GNUmakefile

@@ -57,6 +57,9 @@ test: build
 	tests/cc-compat.sh
 	tests/posix-compat.sh
 
+dj: src/dj.c build_dir
+	$(CC) $(CFLAGS) -o build/bin/dj src/dj.c
+
 false: src/false.rs build_dir
 	$(RUSTC) $(RUSTCFLAGS) -o build/bin/false src/false.rs
 

docs/dj.1

@@ -0,0 +1,155 @@
+.TH dj 1
+
+.SH NAME
+
+dj \(en disk jockey
+
+.SH SYNOPSIS
+
+dj
+.RB ( -AdHnq )
+.RB ( -a
+.RB [ byte ])
+.RB ( -c
+.RB [ count ])
+
+.RB ( -i
+.R [
+.B input file
+.R ])
+.RB ( -b
+.R [
+.B input block size
+.R ])
+.RB ( -s
+.R [
+.B input offset
+.R ])
+
+.RB ( -o
+.R [
+.B output file
+.R ])
+.RB ( -B
+.R [
+.B output block size
+.R ])
+.RB ( -S
+.R [
+.B output offset
+.R ])
+
+.SH USAGE
+
+The
+.B -i
+option takes a path as an argument to open and use in place of standard input.
+The
+.B -o
+option does the same in place of standard output. Dj does not truncate output
+files and instead writes over the bytes in the existing file.
+.PP
+The
+.B -b
+option takes a numeric argument as the size in bytes of the input buffer and
+the
+.B -B
+option does the same for the output buffer, the default for both being 1024
+bytes, or one kibibyte (KiB).
+.PP
+The
+.B -s
+option takes a numeric argument as the number of bytes to skip into the input
+before starting to read, and the
+.B -S
+option skips a number of bytes through the output before starting to write from
+the input. If the input is a stream the bytes are read and discarded. If the
+output is a stream, nul characters are printed.
+.PP
+The
+.B -a
+option takes one argument of one byte in length and pads the input buffer with
+that byte in the event that a read doesn't fill the input buffer, and the
+.B -A
+option takes no arguments and pads with nuls.
+The
+.B -c
+option specifies an amount of reads to make, and if 0 (the default) dj will
+continue reading until a partial or empty read.
+.PP
+On a partial or empty read, dj prints a diagnostic message (unless the
+.B -q
+option is specified) and exits (unless the
+.B -n
+option is specified, in which case only two consecutive empty reads will cause
+dj to exit).
+At exit, usage statistics are printed unless the option
+.B -q
+is specified a second time. The
+.B -H
+option will make these diagnostics human-readable.
+
+.SH DIAGNOSTICS
+
+The
+.B -d
+option prints all information, user-specified or otherwise, before program
+execution.
+.PP
+When dj exits, by default statistics are printed for input and output to
+standard error in the following format:
+.PP
+.R {records read} {ASCII unit separator} {partial records read}
+.R {ASCII record separator} {records written} {ASCII unit separator}
+.R {partial records written} {ASCII group separator} {bytes read}
+.R {ASCII record separator} {bytes written} {ASCII file separator}
+.PP
+If the
+.B -H
+option is specified dj instead uses this following format:
+.PP
+.R {records read} '+' {partial records read} '>' {records written}
+.R '+' {partial records written} ';' {bytes read} '>' {bytes written}
+.R {ASCII line feed}
+.PP
+The
+.B -q
+option suppresses error messages which print when a read or write is partial or
+empty and when used twice suppresses diagnostic output entirely.
+.PP
+In non-recoverable errors that don't pertain to dj's read-write cycle, a
+diagnostic message is printed and dj exits with the appropriate sysexits(3)
+status.
+
+.SH BUGS
+
+If
+.B -n
+is specified along with a specified count, actual byte output may be lower than
+expected (the product of the count multiplied by the input block size). If the
+.B -a
+or
+.B -A
+options are used this could make data written nonsensical.
+.PP
+Many lowercase options have capitalized variants and vice-versa which can be
+confusing. Capitalized options tend to affect output or are more intense
+versions of lowercase options.
