[feature] add verbose mode

[c7h]

I had the problem that for a batch task, I only needed the symbol printed to stdout. Therefore, I made the meta information optional.

[marin-m]

Hello,

Thanks for the PR, having this optional is probably good.

• Having an amount of debugging visible by default is likely useful to understand better the output for the tools, and to debug it if needed, for most users: is it ok if I turn your -v, --verbose flag into a -s, --silent flag?
• I see that you have removed the architecture guess duration counter from the code, and moved the print call out of the original function, likely to be able to more easily reference the self.verbose attribute by keeping it as an attribute of one class. I think I should maybe reintroduce the guess time counter.
• Just out of curiosity, I would like to know how having a --verbose/--silent flag meet your need compared to having the debugging output simply going through stderr?

Thanks again for the time taken for this pull request,

Regards,

[c7h]

Hi @marin-m
Thanks for the quick response.

• Yes, changing the flag to --silent is a good idea to avoid a breaking change. I will update my PR.
• I can also add the duration counter back in. It just wasn't useful for me but I agree that It might be for others.
• We could also just write the debug output to stderr. I would make this optional to avoid a breaking change here. How about in case of --silent, we print the debug output to stderr instead of stdout? This way, we could keep the stdout clean, we would be backwards compatible but still show the debug messages to the user.

Looking forward to hear your opinion,
Cheers,
Chris

[marin-m]

Hello,

We could also just write the debug output to stderr. I would make this optional to avoid a breaking change here. How about in case of --silent, we print the debug output to stderr instead of stdout?

It was just to know. I think that it is fine to keep everything going to stdout as it is currently, as:

• The --silent flag will then actually silence the debugging output, which is a more intuitive behavior than redirecting streams.
• Providing one way to separate both outputs (the flag) is still correct, and fulfill the needs in a simple manner.
• Backwards compatibility is not critical I think as the tool is still recent, but if we can keep it behaving like this it is still good. Especially, both executables output the same debugging output, and the debugging output is also an information output which is the primary output that the user may want to log when running the vmlinux-to-elf executable.

Regards,