-/* Yacc grammar for bash. */
+/* parse.y - Yacc grammar for bash. */
-/* Copyright (C) 1989-2004 Free Software Foundation, Inc.
+/* Copyright (C) 1989-2009 Free Software Foundation, Inc.
This file is part of GNU Bash, the Bourne Again SHell.
- Bash is free software; you can redistribute it and/or modify it under
- the terms of the GNU General Public License as published by the Free
- Software Foundation; either version 2, or (at your option) any later
- version.
+ Bash is free software: you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
- Bash 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 General Public License
- for more details.
+ Bash 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 General Public License for more details.
- You should have received a copy of the GNU General Public License along
- with Bash; see the file LICENSE. If not, write to the Free Software
- Foundation, 59 Temple Place, Suite 330, Boston, MA 02111 USA. */
+ You should have received a copy of the GNU General Public License
+ along with Bash. If not, see <http://www.gnu.org/licenses/>.
+*/
%{
#include "config.h"
#include "parser.h"
#include "mailcheck.h"
#include "test.h"
+#include "builtins.h"
#include "builtins/common.h"
#include "builtins/builtext.h"
extern int eof_encountered;
extern int no_line_editing, running_under_emacs;
extern int current_command_number;
-extern int sourcelevel;
+extern int sourcelevel, parse_and_execute_level;
extern int posixly_correct;
extern int last_command_exit_value;
-extern int interrupt_immediately;
extern char *shell_name, *current_host_name;
extern char *dist_version;
extern int patch_level;
static int yy_string_get __P((void));
static int yy_string_unget __P((int));
+static void rewind_input_string __P((void));
static int yy_stream_get __P((void));
static int yy_stream_unget __P((int));
static int special_case_tokens __P((char *));
static int read_token __P((int));
static char *parse_matched_pair __P((int, int, int, int *, int));
+static char *parse_comsub __P((int, int, int, int *, int));
#if defined (ARRAY_VARS)
static char *parse_compound_assignment __P((int *));
#endif
#endif
static void print_prompt __P((void));
-#if defined (HISTORY)
-char *history_delimiting_chars __P((void));
-#endif
-
#if defined (HANDLE_MULTIBYTE)
static void set_line_mbstate __P((void));
static char *shell_input_line_property = NULL;
/* The number of lines read from input while creating the current command. */
int current_command_line_count;
+/* The token that currently denotes the end of parse. */
+int shell_eof_token;
+
+/* The token currently being read. */
+int current_token;
+
/* Variables to manage the task of reading here documents, because we need to
defer the reading until after a complete command has been collected. */
static REDIRECT *redir_stack[10];
/* The line number in a script at which an arithmetic for command starts. */
static int arith_for_lineno;
+/* The current parser state. */
+static int parser_state;
+
+/* The last read token, or NULL. read_token () uses this for context
+ checking. */
+static int last_read_token;
+
+/* The token read prior to last_read_token. */
+static int token_before_that;
+
+/* The token read prior to token_before_that. */
+static int two_tokens_ago;
+
/* The line number in a script where the word in a `case WORD', `select WORD'
or `for WORD' begins. This is a nested command maximum, since the array
index is decremented after a case, select, or for command is parsed. */
in the case that they are preceded by a list_terminator. Members
of the second group are for [[...]] commands. Members of the
third group are recognized only under special circumstances. */
-%token IF THEN ELSE ELIF FI CASE ESAC FOR SELECT WHILE UNTIL DO DONE FUNCTION
+%token IF THEN ELSE ELIF FI CASE ESAC FOR SELECT WHILE UNTIL DO DONE FUNCTION COPROC
