4b087712af
* Provide some GNUisms as extensions. * Provide dynamically growable string space. * Make define(defn(foo)) work correctly for builtins. (The current version is supposed to be capable of satisfying autoconf.) All still relevant NetBSD changes have been preserved in this version, and formatting and style fixes have been applied in various places. Thanks to Masao Uebayashi <uebayasi@soum.co.jp> for pointing this out.
564 lines
11 KiB
C
564 lines
11 KiB
C
/* $NetBSD: gnum4.c,v 1.1 2001/11/14 06:16:08 tv Exp $ */
|
|
/* $OpenBSD: gnum4.c,v 1.15 2001/10/13 20:18:48 espie Exp $ */
|
|
|
|
/*
|
|
* Copyright (c) 1999 Marc Espie
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions
|
|
* are met:
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
* documentation and/or other materials provided with the distribution.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
* SUCH DAMAGE.
|
|
*/
|
|
|
|
/*
|
|
* functions needed to support gnu-m4 extensions, including a fake freezing
|
|
*/
|
|
|
|
#include <sys/param.h>
|
|
#include <sys/types.h>
|
|
#include <sys/wait.h>
|
|
#include <ctype.h>
|
|
#include <err.h>
|
|
#include <errno.h>
|
|
#include <paths.h>
|
|
#include <regex.h>
|
|
#include <stddef.h>
|
|
#include <stdlib.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include "mdef.h"
|
|
#include "stdd.h"
|
|
#include "extern.h"
|
|
|
|
|
|
int mimic_gnu = 0;
|
|
|
|
/*
|
|
* Support for include path search
|
|
* First search in the the current directory.
|
|
* If not found, and the path is not absolute, include path kicks in.
|
|
* First, -I options, in the order found on the command line.
|
|
* Then M4PATH env variable
|
|
*/
|
|
|
|
struct path_entry {
|
|
char *name;
|
|
struct path_entry *next;
|
|
} *first, *last;
|
|
|
|
static struct path_entry *new_path_entry __P((const char *));
|
|
static void ensure_m4path __P((void));
|
|
static struct input_file *dopath __P((struct input_file *, const char *));
|
|
|
|
static struct path_entry *
|
|
new_path_entry(dirname)
|
|
const char *dirname;
|
|
{
|
|
struct path_entry *n;
|
|
|
|
n = malloc(sizeof(struct path_entry));
|
|
if (!n)
|
|
errx(1, "out of memory");
|
|
n->name = strdup(dirname);
|
|
if (!n->name)
|
|
errx(1, "out of memory");
|
|
n->next = 0;
|
|
return n;
|
|
}
|
|
|
|
void
|
|
addtoincludepath(dirname)
|
|
const char *dirname;
|
|
{
|
|
struct path_entry *n;
|
|
|
|
n = new_path_entry(dirname);
|
|
|
|
if (last) {
|
|
last->next = n;
|
|
last = n;
|
|
}
|
|
else
|
|
last = first = n;
|
|
}
|
|
|
|
static void
|
|
ensure_m4path()
|
|
{
|
|
static int envpathdone = 0;
|
|
char *envpath;
|
|
char *sweep;
|
|
char *path;
|
|
|
|
if (envpathdone)
|
|
return;
|
|
envpathdone = TRUE;
|
|
envpath = getenv("M4PATH");
|
|
if (!envpath)
|
|
return;
|
|
/* for portability: getenv result is read-only */
|
|
envpath = strdup(envpath);
|
|
if (!envpath)
|
|
errx(1, "out of memory");
|
|
for (sweep = envpath;
|
|
(path = strsep(&sweep, ":")) != NULL;)
|
|
addtoincludepath(path);
|
|
free(envpath);
|
|
}
|
|
|
|
static
|
|
struct input_file *
|
|
dopath(i, filename)
|
|
struct input_file *i;
|
|
const char *filename;
|
|
{
|
|
char path[MAXPATHLEN];
|
|
struct path_entry *pe;
|
|
FILE *f;
|
|
|
|
for (pe = first; pe; pe = pe->next) {
|
|
snprintf(path, sizeof(path), "%s/%s", pe->name, filename);
|
|
if ((f = fopen(path, "r")) != 0) {
|
|
set_input(i, f, path);
|
|
