Step 1 - CAST Profiler (optional)
Overview
CAST Profiler is an optional tool you can run before onboarding an application into CAST Imaging. It scans source code and build artifacts on your own machine and returns an inventory: the languages present and their volume, the frameworks and packages in use, the projects and their external dependencies, a high-level architecture preview, and a set of alerts about the source code delivery itself.
Running it is not a requirement. You can go straight to Register and deliver source code and onboard an application without it. It earns its place when you do not yet know whether an application is a good candidate, when the delivery has come from a third party and may be incomplete, or when you want the technology breakdown before committing an analysis cycle to it.
This step answers three questions before you register anything:
- Is the delivery complete? Missing schema files, absent JCL and duplicated trees are all reported as alerts, with the action that resolves each one.
- Will CAST Imaging understand this application? Every language and technology found carries its support level, and the results report the percentage of the application that is not supported.
- How big is the job? The composition figures separate the whole delivery from the part CAST Imaging will actually analyze, which are usually different numbers.
The CAST Profiler website is where you download the tool, upload a result file, view a results page and read the release-notes feed. It also carries the product FAQ and a set of sample results. These pages do not repeat that. What they cover is the reference material the website does not carry: every CLI option and what it does to the output, how each figure on the results page is derived, what every alert means and how to resolve it, and how to run the tool unattended.
How this differs from the fast scan
CAST Imaging runs a fast scan of its own when you register and deliver source code. The two have similar aims but sit at different points, and they are not alternatives:
| CAST Profiler | Fast scan | |
|---|---|---|
| When it runs | Before an application exists, and before CAST Imaging is needed at all | After the application is registered and its source code delivered |
| Where it runs | Your own machine, a build agent or a container | Inside CAST Imaging |
| What it is for | Deciding whether and how to onboard | Verifying the delivered code before a full analysis |
| Required | No | Yes, part of the onboarding flow |
Run CAST Profiler when the decision to onboard has not been made yet. The fast scan happens regardless once you register the application, so a Profiler scan does not replace it.
Assess
- How CAST Profiler fits into an onboarding - what a scan produces, the two run modes, what leaves your machine, and the prerequisites
- Composition and sizing - why the file and line counts are reported twice, and how source lines are calculated
- Technology tree and support levels - how technologies are detected, the role taxonomy, and what each support level means for a deep analysis
- Architecture preview and projects - how the layer diagram is built, and the limits of the preview
- Alerts reference - every alert CAST Profiler raises, why it is raised, and how to resolve it
The support levels reported against each technology come from the full CAST Imaging technology coverage.
Reference
- CLI reference - every option, the files each one produces, and the folders excluded from a scan automatically
- Result files - what each generated file contains and how to gate a pipeline on alerts
Automate
- Run CAST Profiler with Docker - the container image, mounting source, and the file ownership constraint
- Scan a GitHub repository - the GitHub App, and what it commits to your repository
Next steps
- Register and deliver source code - create the application in CAST Imaging and deliver the code you have assessed