Adapt and document KLEVER for usage in CI
Most of our customers consider deployment of KLEVER as a part of existing CI process.
To be ready for such deployment we have to provide a documented way to:
1. run analysis from command line starting from building the target system.
2. wait for results and get statistics for the run.
3. generate stable URLs to result details.
Other details can be extracted during a pilot integration for one of our target systems.
Updated by Evgeny Novikov 2 months ago
- Assignee set to Evgeny Novikov
- Category set to Documentation
I think that it does depend on particular project and particular CI.
There is an example how to build the project and collect data necessary for Klever (https://klever.readthedocs.io/en/latest/tutorial.html#preparing-build-bases). But this example covers just the Linux kernel (one version, one architecture, one configuration).
There is a command-line API to launch Klever, to wait for verification results and to download them in some format. It is used by our own CI. It is not documented anywhere.
#10530 was devoted to stable URLs, so, now URLs are already pretty well and stable, i.e. just a server name/location changes after transferring verification results from one machine to another.