Re: Pointers? Based variables? in CL?

[Scott Klement <midrange-l@xxxxxxxxxxxxxxxx>]

hi Vern,

> Tom, I don't understand your concern about V5R4. What in the article bothers 
> you if it "comes true"? I mean, the stuff is here now.

I think he's referring to the ability to call APIs that return pointers, 
such as string manipulation APIs, record read/write APIs, IFS directory 
reading APIs, etc.

That stuff is currently only available through HLLs, which require the 
customer to buy a WDS license from IBM.  Tom's concern is that now systems 
where they deliberately didn't buy the HLL compilers (possibly to prevent 
people from doing development on the machine because of the potential 
security/stability risks of allowing development) will now have access to 
most or all of the HLL's capabilities through ILE CL.


> One thing NOT mentioned that is even more important, IMO, is the ability to 
> call something using parameters BY VALUE. IOW, we can now call things like 
> the access() function and have it work right.

The *BYVAL stuff that we got in V5R3 is also a big deal. When you combine 
the two (pointers and *BYVAL) this gives you access to pretty much every 
API on the system, and unlocks pretty much all of the functionality on the 
system to a CL program.

And, I think that's Tom's point :)