freebsd-skq/usr.bin/make/str.c
Xin LI 512be259a7 Remove unnecessary includes.
Reviewed by:	rodrigc
2009-12-11 23:20:02 +00:00

560 lines
12 KiB
C

/*-
* Copyright (c) 1988, 1989, 1990, 1993
* The Regents of the University of California. All rights reserved.
* Copyright (c) 1989 by Berkeley Softworks
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Adam de Boor.
*
* 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.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* 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.
*
* @(#)str.c 5.8 (Berkeley) 6/1/90
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <stdlib.h>
#include <string.h>
#include "buf.h"
#include "str.h"
#include "util.h"
/**
* Initialize the argument array object. The array is initially
* eight positions, and will be expaned as neccessary. The first
* position is set to NULL since everything ignores it. We allocate
* (size + 1) since we need space for the terminating NULL. The
* buffer is set to NULL, since no common buffer is alloated yet.
*/
void
ArgArray_Init(ArgArray *aa)
{
aa->size = 8;
aa->argv = emalloc((aa->size + 1) * sizeof(char *));
aa->argc = 0;
aa->argv[aa->argc++] = NULL;
aa->len = 0;
aa->buffer = NULL;
}
/**
* Cleanup the memory allocated for in the argument array object.
*/
void
ArgArray_Done(ArgArray *aa)
{
if (aa->buffer == NULL) {
int i;
/* args are individually allocated */
for (i = 0; i < aa->argc; ++i) {
if (aa->argv[i]) {
free(aa->argv[i]);
aa->argv[i] = NULL;
}
}
} else {
/* args are part of a single allocation */
free(aa->buffer);
aa->buffer = NULL;
}
free(aa->argv);
aa->argv = NULL;
aa->argc = 0;
aa->size = 0;
}
/*-
* str_concat --
* concatenate the two strings, inserting a space or slash between them.
*
* returns --
* the resulting string in allocated space.
*/
char *
str_concat(const char *s1, const char *s2, int flags)
{
int len1, len2;
char *result;
/* get the length of both strings */
len1 = strlen(s1);
len2 = strlen(s2);
/* allocate length plus separator plus EOS */
result = emalloc(len1 + len2 + 2);
/* copy first string into place */
memcpy(result, s1, len1);
/* add separator character */
if (flags & STR_ADDSPACE) {
result[len1] = ' ';
++len1;
} else if (flags & STR_ADDSLASH) {
result[len1] = '/';
++len1;
}
/* copy second string plus EOS into place */
memcpy(result + len1, s2, len2 + 1);
return (result);
}
/**
* Fracture a string into an array of words (as delineated by tabs or
* spaces) taking quotation marks into account. Leading tabs/spaces
* are ignored.
*/
void
brk_string(ArgArray *aa, const char str[], Boolean expand)
{
char inquote;
char *start;
char *arg;
/* skip leading space chars. */
for (; *str == ' ' || *str == '\t'; ++str)
continue;
ArgArray_Init(aa);
aa->buffer = estrdup(str);;
arg = aa->buffer;
start = arg;
inquote = '\0';
/*
* copy the string; at the same time, parse backslashes,
* quotes and build the argument list.
*/
for (;;) {
switch (str[0]) {
case '"':
case '\'':
if (inquote == '\0') {
inquote = str[0];
if (expand)
break;
if (start == NULL)
start = arg;
} else if (inquote == str[0]) {
inquote = '\0';
/* Don't miss "" or '' */
if (start == NULL)
start = arg;
if (expand)
break;
} else {
/* other type of quote found */
if (start == NULL)
start = arg;
}
*arg++ = str[0];
break;
case ' ':
case '\t':
case '\n':
if (inquote) {
if (start == NULL)
start = arg;
*arg++ = str[0];
break;
}
if (start == NULL)
break;
/* FALLTHROUGH */
case '\0':
/*
* end of a token -- make sure there's enough argv
* space and save off a pointer.
*/
if (aa->argc == aa->size) {
aa->size *= 2; /* ramp up fast */
aa->argv = erealloc(aa->argv,
(aa->size + 1) * sizeof(char *));
}
*arg++ = '\0';
if (start == NULL) {
aa->argv[aa->argc] = start;
return;
}
if (str[0] == '\n' || str[0] == '\0') {
aa->argv[aa->argc++] = start;
aa->argv[aa->argc] = NULL;
return;
} else {
aa->argv[aa->argc++] = start;
start = NULL;
break;
}
case '\\':
if (start == NULL)
start = arg;
if (expand) {
switch (str[1]) {
case '\0':
case '\n':
/* hmmm; fix it up as best we can */
*arg++ = '\\';
break;
case 'b':
*arg++ = '\b';
++str;
break;
case 'f':
*arg++ = '\f';
++str;
break;
case 'n':
*arg++ = '\n';
++str;
break;
case 'r':
*arg++ = '\r';
++str;
break;
case 't':
*arg++ = '\t';
++str;
break;
default:
*arg++ = str[1];
++str;
break;
}
} else {
*arg++ = str[0];
if (str[1] != '\0') {
++str;
*arg++ = str[0];
}
}
break;
default:
if (start == NULL)
start = arg;
*arg++ = str[0];
break;
}
++str;
}
}
/*
* Quote a string for appending it to MAKEFLAGS. According to Posix the
* kind of quoting here is implementation-defined. This quoting must ensure
* that the parsing of MAKEFLAGS's contents in a sub-shell yields the same
* options, option arguments and macro definitions as in the calling make.
