems to me that all those applications share a same characteristic: they're relatively bigger than others. For example, Terminal runs, Dictionary runs, Firefox runs, but Ubuntu Software Centre won't run.
So my guess is ... Android system would not allocate too much memory for an application in Ubuntu, thus they get killed before they could start.
I'll give this a one thumb up, because I got it working, got to play around a bit... it looks really nice... but I did a reboot of my phone after shutting it down properly, and now everytime I try to launch ubuntu again, I get a bunch of errors.
I've tried just about everything.... restoring to my nand backup I did before trying this, deleting the files off the sdcard and on my system relating to this, tried to set everything back up... but I still get errors and I can't get past them, even with trying "e" in the term.
FWIW, I'm on CM 6.1.1 and I do use DTa2sd with plenty of space left on all partitions. Here's a pastebin of my errors : http://pastebin.com/adCQ62iF
tehgeekguy said:
I'll give this a one thumb up, because I got it working, got to play around a bit... it looks really nice... but I did a reboot of my phone after shutting it down properly, and now everytime I try to launch ubuntu again, I get a bunch of errors.
I've tried just about everything.... restoring to my nand backup I did before trying this, deleting the files off the sdcard and on my system relating to this, tried to set everything back up... but I still get errors and I can't get past them, even with trying "e" in the term.
FWIW, I'm on CM 6.1.1 and I do use DTa2sd with plenty of space left on all partitions. Here's a pastebin of my errors : http://pastebin.com/adCQ62iF
Click to expand...
Click to collapse
Happy New Year!
Could you please go to Settings-> Application Settings-> Manage Applications-> on SD card and confirm there is no application on SD card?
Also please make sure you execute "b" after "Checking for Errors" notification disappears.
Running this distro does that mean i can use firesheep on my n1?
Awesome! What kind of battery life are you seeing using Ubuntu? Any apps in particular you've enjoyed having on your N1?
rossonza said:
Running this distro does that mean i can use firesheep on my n1?
Click to expand...
Click to collapse
According to my study Firesheep would require WinPcap to work, WinPcap is only available for Windows so unfortunately you can't use it.
Deathwish238 said:
Awesome! What kind of battery life are you seeing using Ubuntu? Any apps in particular you've enjoyed having on your N1?
Click to expand...
Click to collapse
I notice battery drains 2-3 mAh more when Ubuntu is idle, when actually using Ubuntu I notice 30-40mAh higher current draw. So actually it's not much
hey thanx houzuoguo,
working great on my n1....!!
is there any solution for the sound to run...??
I cant seem to get any of the "cat >" letters to work, it just hangs in the terminal.
houzuoguo said:
According to my study Firesheep would require WinPcap to work, WinPcap is only available for Windows so unfortunately you can't use it.
I notice battery drains 2-3 mAh more when Ubuntu is idle, when actually using Ubuntu I notice 30-40mAh higher current draw. So actually it's not much
Click to expand...
Click to collapse
Firesheep works in Ubuntu, you just have to compile it yourself.
1) Download the required dependencies:
Code:
sudo apt-get install autoconf libtool libpcap-dev libboost-all-dev libhal-dev xulrunner-1.9.2-dev
2) Install git:
Code:
sudo apt-get install git
3) Download the Firesheep source:
Code:
git clone https://github.com/codebutler/firesheep.git
4) Change to the Firesheep directory:
Code:
cd firesheep
5) Initialize it:
Code:
git submodule update --init
6) Compile it:
Code:
./autogen.sh && make
7) You'll find the file in the build directory.
hotweiss said:
Firesheep works in Ubuntu, you just have to compile it yourself.
1) Download the required dependencies:
Code:
sudo apt-get install autoconf libtool libpcap-dev libboost-all-dev libhal-dev xulrunner-1.9.2-dev
2) Install git:
Code:
sudo apt-get install git
3) Download the Firesheep source:
Code:
git clone https://github.com/codebutler/firesheep.git
4) Change to the Firesheep directory:
Code:
cd firesheep
5) Initialize it:
Code:
git submodule update --init
6) Compile it:
Code:
./autogen.sh && make
7) You'll find the file in the build directory.
Click to expand...
Click to collapse
Not to threadjack or anything of the sort but:
Does the above method work on getting firesheep installed on a n1 running ubuntu or is this just the way to install it assuming your running on normal hardware?
Second question, what is the resulting file?
Thanks,
Which terminal application are you using, and did you type the "cat >" on screen keyboard or copy&paste?
I'm sorry it took so long, I got lots of stuff to deal with. Good things are worth waiting for (i heard)
First off: You are responsible for any of your actions! Changes that are needed to boot Ubuntu on your folio will void your warranty!
Special Thanks goes to DerArtem for posting his Kernels, Weeds2000 for eMail support, topogigi for wifi hacking and last but not least the Ubuntu-Touch Team on IRC that made it possible for me to tweak the touchscreen driver.
However, the system you are about to create is not fully functional, there are still things missing (a working sound driver for example). Its nice to play with it anyway
Heres a quick walk-through:
We will flash a kernel to the toshiba folio tablet which makes it possible to boot from an sdcard or USB-stick (You have to choose one, we need at least 2GB, I'd recommend 4GB so you can install fancy stuff like OpenOffice etc.).