%token COND_START COND_END COND_ERROR
%token IN BANG TIME TIMEOPT
%token <word_list> ARITH_CMD ARITH_FOR_EXPRS
%token <command> COND_CMD
%token AND_AND OR_OR GREATER_GREATER LESS_LESS LESS_AND LESS_LESS_LESS
-%token GREATER_AND SEMI_SEMI LESS_LESS_MINUS AND_GREATER LESS_GREATER
-%token GREATER_BAR
+%token GREATER_AND SEMI_SEMI SEMI_AND SEMI_SEMI_AND
+%token LESS_LESS_MINUS AND_GREATER AND_GREATER_GREATER LESS_GREATER
+%token GREATER_BAR BAR_AND
/* The types that the various syntactical units return. */
%type <command> arith_command
%type <command> cond_command
%type <command> arith_for_command
+%type <command> coproc
%type <command> function_def function_body if_command elif_clause subshell
%type <redirect> redirection redirection_list
%type <element> simple_command_element
%left '&' ';' '\n' yacc_EOF
%left AND_AND OR_OR
-%right '|'
+%right '|' BAR_AND
%%
inputunit: simple_list simple_list_terminator
global_command = $1;
eof_encountered = 0;
/* discard_parser_constructs (0); */
+ if (parser_state & PST_CMDSUBST)
+ parser_state |= PST_EOFTOKEN;
YYACCEPT;
}
| '\n'
/* Case of regular command, but not a very
interesting one. Return a NULL command. */
global_command = (COMMAND *)NULL;
+ if (parser_state & PST_CMDSUBST)
+ parser_state |= PST_EOFTOKEN;
YYACCEPT;
}
| error '\n'
global_command = (COMMAND *)NULL;
eof_encountered = 0;
/* discard_parser_constructs (1); */
- if (interactive)
+ if (interactive && parse_and_execute_level == 0)
{
YYACCEPT;
}
redir.filename = $2;
$$ = make_redirection (1, r_err_and_out, redir);
}
+ | AND_GREATER_GREATER WORD
+ {
+ redir.filename = $2;
+ $$ = make_redirection (1, r_append_err_and_out, redir);
+ }
| NUMBER LESS_GREATER WORD
{
redir.filename = $3;
}
| function_def
{ $$ = $1; }
+ | coproc
+ { $$ = $1; }
;
shell_command: for_command
{ $$ = make_function_def ($2, $4, function_dstart, function_bstart); }
;
-
function_body: shell_command
{ $$ = $1; }
| shell_command redirection_list
}
;
+coproc: COPROC shell_command
+ {
+ $$ = make_coproc_command ("COPROC", $2);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC shell_command redirection_list
+ {
+ COMMAND *tc;
+
+ tc = $2;
+ if (tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $3;
+ }
+ else
+ tc->redirects = $3;
+ $$ = make_coproc_command ("COPROC", $2);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC WORD shell_command
+ {
+ $$ = make_coproc_command ($2->word, $3);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC WORD shell_command redirection_list
+ {
+ COMMAND *tc;
+
+ tc = $3;
+ if (tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $4;
+ }
+ else
+ tc->redirects = $4;
+ $$ = make_coproc_command ($2->word, $3);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC simple_command
+ {
+ $$ = make_coproc_command ("COPROC", clean_simple_command ($2));
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ ;
+
if_command: IF compound_list THEN compound_list FI
{ $$ = make_if_command ($2, $4, (COMMAND *)NULL); }
| IF compound_list THEN compound_list ELSE compound_list FI
;
case_clause_sequence: pattern_list SEMI_SEMI
+ { $$ = $1; }
| case_clause_sequence pattern_list SEMI_SEMI
{ $2->next = $1; $$ = $2; }
+ | pattern_list SEMI_AND
+ { $1->flags |= CASEPAT_FALLTHROUGH; $$ = $1; }
+ | case_clause_sequence pattern_list SEMI_AND
+ { $2->flags |= CASEPAT_FALLTHROUGH; $2->next = $1; $$ = $2; }
+ | pattern_list SEMI_SEMI_AND
+ { $1->flags |= CASEPAT_TESTNEXT; $$ = $1; }
+ | case_clause_sequence pattern_list SEMI_SEMI_AND
+ { $2->flags |= CASEPAT_TESTNEXT; $2->next = $1; $$ = $2; }
;
pattern: WORD
$$ = $1;
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 '&'
{
$$ = command_connect ($1, (COMMAND *)NULL, '&');
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 ';'
{
$$ = $1;
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
;
{ $$ = $1; }
