On Wed, 28 Oct 2009 at 03:32:16 -0400, Ryan Lortie wrote:
> GVariant type strings can be made "indefinite" by the addition of
> wildcard characters. For example:
>
> "(i*)" - a pair where the first element is a 32bit integer and the
> second element has any valid type
With my D-Bus hacker hat on: in each case where there is a character
that, if it was to appear in D-Bus signatures in future, would break you,
please propose a patch to the D-Bus Specification that adds them to the list
of signature characters.
"* - Reserved for use as a wildcard, as seen in GVariant" would be enough,
IMO. The D-Bus Specification doesn't necessarily have to specify a meaning
fro the characters you're reserving, but it should at least mark them out
as reserved so others don't use them for something incompatible.
I understand you have some similarly magic semantics for "@"?
(When the second user of, say, "*" appears, *then* it's time to require that
the D-Bus Specification documents what it means, for interoperability.)
Simon
Attachment:
signature.asc
Description: Digital signature