+
+.SH RATIONALE
+
+Dj was modeled after the dd utility specified in POSIX but adds additional
+features: typical option formatting, allowing seeks to be specified in bytes
+rather than in blocks, allowing arbitrary bytes as padding, and printing in a
+format that's easy to parse for machines. It also neglects character
+conversion, which may be dd's original intent but is irrelevant to its modern
+use.
+
+.SH COPYRIGHT
+
+Copyright (C) 2023 DTB. License AGPLv3+: GNU AGPL version 3 or later
+<https://gnu.org/licenses/agpl.html>.
+
+.SH SEE ALSO
+
+dd(1)

src/dj.c

@@ -0,0 +1,429 @@
+#include <ctype.h> /* isupper(3), tolower(3) */
+#include <errno.h> /* errno */
+#include <fcntl.h> /* open(2) */
+#include <stdio.h> /* fprintf(3), stderr */
+#include <stdlib.h> /* free(3), malloc(3), strtol(3), size_t */
+#include <string.h> /* memcpy(3), memmove(3), memset(3) */
+#include <sysexits.h> /* EX_OK, EX_USAGE */
+#include <unistd.h> /* close(2), getopt(3), lseek(2), read(2), write(2),
+                       optarg, optind, STDIN_FILENO, STDOUT_FILENO */
+extern int errno;
+
+/* dj uses two structures that respectively correspond to the reading and
+ * writing ends of its jockeyed "pipe". User-configurable members are noted
+ * with their relevant options. */
+struct Io{
+	int bs;			/* buffer size (-bB) */
+	size_t bufuse;	/* buffer usage */
+	char *buf;		/* buffer */
+	int bytes;		/* bytes processed */
+	int fd;			/* file descriptor */
+	int fl;			/* file opening flags */
+	char *fn;		/* file name (may be stdin_name or stdout_name) (-io) */
+	int prec;		/* partial records processed */
+	int rec;		/* records processed */
+	long seek;		/* bytes to seek/skip (will be 0 after skippage) (-sS) */
+} ep[2]; /* "engineered pipe"; also "extended play", for the deejay */
+
+/* Additionally, the following global variables are used to store user options.
+ */
+
+/* (-a)  */ static int align; /* Only the lower 8b are used but align is
+                               * negative if no alignment is being done. */
+
+/* (-c)  */ static int count; /* 0 if dj(1) runs until no more reads are
+                               * possible. */
+
+/* ASCII field separator delimited statistics */
+            static char *fmt_asv   = "%d\037%d\036%d\037%d\035%d\036%d\034";
+/* human-readable statistics */
+            static char *fmt_human = "%d+%d > %d+%d; %d > %d\n";
+/* pointer to chosen formatting */
+/* (-H)  */ static char *fmt_output; /* fmt_asv (default) or fmt_human (-H) */
+
+/* (-dq) */ static char debug;       /*
+ *  -d  increments             dj -qq | 0 - no diagnostic output whatsoever
+ *   -q decrements             dj -q  | 1 - typical output without
+ *                                    |     notifications on partial reads or
+ *                                    |     writes
+ *                             dj     | 2 - typical output (default)
+ *                             dj -d  | 3 - verbose status messages */
+
+/* (-n)  */ static char noerror;     /* 0 - exits on partial reads or writes
+                                      *     (default)
+                                      * 1 - retries on partial reads/writes
+                                      *     (-f) */
+
+/* Non-configurable defaults. */
+#define bs_default 1024 /* GNU dd(1) default; twice POSIX but a neat 2^10 */
+static char *program_name = "<no argv[0]>";
+static char *stdin_name = "<stdin>";
+static char *stdout_name = "<stdout>";
+static int read_flags = O_RDONLY; /* These flags are consistent with Busybox */
+static int write_flags = O_WRONLY | O_CREAT; /* dd(1). */
+
+/* Macro to set defaults for user-configurable options. */
+#define setdefaults do{ \
+	align = -1; \
+	count = 0; \
+	debug = 2; \
+	fmt_output = fmt_asv; \
+	noerror = 0; \
+	ep[0].fl = read_flags; \
+	Io_setdefaults(&ep[0]); \
+	ep[1].fl = write_flags; \
+	Io_setdefaults(&ep[1]); }while(0)
+
+#define MIN(a, b) (((a) < (b)) ? (a) : (b))
+
+/* Macro to check if fd is a std* file, e.g. stdin. */
+#define fdisstd(fd) \
+		((fd) == STDIN_FILENO \
+		|| (fd) == STDOUT_FILENO \
+		|| (fd) == STDERR_FILENO)
+
+/* Macro to call the cleanup functions that operate on struct io on the
+ * particular io[2] used in main. Error conditions are not checked because this
+ * is only used when the program is about to terminate (hence its name). */
+#define terminate(io) do{ \
+	Io_buffree(&(io)[0]); \
+	Io_buffree(&(io)[1]); \
+	Io_fdclose(&(io)[0]); \
+	Io_fdclose(&(io)[1]); }while(0)
+
+/* Allocates *io's buffer. Returns NULL if unsuccessful. */
+static void *
+Io_bufalloc(struct Io *io){
+
+	return (io->buf = malloc(io->bs * (sizeof *io->buf)));
+}
+
+/* Frees *io's buffer. Returns io. */
+static struct Io *
+Io_buffree(struct Io *io){
+
+	free(io->buf);
+
+	return io;
+}
+
+/* Fills the unused portion of io's buffer with padding, updating io->bufuse.
