1
0

simplify and comment

This commit is contained in:
dtb 2024-01-17 11:38:36 -07:00
parent 8ae20904b6
commit 0624c3561b

271
mm/mm.c
View File

@ -1,3 +1,21 @@
/*
* 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),
@ -5,11 +23,29 @@
#include <stdlib.h> /* free(3), realloc(3) */
#include <string.h> /* strerror(3) */
#include <unistd.h> /* getopt(3) */
#if !defined EX_OK || !defined EX_USAGE
#if !defined EX_OSERR || !defined EX_USAGE
# include <sysexits.h>
#endif
extern int errno;
/* mm(1) works similarly to dj(1) in that, fundamentally, they both take bytes
* from input and write them to output, but whereas dj(1)'s utility is in
* controlling the reads and writes themselves across an individual input file
* and individual output file - with regards to quantity of bytes read/written,
* buffering of reads and writes, and where in the input to read and where in
* the output to write - mm(1)'s utility is in handling arbitrary quantities of
* input files and writing to arbitrary quantities of output files, without
* the fine-grained control dj(1) offers. There are design decisions in mm(1)
* that are repeated from dj(1) (the same author wrote each utility's initial
* implementation) but, while dj(1)'s design is modeled around juggling text
* within buffers, mm(1)'s design is modeled around juggling open files. */
/* mm(1) can take arbitrary files but for the sake of resiliency it has a
* 10-file (for input and output independently) pre-allocated array and
* additional dynamically allocated lists of FILE pointers for when those
* arrays are full, with Io_ functions handling necessary operations on both
* the Io struct and its nested Io_ex struct. */
struct Io_ex{
size_t s;
FILE **files;
@ -24,30 +60,26 @@ struct Io{
struct Io_ex *ex;
};
/* 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 *amode = "ab";
static char *rmode = "rb";
static char *wmode = "rb+";
static char *(fmode[]) = { (char []){"rb"}, (char []){"rb+"} };
static char *wharsh = "wb";
/* Opens the file at the path s and puts it in the io struct, returning NULL if
* either the opening or the placement of the open FILE pointer fail. */
#define Io_fopen(io, s) Io_fappend((io), fopen((s), (io)->fmode), (s))
/* Destructs the io[2] used by main. */
#define terminate(io) do{ \
Io_destruct(&(io)[0]); \
Io_destruct(&(io)[1]); }while(0)
static struct Io *
Io_construct(struct Io *io){
io->s = 0;
io->ex->s = 0;
io->ex->files = NULL;
io->ex->names = NULL;
return io;
}
static struct Io *Io_initialize(struct Io *io, struct Io_ex *io_ex);
/* Destructs an io struct by closing its handled files and freeing its extended
* file and name arrays. Returns the destructed io. */
static struct Io *
Io_destruct(struct Io *io){
size_t i;
@ -55,36 +87,46 @@ Io_destruct(struct Io *io){
if(io->s == -1)
io->s = 0;
for(i = 0; i < io->s; ++i)
fclose(io->files[i]);
for(i = 0; i < io->ex->s; ++i)
fclose(io->ex->files[i]);
for(i = 0; i < io->s; fclose(io->files[i++]));
for(i = 0; i < io->ex->s; fclose(io->ex->files[i++]));
free(io->ex->files);
free(io->ex->names);
return Io_construct(io); /* counter-intuitive but NULLs and 0s */
return Io_initialize(io, io->ex);
}
/* Adds the open FILE pointer for the file at the path s to the io struct,
* returning the FILE if successful and NULL if not. */
static FILE *
Io_fappend(struct Io *io, FILE *f, char *s){
if(f == NULL)
return NULL;
if(io->s == -1)
io->s = 0;
if(io->s < (sizeof (io->files)) / (sizeof *(io->files))){
io->names[io->s] = s;
return io->files[io->s++] = f;
}else{
io->ex->files = realloc(io->ex->files,
(sizeof *(io->ex->files)) * ++io->ex->s);
io->ex->names = realloc(io->ex->names,
(sizeof *(io->ex->names)) * io->ex->s);
