2015-09-13 03:59:40 +08:00
|
|
|
/** \file builtin_string.cpp
|
|
|
|
Implementation of the string builtin.
|
|
|
|
*/
|
|
|
|
|
2015-09-13 04:36:22 +08:00
|
|
|
#include "config.h" // IWYU pragma: keep
|
|
|
|
|
2015-09-13 03:59:40 +08:00
|
|
|
#define PCRE2_CODE_UNIT_WIDTH WCHAR_T_BITS
|
|
|
|
#ifdef _WIN32
|
|
|
|
#define PCRE2_STATIC
|
|
|
|
#endif
|
|
|
|
#include "pcre2.h"
|
|
|
|
|
2015-09-13 04:36:22 +08:00
|
|
|
#include "builtin.h"
|
|
|
|
#include "common.h"
|
|
|
|
#include "parser.h"
|
|
|
|
#include "parse_util.h"
|
|
|
|
#include "wgetopt.h"
|
2015-09-13 03:59:40 +08:00
|
|
|
#include "wildcard.h"
|
2015-09-13 04:36:22 +08:00
|
|
|
#include "wutil.h"
|
2015-09-13 16:26:50 +08:00
|
|
|
#include <iterator>
|
|
|
|
#include <algorithm>
|
2015-09-13 04:36:22 +08:00
|
|
|
#include <unistd.h>
|
2015-09-13 03:59:40 +08:00
|
|
|
|
|
|
|
#define MAX_REPLACE_SIZE size_t(1048576) // pcre2_substitute maximum output size in wchar_t
|
|
|
|
#define STRING_ERR_MISSING _(L"%ls: Expected argument\n")
|
|
|
|
|
2015-09-13 04:36:22 +08:00
|
|
|
/* externs from builtin.cpp */
|
|
|
|
extern int builtin_count_args(const wchar_t * const * argv);
|
2015-09-22 02:24:49 +08:00
|
|
|
void builtin_print_help(parser_t &parser, io_streams_t &streams, const wchar_t *cmd, output_stream_t &b);
|
2015-09-13 04:36:22 +08:00
|
|
|
|
|
|
|
|
2015-09-13 03:59:40 +08:00
|
|
|
enum
|
|
|
|
{
|
|
|
|
BUILTIN_STRING_OK = 0,
|
|
|
|
BUILTIN_STRING_NONE = 1,
|
|
|
|
BUILTIN_STRING_ERROR = 2
|
|
|
|
};
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static void string_error(io_streams_t &streams, const wchar_t *fmt, ...)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.err.append(L"string ");
|
2015-09-13 03:59:40 +08:00
|
|
|
va_list va;
|
|
|
|
va_start(va, fmt);
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.err.append_formatv(fmt, va);
|
2015-09-13 03:59:40 +08:00
|
|
|
va_end(va);
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static void string_unknown_option(parser_t &parser, io_streams_t &streams, const wchar_t *subcmd, const wchar_t *opt)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_UNKNOWN, subcmd, opt);
|
|
|
|
builtin_print_help(parser, streams, L"string", streams.err);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
/* We read from stdin if we are the second or later process in a pipeline. */
|
|
|
|
static bool string_args_from_stdin(const io_streams_t &streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
return ! streams.is_first_process_in_pipeline;
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static const wchar_t *string_get_arg_stdin(wcstring *storage, const io_streams_t &streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
std::string arg;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
char ch = '\0';
|
2015-09-22 02:24:49 +08:00
|
|
|
long rc = read_blocked(streams.stdin_fd, &ch, 1);
|
2015-09-13 03:59:40 +08:00
|
|
|
|
|
|
|
if (rc < 0)
|
|
|
|
{
|
|
|
|
// failure
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (rc == 0)
|
|
|
|
{
|
|
|
|
// eof
|
|
|
|
if (arg.empty())
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (ch == '\n')
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
arg += ch;
|
|
|
|
}
|
2015-09-13 07:43:51 +08:00
|
|
|
|
|
|
|
*storage = str2wcstring(arg);
|
|
|
|
return storage->c_str();
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static const wchar_t *string_get_arg_argv(int *argidx, wchar_t **argv)
|
|
|
|
{
|
|
|
|
return (argv && argv[*argidx]) ? argv[(*argidx)++] : 0;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static const wchar_t *string_get_arg(int *argidx, wchar_t **argv, wcstring *storage, const io_streams_t &streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams))
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
return string_get_arg_stdin(storage, streams);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
return string_get_arg_argv(argidx, argv);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_escape(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L"n";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"no-quoted", no_argument, 0, 'n' },
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
escape_flags_t flags = ESCAPE_ALL;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'n':
