Re: [PATCH] stop run-time option processing after "--"

[Peter Bex]

On Tue, Mar 14, 2023 at 02:48:28PM +0100, Mario Domenech Goulart wrote:
 What should the behavior of executables be in case they are compiled
> with the runtime options parser disabled (e.g.,
> -disable-runtime-options, as per Felix' previous patch)?

Assuming we go with the decision to process until we hit either "-:" or
the first non-runtime option, I think programs with the parser disabled
could still consume the first argument if it is identical to "-:".

I prefer that over doing completely nothing, as that makes it easier to
disable runtime parsing in any existing program at any point - all the
existing scripts for that program which already use "-:" to disable
runtime option parsing would continue to work without modification.

Thanks for asking, this is an important point to discuss!

NOTE: Perhaps a simpler alternative to that patch would be to treat
any *non-empty* runtime option as an error (like we do for unknown
options now) and error out with a message saying runtime option parsing
has been disabled.  This would give a better hint to the user that
they could decide to recompile the program with runtime parsing
enabled.

Cheers,
Peter

signature.asc
Description: PGP signature