[feature] did you guys knew that "?" is a valid character for function/command name?

[Ben Rubinstein]

On 07/09/2010 09:51, Ben Rubinstein wrote:
> This is very bad news for my long-held campaign to introduce '@' as an
> optional (but one day might be checked) prefix to a parameter, to indicate
> in the calling line that this is passed by reference.

He was subsequently forced to retract partially, see below.

On 07/09/2010 14:56, Andre Garzia wrote:
> Err, Ben, this is actually in place since forever... we can pass by
> references values with @....

and on 07/09/2010 16:34, Alex Tweedly wrote:
> Ben wants to be able to use the '@' (presumably as an option, with no actual
> effect) in the calling script, to help remind him that the variable will/may
> be changed within the handler.

Andre - yeah, what Alex said.   Both caller and callee need to be fully aware 
when a parameter is by reference - it's a contract, and needs informed consent 
on both sides.  Currently we have only presumed consent on the part of the caller.

On 07/09/2010 17:07, Mark Wieder wrote:
> I went to go vote for this bug report/enhancement request, but I can't
> exactly find it.

At which point Ben has to apologise for his hyperbole, and admit that 
"long-held campaign" means "long-held view, but never done anything about it 
since posting to the now defunct Xtalks mailing list in May 2004".  There was 
a lively (for Xtalks!) discussion, but nothing came of it (except the 
interesting observation from Doug Simons, who maintains SenseTalk, that he has 
the reverse; it's the caller that indicates a variable is passed by reference, 
he allows but ignores @ in the function declaration, only to maintain 
compatibility with Transcript).  At that time I had an idealistic view that 
extensions to the language were better proposed and discussed in Xtalks first, 
before specifically proposing that Rev adopt them.

I've now, better late than never, opened a report at RQCC:
	http://quality.runrev.com/qacenter/show_bug.cgi?id=8945

Ben