We will then create a root filesystem with rootstock, which will hold our Ubuntu system (similar to your hard drive-Ubuntu).
Finally, we have to tweak that filesystem and change passwords, copy wifi driver etc.
Then we're able to boot a beautiful Ubuntu.
Things you will need:
- A Computer running Linux (Debian / Ubuntu would be great. If you're not running Linux on your computer, running it on the folio would make absolutely no sense. Please don't ask, I don't know how to install it in Windows.)
- The files i added to this post
- A SD-Card, 2-4 GB should be fine.
- optional: A second SD-card, 128MB would be enough (for flashing)
- A USB hub (If you want to boot from USB AND use a keyboard.
- Some time... =)
The first thing you should do is flash one of the update.zip files; choose update-sdmmc.zip (this will boot your folio from sdcard) or update-usb.zip (this will boot from a usb pen drive), open it and unpack the content to a sdcard (This has to be a SDcard, we're gonna flash our device from there.)
Backup all your data on your folio (I'm serious! It'll be gone!) and turn it off.
Put the SDcard in its slot and turn the folio back on. Press + hold both power and volume up.
Follow instructions to do a system update (If you've never done this before you shouldn't start now =) )
When it reboots, your folio will show some funny black and white linux text stuff and hang somewhere. Congratulations! You just bricked your tablet. Now lets see how we can fix it... Turn it off (Keep power pressed for about 5 seconds) and take out the sdcard, put it in your PC.
It's getting messy now, so grab a beer and let's go:
Open a Terminal on your linux computer.
Code:
#echo "Hello world!"
If your console says "Hello World" you got it. Awesome! =)
Code:
#sudo apt-get install rootstock gparted
This will install rootstock, the tool we're gonna need to create our filesystem, as well as gparted, a tool for partitioning our flash drive. You can also install rootstock-gtk, the graphical version. Try it if you like:
Code:
#sudo apt-get install rootstock-gtk
I heard that ubuntu-netbook won't work because there are drivers missing, would somebody try that and report please?
If your beer is already empty when this is done, you need a faster computer. Tell your wife the guys at xda-developers told you so
Back to the terminal:
Code:
#sudo rootstock -f MyCoolHostName -l myCoolUserName -p myCoolPassword --seed xubuntu-desktop -i 2G --notarball
This will create our filesystem with a user called "myCoolUserName" and (theoretically) his password "myCoolPassword" and install a basic xubuntu-desktop. I had to change the password manually because it didn't work this way... We'll do that later.
Rootstock will download lots of packages (You could also set it up with "--seed ubuntu-desktop", then youll download even more packages or "--seed ubuntu-minimal" or kubuntu... I still didnt find a list of seeds online :-/
The creating will take quite some time, get another beer.
When this is done, you'll get a file like this : "qemu-armel-201104112120.img". We will now mount this image: (make sure the directory /mnt/ does exist and is empty:
Code:
#ls -la /mnt )
Code:
#sudo su
(its simpler to be root for now)
Code:
#mount -o loop qemu-armel-201104112120.img /mnt/
(You are aware that your filename is different, are you? Try this:
Code:
# mount -o loop qemu-armel-*TAB*
and your terminal will automatically fill in your filename. Awesome linux, huh? =)
)
Code:
#cd /mnt/etc/
#mv fstab fstab.bak
#echo "proc /proc proc defaults 0 0" > fstab
#echo "dev /dev tmpfs rw 0 0" >> fstab
Note that the first time we create a file called fstab (echo asdf > fstab), then we add one line to this file (echo asdfasdf >> fstab)
Now we need to copy our wifi-driver over. Remember I told you to write down the directory you put them in? =)
Code:
#cp /home/YourUserName/Desktop/FolioStuff/firmware /mnt/lib/firmware
We will now do something stupid: We chroot into our arm-based system from an intel-system. But since we're only changing passwords, we should be OK.
Code:
#cd /
#chroot /mnt/
#passwd myCoolUserName
#passwd
#exit
That's it, we're out of the chroot.
It's now time to prepare the SDcard (The steps are the same for a USB Flash drive, I think you'll figure them out.)
Find out how your SDcard can be accessed (You should still be root):
Code:
# fdisk -l
In my card reader, it's "/dev/mmcblk0", it could also be "/dev/sdb1" in your computer.
Now there are two possibilities:
1. You know what youre doing. Then it's easy: You wipe the only partition (or, if there are more than one, the first partition) on your sdcard and format it with ext3:
Code:
#mkfs.ext3 /dev/mmcblk0p1
Note: The device is called /dev/mmcblk0, the partition itself /dev/mmcblk0p1 (Partition 1)
2. You are not sure about this. No problem, start gparted: (as root!)
Code:
#gparted
In the upper right corner, select your flash drive, delete all partitions and create one formatted with ext3. Close gparted.
Mount your freshly created partition somewhere (i chose /mnt2/):
Code:
#mkdir /mnt2/
Code:
#mount /dev/mmcblk0p1 /mnt2
We can now copy our filesystem over. Make sure you use exactly the same flags (rfp for recursive, force, preserve (attributes))
Code:
#cp -rfp /mnt/* /mnt2/
OK, that could work. Unmount the sdcard:
Code:
#umount /mnt2/
Put it in your folio and cross your fingers.
