nRF Connect Install requires a huge amount of packages

[brianreinhold]

I have developed a set of applications using just SoftDevice for Bluetooth and flash read and writes in the framework of nFR51 and nRF52 SDKs. Now I was thinking of moving to nRF Connect. I already have Visual Studio Code installed for a project that is handled remotely from my Windows 64 10 PC  on a Linux Raspberry PI. So I installed the extensions for Nordic. However, I could not find the locations on my computer where they were installed. VS Code itself is installed on a second disk E.

However I was far from finished. It looks like I need to install a great number of third party programs I have never heard of including the nRF Command Line Tools. I do not intend to use the command line tools and have never used this tool to date. Why do I need this? Is it absolutely necessary to install all these third party packages like 'chocolately',  'west', python??? , Zephyr I can understand... This is probably gigabytes of junk. I cannot image all of this stuff ever getting cleaned up. Do I really need all this? Why Python? That's a whole SDK for building Python applications ... that is huge.

How many of these can I get rid of once nRF Connect is installed. What does 'chocolatey and west do? I realize that I can install even more stuff and get the manager to do all of this for me but that is just hiding all the junk getting downloaded and installed.

I went to the Nordic 'Seminar' in Boston (Waltham) about a month ago where I was introduced to nRF Connect.

So the actual goal here was to see if this project would have a smaller footprint using SoftDevice (used now) or the nRF Connect SDK and which was easier to implement. There is interest in this project (which is a generic health device model over a BTLE tunnel) primarily in China at the moment so It would be nice to have an answer. One thing the seminar did not reveal was the massive amount of support software needed to get off the ground.

[ovrebekk]

Hi Brian

However, I could not find the locations on my computer where they were installed. VS Code itself is installed on a second disk E.

VSCode extensions should be installed in the %USERPROFILE%\.vscode\extensions folder. Typically something like c:\Users\USERNAME\.vscode\extensions. I confirmed this on my own machine, and can see 5 folders with the nordic-semiconductor prefix for the various Nordic extensions.

I do not intend to use the command line tools and have never used this tool to date. Why do I need this?

That is odd to hear you have never use this before, the nRF Command Line tools contains the base drivers to interface with the debug chip on the various nRF52 DK's, and is used for flashing and debugging the DK's. I would expect you to have used this when working with the nRF5 SDK and the SoftDevices. 

These tools are listed in the Tools and downloads section of the nRF52 series DK's, based on the 'old' platform (nRF5 SDK and SoftDevice). 

Is it absolutely necessary to install all these third party packages like 'chocolately',  'west', python???

Chocolatey should no longer be necessary. This is a packet manager that was used to install the various toolchain components in the first releases of the nRF Connect SDK, but has now been replaced by the Toolchain Manager (TCM) application in nRF Connect for Desktop. 

When using the TCM it is basically a one click operation to install a specific version of the SDK, and this includes the various toolchain components such as West and Python. 

You define the install directory for any new SDK versions. Then the SDK's will be installed in a subfolder named v2.1.0 for instance, while the toolchain will be installed in the /toolchain/v2.1.0 folder. 

At the moment each SDK install comes with its own version of the toolchain which does contribute to the bloat definitely, but greatly reduces the chance of build errors caused by incorrect toolchain components. 

Do I really need all this? Why Python? That's a whole SDK for building Python applications ... that is huge.

Python is used by a lot of the build scripts in the toolchain and as such is critical, with West being the most important of these Python scripts. West is the meta tool that handles the download and update of the various repositories included under the nRF Connect SDK umbrella, and can also be used to start build, flash and other operations from the command line. 

I doubt the whole Python SDK should be downloaded though. The entire toolchain comes in at about 1GB, and I think the Python binaries are a relatively small portion of this. 

How many of these can I get rid of once nRF Connect is installed. What does 'chocolatey and west do? I realize that I can install even more stuff and get the manager to do all of this for me but that is just hiding all the junk getting downloaded and installed.

I doubt there is much you can get rid of since most of these tools are used for building applications. Possibly you could get rid of chocolatey and leave the scripts in place, but I am not sure. 

Frankly you will be much better of using the Toolchain Manager. As I mentioned earlier all the resulting junk can be found in the [SDK_FOLDER]/toolchain/v2.x.x folder, so using the TCM should not make it any harder to find compared to chocolatey

You can uninstall the TCM and nRF Connect for Desktop once everything is installed, these are not used for building or developing applications. That said there are a lot of useful apps within nRF Connect for Desktop, such as the Bluetooth application and the Power Profiler (for owners of a Power Profiler Kit), so I think this is an application worth keeping. 

If you really want to keep your system as un-affected as possible the best option is probably to install everything in a virtual machine. If you want to go down this route though it is recommended to run a virtual Linux installation, since the build performance is generally faster in Linux than on Windows. 

I went to the Nordic 'Seminar' in Boston (Waltham) about a month ago where I was introduced to nRF Connect.