|
|
|
|
flags |= ESCAPE_NO_QUOTED;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
int nesc = 0;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-13 03:59:40 +08:00
|
|
|
const wchar_t *arg;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(escape(arg, flags));
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
nesc++;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (nesc > 0) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_join(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L"q";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
bool quiet = false;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
|
|
|
const wchar_t *sep;
|
|
|
|
if ((sep = string_get_arg_argv(&i, argv)) == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
int nargs = 0;
|
|
|
|
const wchar_t *arg;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
if (!quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
if (nargs > 0)
|
|
|
|
{
|
|
|
|
streams.out.append(sep);
|
|
|
|
}
|
|
|
|
streams.out.append(arg);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
nargs++;
|
|
|
|
}
|
|
|
|
if (nargs > 0 && !quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.push_back(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return (nargs > 1) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_length(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L"q";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
bool quiet = false;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
const wchar_t *arg;
|
|
|
|
int nnonempty = 0;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
size_t n = wcslen(arg);
|
|
|
|
if (n > 0)
|
|
|
|
{
|
|
|
|
nnonempty++;
|
|
|
|
}
|
|
|
|
if (!quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(to_string(n));
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return (nnonempty > 0) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct match_options_t
|
|
|
|
{
|
|
|
|
bool all;
|
|
|
|
bool ignore_case;
|
|
|
|
bool index;
|
|
|
|
bool quiet;
|
|
|
|
|
|
|
|
match_options_t(): all(false), ignore_case(false), index(false), quiet(false) { }
|
|
|
|
};
|
|
|
|
|
|
|
|
class string_matcher_t
|
|
|
|
{
|
|
|
|
protected:
|
|
|
|
match_options_t opts;
|
2015-09-22 02:24:49 +08:00
|
|
|
io_streams_t &streams;
|
2015-09-13 03:59:40 +08:00
|
|
|
int total_matched;
|
|
|
|
|
|
|
|
public:
|
2015-09-22 02:24:49 +08:00
|
|
|
string_matcher_t(const match_options_t &opts_, io_streams_t &streams_)
|
|
|
|
: opts(opts_), streams(streams_), total_matched(0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{ }
|
|
|
|
|
|
|
|
virtual ~string_matcher_t() { }
|
|
|
|
virtual bool report_matches(const wchar_t *arg) = 0;
|
|
|
|
int match_count() { return total_matched; }
|
|
|
|
};
|
|
|
|
|
|
|
|
class wildcard_matcher_t: public string_matcher_t
|
|
|
|
{
|
|
|
|
wcstring wcpattern;
|
|
|
|
|
|
|
|
public:
|
2015-09-22 02:24:49 +08:00
|
|
|
wildcard_matcher_t(const wchar_t * /*argv0*/, const wchar_t *pattern, const match_options_t &opts, io_streams_t &streams)
|
|
|
|
: string_matcher_t(opts, streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
wcpattern = parse_util_unescape_wildcards(pattern);
|
|
|
|
|
|
|
|
if (opts.ignore_case)
|
|
|
|
{
|
|
|
|
for (int i = 0; i < wcpattern.length(); i++)
|
|
|
|
{
|
|
|
|
wcpattern[i] = towlower(wcpattern[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
virtual ~wildcard_matcher_t() { }
|
|
|
|
|
|
|
|
bool report_matches(const wchar_t *arg)
|
|
|
|
{
|
|
|
|
// Note: --all is a no-op for glob matching since the pattern is always
|
|
|
|
// matched against the entire argument
|
|
|
|
bool match;
|
|
|
|
if (opts.ignore_case)
|
|
|
|
{
|
|
|
|
wcstring s = arg;
|
|
|
|
for (int i = 0; i < s.length(); i++)
|
|
|
|
{
|
|
|
|
s[i] = towlower(s[i]);
|
|
|
|
}
|
|
|
|
match = wildcard_match(s, wcpattern, false);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
match = wildcard_match(arg, wcpattern, false);
|
|
|
|
}
|
|
|
|
if (match)
|
|
|
|
{
|
|
|
|
total_matched++;
|
|
|
|
}
|
|
|
|
if (!opts.quiet)
|
|
|
|
{
|
|
|
|
if (match)
|
|
|
|
{
|
|
|
|
if (opts.index)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append_format(L"1 %lu\n", wcslen(arg));
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(arg);