Wifi drivers will be up soon...
It's recommend to add build-essentials to your rootstock image too..
Because then you can easily build the wireless lan drivers on your devices..
Maybe I'll post a howto this afternoon!
Is this Ubuntu really usable (for daily work) with the folio 100 (touch, response speed, ecc.)?
Thanks.
Yup, it's usable for everyday work. It's quite fast on the Folio 100, I've been pretty impressed when I tested it a few weeks ago.
Thanks ph84
It seems most difficult than expected but, great job !
Only i have 1 question, bricking the tablet sounds dangerous... is possible break out the tablet and get a rock?
Thanksss!!!!!
if ubuntu is installed on sd, does that mean that we can dual boot foliomod from nand and ubuntu from sd/usb?
Hm... I'm using Ubuntu 11.04 and I'm getting this error at the end of the rootfs creation :
E: Sub-process /usr/bin/dpkg returned an error code (1)
E: Second stage build in chroot failed !
E: Please see the log to see what went wrong.
I: Cleaning up...
.....
I: Umounting temporary Image
umount2: Dispositivo o risorsa occupata
umount: /tmp/tmp.zaZ1j1Tgt6/tmpmount/proc: device is busy.
(In some cases useful info about processes that use
the device is found by lsof(8) or fuser(1))
umount2: Dispositivo o risorsa occupata
Click to expand...
Click to collapse
Do you have any idea where I can find the log?
And this procedure will build in the tablet a dual boot? Or dualboot is another procedure?
what is the *.rar????
No it's for singleboot only...
You can only start Ubuntu installed on SD-Card/USB-Stick with this procedure...
I don't know wether the boot-partition is locked when you are booting ubuntu/android..
If not it should be possible to build some applications to flash the boot-partition on the running system to change from ubuntu to android
When I do the rootstock step,I got this error!!!Please help!!!
I: Base system installed successfully.
I: First stage install done
I: Using Chroot for installer
Adding 'local diversion of /usr/sbin/invoke-rc.d to /usr/sbin/invoke-rc.d.rootstock'
Generating locales...
en_GB.UTF-8... done
Generation complete.
Generating locales...
en_US.UTF-8... done
Generation complete.
/bin/installer: line 53: syntax error near unexpected token `+'
E: Second stage build in chroot failed !
E: Please see the log to see what went wrong.
I: Cleaning up...
.....
I: Umounting temporary Image
I: A logfile was saved as //rootstock-201104140706.log
I: Done
Click to expand...
Click to collapse
Could you upload an *.img for us in this forum?
or a databank,etc?
mhmm it always hangs in the xulrunner configuration.. bahh someone know a solution for this? or does i realy have to chroot the rootfst and install everything manualy =(? Cant find any solution on google for this bug =/
sony_tornado said:
what is the *.rar????
Click to expand...
Click to collapse
Nvm, I was talking crap...
schnudergof said:
mhmm it always hangs in the xulrunner configuration.. bahh someone know a solution for this? or does i realy have to chroot the rootfst and install everything manualy =(? Cant find any solution on google for this bug =/
Click to expand...
Click to collapse
Just try rootstock without the parameters on --seed..
Maybe it will work..
You wont have a Desktop Enviroment like Gnome or something but you wont need that for installing the wireless drivers...
And after installing wireless drivers you can install ubuntu-desktop or something with apt-get!
Ok, I have been looking at rootstock, and if you install rootstock-gtk and than click on tasksel there is a whole list of seeds to chose from. I Think im going to try build the netbook build.
On a side note:
We will now do something stupid: We chroot into our arm-based system from an intel-system. But since we're only changing passwords, we should be OK
Click to expand...
Click to collapse
with chroot your only changing the root of your file system, your not running anything from there.
But its its in no way a complaint to your awesome guide! Thanks and keep up the good work!
shidima_101 said:
with chroot your only changing the root of your file system, your not running anything from there.
Click to expand...
Click to collapse
Thats right but when you change your root shell to the rootstock-fs you can also use sudo apt-get update/install and install it into the rootfs!
ph84 said:
Meanwhile, open another terminal. (This is linux! We're gonna multitask since our computer is capable of handling that!) Go to where you stored the attached files, for example
Code:
#cd /home/YourUserName/Desktop/FolioStuff
and unpack them:
Code:
#rar x *.rar
Click to expand...
Click to collapse
Where/what is the "attached files"?
Thanks
Just go to the first page, press STRG+F and type attached into the search window...
I hope you will find the attached files.. Else you dont have to try installing ubuntu on your folio >_<
°EraZoR° said:
Just go to the first page, press STRG+F and type attached into the search window...
I hope you will find the attached files.. Else you dont have to try installing ubuntu on your folio >_<
Click to expand...
Click to collapse
The files I found are only *.zip for update bootloader. Where is *.rar file?
Thanks
Hi XDA developers community!
I think this question was not asked before, because it's too specific.
