You already have parsers that know how to parse REXX comments, so the 
complexity is not that tough to deal with. However, it does require that the 
person updating or creating the file be more careful and, perhaps, use QDI if 
building by hand.

Because of the possibility of having special characters in volsers, it would be 
appropriate to require that comments, however delimited, be isolated 
statements. The complexity is in determining whether what follows the presumed 
delimiter is data or comment. That would be avoided by not allowing comments on 
the command statements. It would also be easier to add to the current code - a 
single check to determine whether it is a comment.

Regards, 
Richard Schuh 

 

> -----Original Message-----
> From: The IBM z/VM Operating System 
> [mailto:[email protected]] On Behalf Of Alan Altmark
> Sent: Thursday, March 11, 2010 1:58 PM
> To: [email protected]
> Subject: Re: Automated DDR funny - has anyone got any ideas
> 
> On Thursday, 03/11/2010 at 03:19 EST, "Daniel P. Martin" 
> <[email protected]> wrote:
> > Well,  yes, the semicolon does work...  but if you're going 
> to have to 
> > do new work to hack the way DDR parses statements anyhow, REXX-ish 
> > syntax would certainly be a desirable approach.  Not that I 
> have any 
> > delusions about the development process being at all 
> democratic, but 
> > if you're considering votes, I vote for:
> > 
> > - "/* blah blah blah */" syntax for comments
> > - ";" (semicolon) for statement continuation
> 
> Ooooh.  Humpty Dumpty strikes again!  You have cleverly 
> changed the definition of "desirable" to mean "the most 
> complicated option possible". 
> REXX comments.  Nesting.  <gack>
> 
> Semicolons, otoh, are simple to describe and simple to implement.
> 
> Alan Altmark
> z/VM Development
> IBM Endicott
> 

Reply via email to