static void do_subst_w_backrefs(const char *line, const char *replace)
{
int i,j;
/* go through the replacement string */
for (i = 0; replace[i]; i++) {
/* if we find a backreference (\1, \2, etc.) print the backref'ed * text */
if (replace[i] == '\\' && replace[i+1]>'0' && replace[i+1]<='9') {
int backref=replace[++i]-'0';
/* print out the text held in regmatch[backref] */
if(regmatch[backref].rm_so != -1)
for (j = regmatch[backref].rm_so; j < regmatch[backref].rm_eo; j++)
pipe_putc(line[j]);
}
/* if we find a backslash escaped character, print the character */
else if (replace[i] == '\\') pipe_putc(replace[++i]);
/* if we find an unescaped '&' print out the whole matched text. */
else if (replace[i] == '&')
for (j = regmatch[0].rm_so; j < regmatch[0].rm_eo; j++)
pipe_putc(line[j]);
/* Otherwise just output the character. */
else pipe_putc(replace[i]);
}
}
size_t pipe_write(struct robject *self, rp_t source, uint8_t *buffer, size_t size, off_t offset) {
struct pipe *pipe;
size_t i;
pipe = robject_data(self, "pipe");
for (i = 0; i < size; i++) {
pipe_putc(pipe, buffer[i]);
}
return size;
}
static void do_subst_w_backrefs(char *line, char *replace)
{
int i, j;
/* go through the replacement string */
for (i = 0; replace[i]; i++) {
/* if we find a backreference (\1, \2, etc.) print the backref'ed text */
if (replace[i] == '\\') {
unsigned backref = replace[++i] - '0';
if (backref <= 9) {
/* print out the text held in G.regmatch[backref] */
if (G.regmatch[backref].rm_so != -1) {
j = G.regmatch[backref].rm_so;
while (j < G.regmatch[backref].rm_eo)
pipe_putc(line[j++]);
}
continue;
}
/* I _think_ it is impossible to get '\' to be
* the last char in replace string. Thus we don't check
* for replace[i] == NUL. (counterexample anyone?) */
/* if we find a backslash escaped character, print the character */
pipe_putc(replace[i]);
continue;
}
/* if we find an unescaped '&' print out the whole matched text. */
if (replace[i] == '&') {
j = G.regmatch[0].rm_so;
while (j < G.regmatch[0].rm_eo)
pipe_putc(line[j++]);
continue;
}
/* Otherwise just output the character. */
pipe_putc(replace[i]);
}
}
static int do_subst_command(sed_cmd_t *sed_cmd, char **line_p)
{
char *line = *line_p;
unsigned match_count = 0;
bool altered = 0;
bool prev_match_empty = 1;
bool tried_at_eol = 0;
regex_t *current_regex;
current_regex = sed_cmd->sub_match;
/* Handle empty regex. */
if (!current_regex) {
current_regex = G.previous_regex_ptr;
if (!current_regex)
bb_error_msg_and_die("no previous regexp");
}
G.previous_regex_ptr = current_regex;
/* Find the first match */
dbg("matching '%s'", line);
if (REG_NOMATCH == regexec(current_regex, line, 10, G.regmatch, 0)) {
dbg("no match");
return 0;
}
dbg("match");
/* Initialize temporary output buffer. */
G.pipeline.buf = xmalloc(PIPE_GROW);
G.pipeline.len = PIPE_GROW;
G.pipeline.idx = 0;
/* Now loop through, substituting for matches */
do {
int start = G.regmatch[0].rm_so;
int end = G.regmatch[0].rm_eo;
int i;
match_count++;
/* If we aren't interested in this match, output old line to
* end of match and continue */
if (sed_cmd->which_match
&& (sed_cmd->which_match != match_count)
) {
for (i = 0; i < end; i++)
pipe_putc(*line++);
/* Null match? Print one more char */
if (start == end && *line)
pipe_putc(*line++);
goto next;
}
/* Print everything before the match */
for (i = 0; i < start; i++)
pipe_putc(line[i]);
/* Then print the substitution string,
* unless we just matched empty string after non-empty one.