So here is the problem:
From the Terminal Emulator I am trying to start Poot like so:
Code:
am start -a android.intent.action.MAIN -n org.giantpune.Poot.Poot/android.intent.action.MAIN
and get this error message:
Code:
Error type 3
Error: Activity class {org.giantpune.Poot.Poot/android.intent.action.MAIN} does not exist.
I inspected the AndroidManifest.xml and looked for useful intent commands, but had no luck and even converted the .apk with dex2jar.
As I don't have any experience with Java, the sole purpose of decompiling it in the first place was an interesting intent entry on line 136 in the QtActivity.java file.
It is either located in org/kde/necessitas/origo/ as QtActivity.class, when converted with dex2jar.
Or in Poot-debug(W100)/smali/org/kde/necessitas/origo/ as QtActivity.smali, when decompiled with apktool.
What I am trying to achieve is not only starting the app from the console, but to make it run the action of the "Press here to poot" button.
The device is rooted, should that matter at all.
How can I achieve this and do I even need intent commands or not?
I hope you can help.
Bump
Can anybody help?
As nobody has answered, here is the information I was able to gather meanwhile:
According to a post (can't post link due to restriction) on Android Enthusiasts I need to enter a classname like this:
Code:
am start -a android.intent.action.MAIN -n <package_name>/<full_class_name>
But as there is no class name specified in the AndroidManifest.xml (Link located in first post above), I am not sure how this problem can be solved.
I entered different variations of the command:
Code:
am start -a android.intent.action.MAIN org.giantpune.Poot
Code:
am start org.giantpune.Poot/android.intent.action.MAIN
Code:
am start -a android.intent.action.LAUNCHER org.giantpune.Poot
Code:
am start org.giantpune.Poot/android.intent.action.LAUCHER
None of them worked.
Can someone answer this?
EDIT: The solution is in post #6.
Hi,
I'm trying to make a kernel modification which would create a file on boot before any partition is mounted. I wrote a simple shell script which tries to create a file in the root directory. The problem is I can't write any file to the / directory from a program or a script called by exec from init.rc. If I run dmesg, I can't find any indication of a problem there. I tried to modify the script to write to the /data directory on post-fs, and that worked, but that is not what I want.
I can't find out where the problem is. Do you have any ideas? Thanks.
simplified init.rc snippet:
Code:
on fs
# some work
exec /sbin/busybox sh /test.sh
# some other work
mount_all /fstab.grouper
# and another work
/test.sh
Code:
#!/sbin/busybox sh
echo "test" > /test
/ mounts as ro pretty early...
try to remount it rw before making files, or change the part of init.rc that mounts it ro.
frantisek.nesveda said:
Hi,
I'm trying to make a kernel modification which would create a file on boot before any partition is mounted. I wrote a simple shell script which tries to create a file in the root directory. The problem is I can't write any file to the / directory from a program or a script called by exec from init.rc. If I run dmesg, I can't find any indication of a problem there. I tried to modify the script to write to the /data directory on post-fs, and that worked, but that is not what I want.
I can't find out where the problem is. Do you have any ideas? Thanks.
simplified init.rc snippet:
Code:
on fs
# some work
exec /sbin/busybox sh /test.sh
# some other work
mount_all /fstab.grouper
# and another work
/test.sh
Code:
#!/sbin/busybox sh
echo "test" > /test
Click to expand...
Click to collapse
I am having the same problem, my script is not called. Did you find the solution?
Khaon said:
I am having the same problem, my script is not called. Did you find the solution?
Click to expand...
Click to collapse
Not yet. But recently I got back to the project where I would use it, so maybe I can take a look again and see if I am any wiser now. I'll try to solve it this evening or tomorrow, I'll be reinstalling my tablet so I'll have a good oportunity. I'll post the results here.
frantisek.nesveda said:
Not yet. But recently I got back to the project where I would use it, so maybe I can take a look again and see if I am any wiser now. I'll try to solve it this evening or tomorrow, I'll be reinstalling my tablet so I'll have a good oportunity. I'll post the results here.
Click to expand...
Click to collapse
That would be great, the only thing I could see is upon booting its permissions are read-only. And chmod 0777 doesn't change its permissions(rootfs is rw)
Khaon said:
That would be great, the only thing I could see is upon booting its permissions are read-only. And chmod 0777 doesn't change its permissions(rootfs is rw)
Click to expand...
Click to collapse
OK, I found the cause of the problem and the solution.
The implementation of the exec command in Android init scripts is missing from the init binary by default - the source looks like this:
Code:
int do_exec(int nargs, char **args)
{
return -1;
}
So there are two options - either write the do_exec function yourself, or use a trick to execute your programs.
Writing it yourself would be probably cleaner, but it would mean compiling the init binary again and not everyone wants to do that, so I'll be using the trick.
The trick is to register your program as a service and then invoke it via the start command.
Example init.rc:
Code:
# service invocation - where you need it
# warning - this will not wait for the finish of the program
start my_awesome_program
.
.
.
# service declaration - along with other declarations at the end of the file
service my_awesome_program /sbin/my_awesome_program
disabled
oneshot
If you want to implement the do_exec function and use the exec command, you can use for example CyanogenMod's implementation.