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
2015-09-13 07:52:38 +08:00
|
|
|
static wcstring pcre2_strerror(int err_code)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 07:52:38 +08:00
|
|
|
wchar_t buf[128];
|
2015-09-13 03:59:40 +08:00
|
|
|
pcre2_get_error_message(err_code, (PCRE2_UCHAR *)buf, sizeof(buf) / sizeof(wchar_t));
|
|
|
|
return buf;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct compiled_regex_t
|
|
|
|
{
|
|
|
|
pcre2_code *code;
|
|
|
|
pcre2_match_data *match;
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
compiled_regex_t(const wchar_t *argv0, const wchar_t *pattern, bool ignore_case, io_streams_t &streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
: code(0), match(0)
|
|
|
|
{
|
|
|
|
// Disable some sequences that can lead to security problems
|
|
|
|
uint32_t options = PCRE2_NEVER_UTF;
|
|
|
|
#if PCRE2_CODE_UNIT_WIDTH < 32
|
|
|
|
options |= PCRE2_NEVER_BACKSLASH_C;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
int err_code = 0;
|
|
|
|
PCRE2_SIZE err_offset = 0;
|
|
|
|
|
|
|
|
code = pcre2_compile(
|
|
|
|
PCRE2_SPTR(pattern),
|
|
|
|
PCRE2_ZERO_TERMINATED,
|
|
|
|
options | (ignore_case ? PCRE2_CASELESS : 0),
|
|
|
|
&err_code,
|
|
|
|
&err_offset,
|
|
|
|
0);
|
|
|
|
if (code == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Regular expression compile error: %ls\n"),
|
2015-09-13 07:52:38 +08:00
|
|
|
argv0, pcre2_strerror(err_code).c_str());
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, L"%ls: %ls\n", argv0, pattern);
|
|
|
|
string_error(streams, L"%ls: %*ls\n", argv0, err_offset, L"^");
|
2015-09-13 03:59:40 +08:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
match = pcre2_match_data_create_from_pattern(code, 0);
|
|
|
|
if (match == 0)
|
|
|
|
{
|
|
|
|
DIE_MEM();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
~compiled_regex_t()
|
|
|
|
{
|
|
|
|
if (match != 0)
|
|
|
|
{
|
|
|
|
pcre2_match_data_free(match);
|
|
|
|
}
|
|
|
|
if (code != 0)
|
|
|
|
{
|
|
|
|
pcre2_code_free(code);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
class pcre2_matcher_t: public string_matcher_t
|
|
|
|
{
|
|
|
|
const wchar_t *argv0;
|
|
|
|
compiled_regex_t regex;
|
|
|
|
|
|
|
|
int report_match(const wchar_t *arg, int pcre2_rc)
|
|
|
|
{
|
|
|
|
// Return values: -1 = error, 0 = no match, 1 = match
|
|
|
|
if (pcre2_rc == PCRE2_ERROR_NOMATCH)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
if (pcre2_rc < 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Regular expression match error: %ls\n"),
|
2015-09-13 07:52:38 +08:00
|
|
|
argv0, pcre2_strerror(pcre2_rc).c_str());
|
2015-09-13 03:59:40 +08:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
if (pcre2_rc == 0)
|
|
|
|
{
|
|
|
|
// The output vector wasn't big enough. Should not happen.
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Regular expression internal error\n"), argv0);
|
2015-09-13 03:59:40 +08:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
PCRE2_SIZE *ovector = pcre2_get_ovector_pointer(regex.match);
|
|
|
|
for (int j = 0; j < pcre2_rc; j++)
|
|
|
|
{
|
|
|
|
PCRE2_SIZE begin = ovector[2*j];
|
|
|
|
PCRE2_SIZE end = ovector[2*j + 1];
|
|
|
|
if (!opts.quiet)
|
|
|
|
{
|
|
|
|
if (begin != PCRE2_UNSET && end != PCRE2_UNSET)
|
|
|
|
{
|
|
|
|
if (opts.index)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append_format(L"%lu %lu", (unsigned long)(begin + 1), (unsigned long)(end - begin));
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else if (end > begin) // may have end < begin if \K is used
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(wcstring(&arg[begin], end - begin));
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
public:
|
2015-09-22 02:24:49 +08:00
|
|
|
pcre2_matcher_t(const wchar_t *argv0_, const wchar_t *pattern, const match_options_t &opts, io_streams_t &streams)
|
|
|
|
: string_matcher_t(opts, streams),
|
2015-09-13 03:59:40 +08:00
|
|
|
argv0(argv0_),
|
2015-09-22 02:24:49 +08:00
|
|
|
regex(argv0_, pattern, opts.ignore_case, streams)
|
2015-09-13 03:59:40 +08:00
|
|
|
{ }
|
|
|
|
|
|
|
|
virtual ~pcre2_matcher_t() { }
|
|
|
|
|
|
|
|
bool report_matches(const wchar_t *arg)
|
|
|
|
{
|
|
|
|
// A return value of true means all is well (even if no matches were
|
|
|
|
// found), false indicates an unrecoverable error.