return i;
|
|
}
|
|
}
|
|
return NULL;
|
|
}
|
|
|
|
struct input_file *
|
|
fopen_trypath(i, filename)
|
|
struct input_file *i;
|
|
const char *filename;
|
|
{
|
|
FILE *f;
|
|
|
|
f = fopen(filename, "r");
|
|
if (f != NULL) {
|
|
set_input(i, f, filename);
|
|
return i;
|
|
}
|
|
if (filename[0] == '/')
|
|
return NULL;
|
|
|
|
ensure_m4path();
|
|
|
|
return dopath(i, filename);
|
|
}
|
|
|
|
void
|
|
doindir(argv, argc)
|
|
const char *argv[];
|
|
int argc;
|
|
{
|
|
ndptr p;
|
|
|
|
p = lookup(argv[2]);
|
|
if (p == NULL)
|
|
errx(1, "undefined macro %s", argv[2]);
|
|
argv[1] = p->defn;
|
|
eval(argv+1, argc-1, p->type);
|
|
}
|
|
|
|
void
|
|
dobuiltin(argv, argc)
|
|
const char *argv[];
|
|
int argc;
|
|
{
|
|
int n;
|
|
argv[1] = NULL;
|
|
n = builtin_type(argv[2]);
|
|
if (n != -1)
|
|
eval(argv+1, argc-1, n);
|
|
else
|
|
errx(1, "unknown builtin %s", argv[2]);
|
|
}
|
|
|
|
|
|
/* We need some temporary buffer space, as pb pushes BACK and substitution
|
|
* proceeds forward... */
|
|
static char *buffer;
|
|
static size_t bufsize = 0;
|
|
static size_t current = 0;
|
|
|
|
static void addchars __P((const char *, size_t));
|
|
static void addchar __P((char));
|
|
static char *twiddle __P((const char *));
|
|
static char *getstring __P((void));
|
|
static void exit_regerror __P((int, regex_t *));
|
|
static void do_subst __P((const char *, regex_t *, const char *, regmatch_t *));
|
|
static void do_regexpindex __P((const char *, regex_t *, regmatch_t *));
|
|
static void do_regexp __P((const char *, regex_t *, const char *, regmatch_t *));
|
|
static void add_sub __P((int, const char *, regex_t *, regmatch_t *));
|
|
static void add_replace __P((const char *, regex_t *, const char *, regmatch_t *));
|
|
#define addconstantstring(s) addchars((s), sizeof(s)-1)
|
|
|
|
static void
|
|
addchars(c, n)
|
|
const char *c;
|
|
size_t n;
|
|
{
|
|
if (n == 0)
|
|
return;
|
|
while (current + n > bufsize) {
|
|
if (bufsize == 0)
|
|
bufsize = 1024;
|
|
else
|
|
bufsize *= 2;
|
|
buffer = realloc(buffer, bufsize);
|
|
if (buffer == NULL)
|
|
errx(1, "out of memory");
|
|
}
|
|
memcpy(buffer+current, c, n);
|
|
current += n;
|
|
}
|
|
|
|
static void
|
|
addchar(c)
|
|
char c;
|
|
{
|
|
if (current +1 > bufsize) {
|
|
if (bufsize == 0)
|
|
bufsize = 1024;
|
|
else
|
|
bufsize *= 2;
|
|
buffer = realloc(buffer, bufsize);
|
|
if (buffer == NULL)
|
|
errx(1, "out of memory");
|
|
}
|
|
buffer[current++] = c;
|
|
}
|
|
|
|
static char *
|
|
getstring()
|
|
{
|
|
addchar('\0');
|
|
current = 0;
|
|
return buffer;
|
|
}
|
|
|
|
|
|
static void
|
|
exit_regerror(er, re)
|
|
int er;
|
|
regex_t *re;
|
|
{
|
|
size_t errlen;
|
|
char *errbuf;
|
|
|
|
errlen = regerror(er, re, NULL, 0);
|
|
errbuf = xalloc(errlen);
|
|
regerror(er, re, errbuf, errlen);
|
|
errx(1, "regular expression error: %s", errbuf);
|
|
}
|
|
|
|
static void
|
|
add_sub(n, string, re, pm)
|
|
int n;
|
|
const char *string;
|
|
regex_t *re;
|
|
regmatch_t *pm;
|
|
{
|
|
if (n > re->re_nsub)
|
|
warnx("No subexpression %d", n);
|
|
/* Subexpressions that did not match are
|
|
* not an error. */
|
|
else if (pm[n].rm_so != -1 &&
|
|
pm[n].rm_eo != -1) {
|
|
addchars(string + pm[n].rm_so,
|
|
pm[n].rm_eo - pm[n].rm_so);
|
|
}
|
|
}
|
|
|
|
/* Add replacement string to the output buffer, recognizing special
|
|
* constructs and replacing them with substrings of the original string.
