dj(1): refactor source

CONTRIBUTING

@@ -97,10 +97,9 @@ their editor or terminal.
 
 For usage text and help messages, do not implement a -h option. Instead, print
 usage information when any erroneous option is specified. Follow the NetBSD
-style guide for the usage text’s output format [0].
+style guide for the usage text’s output format [1].
 
-If committing a new utility, please include tests and documentation (see
-tests/ and docs/) for the new tool.
+[1] <http://cvsweb.netbsd.org/bsdweb.cgi/~checkout~/src/share/misc/style>
 
 If committing a new source file, format the commit message following these
 guidelines:
@@ -129,7 +128,6 @@ $ git commit -m 'tool(1): fix #42 & add feature x'
 
 Commit messages should be written in the present tense.
 
-[0] <http://cvsweb.netbsd.org/bsdweb.cgi/~checkout~/src/share/misc/style>
 --
 This work © 2023–2024 by Emma Tebibyte is licensed under CC BY-SA 4.0. To view a
 copy of this license, visit <http://creativecommons.org/licenses/by-sa/4.0/>

Makefile

@@ -1,6 +1,7 @@
 # Copyright (c) 2023–2024 Emma Tebibyte <emma@tebibyte.media>
 # Copyright (c) 2023–2024 DTB <trinity@trinity.moe>
 # Copyright (c) 2023 Sasha Koshka <sashakoshka@tebibyte.media>
+# Copyright (c) 2024 Aaditya Aryal <aryalaadi123@gmail.com>
 # SPDX-License-Identifier: FSFAP
 #
 # Copying and distribution of this file, with or without modification, are
@@ -14,11 +15,10 @@
 PREFIX=/usr/local
 
 CC=cc
-MAKE=make
 RUSTC=rustc
 
 .PHONY: all
-all: dj false fop hru intcmp rpn scrut str strcmp true
+all: dj false fop hru intcmp mm npc rpn scrut str strcmp swab true
 
 build:
 	# keep build/include until bindgen(1) has stdin support
@@ -40,8 +40,8 @@ install: dist
 	cp -r dist/* $(PREFIX)/
 
 .PHONY: test
-test: all
-	tests/test.sh
+test: build
+	tests/posix-compat.sh
 	$(RUSTC) --test src/getopt-rs/lib.rs -o build/test/getopt
 
 build/o/libsysexits.rlib: build
@@ -87,6 +87,17 @@ intcmp: build/bin/intcmp
 build/bin/intcmp: src/intcmp.c build
 	$(CC) $(CFLAGS) -o $@ src/intcmp.c
 
+.PHONY: mm
+mm: build/bin/mm
+build/bin/mm: src/mm.c build
+	$(CC) $(CFLAGS) -o $@ src/mm.c
+
+
+.PHONY: npc
+npc: build/bin/npc
+build/bin/npc: src/npc.c build
+	$(CC) $(CFLAGAS) -o $@ src/npc.c
+
 .PHONY: rpn
 rpn: build/bin/rpn
 build/bin/rpn: src/rpn.rs build build/o/libsysexits.rlib
@@ -109,6 +120,13 @@ strcmp: build/bin/strcmp
 build/bin/strcmp: src/strcmp.c build
 	$(CC) $(CFLAGS) -o $@ src/strcmp.c
 
+.PHONY: swab
+swab: build/bin/swab
+build/bin/swab: src/swab.rs build build/o/libsysexits.rlib
+	$(RUSTC) $(RUSTFLAGS) --extern getopt=build/o/libgetopt.rlib \
+		--extern sysexits=build/o/libsysexits.rlib \
+		-o $@ src/swab.rs
+
 .PHONY: true
 true: build/bin/true
 build/bin/true: src/true.c build