There may be a third option, but I haven't tested that - you can copy the init binary from a kernel which has the exec command working - I suppose all CM kernels should, but I haven't tried.
It's funny how when you already know the solution, you can find so much info, but when you don't, nothing comes up since you are searching wrong.
frantisek.nesveda said:
OK, I found the cause of the problem and the solution.
The implementation of the exec command in Android init scripts is missing from the init binary by default - the source looks like this:
Code:
int do_exec(int nargs, char **args)
{
return -1;
}
So there are two options - either write the do_exec function yourself, or use a trick to execute your programs.
Writing it yourself would be probably cleaner, but it would mean compiling the init binary again and not everyone wants to do that, so I'll be using the trick.
The trick is to register your program as a service and then invoke it via the start command.
Example init.rc:
Code:
# service invocation - where you need it
start my_awesome_program
.
.
.
# service declaration - along with other declarations at the end of the file
service my_awesome_program /sbin/my_awesome_program
disabled
oneshot
If you want to implement the do_exec function and use the exec command, you can use for example CyanogenMod's implementation.
There may be a third option, but I haven't tested that - you can copy the init binary from a kernel which has the exec command working - I suppose all CM kernels should, but I haven't tried.
It's funny how when you already know the solution, you can find so much info, but when you don't, nothing comes up since you are searching wrong.
Click to expand...
Click to collapse
Oh my, many thanks! So stupid exec always return ;D, wouldn't have thought one sec on checking the implementation of those functions.
Khaon said:
Oh my, many thanks! So stupid exec always return ;D, wouldn't have thought one sec on checking the implementation of those functions.
Click to expand...
Click to collapse
A small warning - starting the program as a service will not wait for the program finish, so its results will not be available immediately after the start call. I learnt this the hard way after a few hours of debugging just now.
I added this info to my previous post too.
frantisek.nesveda said:
A small warning - starting the program as a service will not wait for the program finish, so its results will not be available immediately after the start call. I learnt this the hard way after a few hours of debugging just now.
I added this info to my previous post too.
Click to expand...
Click to collapse
I always use
Code:
class core
user root
in addition to oneshot.
Haven't you com accross SElinux restrictions using busybox? Because I do I need to set permissions.
I have this issue
Hey everyone,
I am looking for a way to activate sleep mode of my chromebook manually in order to save battery without having to log out or power off all the time (it's connected to an external display, so closing the lid doesn't trigger it).
So far I found out that calling "powerd_dbus_suspend" from the shell does the trick, but as you can imagine, using the shell and actually typing the command every time is not really a solution.
Is there a way to create a chrome os app that executes this program call, so I can put it in my shelf and just click it to activate sleep?
I'm not really good at .json, maybe there's a way? Or any other suggestions?
Thanks alot!
Markus
I would like to know it too. Anyone?
Sneets said:
Hey everyone,
I am looking for a way to activate sleep mode of my chromebook manually in order to save battery without having to log out or power off all the time (it's connected to an external display, so closing the lid doesn't trigger it).
So far I found out that calling "powerd_dbus_suspend" from the shell does the trick, but as you can imagine, using the shell and actually typing the command every time is not really a solution.
Is there a way to create a chrome os app that executes this program call, so I can put it in my shelf and just click it to activate sleep?
I'm not really good at .json, maybe there's a way? Or any other suggestions?
Thanks alot!
Markus
Click to expand...
Click to collapse
Run a tiny web server like lighttpd in the shell on a weird port, restrict to localhost. Set up index.php to run system (whatever);
Then bookmark the 127.0.0.1: xxxx URL
Which system call invokes power_dbus_suspend?
Thanks Parrotgeek1. Working off your excellent suggestion I installed a lightweight local server on ChromeOS (Web Server for Chrome with the 200 OK! icon) and built a small framework app that is served from local storage. I'm all set up to call chrome.app.??? from my JavaScript. Which API provides the power_dbus_suspend capability? I looked through the Power and System APIs and I don't see anything. - Thanks!
Sneets said:
Hey everyone,
I am looking for a way to activate sleep mode of my chromebook manually in order to save battery without having to log out or power off all the time (it's connected to an external display, so closing the lid doesn't trigger it).
So far I found out that calling "powerd_dbus_suspend" from the shell does the trick, but as you can imagine, using the shell and actually typing the command every time is not really a solution.
Is there a way to create a chrome os app that executes this program call, so I can put it in my shelf and just click it to activate sleep?
I'm not really good at .json, maybe there's a way? Or any other suggestions?
Thanks alot!
Markus
Click to expand...
Click to collapse
Have you tried this simple extension? Keep Awake
https://chrome.google.com/webstore/...lb?utm_source=chrome-app-launcher-info-dialog
What I do is have the shell tab pinned, and press [up arrow] for the command then hit [enter]. Probably not ideal if you use the shell a lot as you have to step through recently used commands.
Re: Pkt_Lnt's post
Thanks Pkt_Lnt. That app only deals with disabling the normal sleep behavior, not causing it, and in particular not causing it when an external monitor is connected. I'm looking for something like Sneets has requested, to invoke a powerd_dbus_suspend command that will force the system to sleep (and turn off an external monitor). Per Parrotgeek1's comment, I've got a little app now sitting on the shelf that is all set to call some system command to invoke powerd_dbus_suspend, only I'm looking for what command to call, from JavaScript.