|
|
|
|
if (regex.code == 0)
|
|
|
|
{
|
|
|
|
// pcre2_compile() failed
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
int matched = 0;
|
|
|
|
|
|
|
|
// See pcre2demo.c for an explanation of this logic
|
|
|
|
PCRE2_SIZE arglen = wcslen(arg);
|
|
|
|
int rc = report_match(arg, pcre2_match(regex.code, PCRE2_SPTR(arg), arglen, 0, 0, regex.match, 0));
|
|
|
|
if (rc < 0)
|
|
|
|
{
|
|
|
|
// pcre2 match error
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
if (rc == 0)
|
|
|
|
{
|
|
|
|
// no match
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
matched++;
|
|
|
|
total_matched++;
|
|
|
|
|
|
|
|
// Report any additional matches
|
|
|
|
PCRE2_SIZE *ovector = pcre2_get_ovector_pointer(regex.match);
|
|
|
|
while (opts.all || matched == 0)
|
|
|
|
{
|
|
|
|
uint32_t options = 0;
|
|
|
|
PCRE2_SIZE offset = ovector[1]; // Start at end of previous match
|
|
|
|
|
|
|
|
if (ovector[0] == ovector[1])
|
|
|
|
{
|
|
|
|
if (ovector[0] == arglen)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
options = PCRE2_NOTEMPTY_ATSTART | PCRE2_ANCHORED;
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = report_match(arg, pcre2_match(regex.code, PCRE2_SPTR(arg), arglen, offset, options, regex.match, 0));
|
|
|
|
if (rc < 0)
|
|
|
|
{
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
if (rc == 0)
|
|
|
|
{
|
|
|
|
if (options == 0)
|
|
|
|
{
|
|
|
|
// All matches found
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
ovector[1] = offset + 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
matched++;
|
|
|
|
total_matched++;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_match(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L"ainqr";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"all", no_argument, 0, 'a'},
|
|
|
|
{ L"ignore-case", no_argument, 0, 'i'},
|
|
|
|
{ L"index", no_argument, 0, 'n'},
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ L"regex", no_argument, 0, 'r'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
match_options_t opts;
|
|
|
|
bool regex = false;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'a':
|
|
|
|
opts.all = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'i':
|
|
|
|
opts.ignore_case = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'n':
|
|
|
|
opts.index = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
opts.quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'r':
|
|
|
|
regex = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
|
|
|
const wchar_t *pattern;
|
|
|
|
if ((pattern = string_get_arg_argv(&i, argv)) == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
string_matcher_t *matcher;
|
|
|
|
if (regex)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
matcher = new pcre2_matcher_t(argv[0], pattern, opts, streams);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
matcher = new wildcard_matcher_t(argv[0], pattern, opts, streams);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
const wchar_t *arg;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
if (!matcher->report_matches(arg))
|
|
|
|
{
|
|
|
|
delete matcher;
|
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int rc = matcher->match_count() > 0 ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
delete matcher;
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct replace_options_t
|
|
|
|
{
|
|
|
|
bool all;
|
|
|
|
bool ignore_case;
|
|
|
|
bool quiet;
|
|
|
|
|
|
|
|
replace_options_t(): all(false), ignore_case(false), quiet(false) { }
|
|
|
|
};
|
|
|
|
|
|
|
|
class string_replacer_t
|
|
|
|
{
|
|
|
|
protected:
|
|
|
|
const wchar_t *argv0;
|
|
|
|
replace_options_t opts;
|
|
|
|
int total_replaced;
|
2015-09-22 02:24:49 +08:00
|
|
|
io_streams_t &streams;
|
2015-09-13 03:59:40 +08:00
|
|
|
|
|
|
|
public:
|
2015-09-22 02:24:49 +08:00
|
|
|
string_replacer_t(const wchar_t *argv0_, const replace_options_t &opts_, io_streams_t &streams_)
|
|
|
|
: argv0(argv0_), opts(opts_), total_replaced(0), streams(streams_)
|
2015-09-13 03:59:40 +08:00
|
|
|
{ }
|
|
|
|
|
|
|
|
virtual ~string_replacer_t() {}
|
|
|
|
virtual bool replace_matches(const wchar_t *arg) = 0;
|
|
|
|
int replace_count() { return total_replaced; }
|
|
|
|
};
|
|
|
|
|
|
|
|
class literal_replacer_t: public string_replacer_t
|
|
|
|
{
|
|
|
|
const wchar_t *pattern;
|
|
|
|
const wchar_t *replacement;
|
2015-09-13 07:43:51 +08:00
|
|
|
size_t patlen;
|
2015-09-13 03:59:40 +08:00
|
|
|
|
|
|
|
public:
|
|
|
|
literal_replacer_t(const wchar_t *argv0, const wchar_t *pattern_, const wchar_t *replacement_,
|
2015-09-22 02:24:49 +08:00
|
|
|
const replace_options_t &opts, io_streams_t &streams)
|
|
|
|
: string_replacer_t(argv0, opts, streams),
|
2015-09-13 03:59:40 +08:00
|
|
|
pattern(pattern_), replacement(replacement_), patlen(wcslen(pattern))
|
|
|
|
{ }
|
|
|
|
|
|
|
|
virtual ~literal_replacer_t() { }
|
|
|
|
|
|
|
|
bool replace_matches(const wchar_t *arg)
|
|
|
|
{
|
|
|
|
wcstring result;
|
|
|
|
if (patlen == 0)
|
|
|
|
{
|
|
|
|
result = arg;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
int replaced = 0;
|
|
|
|
const wchar_t *cur = arg;
|
|
|
|
while (*cur != L'\0')
|
|
|
|
{
|
|
|
|
if ((opts.all || replaced == 0) &&
|
|
|
|
(opts.ignore_case ? wcsncasecmp(cur, pattern, patlen) : wcsncmp(cur, pattern, patlen)) == 0)
|
|
|
|
{
|
|
|
|
result += replacement;
|
|
|
|
cur += patlen;
|
|
|
|
replaced++;
|
|
|
|
total_replaced++;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
result += *cur;
|
|
|
|
cur++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (!opts.quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(result);
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
class regex_replacer_t: public string_replacer_t
|
|
|
|
{
|
|
|
|
compiled_regex_t regex;
|
|
|
|
wcstring replacement;
|
|
|
|
|
|
|
|
wcstring interpret_escapes(const wchar_t *orig)
|
|
|
|
{
|
|
|
|
wcstring result;
|
|
|
|
|
|
|
|
while (*orig != L'\0')
|
|
|
|
{
|
|
|
|
if (*orig == L'\\')
|
|
|
|
{
|
|
|
|
orig += read_unquoted_escape(orig, &result, true, false);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
result += *orig;
|
|
|
|
orig++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
public:
|
|
|
|
regex_replacer_t(const wchar_t *argv0, const wchar_t *pattern, const wchar_t *replacement_,
|
2015-09-22 02:24:49 +08:00
|
|
|
const replace_options_t &opts, io_streams_t &streams)
|
|
|
|
: string_replacer_t(argv0, opts, streams),
|
|
|
|
regex(argv0, pattern, opts.ignore_case, streams),
|
2015-09-13 03:59:40 +08:00
|
|
|
replacement(interpret_escapes(replacement_))
|
|
|
|
{ }
|
|
|
|
|
|
|
|
virtual ~regex_replacer_t() { }
|
|
|
|
|
|
|
|
bool replace_matches(const wchar_t *arg)
|
|
|
|
{
|
|
|
|
// A return value of true means all is well (even if no replacements
|
|
|
|
// were performed), false indicates an unrecoverable error.