| BANG pipeline
{
- $2->flags |= CMD_INVERT_RETURN;
+ if ($2)
+ $2->flags |= CMD_INVERT_RETURN;
$$ = $2;
}
| timespec pipeline
{
- $2->flags |= $1;
+ if ($2)
+ $2->flags |= $1;
$$ = $2;
}
| timespec BANG pipeline
{
- $3->flags |= $1|CMD_INVERT_RETURN;
+ if ($3)
+ $3->flags |= $1|CMD_INVERT_RETURN;
$$ = $3;
}
| BANG timespec pipeline
{
- $3->flags |= $2|CMD_INVERT_RETURN;
+ if ($3)
+ $3->flags |= $2|CMD_INVERT_RETURN;
$$ = $3;
}
| timespec list_terminator
;
-pipeline:
- pipeline '|' newline_list pipeline
+pipeline: pipeline '|' newline_list pipeline
{ $$ = command_connect ($1, $4, '|'); }
+ | pipeline BAR_AND newline_list pipeline
+ {
+ /* Make cmd1 |& cmd2 equivalent to cmd1 2>&1 | cmd2 */
+ COMMAND *tc;
+ REDIRECTEE rd;
+ REDIRECT *r;
+
+ tc = $1;
+ rd.dest = 1;
+ r = make_redirection (2, r_duplicating_output, rd);
+ if (tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = r;
+ }
+ else
+ tc->redirects = r;
+
+ $$ = command_connect ($1, $4, '|');
+ }
| command
{ $$ = $1; }
;
;
%%
-/* Possible states for the parser that require it to do special things. */
-#define PST_CASEPAT 0x0001 /* in a case pattern list */
-#define PST_ALEXPNEXT 0x0002 /* expand next word for aliases */
-#define PST_ALLOWOPNBRC 0x0004 /* allow open brace for function def */
-#define PST_NEEDCLOSBRC 0x0008 /* need close brace */
-#define PST_DBLPAREN 0x0010 /* double-paren parsing */
-#define PST_SUBSHELL 0x0020 /* ( ... ) subshell */
-#define PST_CMDSUBST 0x0040 /* $( ... ) command substitution */
-#define PST_CASESTMT 0x0080 /* parsing a case statement */
-#define PST_CONDCMD 0x0100 /* parsing a [[...]] command */
-#define PST_CONDEXPR 0x0200 /* parsing the guts of [[...]] */
-#define PST_ARITHFOR 0x0400 /* parsing an arithmetic for command */
-#define PST_ALEXPAND 0x0800 /* OK to expand aliases - unused */
-#define PST_CMDTOKEN 0x1000 /* command token OK - unused */
-#define PST_COMPASSIGN 0x2000 /* parsing x=(...) compound assignment */
-
/* Initial size to allocate for tokens, and the
amount to grow them by. */
#define TOKEN_DEFAULT_INITIAL_SIZE 496
# define expanding_alias() 0
#endif
-/* The token currently being read. */
-static int current_token;
-
-/* The last read token, or NULL. read_token () uses this for context
- checking. */
-static int last_read_token;
-
-/* The token read prior to last_read_token. */
-static int token_before_that;
-
-/* The token read prior to token_before_that. */
-static int two_tokens_ago;
-
-/* The current parser state. */
-static int parser_state;
-
/* Global var is non-zero when end of file has been reached. */
int EOF_Reached = 0;
old_sigint = (SigHandler *)set_signal_handler (SIGINT, sigint_sighandler);
interrupt_immediately++;
}
+ terminate_immediately = 1;
current_readline_line = readline (current_readline_prompt ?
current_readline_prompt : "");
+ terminate_immediately = 0;
if (signal_is_ignored (SIGINT) == 0 && old_sigint)
{
interrupt_immediately--;
init_yy_io (yy_string_get, yy_string_unget, st_string, name, location);
}
+/* Count the number of characters we've consumed from bash_input.location.string
+ and read into shell_input_line, but have not returned from shell_getc.
+ That is the true input location. Rewind bash_input.location.string by
+ that number of characters, so it points to the last character actually
+ consumed by the parser. */
+static void
+rewind_input_string ()
+{
+ int xchars;
+
+ /* number of unconsumed characters in the input -- XXX need to take newlines
+ into account, e.g., $(...\n) */
+ xchars = shell_input_line_len - shell_input_line_index;
+ if (bash_input.location.string[-1] == '\n')
+ xchars++;
+
+ /* XXX - how to reflect bash_input.location.string back to string passed to
+ parse_and_execute or xparse_dolparen? xparse_dolparen needs to know how
+ far into the string we parsed. parse_and_execute knows where bash_input.