|
|
*/
|
|
static void
|
|
add_replace(string, re, replace, pm)
|
|
const char *string;
|
|
regex_t *re;
|
|
const char *replace;
|
|
regmatch_t *pm;
|
|
{
|
|
const char *p;
|
|
|
|
for (p = replace; *p != '\0'; p++) {
|
|
if (*p == '&' && !mimic_gnu) {
|
|
add_sub(0, string, re, pm);
|
|
continue;
|
|
}
|
|
if (*p == '\\') {
|
|
if (p[1] == '\\') {
|
|
addchar(p[1]);
|
|
p++;
|
|
continue;
|
|
}
|
|
if (p[1] == '&') {
|
|
if (mimic_gnu)
|
|
add_sub(0, string, re, pm);
|
|
else
|
|
addchar(p[1]);
|
|
p++;
|
|
continue;
|
|
}
|
|
if (isdigit(p[1])) {
|
|
add_sub(*(++p) - '0', string, re, pm);
|
|
continue;
|
|
}
|
|
}
|
|
addchar(*p);
|
|
}
|
|
}
|
|
|
|
static void
|
|
do_subst(string, re, replace, pm)
|
|
const char *string;
|
|
regex_t *re;
|
|
const char *replace;
|
|
regmatch_t *pm;
|
|
{
|
|
int error;
|
|
int flags = 0;
|
|
const char *last_match = NULL;
|
|
|
|
while ((error = regexec(re, string, re->re_nsub+1, pm, flags)) == 0) {
|
|
if (pm[0].rm_eo != 0) {
|
|
if (string[pm[0].rm_eo-1] == '\n')
|
|
flags = 0;
|
|
else
|
|
flags = REG_NOTBOL;
|
|
}
|
|
|
|
/* NULL length matches are special... We use the `vi-mode'
|
|
* rule: don't allow a NULL-match at the last match
|
|
* position.
|
|
*/
|
|
if (pm[0].rm_so == pm[0].rm_eo &&
|
|
string + pm[0].rm_so == last_match) {
|
|
if (*string == '\0')
|
|
return;
|
|
addchar(*string);
|
|
if (*string++ == '\n')
|
|
flags = 0;
|
|
else
|
|
flags = REG_NOTBOL;
|
|
continue;
|
|
}
|
|
last_match = string + pm[0].rm_so;
|
|
addchars(string, pm[0].rm_so);
|
|
add_replace(string, re, replace, pm);
|
|
string += pm[0].rm_eo;
|
|
}
|
|
if (error != REG_NOMATCH)
|
|
exit_regerror(error, re);
|
|
pbstr(string);
|
|
}
|
|
|
|
static void
|
|
do_regexp(string, re, replace, pm)
|
|
const char *string;
|
|
regex_t *re;
|
|
const char *replace;
|
|
regmatch_t *pm;
|
|
{
|
|
int error;
|
|
|
|
switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
|
|
case 0:
|
|
add_replace(string, re, replace, pm);
|
|
pbstr(getstring());
|
|
break;
|
|
case REG_NOMATCH:
|
|
break;
|
|
default:
|
|
exit_regerror(error, re);
|
|
}
|
|
}
|
|
|
|
static void
|
|
do_regexpindex(string, re, pm)
|
|
const char *string;
|
|
regex_t *re;
|
|
regmatch_t *pm;
|
|
{
|
|
int error;
|
|
|
|
switch(error = regexec(re, string, re->re_nsub+1, pm, 0)) {
|
|
case 0:
|
|
pbunsigned(pm[0].rm_so);
|
|
break;
|
|
case REG_NOMATCH:
|
|
pbnum(-1);
|
|
break;
|
|
default:
|
|
exit_regerror(error, re);
|
|
}
|
|
}
|
|
|
|
/* In Gnu m4 mode, parentheses for backmatch don't work like POSIX 1003.2
|
|
* says. So we twiddle with the regexp before passing it to regcomp.
