Re: New Object Method Invocation Syntax Brokenness [message #76204 is a reply to message #76081] |
Thu, 19 May 2011 09:20   |
JDS
Messages: 94 Registered: March 2009
|
Member |
|
|
Thanks, Chris. This also means that continuing to use "->" will offer the safest way forward, since the dual meaning of a.b(c) is implicit and never reported. I think a new user, i.e. those who are most likely to adopt "." for method invocation, is likely to *mean* a->b(c) here, since they quite likely know nothing of the old () form of indexing, and will therefore also rather reasonably presume the difference is between a.b(c) (method call) and a.b[c] (variable subscript). Without IDL2 in force, and in the likely situation of both a method and class variable named 'b', these are identical.
It would be useful to have a means of enforcing IDL2 across all routines in a file, or even by class (e.g. in class__define), so that you have to include it once and only once.
JD
|
|
|