roku-dev-cli is a command-line tool for loading builds onto a dev-enabled Roku device.

Features:

• build zipping, installation, and logging
• project analysis
• network proxying

roku-dev-cli can be installed from the pypi repository using pip.

pip3 install roku-dev-cli

roku-cli --help

To load a build on the Roku, first make sure the device is on the same network and is in developer mode (how to enter developer mode).

In your terminal, go to the base directory of your Roku project where the manifest file exists. Then run the following command:

roku <IP_ADDRESS>

This will zip up the current directory and send it to the device. If the app is successfully installed the tool will output the application's log to your console as it executes. Hitting Ctrl-C will safely exit the tool without stopping the application.

The script will do some basic checking of the project making sure all the component files are valid XML.

If you already have a zip file you want to load, you can use the zip_file flag instead of zipping the current directory:

roku --zip-file <ZIP_FILE_PATH> <IP_ADDRESS>

This will install the zip file and start logging the application output.

There are two reporting tools available.

One tool provides a graph of nodes and texture memory over time, which is written to a file and viewable in a browser.

roku -n <IP_ADDRESS>

The other parses any logrendezvous output from the telnet port and writes it to a file. This will create three lists:

• rendezvous by frequency (highest to lowest)
• rendezvous by longest duration (highest to lowest)
• rendezvous neighboring other rendezvous lines, which may show an opportunity to use queueFields to mitigate multiple rendezvous

roku --report-rendezvous <IP_ADDRESS>

Before running this tool, make sure the device is printing out rendezvous logs. You can do this by telnetting into port 8080 on the device and running the following command:

> logrendezvous on

The tool has some support for mitmproxy, which allows you to view and control network traffic running through the box.

The Roku OS currently does not support HTTP proxy forwarding, so users must first route traffic through a computer.

On mac, run the following command:

brew install mitmproxy

sudo sysctl -w net.ipv4.ip_forward=1
sudo iptables -t nat -A PREROUTING -i wlan1 -p tcp --dport 443 -j REDIRECT --to-port 8080
sudo iptables -t nat -A PREROUTING -i wlan1 -p tcp --dport 80 -j REDIRECT --to-port 8080

Replace wlan1 with the wifi interface the Roku device is connected to.

Enable port forwarding:

sudo sysctl -w net.inet.ip.forwarding=1

Update pf config. It should look something like this...

...
scrub-anchor "com.apple/*"
nat-anchor "com.apple/*"
rdr-anchor "com.apple/*"
# Port forwarding for mitmproxy
rdr pass on bridge100 inet proto tcp from any to any port 80 -> 127.0.0.1 port 8080
rdr pass on bridge100 inet proto tcp from any to any port 443 -> 127.0.0.1 port 8080
dummynet-anchor "com.apple/*"
anchor "com.apple/*"
...

Notice that the forwarding rules were inserted between some other existing rules, if you append them to the end of the file you will get some warnings.

Finally run the following commands. The first will check file syntax while the second one will restart the service loading the new rules.

sudo pfctl -v -n -f /etc/pf.conf
sudo pfctl -ef /etc/pf.conf

Finally start the tool with the --proxy flag. This will add mitmproxy's certificate to your build and replace certificate instances to use it instead.

roku --proxy <IP_ADDRESS>

At this point the application should be running through your proxy.

TODO: document use of mitmdump scripts

To switch to a local source version:

pip install -e /path/to/your/local/repos/roku-dev-cli

To switch to the published version:

pip uninstall roku-dev-cli && pip install roku-dev-cli

Where relevant, please add unit tests and make sure they are passing before submitting a pull request.

pytest

1. If necessary, create a personal account at pypi.org and ask to be added as a maintainer of https://pypi.org/project/roku-dev-cli/.

2. If necessary, edit ~/.pypirc to include your pypi.org credentials. For example:

   [distutils]
   index-servers =
     pypi
   
   [pypi]
   username: <your-pypi-username>

3. Merge your pull request, including updating version.py.

4. If necessary, install twine and wheel:

   pip3 install twine wheel

5. Publish to the default public pypi repo, you will be prompted to enter your pypi.org password:

   python3 setup.py sdist
   python3 setup.py bdist_wheel --universal
   twine upload dist/*