On Thu, Jan 13, 2011 at 6:37 PM, Glenn Knickerbocker <n...@bestweb.net> wrote: > On 01/13/2011 11:35 AM, SPITZ, HOBART CTR DFAS wrote: >> The fix worked like a charm. > > Until this pops up again in another 12 years . . . > > This "never-expiring" timer is just being used to feed a record to GATE > to shut everything down when PIPMOD STOP is issued. The most > straightforward replacement I see is an IMMCMD stage with a command that > can never actually be entered at the console (so it can't usurp any > other IMMCMDs or console input): > > /* Impose end of stream on loop connections at PIPESERV end */ > pip01 = , > '\ immcmd' '1d1d1d1d'x , /* Wait forever for untypable command */ > '| append literal', /* Dummy record on ending */ > '| g: gate' /* Cut input streams */
I must be more dumb than usual today, but what's the role of a gate that is never triggered? It would never shut the pipeline so you could just take it out? > Naive question: What would it take for DELAY to take advantage of > Extended-TOD-Clock to push this problem out a few millennia? I noticed that you can even have the primary input of gate connected to "hole" so it will never fire. I was thinking the eof on gate's primary would short its secondary input to output, but apparently not... but functionally no difference. | Rob