|
|
*/
|
|
static char *
|
|
twiddle(p)
|
|
const char *p;
|
|
{
|
|
/* This could use strcspn for speed... */
|
|
while (*p != '\0') {
|
|
if (*p == '\\') {
|
|
switch(p[1]) {
|
|
case '(':
|
|
case ')':
|
|
case '|':
|
|
addchar(p[1]);
|
|
break;
|
|
case 'w':
|
|
addconstantstring("[_a-zA-Z0-9]");
|
|
break;
|
|
case 'W':
|
|
addconstantstring("[^_a-zA-Z0-9]");
|
|
break;
|
|
case '<':
|
|
addconstantstring("[[:<:]]");
|
|
break;
|
|
case '>':
|
|
addconstantstring("[[:>:]]");
|
|
break;
|
|
default:
|
|
addchars(p, 2);
|
|
break;
|
|
}
|
|
p+=2;
|
|
continue;
|
|
}
|
|
if (*p == '(' || *p == ')' || *p == '|')
|
|
addchar('\\');
|
|
|
|
addchar(*p);
|
|
p++;
|
|
}
|
|
return getstring();
|
|
}
|
|
|
|
/* patsubst(string, regexp, opt replacement) */
|
|
/* argv[2]: string
|
|
* argv[3]: regexp
|
|
* argv[4]: opt rep
|
|
*/
|
|
void
|
|
dopatsubst(argv, argc)
|
|
const char *argv[];
|
|
int argc;
|
|
{
|
|
int error;
|
|
regex_t re;
|
|
regmatch_t *pmatch;
|
|
|
|
if (argc <= 3) {
|
|
warnx("Too few arguments to patsubst");
|
|
return;
|
|
}
|
|
error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
|
|
REG_NEWLINE | REG_EXTENDED);
|
|
if (error != 0)
|
|
exit_regerror(error, &re);
|
|
|
|
pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
|
|
do_subst(argv[2], &re,
|
|
argc != 4 && argv[4] != NULL ? argv[4] : "", pmatch);
|
|
pbstr(getstring());
|
|
free(pmatch);
|
|
regfree(&re);
|
|
}
|
|
|
|
void
|
|
doregexp(argv, argc)
|
|
const char *argv[];
|
|
int argc;
|
|
{
|
|
int error;
|
|
regex_t re;
|
|
regmatch_t *pmatch;
|
|
|
|
if (argc <= 3) {
|
|
warnx("Too few arguments to regexp");
|
|
return;
|
|
}
|
|
error = regcomp(&re, mimic_gnu ? twiddle(argv[3]) : argv[3],
|
|
REG_EXTENDED);
|
|
if (error != 0)
|
|
exit_regerror(error, &re);
|
|
|
|
pmatch = xalloc(sizeof(regmatch_t) * (re.re_nsub+1));
|
|
if (argv[4] == NULL || argc == 4)
|
|
do_regexpindex(argv[2], &re, pmatch);
|
|
else
|
|
do_regexp(argv[2], &re, argv[4], pmatch);
|
|
free(pmatch);
|
|
regfree(&re);
|
|
}
|
|
|
|
void
|
|
doesyscmd(cmd)
|
|
const char *cmd;
|
|
{
|
|
int p[2];
|
|
pid_t pid, cpid;
|
|
char *argv[4];
|
|
int cc;
|
|
int status;
|
|
|
|
/* Follow gnu m4 documentation: first flush buffers. */
|
|
fflush(NULL);
|
|
|
|
argv[0] = "sh";
|
|
argv[1] = "-c";
|
|
argv[2] = (char *)cmd;
|
|
argv[3] = NULL;
|
|
|
|
/* Just set up standard output, share stderr and stdin with m4 */
|
|
if (pipe(p) == -1)
|
|
err(1, "bad pipe");
|
|
switch(cpid = fork()) {
|
|
case -1:
|
|
err(1, "bad fork");
|
|
/* NOTREACHED */
|
|
case 0:
|
|
(void) close(p[0]);
|
|
(void) dup2(p[1], 1);
|
|
(void) close(p[1]);
|
|
execv(_PATH_BSHELL, argv);
|
|
exit(1);
|
|
default:
|
|
/* Read result in two stages, since m4's buffer is
|
|
* pushback-only. */
|
|
(void) close(p[1]);
|
|
do {
|
|
char result[BUFSIZE];
|
|
cc = read(p[0], result, sizeof result);
|
|
if (cc > 0)
|
|
addchars(result, cc);
|
|
} while (cc > 0 || (cc == -1 && errno == EINTR));
|
|
|
|
(void) close(p[0]);
|
|
while ((pid = wait(&status)) != cpid && pid >= 0)
|
|
continue;
|
|
pbstr(getstring());
|
|
}
|
|
}
|