+ * Returns io. */
+static struct Io *
+Io_bufrpad(struct Io *io, int padding){
+	
+	memset(io->buf + io->bufuse, padding, io->bs - io->bufuse);
+	io->bufuse = io->bs;
+
+	return io;
+}
+
+/* Copies from the buffer in src as much as possible to the free space in the
+ * dest buffer, removing the copied units from src and permuting the remaining
+ * units in the src buffer to the start of the buffer, modifying both the src
+ * and dest bufuse and returning dest. */
+static struct Io*
+Io_bufxapp(struct Io *dest, struct Io *src){
+	int n;
+
+	n = MIN(src->bufuse, dest->bs - dest->bufuse);
+	memcpy(dest->buf + dest->bufuse, src->buf, n);
+	dest->bufuse += n;
+	memmove(src->buf, src->buf + n, src->bs - n);
+	src->bufuse -= n;
+
+	return dest;
+}
+
+/* Copies from the buffer in src to the buffer in dest no more than n units,
+ * removing the copied units from src and permuting the remaining units in the
+ * src buffer to the start of the buffer, modifying both the src and dest
+ * bufuse and returning dest. */
+static struct Io*
+Io_bufxfer(struct Io *dest, struct Io *src, int n){
+
+	memcpy(dest->buf, src->buf, (dest->bufuse = n));
+	memmove(src->buf, src->buf + n, (src->bufuse -= n));
+
+	return dest;
+}
+
+/* Closes io->fn and returns -1 on error, otherwise io->fd. */
+static int
+Io_fdclose(struct Io *io){
+
+	return fdisstd(io->fd)
+		? 0
+		: close(io->fd);
+}
+
+/* Opens io->fn and saves the file descriptor into io->fd. Returns io->fd,
+ * which will be -1 if an error occured. */
+static int
+Io_fdopen(struct Io *io, char *fn){
+	int fd;
+	
+	if((fd = open(fn, io->fl,
+					/* these are the flags used by touch(1p) */
+					S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP | S_IROTH | S_IWOTH))
+				!= -1
+			&& Io_fdclose(io) == 0){
+		io->fd = fd;
+		io->fn = fn;
+	}
+
+	return fd;
+}
+
+/* Seeks io->seek bytes through *io's file descriptor, (counter-intuitively)
+ * returning -1 if successful and a sysexits.h exit code if an unrecoverable
+ * error occurred. io->buf will be cleared of useful bytes and io->seek will
+ * be set to zero to indicate the seek occurred. */
+static int
+Io_fdseek(struct Io *io){
+	int (*op)(int, void *, size_t);
+	
+	if(!fdisstd(io->fd) && lseek(io->fd, io->seek, SEEK_SET) != -1)
+		return -1;
+	else if(io->fl == write_flags){
+		memset(io->buf, '\0', io->bs);
+		/* This is a dirty trick; rather than testing conditions and operating
+		 * likewise, because the parameters to read or write are going to be
+		 * the same either way, just use a function pointer to keep track of
+		 * the intended operation. */
+		op = (int (*)(int, void *, size_t))&write;
+		/* Function pointer casts are risky; this works because the difference
+		 * is in the second parameter and only that write(2) makes the buffer
+		 * const whereas read(2) does not. To avoid even the slightest
+		 * undefined behavior comment out the cast, just be ready for a
+		 * -Wincompatible-function-pointer-types if your compiler notices it.