I remember, I was there!

I think we had a chat at the end, regarding the work you have been doing with the SoftDevice

I didn't represent nRF Connect though, so any promises made about it I will put on Karl's shoulders

So the actual goal here was to see if this project would have a smaller footprint using SoftDevice (used now) or the nRF Connect SDK and which was easier to implement. There is interest in this project (which is a generic health device model over a BTLE tunnel) primarily in China at the moment so It would be nice to have an answer. One thing the seminar did not reveal was the massive amount of support software needed to get off the ground.

Are you talking about memory footprint on the device, or the size of the toolchain?

What is more easy to implement is a complicated question to answer. The toolchain and build system is definitely more complicated, but making applications is typically not because of the advantages the build system offers. As an example the Bluetooth heart rate example in Zephyr is about 130 lines of code i main.c, compared to a much more complex example in the nRF5 SDK. And exactly the same main file can run fine on the nRF51, nRF52 or nRF53 devices by only changing the board file in the build, which is something the nRF5 SDK could only dream off...

Your development model is a bit different than most customers though, as you don't really use any of the libraries in the SDK and just program for the SoftDevice directly. I am not sure this will work as well in the new SDK, since we don't really have a SoftDevice anymore, but have split the BLE controller and host into two separate pieces of software. 

One thing the seminar did not reveal was the massive amount of support software needed to get off the ground.

To be fair, I think we mentioned most of these tools in the seminar, but it is true that Python isn't mentioned in the slides. This I can take as feedback to future iterations of the slides. Chocolatey is not mentioned either, but as I said we recommend using the TCM instead, rather than having to install all the tools manually. 

Finally I hope you have taken the time to go through the DevAcademy? 
It really makes a big difference in getting started with the nRF Connect SDK

Best regards
Torbjørn

[Nguyen Hoan Hoang]

And exactly the same main file can run fine on the nRF51, nRF52 or nRF53 devices by only changing the board file in the build, which is something the nRF5 SDK could only dream off...

Wrong, it exists 5 years ago without any scripting software only gcc compiler. It has been posted on this forum since.  

https://embeddedsoftdev.blogspot.com/2017/12/bluetooth-le-with-nordic-nrf51-nrf52.html

https://embeddedsoftdev.blogspot.com/2018/02/bluetooth-le-with-nordic-nrf51-nrf52.html

With only a few line of code, no RTOS, no 10K line of defines.

Original source code here

https://github.com/I-SYST/EHAL

new version moved here

https://github.com/IOsonata/IOsonata

[brianreinhold]

The non-blocking restriction does make life more difficult but most non-OS based embedded systems run in that manner (at least that is what I am led to believe). In my project I have a situation where the characteristic value can be longer than the MTU. So it has to be sent fragmented. So when notifying, one would send 6 notifications but then one needs to wait for the ack event before one can send more. So I would exit the send routine and basically wait on sd_app_evt_wait() and handle the events myself when it woke.

I have never really fully understood that call, but I assumed it could only be called in one place in your program and your entire program ran around this call. When it would exit, you would check everything you might need to do and return to the sd_app_evt_wait().  Anything else, like a sleep, is just a busy loop and a horrible way to wait for something to happen before continuing,

So that's what I have done. When I used the pc-driver-btle library on the PC, I could use mutexes and semahores which DID dramatically simplify the program - but at the cost of having a multi-tasking OS which diminishes real time implementations.

[Nguyen Hoan Hoang]

ovrebekk said:

Just to be clear, Nordic doesn't force you to do anything

It has been said time over by Nordic people that if you want to use the nRF53 or nRF91 you have use Zephyr!!!  SDK not available otherwise.  DUH !!!

[ovrebekk]

Nguyen Hoan Hoang said:

It has been said time over by Nordic people that if you want to use the nRF53 or nRF91 you have use Zephyr!!!  SDK not available otherwise.  DUH !!!

Then I am glad I could clear up that misunderstanding

[Brian]

ovrebekk said:

Case in point, if I download the latest and greatest version of the nRF Connect SDK I can still build many of the Bluetooth samples for the nRF51, a device released 12 years ago!

Hi Torbjørn, I am a teacher and I have a bunch of nRF51822s that I want to use to set up a lab for my students to use. If I can get help using the latest nRF Connect SDK to build apps on the older chips, I would certainly make sure my students get an extra positive review of Nordic Semi. 

[ovrebekk]

Hi Brian

Please open a new ticket if you have a question not directly related to an older case, the devzone is not really designed do handle long running cases with multiple different questions. 

The short answer is that the nRF51 series is not officially supported in the new SDK, but many samples will still work by using the nrf51dk_nrf51422 board definition, including many of the Bluetooth samples. 

In other words it should be possible to put together a lab session based on the nRF51 boards and the new SDK, if you find a set of features and libraries in the SDK that work fine on the older boards. 

Best regards
Torbjørn