Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!spool.mu.edu!uunet!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.c
Subject: Re: FORTRAN type I/O in C - Help!
Message-ID: <6535@goanna.cs.rmit.oz.au>
Date: 27 Jun 91 07:27:12 GMT
Article-I.D.: goanna.6535
References: <31839@hydra.gatech.EDU>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 34

In article <31839@hydra.gatech.EDU>, griffin@prism.gatech.EDU (GRIFFIN,JEFFREY A) writes:
> I am looking for suggestions for reading a data file of which two 
> records are listed below.  Each record consists of 512 characters 
> plus a new-line.  There is an initial byte count of 0133 then 32 
> integer values, another byte count of 0133 and 32 integers, the
> last byte count of 0133 and the last 32 integers.  The remainder of
> the record is padded with the '^' character.

What you have here is ANSI format for a tape with variable-length
records stored in 512-byte blocks.  The new-line is probably someone
trying to be helpful.  I suggest attacking this in two stages.

Stage 1:  break the file into records.

    int read_Record(FILE *f, char *buffer)	/* returns EOF */
	{					/* or strlen(buffer) */
	    int c;				/* EOF also means "error" */

	    /*  We should be at the beginning of a record. */
	    /*  Padding ^ characters and extra new-lines may intervene. */
	    do { do c = getc(f); while (c == '^'); } while (c == '\n');
	    /*  By now, only EOF and <four decimal digits> are ok.  */
	    if (c == EOF) return EOF;		/* real EOF; others=error */
	    ungetc(c, f);
	    if (fscanf(f, "%4d", &c) != 1 || c <= 0) return EOF;
	    if (fread(buffer, 1, c, f) != c) return EOF;
	    buffer[c] = '\0';
	    return c;
	}

Stage 2: parse the records using sscanf(), or, perhaps easier in this case,
strtol().
-- 
I agree with Jim Giles about many of the deficiencies of present UNIX.