+		 */
+	}else
+		op = &read;
+
+	/* We're going to cheat and use bufuse as the retval for write(2), which is
+	 * fine because it'll be zeroed as this function returns anyway. */
+	do{	if(	(io->bufuse = (*op)(io->fd, io->buf, MIN(io->bs, io->seek))) == 0)
+			/* second chance */
+			io->bufuse = (*op)(io->fd, io->buf, MIN(io->bs, io->seek));
+	}while((io->seek -= io->bufuse) > 0 && io->bufuse != 0);
+
+	io->bufuse = 0;
+
+	return -1;
+}
+
+/* Reads io->bs bytes from *io's file descriptor into io->buf, storing the
+ * number of read bytes in io->bufuse and updating io->bytes. If io->bufuse is
+ * 0, errno will probably be set. Returns io. */
+static struct Io *
+Io_read(struct Io *io){
+
+	io->bytes += (io->bufuse = read(io->fd, io->buf, io->bs));
+
+	return io;
+}
+
+/* Sets the variables in a struct *io to the defaults. Identifies the read/
+ * write ends of the "pipe" by checking io->fl. Returns io. */
+static struct Io *
+Io_setdefaults(struct Io *io){
+
+	io->bs = bs_default;
+	io->buf = NULL;
+	io->bytes = 0;
+	io->fd = (io->fl == read_flags) ? STDIN_FILENO : STDOUT_FILENO;
+	io->fn = (io->fl == read_flags) ? stdin_name   : stdout_name;
+	io->prec = 0;
+	io->rec = 0;
+	io->seek = 0;
+
+	return io;
+}
+
+/* Writes io->bufuse units from io->buf to io->fd, permuting any unwritten
+ * bytes to the start of io->buf and updating io->bufuse. If io->bufuse doesn't
+ * change, errno will probably be set. Returns io. */
+static struct Io *
+Io_write(struct Io *io){
+	int t;
+
+	if((t = write(io->fd, io->buf, io->bufuse)) > 0)
+		memmove(io->buf, io->buf + t, (io->bufuse -= t));
+	io->bytes += t;
+
+	return io;
+}
+
+/* Prints an error message suitable for the event of an operating system error,
+ * with the error itself to be described in the string s. */
+static int
+oserr(char *s){
+
+	fprintf(stderr, "%s: %s: %s\n", program_name, s, strerror(errno));
+
+	return EX_OSERR;
+}
+
+/* Prints statistics regarding the use of dj, particularly partially and
+ * completely read and written records, accessing debug, ep, and fmt_output. */
+static void
+output(void){
+
+	if(debug >= 1)
+		fprintf(stderr, fmt_output,
+			ep[0].rec, ep[0].prec, ep[1].rec, ep[1].prec,
+			ep[0].bytes, ep[1].bytes);
+
+	return;
+}
+
+/* Parses the string s to an integer, returning either the integer or in the
+ * case of an error a negative integer. This is used for argument parsing
+ * (e.g. -B [int]) in dj and no negative integer would be valid anyway. */
+static long
+parse(char *s){
+	long r;
+
+	errno = 0;
+	r = strtol(s, &s, 0);
+	return (*s == '\0' /* no chars left unparsed */ && errno == 0)
+		? r
+		: -1;
+}
+
+static int
+usage(void){
+
+	fprintf(stderr, "Usage: %s (-AdfHqQ) (-a [byte]) (-c [count])\n"
+		"\t(-i [input file]) (-b [input block size]) (-s [input offset])\n"
+		"\t(-o [output file]) (-B [output block size]) (-S [output offset])\n",
+		program_name);
+
+	return EX_USAGE;
+}
+
+int main(int argc, char *argv[]){
+	int c;
+	int i;
+
