More info

[marcelloh]

When I run it, I see some info at the end , like:

CAPABILITY_ARBITRARY_EXECUTION: 3 references
CAPABILITY_FILES: 3 references
CAPABILITY_MODIFY_SYSTEM_STATE: 3 references
CAPABILITY_NETWORK: 3 references
CAPABILITY_OPERATING_SYSTEM: 2 references
CAPABILITY_READ_SYSTEM_STATE: 3 references
CAPABILITY_REFLECT: 3 references
CAPABILITY_SYSTEM_CALLS: 3 references
CAPABILITY_UNANALYZED: 3 references
CAPABILITY_UNSAFE_POINTER: 3 references

It would be nice if there was a flag (like -v or something) that would tell me more in detail about what the files and the positions of there references are.
Example
Like/my/dir/to/file.go:14:28
(this makes it easy to go to the position with just 1 click)

[jcd2]

Thanks for trying Capslock! Output formats are definitely something we're still experimenting with to learn what's most useful. We have the command-line flag -output to change how the information is presented. -output=v is similar to what you suggest, although we should probably add the full path for source files instead of just the final filename.

[gedw99]

Outputs format are all good.
but I was hoping to see discover of where is food some aspects

For example:
CAPABILITY_NETWORK: 3 references could pump back a link to the path to the file with line numbers for example.

[lyda]

Yes, I just ran this on a module and see CAPABILITY_ARBITRARY_EXECUTION: 2 references. I'm not clear which modules can do this. I'm also not clear if I'm using the parts that can do this. If it helps, I was running it against this.

https://gitlab.com/lyda/external-dns-openwrt

[jcd2]

Using -output=json gives you example call paths showing why functions have particular capabilities.

There's also -output=graph which outputs the full set of paths from functions to capabilities in a format suitable for use with graphviz, but this can often be very large.

[lyda]

Perfect. The json output meets my needs.