jayy001 / codexctl Goto Github PK

Just tried to use this using the latest windows release, and everytime I ran the install command, it told me that it could not find the firmware and downloaded it again, even though it already did.

Cannot roll back from beta version 3.8.2.1961 Error Updating

You can use the following to only run code if it's on windows:

import os

if os.name == 'nt':

You can then use powershell to check the status of a network, and mark is as private:

Get-NetConnectionProfile
Set-NetConnectionProfile -InterfaceIndex 13 -NetworkCategory Private

Python can run powershell in a subprocess, which is probably easiest. I've also written code in the past to use Python.NET to run powershell code within a .NET context that I could give you some pointers from, but that adds more dependencies to your code that you probably don't need.

It would be a good idea to verify that the file contents match what is expected after downloading it, and before installing it. A simple hash of the file should be enough. You can then store that in the file version/id table. Maybe store it something like this:

self.id_lookups_rm1 = {
    "3.6.1.1894": ("CGaIOXfzAA", "HASH"),
}

You can then get the id and hash from it with the following:

id, hash = self.id_lookups_rm1[version]

Since there is a separate EULA for the beta, that you can't confirm that the user has agreed to, you shouldn't include beta versions in the list.

It would be nice to add an extract command that can pull the raw image from a downloaded upgrade file.

Are there instructions for putting this on the RM2? I have a bit of a chicken-and-egg problem: no python/git on the RM2. Can you provide instructions for basic installations on vanilla RM2s?

This currently downloads update files to the current working directory. Instead, they should be put in a temporary directory.

https://docs.python.org/3/library/tempfile.html#tempfile.TemporaryDirectory

You'll need to disable the update-engine service on the other partition, which is possible by removing the symlink systemd creates.

This way you avoid the issue where someone downgrades, forgets to turn off auto update and then gets updated again.

I was asked by Eeems to submit an Issue request. I have never added anything to github, so my apologies. Running the main.py script was opening cmd prompt in Windows 10, displaying about 4 lines and then closing. Checking the code in IDLE I got this when running the code.
Getting latest toltec supported version Traceback (most recent call last): File "C:\Users\<USER>\Desktop\rm-update-helper-1.0.0\main.py", line 156, in <module> driver = updates_manager() #TODO: Delete file after its done File "C:\Users\<USER>\Desktop\rm-update-helper-1.0.0\main.py", line 42, in __init__ self.latest_toltec_version = self.get_latest_toltec_supported() File "C:\Users\<USER>\Desktop\rm-update-helper-1.0.0\main.py", line 62, in get_latest_toltec_supported return re.search('Toltec only supports OS builds up to and including OS(.*)\. You will soft-brick', site_body_html).group(1).strip() AttributeError: 'NoneType' object has no attribute 'group'
At this point the code would cease. I checked line 62 in the code and found the following:
return re.search('Toltec only supports OS builds up to and including OS(.*)\. You will soft-brick', site_body_html).group(1).strip()
As far as I could tell the .group(1).strip() did not appear to be used anywhere else (I know very little about Python, so I may have missed something). I removed that part of the code and the module ran correctly when running it again.

The current data structure for latest doesn't work properly when there is a stable release for one device, but not the other.

Ideally, you would just compare the listed version numbers and grab the highest one for latest. That way, there would be no need to keep the value updated.

The toltec value should allow splitting rm1/rm2 values out though, if there is ever another release only for one device.

Currently this tool doesn't allow you to use a SSH key to connect to the device, only the SSH password.

Hello, thanks for the awesome software. It saved me several months ago when I wanted to go back to 2.x version :)
I downloaded the binary for windows today, and after unzipping it, Windows Defender gave me a notification that a threat was found. It's really weird because in Windows Defender main page, it would change to "no threat" in a matter of seconds without me doing anything. Every time the folder is unzipped, the notification pops again, only for the threat to disappear again.

• The message I get is:
  "Threats found. Start the recommended actions.
  Trojan:Win32/Wacatc.B!ml
  1/8/2024 2:31PM (Active) Severe
  Start actions"
• Windows 11

+after unzipping several times, it stopped giving me threat notification
+now the threat doesn't go away automatically

During the status checking, REQUIREMENTS_MET is not checked.

Instead of requiring the user to pass the --rm1 flag, it would be preferable to auto-detect the device where possible.

I'm having an issue where the updater hangs when it tries to start the fake updater.

My remarkable 2 is currently running v3.5.1.1798 and I'm trying to roll it back to v2.15.1.1189.

It would be nice to have a command to bind mount an update file to allow exploring the contents. This would likely depend on #31

When getting it to run I had to install both 'requests' and 'netifaces'.

Thanks for the effort!
I have problems with the macOS binary (latest codexctl release, macOS 12.7 M1 mac)

• binary needs chmod u+x to be executed; ok easy
• still wont run even with sudo; it's killed by the OS with a security warning; maybe a code signing issue

Currently, we have to manually add the update file IDs for each version, we should have a service to periodically check for new updates and then commit them to the repo

https://github.com/al45tair/netifaces has been archived and is no longer maintained. It also breaks on 3.11. This should be replaced with a new maintained alternative.

First of all thanks a lot for the great work and collaboration I'm witnessing !

I just downgraded from 3.9.x to 3.8.x and I have a few questions :

• I noticed the "Check for updates" button doesn't do anything now. Is this to be expected? (no logs are generated)
• If I want to update the device normally in the future is it possible or am I "stuck" with using this tool?
• If I "revert" using the switch shell script, am I back to normal or is there still information I should take into account?

Once again thanks a lot!

It would be interesting to have commands to let you quickly explore the contents of an update image. The couple I could think of would be cat and ls. This would depend on #31 and #32 most likely, but in this scenario, codexctl would automate extracting and mounting to a temporary location, and then cat/ls the output and unmount. The extracted image would be cleaned up at some point too, but probably not right away, unless that performs well.

Performance on a rM device would not be a good idea for mounting to tmpfs though due to the limited ram, so maybe only extract/mount to a specific location on disk under /opt on device?

You can use https://github.com/Eeems-Org/remarkable-update-fuse to mount an update file to access the image inside without having to extract it. From there you can either do a loopback mount, or use ext4fuse to mount the image. This will avoid spending the time required to extract the image.