|
|
|
|
if (regex.code == 0)
|
|
|
|
{
|
|
|
|
// pcre2_compile() failed
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
uint32_t options = opts.all ? PCRE2_SUBSTITUTE_GLOBAL : 0;
|
2015-09-13 07:43:51 +08:00
|
|
|
size_t arglen = wcslen(arg);
|
2015-09-13 03:59:40 +08:00
|
|
|
PCRE2_SIZE outlen = (arglen == 0) ? 16 : 2 * arglen;
|
|
|
|
wchar_t *output = (wchar_t *)malloc(sizeof(wchar_t) * outlen);
|
|
|
|
if (output == 0)
|
|
|
|
{
|
|
|
|
DIE_MEM();
|
|
|
|
}
|
|
|
|
int pcre2_rc = 0;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
pcre2_rc = pcre2_substitute(
|
|
|
|
regex.code,
|
|
|
|
PCRE2_SPTR(arg),
|
|
|
|
arglen,
|
|
|
|
0, // start offset
|
|
|
|
options,
|
|
|
|
regex.match,
|
|
|
|
0, // match context
|
|
|
|
PCRE2_SPTR(replacement.c_str()),
|
|
|
|
PCRE2_ZERO_TERMINATED,
|
|
|
|
(PCRE2_UCHAR *)output,
|
|
|
|
&outlen);
|
|
|
|
|
|
|
|
if (pcre2_rc == PCRE2_ERROR_NOMEMORY)
|
|
|
|
{
|
|
|
|
if (outlen < MAX_REPLACE_SIZE)
|
|
|
|
{
|
|
|
|
outlen = std::min(2 * outlen, MAX_REPLACE_SIZE);
|
|
|
|
output = (wchar_t *)realloc(output, sizeof(wchar_t) * outlen);
|
|
|
|
if (output == 0)
|
|
|
|
{
|
|
|
|
DIE_MEM();
|
|
|
|
}
|
|
|
|
continue;
|
|
|
|
}
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Replacement string too large\n"), argv0);
|
2015-09-13 03:59:40 +08:00
|
|
|
free(output);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool rc = true;
|
|
|
|
if (pcre2_rc < 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Regular expression substitute error: %ls\n"),
|
2015-09-13 07:52:38 +08:00
|
|
|
argv0, pcre2_strerror(pcre2_rc).c_str());
|
2015-09-13 03:59:40 +08:00
|
|
|
rc = false;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
if (!opts.quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(output);
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
total_replaced += pcre2_rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
free(output);
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_replace(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L"aiqr";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"all", no_argument, 0, 'a'},
|
|
|
|
{ L"ignore-case", no_argument, 0, 'i'},
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ L"regex", no_argument, 0, 'r'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
replace_options_t opts;
|
|
|
|
bool regex = false;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'a':
|
|
|
|
opts.all = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'i':
|
|
|
|
opts.ignore_case = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
opts.quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'r':
|
|
|
|
regex = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
|
|
|
const wchar_t *pattern, *replacement;
|
|
|
|
if ((pattern = string_get_arg_argv(&i, argv)) == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
if ((replacement = string_get_arg_argv(&i, argv)) == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
string_replacer_t *replacer;
|
|
|
|
if (regex)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
replacer = new regex_replacer_t(argv[0], pattern, replacement, opts, streams);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
replacer = new literal_replacer_t(argv[0], pattern, replacement, opts, streams);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
const wchar_t *arg;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
if (!replacer->replace_matches(arg))
|
|
|
|
{
|
|
|
|
delete replacer;
|
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int rc = replacer->replace_count() > 0 ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
delete replacer;
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
2015-09-13 16:11:49 +08:00
|
|
|
// Given iterators into a string (forward or reverse), splits the haystack iterators
|
|
|
|
// about the needle sequence, up to max times. Inserts splits into the output array
|
|
|
|
// If the iterators are forward, this does the normal thing.
