Wireshark-dev: Re: [Wireshark-dev] Future of extcap "API"

From: Jirka Novak <j.novak@xxxxxxxxxxxx>
Date: Tue, 23 Aug 2022 18:30:41 +0200
Hi,

Adding another helper may be helpful, as it would probably gives us greater control, and maybe also solve the "helper-script" issue in the future by putting that stuff inside Wireshark? I am just wondering if it is worth the effort. We can obviously strive for a perfect - no user interaction required - solution, but do we really need to be perfect here?

In my experience, as long as we can solve the real issue - the zombie processes - and have minimal interaction by the creators of the original extcaps we should be fine. Now as I understand it, we can achieve that at some level with one of the proposed solutions above, just not in an ideal way, right?

I wrote my notes and thoughts in previous email.

I am fine with having developers adapt their script, as long as there is some form of compatibility mode, and maybe some warning displayed before starting a non-converted extcap

API with pipe is backward compatible. Therefore non-converted extcaps can run as before. I just think that warning about non-converted extcap is not required. a extcap was working before, without graceful shutdown, therefore it is not required for the extcap. When author learns that graceful shutdown is available and is useful for the extcap, they can adapt it.
But it is optional from my point of view.

						Best regards,

							Jirka