docs/mm.1

@@ -0,0 +1,76 @@
+.\" Copyright (c) 2024 DTB <trinity@trinity.moe>
+.\"
+.\" This work is licensed under CC BY-SA 4.0. To see a copy of this license,
+.\" visit <http://creativecommons.org/licenses/by-sa/4.0/>.
+
+.TH mm 1
+
+.SH NAME
+
+mm \(en middleman
+
+.SH SYNOPSIS
+
+mm
+.RB ( -aenu )
+.RB ( -i
+.RB [ input ])
+.RB ( -o
+.RB [ output ])
+
+.SH DESCRIPTION
+
+Mm catenates input files and writes them to the start of each output file.
+
+.SH OPTIONS
+
+Mm, upon receiving the
+.B -a
+option, will open subsequent outputs for appending rather than updating.
+.PP
+The
+.B -i
+option opens a path as an input. Without any inputs specified mm will use
+standard input. Standard input itself can be specified by giving the path '-'.
+.PP
+The
+.B -o
+option opens a path as an output. Without any outputs specified mm will use
+standard output. Standard output itself can be specified by giving the
+path '-'. Standard error itself can be specified with the
+.B -e
+option.
+.PP
+The
+.B -u
+option ensures neither input or output will be buffered.
+.PP
+The
+.B -n
+option tells mm to ignore SIGINT signals.
+
+.SH DIAGNOSTICS
+
+If an output can no longer be written mm prints a diagnostic message, ceases
+writing to that particular output, and if there are more outputs specified,
+continues, eventually exiting unsuccessfully.
+.PP
+On error mm prints a diagnostic message and exits with the appropriate
+sysexits.h(3) status.
+
+.SH BUGS
+
+Mm does not truncate existing files, which may lead to unexpected results.
+
+.SH RATIONALE
+
+Mm was modeled after the cat and tee utilities specified in POSIX.
+
+.SH COPYRIGHT
+
+Copyright (c) 2024 DTB. License AGPLv3+: GNU AGPL version 3 or later
+<https://gnu.org/licenses/agpl.html>.
+
+.SH SEE ALSO
+
+cat(1p), dd(1), dj(1), tee(1p)

docs/swab.1

@@ -0,0 +1,71 @@
+.\" Copyright (c) 2024 DTB <trinity@trinity.moe>
+.\"
+.\" This work is licensed under CC BY-SA 4.0. To see a copy of this license,
+.\" visit <http://creativecommons.org/licenses/by-sa/4.0/>.
+
+.TH swab 1
+
+.SH NAME
+
+swab \(en swap bytes
+
+.SH SYNOPSIS
+
+swab
+.RB ( -f )
+.RB ( -w
+.R [
+.B word size
+.R ])
+
+.SH USAGE
+
+Swab swaps the latter and former halves of a block of bytes.
+
+.SH EXAMPLES
+
+The following sh(1p) line:
+
+.R printf 'hello world!\n' | swab
+
+Produces the following output:
+
+.R ehll oowlr!d
+
+.SH OPTIONS
+
+The
+.B -f
+option ignores system call interruptions.
+.PP
+The
+.B -w
+option configures the word size; that is, the size in bytes of the block size
+on which to operate. By default the word size is 2. The word size must be
+cleanly divisible by 2, otherwise the block of bytes being processed can't be
+halved.
+
+.SH DIAGNOSTICS
+
+If an error is encountered in input, output, or invocation, a diagnostic
+message will be written to standard error and swab will exit with the
+appropriate status from sysexits.h(3).
+
+.SH RATIONALE
+
+Swab was modeled after the
+.R conv=swab
+functionality specified in the POSIX dd utility but additionally allows the
+word size to be configured.
+.PP
+Swab is useful for fixing the endianness of binary files produced on other
+machines.
+
+.SH COPYRIGHT
+
+Copyright (c) 2024 DTB. License AGPLv3+: GNU AGPL version 3 or later
+<https://gnu.org/licenses/agpl.html>.
+
+.SH SEE ALSO
+
+dd(1p)

src/dj.c

@@ -46,31 +46,11 @@ struct Io{
 /* 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) */
+static char *fmt_asv   = "%d\037%d\036%d\037%d\035%d\036%d\034";
 