I managed to find a way to do something similar with the Secure Shell app. During the process of figuring it out, I also found that I was able to create a homescreen shortcut on my Android tablet which when tapped sends a shell command to the Chromebook.
This method might be overkill somewhat, but, anyway, what I did was as follows:
Installed Secure Shell extension; setup sshd on the Chromebook (by making a symlink in /etc/init/ pointing to /usr/share/chromeos-ssh-config/init/openssh-server-conf); rebooted; added public key from ConnectBot (Android app) to ~/.ssh/authorized_keys in order to test connection; setup Android homescreen shortcut to send shell command; created profile in Secure Shell app on Chromebook to connect to [email protected]/my local IP address; generated keys on Chromebook (in ~/.ssh); added generated public key to ~/.ssh/authorized_keys; imported newly-generated private/public key pair within Secure Shell app (one by one); created bookmark link to [email protected] profile in shelf, and, finally, added required command to profile in Secure Shell.
It works well! It looks like any command can be added to the 'Arguments' field in Secure Shell, with the syntax:
Code:
-- thecommand
Although I guess commands that need sudo would require a connection to [email protected], instead of chronos. (-t argument lets you use sudo, but then sending a line break/ENTER seems to be needed)
The window opened with the click does hang around - I haven't figured out if it's possible to avoid this - but it does at least seem to persistently remember the size and position to which it's set.
I posted up some further details at http://nolirium.blogspot.com/shell-command-shelf-shortcut-chromeos.html
Impressive! But a bit more than I'm looking to do
@Nolirum - I read your blogspot doc and your procedure's pretty impressive. Frankly it's a little beyond me and beyond what I'm looking for. If you ever run across a way to invoke powerd_dbus_suspend a little more directly please post here as well. You clearly have a strong grasp of the issues!
Nolirum said:
I managed to find a way to do something similar with the Secure Shell app. During the process of figuring it out, I also found that I was able to create a homescreen shortcut on my Android tablet which when tapped sends a shell command to the Chromebook.
This method might be overkill somewhat, but, anyway, what I did was as follows:
Installed Secure Shell extension; setup sshd on the Chromebook (by making a symlink in /etc/init/ pointing to /usr/share/chromeos-ssh-config/init/openssh-server-conf); rebooted; added public key from ConnectBot (Android app) to ~/.ssh/authorized_keys in order to test connection; setup Android homescreen shortcut to send shell command; created profile in Secure Shell app on Chromebook to connect to [email protected]/my local IP address; generated keys on Chromebook (in ~/.ssh); added generated public key to ~/.ssh/authorized_keys; imported newly-generated private/public key pair within Secure Shell app (one by one); created bookmark link to [email protected] profile in shelf, and, finally, added required command to profile in Secure Shell.
It works well! It looks like any command can be added to the 'Arguments' field in Secure Shell, with the syntax:
Code:
-- thecommand
Although I guess commands that need sudo would require a connection to [email protected], instead of chronos. (-t argument lets you use sudo, but then sending a line break/ENTER seems to be needed)
The window opened with the click does hang around - I haven't figured out if it's possible to avoid this - but it does at least seem to persistently remember the size and position to which it's set.
Click to expand...
Click to collapse
This really is impressive! Thanks for the extremely detailed explanation.
I was able to follow along completely and get the SSH argument "powerd_dbus_suspend" working.
However, I got to thinking. I was wondering if this could be extended to use an alias stored in ~/.bashrc?
After doing some research of my own and using various parameters and arguments, I was unsuccessful.
The goal is to have a linux app (through crouton) be run using xiwi with just a single bookmark by utilizing an alias.
Any thoughts?
UPDATE:
Of course just a few moments later I figured this out! Rather than dealing with an alias stored in ~/.bashrc and ensuring ssh had access, typing the entire alias command as an argument worked. Now I can run Steam directly from a shelf bookmark. All it took was replacing "powerd_dbus_suspend" with "-- sudo enter-chroot xiwi steam" in the argument.
Thanks again!
DandyRandyMarsh said:
This really is impressive! Thanks for the extremely detailed explanation.
I was able to follow along completely and get the SSH argument "powerd_dbus_suspend" working.
Click to expand...
Click to collapse
No problem! I think it's perhaps overly detailed if anything. Might be easier to follow if I cut it down a bit.
DandyRandyMarsh said:
Now I can run Steam directly from a shelf bookmark. All it took was replacing "powerd_dbus_suspend" with "-- sudo enter-chroot xiwi steam" in the argument.
Click to expand...
Click to collapse
Nice!
Adding crouton single app shortcuts is a cool idea which I hadn't considered. Does it work OK for you with sudo in the command, even straight after booting up? ...You didn't have to setup sudoing without a password via sudoers.d or anything?
Inspired by your post I added xiwi to my chroot to try it out. I seem to remember experiencing quite unbearable lag and slowness when running a full desktop via xiwi in the past, but with single apps there doesn't seem to be any noticeable performance hit (this is on an armv7 Asus Flip, with the chroot on a slooow USB drive).
