Right. Exactly. This is not the only requirement. The problem space is not 
"demonstrate how one might ..."

I have other build issues or requirements or gotchas or whatever (habits?) that 
also come into consideration. I have multiple processes in place. And my hope 
had been that this was not a "rearrange my world" project but a simple "compile 
most of the product ARCH(9). Compile two modules ARCH(5) and use them (only) to 
catch a customer environment that is downlevel from the build target."

For your specific "omit the problem file from the JCL" suggestion there are no 
files in the JCL. //SYSIN specifies a UNIX folder and the compiler grabs all of 
the .C or .c files.

Yes, I could put them in a separate folder and have two compile steps. That 
starts to get off into the "rearrange my world" area. I also have a TSO command 
that compiles a single file from the "current" (and that is a whole requirement 
or gotcha or whatever of its own ...) code set folder. That would have to get 
improved somehow.

I guess my point is not that the problem as stated is impossible to solve but 
rather "there's got to be a better way."

Charles

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf 
Of Tony Harminc
Sent: Thursday, November 26, 2015 10:35 AM
To: [email protected]
Subject: Re: Any clever way to defeat the C compiler's options precedence?

On 26 November 2015 at 13:28, Jack J. Woehr <[email protected]> wrote:
> Okay, you're right. Can you just omit the problem file from the JCL, 
> compile it separately, put it in the right dataset of output objects, 
> and then run the JCL and link it with the others?

For some value of "just"...

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to