-/* (-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
-                                      *     (-n) */
+/* (-H) Human-readable statistics */
+static char *fmt_human = "%d+%d > %d+%d; %d > %d\n";
 
 /* Non-configurable defaults. */
 #define bs_default 1024 /* GNU dd(1) default; twice POSIX but a neat 2^10 */
@@ -80,19 +60,8 @@ 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))
+#define MAX(a, b) (((a) > (b)) ? (a) : (b))
 
 /* Macro to check if fd is a std* file, e.g. stdin. */
 #define fdisstd(fd) \
@@ -104,43 +73,16 @@ static int write_flags = O_WRONLY | O_CREAT; /* dd(1). */
  * 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]); \
+	free((io)[0].buf); \
+	free((io)[1].buf); \
 	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*
+static struct Io *
 Io_bufxapp(struct Io *dest, struct Io *src){
 	int n;
 
@@ -157,7 +99,7 @@ Io_bufxapp(struct Io *dest, struct Io *src){
  * 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*
+static struct Io *
 Io_bufxfer(struct Io *dest, struct Io *src, int n){
 
 	memcpy(dest->buf, src->buf, (dest->bufuse = n));
@@ -167,13 +109,10 @@ Io_bufxfer(struct Io *dest, struct Io *src, int n){
 }
 
 /* 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);
-}
+#define Io_fdclose(io) \
+	(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. */
@@ -199,32 +138,31 @@ Io_fdopen(struct Io *io, char *fn){
  * 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);
+	/* repeated code to get the condition out of the loop */
+	if(io->fl == write_flags){
+		memset(io->buf, '\0', io->bs);
+		/* 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 = write(io->fd, io->buf, MIN(io->bs, io->seek)))
+					== 0)
+				/* second chance */
+				io->bufuse = write(io->fd, io->buf, MIN(io->bs, io->seek));
+		}while((io->seek -= io->bufuse) > 0 && io->bufuse != 0);
+	}else if(io->fl == read_flags){
+		do{
+			if((io->bufuse = read(io->fd, io->buf, MIN(io->bs, io->seek)))
+					== 0)
+				/* second chance */
+				io->bufuse = read(io->fd, io->buf, MIN(io->bs, io->seek));
+		}while((io->seek -= io->bufuse) > 0 && io->bufuse != 0);
+	}else
+		return EX_SOFTWARE;
 
 	io->bufuse = 0;
 
@@ -283,19 +221,6 @@ oserr(char *s){
 	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. */
@@ -321,11 +246,30 @@ usage(void){
 	return EX_USAGE;
 }
 
-int main(int argc, char *argv[]){
-	int c;
-	int i;
+/* For use in main only.
+ * Prints statistics regarding the use of dj, particularly partially and
+ * completely read and written records, accessing ep, and fmt_output. */
+#define output fprintf(stderr, fmt_output, \
+	ep[0].rec, ep[0].prec, ep[1].rec, ep[1].prec, ep[0].bytes, ep[1].bytes \
+)
 
-	setdefaults;
+int main(int argc, char *argv[]){
+	int align; /* Only the lower 8b are used, negative if no alignment. */
+	int count;
+	int c, i;
+	enum { QUIETER = 0, TYPICAL = 1, VERBOSE = 2 } debug;
+	enum { GIVEUP = 0, RETRY = 1 } error;
+	char *fmt_output = fmt_asv;
+
+	ep[0].fl = read_flags;
+	Io_setdefaults(&ep[0]);
+	ep[1].fl = write_flags;
+	Io_setdefaults(&ep[1]);
+
+	align = -1;
+	count = 0;
+	debug = TYPICAL;
+	error = GIVEUP;
 
 	if(argc > 0){
 		program_name = argv[0];
@@ -342,10 +286,10 @@ int main(int argc, char *argv[]){
 				terminate(ep);
 				return oserr(optarg);
 			case 'A': align = '\0'; break;
-			case 'd': ++debug; break;
-			case 'n': noerror = 1; break;
+			case 'd': debug = VERBOSE; break;
+			case 'n': error = RETRY; break;
 			case 'H': fmt_output = fmt_human; break;
-			case 'q': --debug; break;
+			case 'q': debug = QUIETER; break;
 			case 'a':
 				if(optarg[0] != '\0' && optarg[1] == '\0'){
 					align = optarg[0];
@@ -367,14 +311,15 @@ int main(int argc, char *argv[]){
 			}
 	}
 
