Leaving your home invented terminology aside, First time a label is used, it marks the stage that it refers to. Subsequent references are to additional streams for the stage that declared the label. Since the program to run has already been specified, subsequent references to the label must be bare.
On 09/28/2016 08:58 PM, Stanislawski, Shawn (National VM Capability) wrote:
Would someone please explain to me the internal mechanism(s) behind why the following causes an error? PIPE LITERAL PICKLES CAKE | SPLIT | A: DROP 1 | SPEC /CHEESE/ 1 W1 N | A: FANINANY | CONSOLE " Label A is already declared ... Scan at position 70; previous data "CHEESE/ 1 W1 N | A:" Ready (-0047); T=0.01/0.01 11:44:14 " With regard to streams, the DROP stage has a Primary Input, Primary Output, and one optional Secondary Output. Therefore, a connected Secondary stream could only be an Output. With regard to streams, the FANINANY stage has a Primary Input, Primary Output, and any number of optional Secondary Inputs. Therefore, a connected Secondary stream could only be an Input. I'm aware I can make this work by utilizing multistreaming, such as below, but I am curious why unambiguous connections are not tolerated. PIPE (end ?) LITERAL PICKLES CAKE | SPLIT | A: DROP 1 | SPEC /CHEESE/ 1 W1 N | B: FANINANY | CONSOLE ?A: | B: To be clear: I am not attempting to advocate for unambiguous connections to be tolerated, I am merely curious about the current mechanism(s) in play. Best regards, Shawn S.
