The Saleae Analyzer SDK is used to create Low Level Analyzers (LLA) for the Saleae Logic software via a plugin architecture. These plugins are used to decode protocol data from captured waveforms. In many cases you can use a [High Level Analyzer Extension](https://support.saleae.com/extensions/high-level-analyzer-quickstart) to process data from an existing protocol decoder instead of building a LLA.
Once downloaded, first run the script rename_analyzer.py. This script is used to rename the sample analyzer automatically. Specifically, it changes the class names in the source code, it changes the text name that will be displayed once the custom analyzer has been loaded into the Saleae Logic software, and it updates the visual studio project.
There are two names you need to provide to rename_analyzer. The first is the class name. For instance, if you are developing a SPI analyzer, the class names would be SPIAnalyzer, SPIAnalyzerResults, SPIAnalyzerSettings, etc.
The file names would be similar, like SPIAnalyzer.cpp, etc.
All analyzer classes should end with "Analyzer," so the rename script will add that for you. In the first prompt after starting the script, enter "SPI". The analyzer suffix will be added for you. This needs to be a valid C++ class name - no spaces, it can't start with a number, etc.
Second, the script will prompt you for the display name. This will appear in the software in the list of analyzers after the plugin has loaded. This string can have spaces, since it will always be treated as a string, and not as the name of a class.
After that, the script will complete the renaming process and exit.
This example repository includes support for GitHub actions, which is a continuous integration service from GitHub. The file located at `.github\workflows\build.yml` contains the configuration.
When building in CI, the release version of the analyzer is built for Windows, Linux, and MacOS. The built analyzer files are available for every CI build. Additionally, GitHub releases are automatically created for any tagged commits, making it easy to share pre-built binaries with others once your analyzer is complete.
This section only applies to downloaded pre-built protocol analyzer binaries on MacOS. If you build the protocol analyzer locally, or acquire it in a different way, this section does not apply.
Any time you download a binary from the internet on a Mac, wether it be an application or a shared library, MacOS will flag that binary for "quarantine". MacOS then requires any quarantined binary to be signed and notarized through the MacOS developer program before it will allow that binary to be executed.
Because of this, when you download a pre-compiled protocol analyzer plugin from the internet and try to load it in the Saleae software, you will most likely see an error message like this:
> "libSimpleSerialAnalyzer.so" cannot be opened because th developer cannot be verified.
Signing and notarizing of open source software can be rare, because it requires an active paid subscription to the MacOS developer program, and the signing and notarization process frequently changes and becomes more restrictive, requiring frequent updates to the build process.
The quickest solution to this is to simply remove the quarantine flag added by MacOS using a simple command line tool.
Note - the purpose of code signing and notarization is to help end users be sure that the binary they downloaded did indeed come from the original publisher and hasn't been modified. Saleae does not create, control, or review 3rd party analyzer plugins available on the internet, and thus you must trust the original author and the website where you are downloading the plugin. (This applies to all software you've ever downloaded, essentially.)
To remove the quarantine flag on MacOS, you can simply open the terminal and navigate to the directory containing the downloaded shared library.
This will show what flags are present on the binary:
The included Github Actions integration, documented [here](#cloud-building---publishing), will automatically build your analyzer for both x86_64 and arm64 architectures, for MacOS.
When you build your custom analyzer on a Mac, by default it will compile for the architecture of the system.
Unfortunately, universal binaries are not currently supported.
You can optionally cross build your analyzer using an x86_64 host system targeting arm64, or from a arm64 host system targeting x86_64.
To cross build, you will need to create a new build directory (for example `build/x86_64` or `build/arm64`). Then use the CMake variable [CMAKE_OSX_ARCHITECTURES](https://cmake.org/cmake/help/latest/variable/CMAKE_OSX_ARCHITECTURES.html).
Although the exact debugging process varies slightly from platform to platform, part of the process is the same for all platforms.
First, build your analyzer. Then, in the Logic 2 software, load your custom analyzer, and then restart the software. Instructions can be found here: https://support.saleae.com/faq/technical-faq/setting-up-developer-directory
Once restarted, the software should show your custom analyzer in the list of available analyzers.