* Example: string "cccd", pattern "c*", repl "R":
* result is "RdR", not "RRdR": first match "ccc",
* second is "" before "d", third is "" after "d".
* Second match is NOT replaced!
*/
if (prev_match_empty || start != 0 || start != end) {
//dbg("%d %d %d", prev_match_empty, start, end);
dbg("inserting replacement at %d in '%s'", start, line);
do_subst_w_backrefs(line, sed_cmd->string);
/* Flag that something has changed */
altered = 1;
} else {
dbg("NOT inserting replacement at %d in '%s'", start, line);
}
/* If matched string is empty (f.e. "c*" pattern),
* copy verbatim one char after it before attempting more matches
*/
prev_match_empty = (start == end);
if (prev_match_empty) {
if (!line[end]) {
tried_at_eol = 1;
} else {
pipe_putc(line[end]);
end++;
}
}
/* Advance past the match */
dbg("line += %d", end);
line += end;
/* if we're not doing this globally, get out now */
if (sed_cmd->which_match != 0)
break;
next:
/* Exit if we are at EOL and already tried matching at it */
if (*line == '\0') {
if (tried_at_eol)
break;
tried_at_eol = 1;
}
//maybe (end ? REG_NOTBOL : 0) instead of unconditional REG_NOTBOL?
} while (regexec(current_regex, line, 10, G.regmatch, REG_NOTBOL) != REG_NOMATCH);
/* Copy rest of string into output pipeline */
while (1) {
char c = *line++;
pipe_putc(c);
if (c == '\0')
break;
}
free(*line_p);
*line_p = G.pipeline.buf;
return altered;
}
static int do_subst_command(sed_cmd_t *sed_cmd, char **line)
{
char *oldline = *line;
int altered = 0;
unsigned match_count = 0;
regex_t *current_regex;
/* Handle empty regex. */
if (sed_cmd->sub_match == NULL) {
current_regex = G.previous_regex_ptr;
if (!current_regex)
bb_error_msg_and_die("no previous regexp");
} else
G.previous_regex_ptr = current_regex = sed_cmd->sub_match;
/* Find the first match */
if (REG_NOMATCH == regexec(current_regex, oldline, 10, G.regmatch, 0))
return 0;
/* Initialize temporary output buffer. */
G.pipeline.buf = xmalloc(PIPE_GROW);
G.pipeline.len = PIPE_GROW;
G.pipeline.idx = 0;
/* Now loop through, substituting for matches */
do {
int i;
/* Work around bug in glibc regexec, demonstrated by:
echo " a.b" | busybox sed 's [^ .]* x g'
The match_count check is so not to break
echo "hi" | busybox sed 's/^/!/g' */
if (!G.regmatch[0].rm_so && !G.regmatch[0].rm_eo && match_count) {
pipe_putc(*oldline++);
continue;
}
match_count++;
/* If we aren't interested in this match, output old line to
end of match and continue */
if (sed_cmd->which_match
&& (sed_cmd->which_match != match_count)
) {
for (i = 0; i < G.regmatch[0].rm_eo; i++)
pipe_putc(*oldline++);
continue;
}
/* print everything before the match */
for (i = 0; i < G.regmatch[0].rm_so; i++)
pipe_putc(oldline[i]);
/* then print the substitution string */
do_subst_w_backrefs(oldline, sed_cmd->string);
/* advance past the match */
oldline += G.regmatch[0].rm_eo;
/* flag that something has changed */
altered++;
/* if we're not doing this globally, get out now */
if (sed_cmd->which_match)
break;
} while (*oldline && (regexec(current_regex, oldline, 10, G.regmatch, 0) != REG_NOMATCH));
/* Copy rest of string into output pipeline */
while (*oldline)
pipe_putc(*oldline++);
pipe_putc(0);
free(*line);
*line = G.pipeline.buf;
return altered;
}