-	if(debug >= 3)
+	if(debug >= VERBOSE)
 		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",
+			"out=%s\tobs=%d\tseek=%ld\tdebug=%2d\terror=%s\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);
+			ep[1].fn, ep[1].bs, ep[1].seek, debug,
+		    error == GIVEUP ? "GIVEUP" : "RETRY");
 
 	if(argc > optind){
 		terminate(ep);
@@ -382,36 +327,37 @@ int main(int argc, char *argv[]){
 	}
 
 	for(i = 0; i <= 1; ++i){
-		if(Io_bufalloc(&ep[i]) == NULL){
+		if((ep[i].buf = malloc(ep[i].bs * (sizeof ep[i].buf))) == NULL){
 			fprintf(stderr, "%s: Failed to allocate %d bytes\n",
-				program_name, ep[i].bs);
+				program_name, ep[i].bs * (sizeof ep[i].buf));
 			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;
-			}
+		}else if(ep[i].seek > 0 && (c = Io_fdseek(&ep[i])) != -1){
+			output;
+			terminate(ep);
+			return c;
+		}
 	}
 
 	do{	/* read */
 		Io_read(&ep[0]);
-		if(!noerror && ep[0].bufuse == 0)
+		if(error == RETRY && 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){
+			if(debug >= TYPICAL){
 				fprintf(stderr, "%s: Partial read:\n\t", program_name);
-				output();
+				output;
 			}
-			if(!noerror)
+			if(error == GIVEUP)
 				count = 1;
-			if(align >= 0)
-				Io_bufrpad(&ep[0], align);
+			if(align >= 0){
+				memset(ep[0].buf + ep[0].bufuse, align,
+					ep[0].bs - ep[0].bufuse);
+				ep[0].bufuse = ep[0].bs;
+			}
 		}else
 			++ep[0].rec;
 
@@ -425,18 +371,18 @@ int main(int argc, char *argv[]){
 
 			c = ep[1].bufuse;
 			Io_write(&ep[1]);
-			if(!noerror && ep[1].bufuse == c)
+			if(error == GIVEUP && 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){
+				if(debug >= TYPICAL){
 					fprintf(stderr, "%s: Partial write:\n\t", program_name);
-					output();
+					output;
 				}
-				if(!noerror)
+				if(error == GIVEUP)
 					count = 1;
 			}else if(ep[1].bufuse == 0 && c < ep[1].bs)
 				++ep[1].prec;
@@ -445,7 +391,7 @@ int main(int argc, char *argv[]){
 		}while(ep[0].bufuse > 0);
 	}while(count == 0 || --count > 0);
 
-	output();
+	output;
 	terminate(ep);
 
 	return EX_OK;