+ location.string is, and how far from orig_string that is -- that's the
+ number of characters the command consumed. */
+
+ /* bash_input.location.string - xchars should be where we parsed to */
+ /* need to do more validation on xchars value for sanity -- test cases. */
+ bash_input.location.string -= xchars;
+}
+
/* **************************************************************** */
/* */
/* Let input come from STREAM. */
result = EOF;
if (bash_input.location.file)
- result = getc_with_restart (bash_input.location.file);
-
+ {
+ if (interactive)
+ {
+ interrupt_immediately++;
+ terminate_immediately++;
+ }
+ result = getc_with_restart (bash_input.location.file);
+ if (interactive)
+ {
+ interrupt_immediately--;
+ terminate_immediately--;
+ }
+ }
return (result);
}
pass_next = 0;
while (1)
{
- c = yy_getc ();
-
/* Allow immediate exit if interrupted during input. */
QUIT;
+ c = yy_getc ();
+
/* Ignore null bytes in input. */
if (c == 0)
{
read_secondary_line (remove_quoted_newline)
int remove_quoted_newline;
{
+ char *ret;
+ int n, c;
+
prompt_string_pointer = &ps2_prompt;
if (SHOULD_PROMPT())
prompt_again ();
- return (read_a_line (remove_quoted_newline));
+ ret = read_a_line (remove_quoted_newline);
+#if defined (HISTORY)
+ if (remember_on_history && (parser_state & PST_HEREDOC))
+ {
+ /* To make adding the the here-document body right, we need to rely
+ on history_delimiting_chars() returning \n for the first line of
+ the here-document body and the null string for the second and
+ subsequent lines, so we avoid double newlines.
+ current_command_line_count == 2 for the first line of the body. */
+
+ current_command_line_count++;
+ maybe_add_history (ret);
+ }
+#endif /* HISTORY */
+ return ret;
}
/* **************************************************************** */
{ "[[", COND_START },
{ "]]", COND_END },
#endif
+#if defined (COPROCESS_SUPPORT)
+ { "coproc", COPROC },
+#endif
{ (char *)NULL, 0}
};
{ "<&", LESS_AND },
{ ">&", GREATER_AND },
{ ";;", SEMI_SEMI },
+ { ";&", SEMI_AND },
+ { ";;&", SEMI_SEMI_AND },
{ "<<-", LESS_LESS_MINUS },
{ "<<<", LESS_LESS_LESS },
{ "&>", AND_GREATER },
+ { "&>>", AND_GREATER_GREATER },
{ "<>", LESS_GREATER },
{ ">|", GREATER_BAR },
+ { "|&", BAR_AND },
{ "EOF", yacc_EOF },
/* Tokens whose value is the character itself */
{ ">", '>' },
register int i;
int c;
unsigned char uc;
- static int mustpop = 0;
QUIT;
+ if (sigwinch_received)
+ {
+ sigwinch_received = 0;
+ get_new_window_size (0, (int *)0, (int *)0);
+ }
+
if (eol_ungetc_lookahead)
{
c = eol_ungetc_lookahead;
if (uc)
shell_input_line_index++;
- if MBTEST(uc == '\\' && remove_quoted_newline && shell_input_line[shell_input_line_index] == '\n')
- {
- if (SHOULD_PROMPT ())
- prompt_again ();
- line_number++;
- goto restart_read;
- }
-
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
/* If UC is NULL, we have reached the end of the current input string. If
pushed_string_list is non-empty, it's time to pop to the previous string
}
#endif /* ALIAS || DPAREN_ARITHMETIC */
+ if MBTEST(uc == '\\' && remove_quoted_newline && shell_input_line[shell_input_line_index] == '\n')
+ {
+ if (SHOULD_PROMPT ())
+ prompt_again ();
+ line_number++;
+ goto restart_read;
+ }
+
if (!uc && shell_input_line_terminator == EOF)
return ((shell_input_line_index != 0) ? '\n' : EOF);
}
void
-execute_prompt_command (command)
- char *command;
+execute_variable_command (command, vname)
+ char *command, *vname;
{
char *last_lastarg;
sh_parser_state_t ps;
if (last_lastarg)
last_lastarg = savestring (last_lastarg);
- parse_and_execute (savestring (command), "PROMPT_COMMAND", SEVAL_NONINT|SEVAL_NOHIST);
+ parse_and_execute (savestring (command), vname, SEVAL_NONINT|SEVAL_NOHIST);
restore_parser_state (&ps);