In order to not have to enter my sudo password, in the argument I put:
Code:
-t -- echo mypassword | sudo -S sh /media/removable/3/bin/enter-chroot xiwi firefox
Maybe I'm missing something? Is there a better way, perhaps?
Anyway, since I have rootfs verification switched off, I also tried saving a similar command to a file in /usr/bin, and putting the filename in the argument instead. This works, too, and I suppose could potentially be useful e.g. in the case of needing to run a more convoluted sequence of commands with a shortcut.
Nolirum said:
Adding crouton single app shortcuts is a cool idea which I hadn't considered. Does it work OK for you with sudo in the command, even straight after booting up? ...You didn't have to setup sudoing without a password via sudoers.d or anything?
Click to expand...
Click to collapse
I'm not entirely familiar with exactly how my process worked, but I do not have any sudo password required within shell.
Here's the list of steps I followed allowing me to use sudo commands in the SSH Arguments box.
As a side note, this is the crouton installation I am running on an ASUS c302.
sudo sh ~/Downloads/crouton -r trusty -t unity,touch,xiwi,extension
Setup Bookmarks that Run Shell Commands
1. Installed Secure Shell extension
2. Open crosh with Ctrl+Atl+T
3. Enter Command: shell
4. Enter Command: sudo su -
5. Enter Command: sudo /usr/share/vboot/bin/make_dev_ssd.sh --force --remove_rootfs_verification
6. Enter Command: reboot
7. Let Chrome OS reboot
8. Open crosh with Ctrl+Atl+T
9. Enter Command: shell
10. Enter Command: sudo su -
11. Enter Command: mount –o remount rw /
12. Close current crosh window
13. Open a new crosh with Ctrl+Atl+T
14. Enter Command: shell
15. Enter Command: sudo ln -s /usr/share/chromeos-ssh-config/init/openssh-server.conf /etc/init/openssh-server.conf
16. Enter Command: sudo initctl reload-configuration
17. Enter Command: sudo initctl start openssh-server
18. Enter Command: cd ~/.ssh/
19. Enter Command: ssh-keygen
“Enter file in which…” *LEAVE BLANK*
“Enter passphrase” *LEAVE BLANK*
“Enter same passphrase” * LEAVE BLANK*20. Enter Command: cat /home/chronos/user/.ssh/id_rsa.pub >> /home/chronos/user/.ssh/authorized_keys
21. Enter Command: cp -a /home/chronos/user/.ssh/id_rsa* /home/chronos/user/Downloads
22. Close current crosh terminal and open Secure Shell extension
23. Select [New Connection]
24. Click Import… and select “id_rsa”
25. Click Import… again and select “id_rsa.pub”
26. Enter chronos as the username and localhost as the hostname
27. Enter a shorthand name replacing the text in the top cell “[email protected]”
28. Click Enter. If prompted, enter “yes” to continue connecting
29. Bookmark the page that is open which should end in, "#profile-id:_ _ _ _"
DONE. Now time to add some commands to automate.
Open Secure Shell and select the connection created. In the arguments box, start with two dashes -- and follow it your desired shell command.
Examples:
-- sudo startunity
-- sudo startgnome
-- sudo enter-chroot xiwi steam
-- sudo enter-chroot xiwi wine microWord
Next Steps... Getting Custom Shelf Icons
Now here I am running into some issues. Rather than have all these icons be the same, I was hoping to customize them.
Followed a strategy found on reddit, Change Shelf Icon for Website, but this only works if the site already had an "apple-touch-icon" set. Also, tried to inject the entire code myself which I couldn't get to work. Unfortunately, I think that since this bookmark is created from a Secure Shell, there are some issues.
I have already tried adding in code for this into the site, but it is having no effect. If someone could make this work then we're really looking good here.
DandyRandyMarsh said:
I'm not entirely familiar with exactly how my process worked, but I do not have any sudo password required within shell.
Click to expand...
Click to collapse
Oh, I see what you mean. You are saying that you can sudo in the shell generally, without being prompted for a password, right?
Was that the default setup on your CB (after switching to Dev mode)? Or did you have to configure it manually for passwordless sudoing...
DandyRandyMarsh said:
Here's the list of steps I followed allowing me to use sudo commands in the SSH Arguments box.
Click to expand...
Click to collapse
Nice work summarizing the steps. :highfive: It's probably easier to follow, listed out like that. One slight addendum, perhaps - at step 29, it might be good to clarify the "add to shelf" procedure. Especially because of the following...
DandyRandyMarsh said:
Next Steps... Getting Custom Shelf Icon
Now here I am running into some issues. Rather than have all these icons be the same, I was hoping to customize them.
Click to expand...
Click to collapse
Happily it turns out that, when you add shortcuts to the shelf on Chrome OS like this, a minimal app is created for each one, with its own manifest.json, and icon.pngs. Our shell shortcuts were getting created with blank pngs by default (the black box with P in the center).
So, all we need to do, is get some icons, find the right folder for our shelf shortcut apps, then copy our new icons into them!
For instance, here's how I made a nice shiny icon for my single app shelf shortcut to Firefox.
Prepared icons:
Created a temporary folder in ~/Downloads to store downloaded icons.
