Remote debugging: Embedded Linux Target: Why do we need program?

[paulo-a-junior]

I am using cpptools to perform remote debugging on Linux embedded targets.

I have had little success using the extension to accomplish this in the past. However, in my lastest trials, I wrote a launch request, which worked incredibly well, except for not getting a console nor the application stdout on the debug console, even though I had the logging options apparently appropriately set. I even tried with the externalConsole option set to true, also with no success.

I then moved on to attaching to a running process, which after a few hours of trial and error, I got working.

Now, to the question: why do we need a mandatory program field specified when attaching to a running process? I don't see why/how the file is used - maybe at most to define the target architecture?

For reference, here is my launch.json entry:

        {
            "name": "Attach to Process (CPPDBG)",
            "type": "cppdbg",
            "request": "attach",
            "program": "/path/to/a/program",
            "processId": "${command:pickRemoteProcess}",
            "useExtendedRemote": true,
            "miDebuggerServerAddress": "192.168.10.11:1234",
            "miDebuggerPath": "gdb-multiarch",
            "miDebuggerArgs": "",
            "setupCommands": [
                {"text": "set sysroot /path/to/sysroot"},
                {"text": "set print pretty on"},
                {"text": "set substitute-path /usr/path/to/comoponent ${workspaceFolder}"},
                {"text": "set substitute-path /usr /path/to/debugfs/usr"},
            ],
        }

Thanks for any explanation on this =)

PS: Before submitting, I did the required search, and found this as a similar discussion, with no answers.