- bind_variable ("_", last_lastarg);
+ bind_variable ("_", last_lastarg, 0);
FREE (last_lastarg);
if (token_to_read == '\n') /* reset_parser was called */
token_before_that = last_read_token;
last_read_token = current_token;
current_token = read_token (READ);
+
+ if ((parser_state & PST_EOFTOKEN) && current_token == shell_eof_token)
+ {
+ current_token = yacc_EOF;
+ if (bash_input.type == st_string)
+ rewind_input_string ();
+ }
+ parser_state &= ~PST_EOFTOKEN;
+
return (current_token);
}
void
gather_here_documents ()
{
- int r = 0;
+ int r;
+
+ r = 0;
while (need_here_doc)
{
- make_here_document (redir_stack[r++]);
+ parser_state |= PST_HEREDOC;
+ make_here_document (redir_stack[r++], line_number);
+ parser_state &= ~PST_HEREDOC;
need_here_doc--;
}
}
#define command_token_position(token) \
(((token) == ASSIGNMENT_WORD) || \
- ((token) != SEMI_SEMI && reserved_word_acceptable(token)))
+ ((token) != SEMI_SEMI && (token) != SEMI_AND && (token) != SEMI_SEMI_AND && reserved_word_acceptable(token)))
#define assignment_acceptable(token) \
(command_token_position(token) && ((parser_state & PST_CASEPAT) == 0))
#endif /* ALIAS */
/* Read a single word from input. Start by skipping blanks. */
- while ((character = shell_getc (1)) != EOF && whitespace (character))
+ while ((character = shell_getc (1)) != EOF && shellblank (character))
;
if (character == EOF)
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
+ parser_state &= ~PST_ASSIGNOK;
+
return (character);
}
+ if (parser_state & PST_REGEXP)
+ goto tokword;
+
/* Shell meta-characters. */
if MBTEST(shellmeta (character) && ((parser_state & PST_DBLPAREN) == 0))
{
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
+ parser_state &= ~PST_ASSIGNOK;
+
peek_char = shell_getc (1);
if (character == peek_char)
{
/* If '<' then we could be at "<<" or at "<<-". We have to
look ahead one more character. */
peek_char = shell_getc (1);
- if (peek_char == '-')
+ if MBTEST(peek_char == '-')
return (LESS_LESS_MINUS);
- else if (peek_char == '<')
+ else if MBTEST(peek_char == '<')
return (LESS_LESS_LESS);
else
{
#if defined (ALIAS)
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
- return (SEMI_SEMI);
+
+ peek_char = shell_getc (1);
+ if MBTEST(peek_char == '&')
+ return (SEMI_SEMI_AND);
+ else
+ {
+ shell_ungetc (peek_char);
+ return (SEMI_SEMI);
+ }
case '&':
return (AND_AND);
return (LESS_GREATER);
else if MBTEST(character == '>' && peek_char == '|')
return (GREATER_BAR);
- else if MBTEST(peek_char == '>' && character == '&')
- return (AND_GREATER);
+ else if MBTEST(character == '&' && peek_char == '>')
+ {
+ peek_char = shell_getc (1);
+ if MBTEST(peek_char == '>')
+ return (AND_GREATER_GREATER);
+ else
+ {
+ shell_ungetc (peek_char);
+ return (AND_GREATER);
+ }
+ }
+ else if MBTEST(character == '|' && peek_char == '&')
+ return (BAR_AND);
+ else if MBTEST(character == ';' && peek_char == '&')
+ {
+ parser_state |= PST_CASEPAT;
+#if defined (ALIAS)
+ parser_state &= ~PST_ALEXPNEXT;
+#endif /* ALIAS */
+ return (SEMI_AND);
+ }
shell_ungetc (peek_char);
if MBTEST(character == '-' && (last_read_token == LESS_AND || last_read_token == GREATER_AND))
return (character);
+tokword:
/* Okay, if we got this far, we have to read a word. Read one,
and then check it against the known ones. */
result = read_token_word (character);
/*
* Match a $(...) or other grouping construct. This has to handle embedded
* quoted strings ('', ``, "") and nested constructs. It also must handle
- * reprompting the user, if necessary, after reading a newline (unless the
- * P_NONL flag is passed), and returning correct error values if it reads
- * EOF.
+ * reprompting the user, if necessary, after reading a newline, and returning
+ * correct error values if it reads EOF.