|
|
|
|
// If the iterators are backward, this returns reversed strings, in reversed order!
|
|
|
|
// If the needle is empty, split on individual elements (characters)
|
|
|
|
template<typename ITER>
|
|
|
|
void split_about(ITER haystack_start, ITER haystack_end,
|
|
|
|
ITER needle_start, ITER needle_end,
|
|
|
|
wcstring_list_t *output, long max)
|
|
|
|
{
|
|
|
|
long remaining = max;
|
|
|
|
ITER haystack_cursor = haystack_start;
|
|
|
|
while (remaining > 0 && haystack_cursor != haystack_end)
|
|
|
|
{
|
|
|
|
ITER split_point;
|
|
|
|
if (needle_start == needle_end)
|
|
|
|
{
|
|
|
|
// empty needle, we split on individual elements
|
|
|
|
split_point = haystack_cursor + 1;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
split_point = std::search(haystack_cursor, haystack_end, needle_start, needle_end);
|
|
|
|
}
|
|
|
|
if (split_point == haystack_end)
|
|
|
|
{
|
|
|
|
// not found
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
output->push_back(wcstring(haystack_cursor, split_point));
|
|
|
|
remaining--;
|
|
|
|
// need to skip over the needle for the next search
|
|
|
|
// note that the needle may be empty
|
|
|
|
haystack_cursor = split_point + std::distance(needle_start, needle_end);
|
|
|
|
}
|
|
|
|
// trailing component, possibly empty
|
|
|
|
output->push_back(wcstring(haystack_cursor, haystack_end));
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_split(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L":m:qr";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"max", required_argument, 0, 'm'},
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ L"right", no_argument, 0, 'r'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
long max = LONG_MAX;
|
|
|
|
bool quiet = false;
|
|
|
|
bool right = false;
|
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'm':
|
|
|
|
{
|
|
|
|
errno = 0;
|
|
|
|
wchar_t *endptr = 0;
|
|
|
|
max = wcstol(w.woptarg, &endptr, 10);
|
|
|
|
if (*endptr != L'\0' || errno != 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_NOT_NUMBER, argv[0], w.woptarg);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'r':
|
|
|
|
right = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case ':':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
|
|
|
const wchar_t *sep;
|
2015-09-13 16:11:49 +08:00
|
|
|
if ((sep = string_get_arg_argv(&i, argv)) == NULL)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
const wchar_t *sep_end = sep + wcslen(sep);
|
2015-09-13 03:59:40 +08:00
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
|
|
|
|
wcstring_list_t splits;
|
|
|
|
size_t arg_count = 0;
|
|
|
|
wcstring storage;
|
2015-09-13 03:59:40 +08:00
|
|
|
const wchar_t *arg;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 16:11:49 +08:00
|
|
|
const wchar_t *arg_end = arg + wcslen(arg);
|
|
|
|
if (right)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 16:11:49 +08:00
|
|
|
typedef std::reverse_iterator<const wchar_t *> reverser;
|
|
|
|
split_about(reverser(arg_end), reverser(arg),
|
|
|
|
reverser(sep_end), reverser(sep),
|
|
|
|
&splits, max);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
split_about(arg, arg_end,
|
|
|
|
sep, sep_end,
|
|
|
|
&splits, max);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
arg_count++;
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
|
|
|
|
// If we are from the right, split_about gave us reversed strings, in reversed order!