Next, in order to attach your debugger, you will need to find the process ID of the Logic 2 software. To make this easy, we display the process ID of the correct process in the About dialog in the software, which you can open from the main menu. It's the last item in the "Build Info" box, labeled "PID". Note that this is not the correct PID when using an ARM based M1 Mac. (Please contact support for details on debugging on M1 Macs.)
Note, we strongly recommend only debugging your analyzer on existing captures, and not while making new recordings. The act of pausing the application with the debugger while recording data will cause the recording to fail once the application is resumed. To make development smooth, we recommend saving the capture you wish to debug with before starting the debugging process, so you can easily re-load it later.
when `cmake .. -A x64` was run, a Visual Studio solution file was created automatically in the build directory. To debug your analyzer, first open that solution in visual studio.
Then, open the Debug menu, and select "attach to process...".
In the "Attach to:" selection box, ensure that "Native code" is selected. This is usually selected by default. If not, you will need to select it manually by clicking the "Select..." button to the right. After clicking "Select...", a new window will appear like shown below. Select "Native code" from there.
However, before you can attach a debugger to the Logic 2 process on MacOS, you will need to add an additional [entitlement](https://developer.apple.com/documentation/bundleresources/entitlements) to the Logic 2 app packages.
This is because in order to distribute applications for MacOS, these applications must be [signed and notarized](https://developer.apple.com/documentation/security/notarizing_macos_software_before_distribution). One requirement for notarization is that debugging support is disabled.
> error: attach failed: attach failed (Not allowed to attach to process. Look in the console messages (Console.app), near the debugserver entries, when the attach failed. The subsystem that denied the attach permission will likely have logged an informative message about why it was denied.)
Checking the output in Console.app, you will likely find logs like this:
> macOSTaskPolicy: (com.apple.debugserver) may not get the task control port of (Logic2 Helper (R) (pid: 95234): (Logic2 Helper (R) is hardened, (Logic2 Helper (R) doesn't have get-task-allow, (com.apple.debugserver) is a declared debugger(com.apple.debugserver) is not a declared read-only debugger
To fix this, you will need to add the [get-task-allow](https://developer.apple.com/documentation/bundleresources/entitlements/com_apple_security_cs_debugger) entitlement to all of the Logic 2 app packages.
You can use this 3rd party convenience script with the instructions below to add the entitlements. This process needs to be completed once per computer, and will need to be repeated after updating Logic 2.
Now you're all set! To debug with command line lldb, simply launch the Logic 2 software and check the PID as explained above. Then run this from the terminal:
```bash
lldb
attach <pid>
```
Please see the Linux instructions below for more gdb command examples, which _mostly_ translate to lldb 1:1.
Once complete, you should also be able to attach other debugger GUIs like xcode or CLion to Logic 2 using the same PID.
(Note, this section needs to be tested and updated if needed)
On Linux, you can debug your custom analyzer using GDB. This can be done from the console, however we recommend using a GUI tool like Visual Studio Code, with the C++ extension installed.
To debug from the command line, once you have loaded your analyzer into the logic software and have checked the process ID, you can attach gdb like so:
finally, attaching to the process will have paused the Logic application execution. Resume it with the continue command:
```bash
continue
```
If your analyzer hasn't been loaded yet, GDB will notify you that it can't find this function, and ask if you want to automatically set this breakpoint if a library with a matching function is loaded in the future. Type `y <enter>`
2. Copy the contents of this sample repository into the existing analyzer, except for the src and docs directories, or the rename_analyzer.py script. The `.clang-format` file is optional, it would allow you to auto-format your code to our style using [clang-format](https://clang.llvm.org/docs/ClangFormat.html).
3. Rename the existing source directory to src. This is optional, but it might make future updates from this sample analyzer easier to roll out. Make sure the CMakeLists.txt file reflects your source path.
- In the line `project(SimpleSerialAnalyzer)`, replace `SimpleSerialAnalyzer` with the name of the existing analyzer, for example `project(I2CAnalyzer)`
- In the section `set(SOURCES`, replace all of the existing source code file names with the file names of the existing source code.