Prepared a set of icons in ~/Downloads/firefox. (I downloaded mine from findicons.com).
Six sizes are required - from 32px to 256. To avoid any manifest editing, saved them as *size*.png
e.g. 32.png; 48.png; 64.png; 96.png; 128.png; 256.png
Located folder to copy icons to:
It is easiest to do this directly after clicking "add to shelf"
EITHER:
To change dir into the most recently modified extension icon subfolder automatically, the following one-liner worked for me*.
Code:
shell
cd ~/Extensions/&&cd $(ls -v1td */ | head -1)/*/icons&&ls
A list of blank icons (*.png) should be visible. If so, the new icons can simply be copied over these, e.g..
Code:
sudo cp ~/Downloads/firefox/* .
*YMMV with this particular one-liner - sometimes there's an extra 'temp' folder in ~/Extensions, which stops it working as expected. If you get an error after running the command, simply use the slightly more manual method to locate the folder, as detailed below.
OR:
(slightly more manual method).
Code:
shell
ls -ltr ~/Extensions
Checked the date/times - the most recently added folder is at the bottom. Then,
Code:
cd themostrecentlyaddedfolder
(or, the folder modified at the date/time that the shortcut was "added to shelf"). Then,
Code:
cd */icons
ls
sudo cp ~/Downloads/firefox/* .
I then repeated the above with a new SSH argument and a set of Ubuntu icons, for a custom startxfce4 shortcut, too.
Then just did sudo restart ui, and enjoyed the shiny new shelf icons!
I am hopeful that these changes will get picked up by CrOS's sync, thus getting backed-up and restored automatically. At this point I am not sure if that will be the case, but I don't see why it wouldn't.
Another way?
Another way to approach this, since you're modifying system stuff anyway, might be to edit crosh at /usr/bin/crosh. It's a regular ol' dash/bash shell program, and you can create a shortcut to it on the shelf by opening crosh (ctrl-alt-T), then right click the top right of the browser and select More Tools --> Add to shelf...
Clicking the newly-created link (which looks like an "N") will launch /usr/bin/crosh in a tab. Changing the default behavior of /usr/bin/crosh could accomplish your goal to do whatever you want. With some experimenting, you might even be able to pass arguments to make it do different things depending on how it's opened (via ALT-CTRL-T vs from the shelf-- dunno if it can read the url used to call crosh), but at the very least you could do something like have it run whatever you want it to run if a second modifier key is held down (or if it isn't...), otherwise return to normal operation. A menu w/1 sec timeout could do this as well, ie--
Doing [THE THING YOU WANT] in 2 seconds (press SPACE for crosh)... /
Note there's no proof-of-concept here. Just throwing it out as an idea that I'm pretty sure would work... Someone might even create a neat extension that allows crosh to use scripts/plugins to execute different system operations beyond the stock ones that are built-in via single-clicks... If it worked, it might be a little simplier than via the SSH extension and you wouldn't have to run any extra daemons.
ft
This is probably harder, but a third way, which probably is the "right" way is to see how the crosh extension works and learn about opening the hterm.Terminal instance chrome.terminalprivate object directly, sending it output, and closing it in the same way crosh does it. There's a file in the crosh extension at js/crosh.js that shows the basics of how it's done:
var terminal = new hterm.Terminal(profileName);
And then you can do stuff with it. You could then create an extension that could handle multiple shelf-icons (say by appending ?COMMAND=ls to the URL to do different shell commands(. The security implications for this would probably be huge obviously.
Update: More on embedding hterm is here.
Nolirum said:
I managed to find a way to do something similar with the Secure Shell app. During the process of figuring it out, I also found that I was able to create a homescreen shortcut on my Android tablet which when tapped sends a shell command to the Chromebook.
I posted up some further details at http://nolirium.blogspot.com/shell-command-shelf-shortcut-chromeos.html
Click to expand...
Click to collapse
Nice post - Some good food for thought, changing up the contents of crosh etc directly is an interesting idea.
fattire said:
Another way to approach this, since you're modifying system stuff anyway, might be to edit crosh at /usr/bin/crosh. It's a regular ol' dash/bash shell program, and you can create a shortcut to it on the shelf by opening crosh (ctrl-alt-T), then right click the top right of the browser and select More Tools --> Add to shelf...
Click to expand...
Click to collapse
Interesting! ....Aaaaand, now I've broken crosh. lol. Lucky I had my SSH shortcut setup... (actually, we can still get into the VTs with CtrlAltF2 etc anyway).
Some good potential for experimentation with this, definitely.
fattire said:
This is probably harder, but a third way, which probably is the "right" way is to see how the crosh extension works and learn about opening the hterm.Terminal instance chrome.terminalprivate object directly, sending it output, and closing it in the same way crosh does it. There's a file in the crosh extension at js/crosh.js that shows the basics of how it's done:
Click to expand...
Click to collapse
Again, very interesting. Now you come to mention it, an approach such as this does seem like it might be do-able without too much messing around (maybe)...
Any possible method we can get e.g. an extension to 'talk' to the shell is worth looking into I think; as you say, if we can cut out the middleman (ssh) and maybe do it more elegantly, that might be pretty useful indeed.