Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!umich!yale!cs.utexas.edu!usc!snorkelwacker!ira.uka.de!smurf!urlichs
From: urlichs@smurf.ira.uka.de
Newsgroups: comp.sys.mac.programmer
Subject: Re: For GURU'S Only
Message-ID: <1546@smurf.ira.uka.de>
Date: 16 Feb 90 10:17:54 GMT
References: <4073@ccncsu.ColoState.EDU> <34253@ucbvax.BERKELEY.EDU> <6711@ucdavis.ucdavis.edu> <1533@smurf.ira.uka.de> <6747@ucdavis.ucdavis.edu>
Organization: University of Karlsruhe, FRG
Lines: 31

In comp.sys.mac.programmer, article <6747@ucdavis.ucdavis.edu>,
  vmrad@pollux (Bernard Littau) writes:
< In article <1533@smurf.ira.uka.de> urlichs@smurf.ira.uka.de (Matthias Urlichs) writes:
< >< 
< >I don't think you need a disk write buffer if you write to the hard disk.
< >What I did was to call PBRead(serial port, (16 kByte or end of picture))
< >asynchronously, call WaitNextEvent with a reasonable timeout (based on the
< >serial speed, how much you want to read, and how much is already there), and
< >call PBWrite to your disk file if that call completes.
< >Allocate a nice big serial buffer (maybe 48k)...
< 
< As I read your reply, you are doing:
< SerSetBuf( a 48k buffer)
< PBRead(serial port, serial port buffer)
< PBWrite(to disk, serial port buffer)
Replace this with:
PBRead (serial port, private 16kB buffer)
PBWrite (Disk, same buffer)
< 
< This I am sure will work, once.  Maybe more often.  It is however very
< dangerous.  The serial port buffer is a circular buffer.  You could put
< yourself in the situation of clobbering your data as you do a PBRead
< from the serial port buffer to the serial port buffer. [...]
< 
After I say SerSetBuf(), said buffer belongs to the serial driver until
the next SerSetBuf. I do not assume anything about what the serial driver does
with that buffer.
I never understood why anyone would be reading from the serial buffer to the
serial buffer, much less assume that I implied this...
So, I don't see what could possibly go wrong with the above method.
-- 
Matthias Urlichs