|
|
|
|
if (right)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 16:11:49 +08:00
|
|
|
for (size_t i=0; i < splits.size(); i++)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 16:11:49 +08:00
|
|
|
std::reverse(splits[i].begin(), splits[i].end());
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
std::reverse(splits.begin(), splits.end());
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 16:11:49 +08:00
|
|
|
|
2015-09-13 03:59:40 +08:00
|
|
|
if (!quiet)
|
|
|
|
{
|
2015-09-13 16:11:49 +08:00
|
|
|
for (wcstring_list_t::const_iterator si = splits.begin(); si != splits.end(); ++si)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(*si);
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-09-13 16:11:49 +08:00
|
|
|
// we split something if we have more split values than args
|
|
|
|
return (splits.size() > arg_count) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_sub(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L":l:qs:";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"length", required_argument, 0, 'l'},
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ L"start", required_argument, 0, 's'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
2015-09-13 17:15:37 +08:00
|
|
|
long start = 0;
|
|
|
|
long length = -1;
|
2015-09-13 03:59:40 +08:00
|
|
|
bool quiet = false;
|
|
|
|
wgetopter_t w;
|
2015-09-13 17:15:37 +08:00
|
|
|
wchar_t *endptr = NULL;
|
2015-09-13 03:59:40 +08:00
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'l':
|
|
|
|
errno = 0;
|
2015-09-13 17:15:37 +08:00
|
|
|
length = wcstol(w.woptarg, &endptr, 10);
|
|
|
|
if (*endptr != L'\0' || (errno != 0 && errno != ERANGE))
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_NOT_NUMBER, argv[0], w.woptarg);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
2015-09-13 17:15:37 +08:00
|
|
|
if (length < 0 || errno == ERANGE)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Invalid length value '%ls'\n"), argv[0], w.woptarg);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 's':
|
|
|
|
errno = 0;
|
2015-09-13 17:15:37 +08:00
|
|
|
start = wcstol(w.woptarg, &endptr, 10);
|
|
|
|
if (*endptr != L'\0' || (errno != 0 && errno != ERANGE))
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_NOT_NUMBER, argv[0], w.woptarg);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
2015-09-13 17:15:37 +08:00
|
|
|
if (start == 0 || start == LONG_MIN || errno == ERANGE)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, _(L"%ls: Invalid start value '%ls'\n"), argv[0], w.woptarg);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
case ':':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
int nsub = 0;
|
|
|
|
const wchar_t *arg;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != NULL)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 17:15:37 +08:00
|
|
|
typedef wcstring::size_type size_type;
|
|
|
|
size_type pos = 0;
|
|
|
|
size_type count = wcstring::npos;
|
2015-09-13 03:59:40 +08:00
|
|
|
wcstring s(arg);
|
|
|
|
if (start > 0)
|
|
|
|
{
|
2015-09-13 17:15:37 +08:00
|
|
|
pos = static_cast<size_type>(start - 1);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
else if (start < 0)
|
|
|
|
{
|
2015-09-13 17:15:37 +08:00
|
|
|
assert(start != LONG_MIN); // checked above
|
|
|
|
size_type n = static_cast<size_type>(-start);
|
2015-09-13 03:59:40 +08:00
|
|
|
pos = n > s.length() ? 0 : s.length() - n;
|
|
|
|
}
|
|
|
|
if (pos > s.length())
|
|
|
|
{
|
|
|
|
pos = s.length();
|
|
|
|
}
|
|
|
|
|
|
|
|
if (length >= 0)
|
|
|
|
{
|
2015-09-13 17:15:37 +08:00
|
|
|
count = static_cast<size_type>(length);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
|
2015-09-13 17:15:37 +08:00
|
|
|
// note that std::string permits count to extend past end of string
|
2015-09-13 03:59:40 +08:00
|
|
|
if (!quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(s.substr(pos, count));
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
nsub++;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (nsub > 0) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
}
|
|
|
|
|
2015-09-22 02:24:49 +08:00
|
|
|
static int string_trim(parser_t &parser, io_streams_t &streams, int argc, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
const wchar_t *short_options = L":c:lqr";
|
|
|
|
const struct woption long_options[] =
|
|
|
|
{
|
|
|
|
{ L"chars", required_argument, 0, 'c'},
|
|
|
|
{ L"left", no_argument, 0, 'l'},
|
|
|
|
{ L"quiet", no_argument, 0, 'q'},
|
|
|
|
{ L"right", no_argument, 0, 'r'},
|
|
|
|
{ 0, 0, 0, 0 }
|
|
|
|
};
|
|
|
|
|
2015-09-13 08:15:08 +08:00
|
|
|
bool do_left = 0, do_right = 0;
|
2015-09-13 03:59:40 +08:00
|
|
|
bool quiet = false;
|
2015-09-13 08:15:08 +08:00
|
|
|
wcstring chars_to_trim = L" \f\n\r\t";
|
2015-09-13 03:59:40 +08:00
|
|
|
wgetopter_t w;
|
|
|
|
for (;;)
|
|
|
|
{
|
|
|
|
int c = w.wgetopt_long(argc, argv, short_options, long_options, 0);
|
|
|
|
|
|
|
|
if (c == -1)
|
|
|
|
{
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (c)
|
|
|
|
{
|
|
|
|
case 0:
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'c':
|
2015-09-13 08:15:08 +08:00
|
|
|
chars_to_trim = w.woptarg;
|
2015-09-13 03:59:40 +08:00
|
|
|
break;
|
|
|
|
|
|
|
|
case 'l':
|
2015-09-13 08:15:08 +08:00
|
|
|
do_left = true;
|
2015-09-13 03:59:40 +08:00
|
|
|
break;
|
|
|
|
|
|
|
|
case 'q':
|
|
|
|
quiet = true;
|
|
|
|
break;
|
|
|
|
|
|
|
|
case 'r':
|
2015-09-13 08:15:08 +08:00
|
|
|
do_right = true;
|
2015-09-13 03:59:40 +08:00
|
|
|
break;
|
|
|
|
|
|
|
|
case ':':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, STRING_ERR_MISSING, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
|
|
|
|
case '?':
|
2015-09-22 02:24:49 +08:00
|
|
|
string_unknown_option(parser, streams, argv[0], argv[w.woptind - 1]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
int i = w.woptind;
|
2015-09-22 02:24:49 +08:00
|
|
|
if (string_args_from_stdin(streams) && argc > i)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
string_error(streams, BUILTIN_ERR_TOO_MANY_ARGUMENTS, argv[0]);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
2015-09-13 08:15:08 +08:00
|
|
|
|
|
|
|
/* if neither left or right is specified, we do both */
|
|
|
|
if (! do_left && ! do_right)
|
|
|
|
{
|
|
|
|
do_left = true;
|
|
|
|
do_right = true;
|
|
|
|
}
|
2015-09-13 03:59:40 +08:00
|
|
|
|
|
|
|
const wchar_t *arg;
|
2015-09-13 08:15:08 +08:00
|
|
|
size_t ntrim = 0;
|
|
|
|
|
|
|
|
wcstring argstr;
|
2015-09-13 07:43:51 +08:00
|
|
|
wcstring storage;
|
2015-09-22 02:24:49 +08:00
|
|
|
while ((arg = string_get_arg(&i, argv, &storage, streams)) != 0)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 08:15:08 +08:00
|
|
|
argstr = arg;
|
|
|
|
// begin and end are respectively the first character to keep on the left,
|
|
|
|
// and first character to trim on the right. The length is thus end - start.
