| Re: New Object Method Invocation Syntax Brokenness [message #76129 is a reply to message #76123] |
Mon, 16 May 2011 18:09   |
penteado
Messages: 866
Registered: February 2018
|
Senior Member
Administrator |
|
|
On May 16, 7:35 pm, JDS <jdtsmith.nos...@yahoo.com> wrote:
> My understanding is compile_opt idl2 alleviates this particular undesirable feature of the newly ambiguated syntax, though I'm not eager to go back and insert it for thousands of methods, or develop methods to discover potential brokenness. I suppose talk of enabling idl2 by default died off; I'd be in favor of it. It would break old code, but in a well-controlled and presumably well-advertised way.
>
> By the way, this behavior seems to have been introduced in 8.0.1. I believe the short lived IDL 8.0 respected the arrow operator.
I just dug out an old idl 8.0 and can confirm it already had that
behavior. In a more obvious manifestation of the problem:
IDL> a={b:0,c:1.0}
IDL> help,a->c(0)
<Expression> FLOAT = 1.00000
Since there was never any discussion about using -> for structures,
and this only seems to be undesirable, as JD mentioned, to me it looks
like a bug. The main argument against making idl2 assumed by default
was to keep compatibility, so this seems unintended. A side effect of
some other change to the parser, which went unnoticed.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