src/mm.c

@@ -0,0 +1,224 @@
+/*
+ * Copyright (c) 2024 DTB <trinity@trinity.moe>
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ *
+ * This program is free software: you can redistribute it and/or modify it under
+ * the terms of the GNU Affero General Public License as published by the Free
+ * Software Foundation, either version 3 of the License, or (at your option) any
+ * later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see https://www.gnu.org/licenses/.
+ */
+
+#include <errno.h> /* errno */
+#include <signal.h> /* signal(2), SIG_ERR, SIG_IGN, SIGINT */
+#include <stdio.h> /* fclose(3), fopen(3), fprintf(3), getc(3), putc(3),
+                    * setvbuf(3), size_t, _IONBF, NULL */
+#include <stdlib.h> /* free(3), realloc(3) */
+#include <string.h> /* strcmp(3), strerror(3) */
+#include <unistd.h> /* getopt(3) */
+#if !defined EX_IOERR || !defined EX_OK || !defined EX_OSERR \
+		|| !defined EX_USAGE
+#	include <sysexits.h>
+#endif
+extern int errno;
+
+/* This structure is how open files are tracked. */
+struct Files{
+	size_t a;	/* allocation */
+	size_t s;	/* used size */
+	char *mode;	/* file opening mode */
+	char **names;	/* file names */
+	FILE **files;	/* file pointers */
+};
+
+/* How much to grow the allocation when it's saturated. */
+#ifndef ALLOC_INCREMENT
+#	define ALLOC_INCREMENT 1
+#endif
+
+/* How much to grow the allocation at program start. */
+#ifndef ALLOC_INITIAL
+#	define ALLOC_INITIAL 10
+#endif
+
+/* pre-allocated strings */
+static char *program_name = "<no argv[0]>";
+static char *stdin_name = "<stdin>";
+static char *stdout_name = "<stdout>";
+static char *stderr_name = "<stderr>";
+static char *(fmode[]) = { (char []){"rb"}, (char []){"rb+"} };
+static char *wharsh = "wb";
+
+/* Adds the open FILE pointer for the file at the path s to the files struct,
+ * returning the FILE if successful and NULL if not, allocating more memory in
+ * the files buffers as needed. */
+static FILE *
+Files_append(struct Files *files, FILE *file, char *name){
+
+	if(file == NULL || (files->s == files->a
+			&& ((files->files = realloc(files->files,
+					(files->a += (files->a == 0)
+						? ALLOC_INITIAL
+						: ALLOC_INCREMENT)
+					* sizeof *(files->files))) == NULL
+				|| (files->names = realloc(files->names,
+					files->a * sizeof *(files->names))) == NULL)))
+		return NULL;
+
+	files->names[files->s] = name;
+	return files->files[files->s++] = file;
+}
+
+/* Opens the file at the path p and puts it in the files struct, returning NULL
+ * if either the opening or the placement of the open FILE pointer fail. */
+#define Files_open(files, p) \
+	Files_append((files), fopen((p), (files)->mode), (p))
+
+/* Prints a diagnostic message based on errno and returns an exit status
+ * appropriate for an OS error. */
+static int
+oserr(char *s, char *r){
+
+	fprintf(stderr, "%s: %s: %s\n", s, r, strerror(errno));
+
+	return EX_OSERR;
+}
+
+/* Hijacks i and j from main and destructs the files[2] struct used by main by
+ * closing its files and freeing its files and names arrays, returning retval
+ * from main. */
+#define terminate \
+	for(i = 0; i < 2; ++i){ \
+		for(j = 0; j < files[i].s; ++j) \
+			if(files[i].files[j] != stdin \
+					&& files[i].files[j] != stdout \
+					&& files[i].files[j] != stderr) \
+				fclose(files[i].files[j]); \