* We simply quote all blanks, which according to Posix are space and tab
* in the POSIX locale. Don't use isblank because in that case makes with
* different locale settings could not communicate. We must also quote
* backslashes obviously.
*/
char *
MAKEFLAGS_quote(const char *str)
{
char *ret, *q;
const char *p;
/* assume worst case - everything has to be quoted */
ret = emalloc(strlen(str) * 2 + 1);
p = str;
q = ret;
while (*p != '\0') {
switch (*p) {
case ' ':
case '\t':
*q++ = '\\';
break;
default:
break;
}
*q++ = *p++;
}
*q++ = '\0';
return (ret);
}
void
MAKEFLAGS_break(ArgArray *aa, const char str[])
{
char *arg;
char *start;
ArgArray_Init(aa);
aa->buffer = strdup(str);
arg = aa->buffer;
start = NULL;
for (;;) {
switch (str[0]) {
case ' ':
case '\t':
/* word separator */
if (start == NULL) {
/* not in a word */
str++;
continue;
}
/* FALLTHRU */
case '\0':
if (aa->argc == aa->size) {
aa->size *= 2;
aa->argv = erealloc(aa->argv,
(aa->size + 1) * sizeof(char *));
}
*arg++ = '\0';
if (start == NULL) {
aa->argv[aa->argc] = start;
return;
}
if (str[0] == '\0') {
aa->argv[aa->argc++] = start;
aa->argv[aa->argc] = NULL;
return;
} else {
aa->argv[aa->argc++] = start;
start = NULL;
str++;
continue;
}
case '\\':
if (str[1] == ' ' || str[1] == '\t')
str++;
break;
default:
break;
}
if (start == NULL)
start = arg;
*arg++ = *str++;
}
}
/*
* Str_Match --
*
* See if a particular string matches a particular pattern.
*
* Results: Non-zero is returned if string matches pattern, 0 otherwise. The
* matching operation permits the following special characters in the
* pattern: *?\[] (see the man page for details on what these mean).
*
* Side effects: None.
*/
int
Str_Match(const char *string, const char *pattern)
{
char c2;
for (;;) {
/*
* See if we're at the end of both the pattern and the
* string. If, we succeeded. If we're at the end of the
* pattern but not at the end of the string, we failed.
*/
if (*pattern == 0)
return (!*string);
if (*string == 0 && *pattern != '*')
return (0);
/*
* Check for a "*" as the next pattern character. It matches
* any substring. We handle this by calling ourselves
* recursively for each postfix of string, until either we
* match or we reach the end of the string.
*/
if (*pattern == '*') {
pattern += 1;
if (*pattern == 0)
return (1);
while (*string != 0) {
if (Str_Match(string, pattern))
return (1);
++string;
}
return (0);
}
/*
* Check for a "?" as the next pattern character. It matches
* any single character.
*/
if (*pattern == '?')
goto thisCharOK;
/*
* Check for a "[" as the next pattern character. It is
* followed by a list of characters that are acceptable, or
* by a range (two characters separated by "-").
*/
if (*pattern == '[') {
++pattern;
for (;;) {
if ((*pattern == ']') || (*pattern == 0))
return (0);
if (*pattern == *string)
break;
if (pattern[1] == '-') {
c2 = pattern[2];
if (c2 == 0)
return (0);
if ((*pattern <= *string) &&
(c2 >= *string))
break;
if ((*pattern >= *string) &&
(c2 <= *string))
break;
pattern += 2;
}
++pattern;
}
while ((*pattern != ']') && (*pattern != 0))
++pattern;
goto thisCharOK;
}
/*
* If the next pattern character is '/', just strip off the
* '/' so we do exact matching on the character that follows.
*/
if (*pattern == '\\') {
++pattern;
if (*pattern == 0)
return (0);
}
/*
* There's no special character. Just make sure that the
* next characters of each string match.
*/
if (*pattern != *string)
return (0);
thisCharOK: ++pattern;
++string;
}
}
/**
* Str_SYSVMatch
* Check word against pattern for a match (% is wild),
*
* Results:
* Returns the beginning position of a match or null. The number
* of characters matched is returned in len.
*/
const char *
Str_SYSVMatch(const char *word, const char *pattern, int *len)
{
const char *m, *p, *w;
p = pattern;
w = word;
if (*w == '\0') {
/* Zero-length word cannot be matched against */
*len = 0;
return (NULL);
}
if (*p == '\0') {
/* Null pattern is the whole string */
*len = strlen(w);
return (w);
}
if ((m = strchr(p, '%')) != NULL) {
/* check that the prefix matches */
for (; p != m && *w && *w == *p; w++, p++)
continue;
if (p != m)
return (NULL); /* No match */
if (*++p == '\0') {
/* No more pattern, return the rest of the string */
*len = strlen(w);
return (w);
}
}
m = w;
/* Find a matching tail */
do
if (strcmp(p, w) == 0) {
*len = w - m;
return (m);
}
while (*w++ != '\0');
return (NULL);
}
/**
* Str_SYSVSubst
* Substitute '%' on the pattern with len characters from src.
* If the pattern does not contain a '%' prepend len characters
* from src.
*
* Side Effects:
* Places result on buf
*/
void
Str_SYSVSubst(Buffer *buf, const char *pat, const char *src, int len)
{
const char *m;
if ((m = strchr(pat, '%')) != NULL) {
/* Copy the prefix */
Buf_AppendRange(buf, pat, m);
/* skip the % */
pat = m + 1;
}
/* Copy the pattern */
Buf_AddBytes(buf, len, (const Byte *)src);
/* append the rest */
Buf_Append(buf, pat);
}