Stephan Bergmann wrote:
Frank Schönheit wrote:
[...]
(We should switch to non-default mode then where idlc *does* complain).
(Which we unfortunately cannot do easily for backwards-compatibility
issues...)
Hmm. Can we switch to a slightly-non-default-mode which only finds those
identifiers, and behaves like in default mode for all other issues?
Jürgen, want to pick up on this?
We should use it as default asap. I suggest a cleanup of those
identifiers where the idlc does complain.
It would be incompatible in principal as Stephan mentioned but we can do
that now because we don't have a language binding which depends on
parameter names. Types as structs where a change of member names would
be necessary will be either removed if possible (e.g. deprecated) or
will be handled by a negative list in the future.
If nobody does complain about this strategy i will do that asap.
Juergen
We could just change the (argument?) identifiers then which currently
already start with -. Or is renaming an argument identifier considered
incompatible?
Yes, incompatible.
[...]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