+		free(files[i].files); \
+		free(files[i].names); \
+	} \
+	return retval
+
+int main(int argc, char *argv[]){
+	int c;
+	struct Files files[2]; /* {read, write} */
+	size_t i;
+	size_t j;
+	size_t k; /* loop index but also unbuffer status */
+	int retval;
+
+	/* Initializes the files structs with their default values, standard
+	 * input and standard output. If an input or an output is specified
+	 * these initial values will be overwritten, so to, say, use mm(1)
+	 * equivalently to tee(1p), -o - will need to be specified before
+	 * additional files to ensure standard output is still written. */
+	for(i = 0; i < 2; ++i){
+		files[i].a = 0;
+		files[i].s = 0;
+		files[i].mode = fmode[i];
+		files[i].files = NULL;
+		files[i].names = NULL;
+		Files_append(&files[i], i == 0 ? stdin : stdout,
+			i == 0 ? stdin_name : stdout_name);
+		files[i].s = 0;
+	}
+
+	k = 0;
+
+	if(argc > 0)
+		program_name = argv[0];
+
+	if(argc > 1)
+		while((c = getopt(argc, argv, "aehi:no:u")) != -1)
+			switch(c){
+			case 'a': /* "rb+" -> "ab" */
+				files[1].mode[0] = 'a';
+				files[1].mode[2] = '\0';
+				break;
+			case 'e':
+				if(Files_append(&files[1], stderr, stderr_name) != NULL)
+					break;
+				retval = oserr(argv[0], "-e");
+				terminate;
+			case 'i':
+				if((strcmp(optarg, "-") == 0 && Files_append(&files[0],
+							stdin, stdin_name) != NULL)
+						|| Files_open(&files[0], optarg) != NULL)
+					break;
+				retval = oserr(argv[0], optarg);
+				terminate;
+			case 'o':
+				if((strcmp(optarg, "-") == 0 && Files_append(&files[1],
+							stdout, stdout_name) != NULL)
+						|| Files_open(&files[1], optarg) != NULL)
+					break;
+				/* does not exist, so try to create it */
+				if(errno == ENOENT){
+					files[1].mode = wharsh;
+					if(Files_open(&files[1], optarg) != NULL){
+						files[1].mode = fmode[1];
+						break;
+					}
+				}
+				retval = oserr(argv[0], optarg);
+				terminate;
+			case 'n':
+				if(signal(SIGINT, SIG_IGN) != SIG_ERR)
+					break;
+				retval = oserr(argv[0], "-n");
+				terminate;
+			case 'u':
+				k = 1;
+				break;
+			default:
+				fprintf(stderr, "Usage: %s (-aenu) (-i [input])..."
+					" (-o [output])...\n", argv[0]);
+				retval = EX_USAGE;
+				terminate;
+			}
+
+	files[0].s += files[0].s == 0;
+	files[1].s += files[1].s == 0;
+	
+	/* Unbuffer files. */
+	if(k){
+		for(i = 0;
+			i < files[0].s;
+			setvbuf(files[0].files[i++], NULL, _IONBF, 0));
+		for(i = 0;
+			i < files[1].s;
+			setvbuf(files[1].files[i++], NULL, _IONBF, 0));
+	}
+
+	retval = EX_OK;
+
+	/* Actual program loop. */
+	for(i = 0; i < files[0].s; ++i) /* iterate ins */
+		while((c = getc(files[0].files[i])) != EOF) /* iterate chars */
+			for(j = 0; j < files[1].s; ++j) /* iterate outs */
+				if(putc(c, files[1].files[j]) == EOF){
+					/* notebook's full */
+					retval = EX_IOERR;
+					fprintf(stderr, "%s: %s: %s\n",
+						program_name, files[1].names[j], strerror(errno));
+					if(fclose(files[1].files[j]) == EOF)
+						fprintf(stderr, "%s: %s: %s\n",
+							program_name, files[1].names[j], strerror(errno));
+					/* massage out the tense muscle */
+					for(k = j--; k < files[1].s - 1; ++k){
+						files[1].files[k] = files[1].files[k+1];
+						files[1].names[k] = files[1].names[k+1];
+					}
+					if(--files[1].s == 0)
+						terminate;
+				}
+
+	terminate;
+}