+	setdefaults;
+
+	if(argc > 0){
+		program_name = argv[0];
+		while((c = getopt(argc, argv, "a:Ab:B:c:di:fhHqs:S:o:")) != -1)
+			switch(c){
+			case 'i': case 'o':
+				if(Io_fdopen(&ep[c == 'o'], optarg) != -1)
+					break;
+				terminate(ep);
+				return oserr(optarg);
+			case 'A': align = '\0'; break;
+			case 'd': ++debug; break;
+			case 'n': noerror = 1; break;
+			case 'H': fmt_output = fmt_human; break;
+			case 'q': --debug; break;
+			case 'a':
+				if(optarg[0] != '\0' && optarg[1] == '\0'){
+					align = optarg[0];
+					break;
+				}
+				/* FALLTHROUGH */
+			case 'c': case 'b': case 's': case 'B': case 'S':
+				if(c == 'c' && (count = parse(optarg)) >= 0)
+					break;
+				i = isupper(c);
+				c = tolower(c);
+				if((c == 'b' && (ep[i].bs = parse(optarg)) > 0)
+						|| (c == 's' && (ep[i].seek = parse(optarg)) >= 0))
+					break;
+				/* FALLTHROUGH */
+			default:
+				terminate(ep);
+				return usage();
+			}
+	}
+
+	if(debug >= 3)
+		fprintf(stderr,
+			"argv0=%s\n"
+			"in=%s\tibs=%d\tskip=%ld\talign=%hhx\tcount=%d\n"
+			"out=%s\tobs=%d\tseek=%ld\tdebug=%2d\tnoerror=%d\n",
+			program_name,
+			ep[0].fn, ep[0].bs, ep[0].seek, align, count,
+			ep[1].fn, ep[1].bs, ep[1].seek, debug, noerror);
+
+	if(argc > optind){
+		terminate(ep);
+		return usage();
+	}
+
+	for(i = 0; i <= 1; ++i){
+		if(Io_bufalloc(&ep[i]) == NULL){
+			fprintf(stderr, "%s: Failed to allocate %d bytes\n",
+				program_name, ep[i].bs);
+			terminate(ep);
+			return EX_OSERR;
+		}else if(ep[i].seek > 0)
+			switch(Io_fdseek(&ep[i])){
+			case EX_OK:
+				output();
+				terminate(ep);
+				return EX_OK;
+			}
+	}
+
+	do{	/* read */
+		Io_read(&ep[0]);
+		if(!noerror && ep[0].bufuse == 0)
+			Io_read(&ep[0]); /* second chance */
+		if(ep[0].bufuse == 0) /* that's all she wrote */
+			break;
+		else if(ep[0].bufuse < ep[0].bs){
+			++ep[0].prec;
+			if(debug >= 2){
+				fprintf(stderr, "%s: Partial read:\n\t", program_name);
+				output();
+			}
+			if(!noerror)
+				count = 1;
+			if(align >= 0)
+				Io_bufrpad(&ep[0], align);
+		}else
+			++ep[0].rec;
+
+		/* write */
+		do{	if(ep[1].bs > ep[0].bs){ /* io[1].bs > io[0].bs */
+				Io_bufxapp(&ep[1], &ep[0]);
+				if(ep[0].bs + ep[1].bufuse <= ep[1].bs && count != 1)
+					continue; /* we could write more */
+			}else
+				Io_bufxfer(&ep[1], &ep[0], MIN(ep[0].bufuse, ep[1].bs));
+
+			c = ep[1].bufuse;
+			Io_write(&ep[1]);
+			if(!noerror && ep[1].bufuse == c)
+				Io_write(&ep[1]); /* second chance */
+			if(c == ep[1].bufuse){ /* no more love */
+				count = 1;
+				break;
+			}else if(c > ep[1].bufuse && ep[1].bufuse > 0){
+				ep[1].prec += 1;
+				if(debug >= 2){
+					fprintf(stderr, "%s: Partial write:\n\t", program_name);
+					output();
+				}
+				if(!noerror)
+					count = 1;
+			}else if(ep[1].bufuse == 0 && c < ep[1].bs)
+				++ep[1].prec;
+			else
+				++ep[1].rec;
+		}while(ep[0].bufuse > 0);
+	}while(count == 0 || --count > 0);
+
+	output();
+	terminate(ep);
+
+	return EX_OK;
+}