if(io->s >= (sizeof io->files) / (sizeof *io->files)){
if((io->ex->files = realloc(io->ex->files,
(sizeof *(io->ex->files))
* ++io->ex->s))
== NULL
|| (io->ex->names = realloc(io->ex->names,
(sizeof *(io->ex->names))
* io->ex->s))
== NULL)
return NULL;
io->ex->names[io->ex->s] = s;
return io->ex->files[io->ex->s++] = f;
}
io->names[io->s] = s;
return io->files[io->s++] = f;
}
/* Returns the first index within io->files of the given FILE pointer f. If f
* isn't within io->files but is within io->ex->files, returns the sum of io->s
* and the index of the FILE pointer in io->ex->files. If f is within neither
* io->files or io->ex->files, returns -1. */
static int
Io_fileindex(struct Io *io, FILE *f){
size_t i;
@ -100,6 +142,8 @@ Io_fileindex(struct Io *io, FILE *f){
return -1;
}
/* Returns the corresponding file name in the io struct to the FILE pointer f,
* or NULL if f isn't in io->files or io->ex->files. */
static char *
Io_filename(struct Io *io, FILE *f){
int i;
@ -112,19 +156,11 @@ Io_filename(struct Io *io, FILE *f){
return io->ex->names[i - io->s];
}
static FILE *
Io_fopen(struct Io *io, char *s){
FILE *f;
if((f = fopen(s, io->fmode)) == NULL)
return NULL;
else
return Io_fappend(io, f, s);
}
static struct Io *Io_fremove(struct Io *io, FILE *f);
static FILE *Io_nextfile(struct Io *io, FILE *f);
/* Prints c to all io->files, returning c. If a stream can't be written, a
* message is printed to standard error and the stream is closed and removed
* from io. If no stream can be written, returns EOF. */
static int
Io_fputc(struct Io *io, int c){
FILE *f;
@ -133,25 +169,24 @@ Io_fputc(struct Io *io, int c){
f = NULL;
lf = NULL;
while((f = Io_nextfile(io, f)) != NULL){
for(f = NULL; (f = Io_nextfile(io, f)) != NULL;){
if(lf != NULL){
if(fclose(lf) == EOF)
fprintf(stderr, "%s: %s: %s\n",
program_name, ln, strerror(errno));
Io_fremove(io, lf);
if(io->s == 0)
if(Io_fremove(io, lf)->s == 0)
return EOF;
lf = NULL;
}
if(putc(c, f) == EOF){
lf = f;
ln = Io_filename(io, lf);
}
if(putc(c, f) == EOF)
ln = Io_filename(io, lf = f);
}
return c;
}
/* Removes a given file pointer from the given io struct, returning NULL if the
* file was already absent from io and otherwise io. */
static struct Io *
Io_fremove(struct Io *io, FILE *f){
size_t i;
@ -185,46 +220,49 @@ Io_fremove(struct Io *io, FILE *f){
return io;
}
static FILE *
Io_nextfile(struct Io *io, FILE *f){
size_t i;
/* Initializes the members of the given io and io_ex structs (including setting
* io->ex to io_ex) and returns io. */
static struct Io *
Io_initialize(struct Io *io, struct Io_ex *io_ex){
if(f == NULL)
return io->files[0];
io->s = 0;
io->ex = io_ex;
io->ex->s = 0;
io->ex->files = NULL;
io->ex->names = NULL;
for(i = 0; i < io->s; ++i)
if(io->files[i] == f){
if(i == io->s - 1 && io->ex->s == 0)
return NULL;
else if(i == io->s - 1)
return io->ex->files[0];
else
return io->files[i + 1];
}
for(i = 0; i < io->ex->s; ++i)
if(io->ex->files[i] == f){
if(i == io->ex->s - 1)
break;
else
return io->ex->files[i + 1];
}
return NULL;
return io;
}
/* Returns the first file in io if f is NULL, NULL if the file is the last in
* or absent from io, or the file after f in io. */
static FILE *
Io_nextfile(struct Io *io, FILE *f){
int i;
return (f == NULL)
? io->files[0]
: ((i = Io_fileindex(io, f)) == -1
|| i == io->s + io->ex->s - 1)
? NULL
: (i < io->s - 1)
? io->files[i + 1]
: io->ex->files[i - io->s + 1];
}
/* Stops stdio from buffering io. Returns io. */
static struct Io *
Io_unbuffer(struct Io *io){
size_t i;
for(i = 0; i < io->s; ++i)
setvbuf(io->files[i], NULL, _IONBF, 0);
for(i = 0; i < io->ex->s; ++i)
setvbuf(io->ex->files[i], NULL, _IONBF, 0);
