Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!netnews.upenn.edu!grad2.cis.upenn.edu!ranjit From: ranjit@grad2.cis.upenn.edu (Ranjit Bhatnagar) Newsgroups: comp.sys.amiga Subject: Is ARP addbuffers broken? Message-ID: <11127@netnews.upenn.edu> Date: 15 May 89 06:41:42 GMT Sender: news@netnews.upenn.edu Reply-To: ranjit@grad2.cis.upenn.edu.UUCP (Ranjit Bhatnagar) Distribution: na Organization: University of Pennsylvania Lines: 42 I've noticed a strange phenomenon: when I ADDBUFFERS on df0: or df1:, the buffers appear, but are never re-used unless I eject the disk: that is, if I allocate 40 buffers, as soon as I've read 40 sectors off the disk, I don't get any buffering any more except from those 40 sectors. Sample session: command ------- ---------------- date date addbuffers df0: 40 date date ls c: date echo echo even though the echo command clearly should have been buffered, it wasn't, apparently because the date and ls commands used up all the buffers. NOTE: I am not sure that my explanation is correct, but the behavior in the above sample session has been observed consistently. Similarly, when I use a vt100 script that contains a loop, the disk runs continuously during the loop (if the loop crosses a sector boundary) unless I add fresh new buffers just before starting the script. This seems to happen under arp 1.2 and 1.3, on a 1 meg, 2 drive 2000 and a 2.5 meg, 1 drive 1000. I do not remember it happening when I used the commodore commands. Am I hallucinating? Does this happen to anyone else? Why should ARP have anything to do with it? (Maybe it doesn't.) Should I give up and buy FACC? (I probably should anyway, now that I have 2.5 megs to burn.) - - ranjit "Trespassers w" ranjit@eniac.seas.upenn.edu mailrus!eecae!netnews!eniac!... Near the sides of tall buildings, how dare they kiss goodbye? Those buildings that saw the airplanes that kiss the air in their fantasy.