src/swab.rs

@@ -0,0 +1,90 @@
+/*
+ * Copyright (c) 2024 DTB <trinity@trinity.moe>
+ * SPDX-License-Identifier: AGPL-3.0-or-later
+ *
+ * This program is free software: you can redistribute it and/or modify it under
+ * the terms of the GNU Affero General Public License as published by the Free
+ * Software Foundation, either version 3 of the License, or (at your option) any
+ * later version.
+ * 
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+ * FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more
+ * details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program. If not, see https://www.gnu.org/licenses/.
+ */
+
+use std::{
+	env::args,
+	io::{ stdin, stdout, Error, ErrorKind, Read, Write },
+	process::ExitCode,
+	vec::Vec
+};
+
+extern crate getopt;
+use getopt::{ Opt, Parser };
+
+extern crate sysexits;
+use sysexits::{ EX_OK, EX_OSERR, EX_USAGE };
+
+fn oserr(s: &str, e: Error) -> ExitCode {
+	eprintln!("{}: {}", s, e);
+	ExitCode::from(EX_OSERR as u8)
+}
+
+fn usage(s: &str) -> ExitCode {
+	eprintln!("Usage: {} (-f) (-w [wordsize])", s);
+	ExitCode::from(EX_USAGE as u8)
+}
+
+fn main() -> ExitCode {
+	let argv = args().collect::<Vec<String>>();
+	let mut buf: Vec<u8> = Vec::new();
+	let mut input = stdin();
+	let mut output = stdout().lock();
+
+	let mut opts = Parser::new(&argv, "fw:");
+	let mut force = false;
+	let mut wordsize: usize = 2;
+
+	loop {
+		match opts.next() {
+			None => break,
+			Some(opt) =>
+				match opt {
+					Ok(Opt('f', None)) => force = true,
+					Ok(Opt('w', Some(arg))) => {
+						match arg.parse::<usize>() {
+							Ok(w) if w % 2 == 0 => { wordsize = w; () },
+							_ => { return usage(&argv[0]); },
+						}
+					},
+					_ => { return usage(&argv[0]); }
+				}
+		}
+	}
+
+	buf.resize(wordsize, 0);
+
+	loop {
+		match input.read(&mut buf) {
+			Ok(0) => break ExitCode::from(EX_OK as u8),
+			Ok(v) if v == wordsize => {
+				let (left, right) = buf.split_at(v/2);
+				if let Err(e) = output.write(&right)
+						.and_then(|_| output.write(&left)) {
+					break oserr(&argv[0], e)
+				}
+			},
+			Ok(v) => {
+				if let Err(e) = output.write(&buf[..v]) {
+					break oserr(&argv[0], e)
+				}
+			},
+			Err(e) if e.kind() == ErrorKind::Interrupted && force => continue,
+			Err(e) => break oserr(&argv[0], e)
+		}
+	}
+}

tests/bonsai/dj.sh

@@ -1,27 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-! dj -h
-
-# This test is theoretically Linux-dependent; write(2) should return -1 on
-# error.
-# Right now dj(1) interprets the return value of write(2) as the amount of
-# bytes written. This can decrement the stored quantity of bytes written,
-# which is an int, so doesn't underflow but goes negative. dj(1) tries to
-# again to write(2) if an error occurs in which no bytes are written, so in
-# total two write(2)s are attempted and so the written byte quantity is -2.
-# This is a bug and will change, but for now is at least documented.
-dj -Hi /dev/zero -o /dev/full \
-	| xargs -I out "$BIN/strcmp" '1+0 > 0+0; 1024 > -2' out
-
-# Read nothing from /dev/null, write nothing to /dev/null.
-dj -Hi /dev/null -o /dev/null \
-	| xargs -I out "$BIN/strcmp" '0+0 > 0+0; 0 > 0' out

tests/bonsai/false.sh

@@ -1,13 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-! false
-! false -h

tests/bonsai/fop.sh

@@ -1,21 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-! fop -h
-
-"$BIN/strcmp" "$(printf 'test0␞test1␞test2\n' | fop 1 sed 's/1/4/g')" \
-	'test0␞test4␞test2'
-
-"$BIN/strcmp" "$(printf 'test0 test1 test2\n' | fop -d' ' 2 sed 's/2/4/g')" \
-	'test0 test1 test4'
-
-! printf 'test\n' | fop 1 cat
-! printf 'test\n' | fop 'test' cat
-! printf 'test\n' | fop -d'test' cat

tests/bonsai/hru.sh

@@ -1,26 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-alias strcmp="$BIN/strcmp"
-alias rpn="$BIN/rpn"
-
-strcmp "$(printf '1234\n' | hru)" '1.2 kB'
-strcmp "$(printf '0\n' | hru)" '0 B'
-
-# doesn’t currently work but would be useful for testing for regressions
-#n=1
-#while "$BIN/true"; do
-#	n="$(rpn "$n" 10 ×)"
-#
-#	printf '%s\n' "$n" | hru || break
-#done
-#printf 'integer limit: ~%s\n' "$(rpn "$n" 10 ÷)"
-
-! printf '%s\n' '-1' | hru