for(i = 0; i < io->s; setvbuf(io->files[i++], NULL, _IONBF, 0));
for(i = 0; i < io->ex->s; setvbuf(io->ex->files[i++], NULL,_IONBF, 0));
return io;
}
/* Prints a diagnostic message based on errno and returns an exit status
* appropriate for an OS error. */
static int
oserr(char *s, char *r){
@ -233,6 +271,8 @@ oserr(char *s, char *r){
return EX_OSERR;
}
/* Prints a diagnostic message synopsizing usage and returns an exit status
* appropriate for a usage error. */
static int
usage(char *s){
@ -245,47 +285,59 @@ usage(char *s){
int main(int argc, char *argv[]){
int c;
size_t i;
struct Io io[2]; /* {read, write, error} */
struct Io io[2]; /* {read, write} */
struct Io_ex io_ex[2];
int o;
char unbuffered;
int s; /* scratch variable */
/* The simple invocation (without given arguments) requires no memory
* allocations or use of the io structs and is therefore treated as a
* special case. To do the same with all bells and whistles call
* $ mm -i - -o - */
if(argc < 2){ /* simple invocation */
while((c = getc(stdin)) != EOF)
if(putc(c, stdout) == EOF)
break;
return EX_OK;
return 1;
return 0;
}
io[0].ex = &io_ex[0];
io[1].ex = &io_ex[1];
Io_construct(&io[0]);
Io_construct(&io[1]);
io[0].fmode = rmode;
io[1].fmode = wmode;
/* Initializes the io structs with their default values, standard input
* and standard output. Both io->s are set to -1, a special value
* indicating that they're at their defaults and that their FILE
* pointers won't need to be closed. 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){
Io_initialize(&io[i], &io_ex[i])->fmode = fmode[i];
Io_fappend(&io[i], i == 0 ? stdin : stdout,
i == 0 ? stdin_name : stdout_name);
io[i].s = -1;
}
Io_fappend(&io[0], stdin, stdin_name);
Io_fappend(&io[1], stdout, stdout_name);
io[0].s = -1;
io[1].s = -1;
unbuffered = 0;
s = 0; /* Refers to whether or not io will be unbuffered. */
while((c = getopt(argc, argv, "aehi:no:u")) != -1)
switch(c){
case 'a':
io[1].fmode = amode;
case 'a': /* "rb+" -> "ab" */
io[1].fmode[0] = 'a';
io[1].fmode[2] = '\0';
break;
case 'e':
Io_fappend(&io[1], stderr, stderr_name);
break;
if(Io_fappend(&io[1], stderr, stderr_name) != NULL)
break;
terminate(io);
return oserr(argv[0], "-e");
case 'i': case 'o':
if(optarg[0] == '-' && optarg[1] == '\0'){
/* "-" */
Io_fappend(&io[c == 'o'],
c == 'i' ? stdin : stdout,
c == 'i' ? stdin_name : stdout_name);
break;
if(Io_fappend(&io[c == 'o'],
c == 'i' ? stdin : stdout,
c == 'i'
? stdin_name
: stdout_name) != NULL)
break;
terminate(io);
return oserr(argv[0], optarg);
}else if(Io_fopen(&io[c == 'o'], optarg)
!= NULL)
break;
@ -293,7 +345,7 @@ int main(int argc, char *argv[]){
if(c == 'o' && errno == ENOENT){
io[1].fmode = wharsh;
if(Io_fopen(&io[1], optarg) != NULL){
io[1].fmode = wmode;
io[1].fmode = fmode[1];
break;
}
}
@ -305,14 +357,14 @@ int main(int argc, char *argv[]){
terminate(io);
return oserr(argv[0], "-n");
case 'u':
unbuffered = 1;
s = 1;
break;
default:
terminate(io);
return usage(argv[0]);
}
if(unbuffered){
if(s){
Io_unbuffer(&io[0]);
Io_unbuffer(&io[1]);
}
@ -321,7 +373,10 @@ int main(int argc, char *argv[]){
io[0].s = 1;
if(io[1].s == -1)
io[1].s = 1;
o = io[1].s;
/* Refers to the amount of outputs prior to writing; Io_fputc will
* remove any to which it can't write. */
s = io[1].s;
for(i = 0; i < io[0].s; ++i){
while((c = getc(io[0].files[i])) != EOF)
@ -331,8 +386,12 @@ int main(int argc, char *argv[]){
}
}
o = io[1].s != o;
/* Reflects whether (s==0) or not (s==1) the full length of the inputs
* could be written to all outputs, which is the return value for main
* to maintain feature parity with tee(1p). */
s = io[1].s != s;
terminate(io);
return o;
return s;
}