|
|
|
|
size_t begin = 0, end = argstr.size();
|
|
|
|
if (do_right)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 08:15:08 +08:00
|
|
|
size_t last_to_keep = argstr.find_last_not_of(chars_to_trim);
|
|
|
|
end = (last_to_keep == wcstring::npos) ? 0 : last_to_keep + 1;
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 08:15:08 +08:00
|
|
|
if (do_left)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
2015-09-13 08:15:08 +08:00
|
|
|
size_t first_to_keep = argstr.find_first_not_of(chars_to_trim);
|
|
|
|
begin = (first_to_keep == wcstring::npos ? end : first_to_keep);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
2015-09-13 08:15:08 +08:00
|
|
|
assert(begin <= end && end <= argstr.size());
|
|
|
|
ntrim += argstr.size() - (end - begin);
|
2015-09-13 03:59:40 +08:00
|
|
|
if (!quiet)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
streams.out.append(wcstring(argstr, begin, end - begin));
|
|
|
|
streams.out.append(L'\n');
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return (ntrim > 0) ? BUILTIN_STRING_OK : BUILTIN_STRING_NONE;
|
|
|
|
}
|
|
|
|
|
|
|
|
static const struct string_subcommand
|
|
|
|
{
|
|
|
|
const wchar_t *name;
|
2015-09-22 02:24:49 +08:00
|
|
|
int (*handler)(parser_t &, io_streams_t &, int argc, wchar_t **argv);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|
|
|
|
string_subcommands[] =
|
|
|
|
{
|
|
|
|
{ L"escape", &string_escape },
|
|
|
|
{ L"join", &string_join },
|
|
|
|
{ L"length", &string_length },
|
|
|
|
{ L"match", &string_match },
|
|
|
|
{ L"replace", &string_replace },
|
|
|
|
{ L"split", &string_split },
|
|
|
|
{ L"sub", &string_sub },
|
|
|
|
{ L"trim", &string_trim },
|
|
|
|
{ 0, 0 }
|
|
|
|
};
|
|
|
|
|
|
|
|
/**
|
|
|
|
The string builtin, for manipulating strings.
|
|
|
|
*/
|
2015-09-22 02:24:49 +08:00
|
|
|
int builtin_string(parser_t &parser, io_streams_t &streams, wchar_t **argv)
|
2015-09-13 03:59:40 +08:00
|
|
|
{
|
|
|
|
int argc = builtin_count_args(argv);
|
|
|
|
if (argc <= 1)
|
|
|
|
{
|
2015-09-25 09:21:09 +08:00
|
|
|
streams.err.append_format(_(L"string: Expected subcommand\n"));
|
2015-09-22 02:24:49 +08:00
|
|
|
builtin_print_help(parser, streams, L"string", streams.err);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (wcscmp(argv[1], L"-h") == 0 || wcscmp(argv[1], L"--help") == 0)
|
|
|
|
{
|
2015-09-22 02:24:49 +08:00
|
|
|
builtin_print_help(parser, streams, L"string", streams.err);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
const string_subcommand *subcmd = &string_subcommands[0];
|
|
|
|
while (subcmd->name != 0 && wcscmp(subcmd->name, argv[1]) != 0)
|
|
|
|
{
|
|
|
|
subcmd++;
|
|
|
|
}
|
|
|
|
if (subcmd->handler == 0)
|
|
|
|
{
|
2015-09-25 09:21:09 +08:00
|
|
|
streams.err.append_format(_(L"string: Unknown subcommand '%ls'\n"), argv[1]);
|
2015-09-22 02:24:49 +08:00
|
|
|
builtin_print_help(parser, streams, L"string", streams.err);
|
2015-09-13 03:59:40 +08:00
|
|
|
return BUILTIN_STRING_ERROR;
|
|
|
|
}
|
|
|
|
|
|
|
|
argc--;
|
|
|
|
argv++;
|
2015-09-22 02:24:49 +08:00
|
|
|
return subcmd->handler(parser, streams, argc, argv);
|
2015-09-13 03:59:40 +08:00
|
|
|
}
|