tests/bonsai/intcmp.sh

@@ -1,25 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-intcmp -e     3 3 3
-intcmp -g     3 2 1
-intcmp -l     1 2 3
-intcmp -ge    3 3 1
-intcmp -le    1 3 3
-intcmp -gl    1 2 3
-intcmp -egl   3 1 1 2
-! intcmp -e   1 2 3
-! intcmp -g   1 3 3
-! intcmp -l   3 3 1
-! intcmp -ge  1 2 3
-! intcmp -le  3 2 1
-! intcmp -gl  3 3 3
-! intcmp -egl foo

tests/bonsai/mm.sh

@@ -1,19 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-exec 3>&1
-
-! mm -h
-
-# mm(1) will error if positional arguments are given without -i or -o
-! mm argument
-
-# check if stderr is empty upon specifying -e
-! "$BIN/strcmp" "$(printf 'test\n' | mm -i - -e 2>&1 1>&3)" ''

tests/bonsai/strcmp.sh

@@ -1,16 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-strcmp equals equals
-! strcmp inequals equals
-strcmp - -
-strcmp -h
-! strcmp nocmp

tests/bonsai/test_env

@@ -1,5 +0,0 @@
-#!/bin/sh
-
-set -x
-
-alias "$UTIL=$BIN/$UTIL"

tests/bonsai/true.sh

@@ -1,13 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-. tests/bonsai/test_env
-
-true
-true -h

tests/posix-compat.sh

@@ -9,30 +9,16 @@
 
 set -e
 
-export BIN=build/bin
-
 if ! ls Makefile >/dev/null 2>&1
 then
 	printf '%s: Run this script in the root of the project.\n' "$0" 1>&2 
 	exit 1
 fi
 
-printf "Starting Bonsai testing.\n\n"
+printf "Starting POSIX compatibility testing.\n"
 
-for script in tests/bonsai/*.sh; do
-	export UTIL="$(printf '%s\n' "$script" \
-		| sed -e 's/\.sh//g' -e 's;tests\/bonsai\/;;g')"
-
-	printf '%s: %s: Testing utility.\n' "$0" "$UTIL"
-	"$script"
-	printf '\n'
-done
-
-printf "Starting POSIX compatibility testing.\n\n"
-
-for test in tests/posix/*.sh; do
-	export PATH="$BIN:$PATH"
-	printf '%s: %s: Testing utility.\n' "$0" "$test"
+for utility in tests/posix/*; do
+	printf '%s: %s: Testing utility.\n' "$0" "$utility"
 	"$utility"
 	printf '\n'
 done

tests/posix/bin/cat

@@ -1,22 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-# Strictly POSIX-compliant cat(1) implementation. See cat(1p)
-
-for arg in "$@"; do
-	case "$arg" in
-		-u) args="$(printf '%s	%s\n' "$args" "$arg")" ;;
-		*) args="$(printf -- '%s	-i	%s\n' "$args" "$arg")" ;;
-	esac
-done
-
-# See IEEE Std 1003.1-2017 3.282
-# https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap03.html#tag_03_282
-IFS='	'
-
-mm $args

tests/posix/bin/false

@@ -1,12 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-# Strictly POSIX-compliant false(1) implementation. See false(1p)
-
-false "$@"

tests/posix/bin/true

@@ -1,11 +0,0 @@
-#!/bin/sh
-# Copyright (c) 2024 DTB <trinity@trinity.moe>
-# Copyright (c) 2024 Emma Tebibyte <emma@tebibyte.media>
-# SPDX-License-Identifier: FSFAP
-#
-# Copying and distribution of this file, with or without modification, are
-# permitted in any medium without royalty provided the copyright notice and this
-# notice are preserved.  This file is offered as-is, without any warranty.
-
-# Strictly POSIX-compliant true(1) implementation. See true(1p)
-true "$@"