*/
#define P_FIRSTCLOSE 0x01
#define P_ALLOWESC 0x02
#define P_DQUOTE 0x04
+#define P_COMMAND 0x08 /* parsing a command, so look for comments */
+#define P_BACKQUOTE 0x10 /* parsing a backquoted command substitution */
+
+/* Lexical state while parsing a grouping construct or $(...). */
+#define LEX_WASDOL 0x001
+#define LEX_CKCOMMENT 0x002
+#define LEX_INCOMMENT 0x004
+#define LEX_PASSNEXT 0x008
+#define LEX_RESWDOK 0x010
+#define LEX_CKCASE 0x020
+#define LEX_INCASE 0x040
+#define LEX_INHEREDOC 0x080
+#define LEX_HEREDELIM 0x100 /* reading here-doc delimiter */
+#define LEX_STRIPDOC 0x200 /* <<- strip tabs from here doc delim */
+
+#define COMSUB_META(ch) ((ch) == ';' || (ch) == '&' || (ch) == '|')
+
+#define CHECK_NESTRET_ERROR() \
+ do { \
+ if (nestret == &matched_pair_error) \
+ { \
+ free (ret); \
+ return &matched_pair_error; \
+ } \
+ } while (0)
+
+#define APPEND_NESTRET() \
+ do { \
+ if (nestlen) \
+ { \
+ RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64); \
+ strcpy (ret + retind, nestret); \
+ retind += nestlen; \
+ } \
+ } while (0)
static char matched_pair_error;
+
static char *
parse_matched_pair (qc, open, close, lenp, flags)
int qc; /* `"' if this construct is within double quotes */
int open, close;
int *lenp, flags;
{
- int count, ch, was_dollar;
- int pass_next_character, nestlen, ttranslen, start_lineno;
+ int count, ch, tflags;
+ int nestlen, ttranslen, start_lineno;
char *ret, *nestret, *ttrans;
int retind, retsize, rflags;
+/* itrace("parse_matched_pair: open = %c close = %c flags = %d", open, close, flags); */
count = 1;
- pass_next_character = was_dollar = 0;
+ tflags = 0;
+
+ if ((flags & P_COMMAND) && qc != '`' && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0)
+ tflags |= LEX_CKCOMMENT;
/* RFLAGS is the set of flags we want to pass to recursive calls. */
rflags = (qc == '"') ? P_DQUOTE : (flags & P_DQUOTE);
start_lineno = line_number;
while (count)
{
- ch = shell_getc ((qc != '\'' || (flags & P_ALLOWESC)) && pass_next_character == 0);
+ ch = shell_getc (qc != '\'' && (tflags & LEX_PASSNEXT) == 0);
+
if (ch == EOF)
{
free (ret);
if (ch == '\n' && SHOULD_PROMPT ())
prompt_again ();
- if (pass_next_character) /* last char was backslash */
+ /* Don't bother counting parens or doing anything else if in a comment
+ or part of a case statement */
+ if (tflags & LEX_INCOMMENT)
{
- pass_next_character = 0;
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ if (ch == '\n')
+ tflags &= ~LEX_INCOMMENT;
+
+ continue;
+ }
+
+ /* Not exactly right yet, should handle shell metacharacters, too. If
+ any changes are made to this test, make analogous changes to subst.c:
+ extract_delimited_string(). */
+ else if MBTEST((tflags & LEX_CKCOMMENT) && (tflags & LEX_INCOMMENT) == 0 && ch == '#' && (retind == 0 || ret[retind-1] == '\n' || shellblank (ret[retind - 1])))
+ tflags |= LEX_INCOMMENT;
+
+ if (tflags & LEX_PASSNEXT) /* last char was backslash */
+ {
+ tflags &= ~LEX_PASSNEXT;
if (qc != '\'' && ch == '\n') /* double-quoted \<newline> disappears. */
{
- if (retind > 0) retind--; /* swallow previously-added backslash */
+ if (retind > 0)
+ retind--; /* swallow previously-added backslash */
continue;
}
ret[retind++] = ch;
continue;
}
+ /* If we're reparsing the input (e.g., from parse_string_to_word_list),
+ we've already prepended CTLESC to single-quoted results of $'...'.
+ We may want to do this for other CTLESC-quoted characters in
+ reparse, too. */
+ else if MBTEST((parser_state & PST_REPARSE) && open == '\'' && (ch == CTLESC || ch == CTLNUL))
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ continue;
+ }
else if MBTEST(ch == CTLESC || ch == CTLNUL) /* special shell escapes */
{
RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
}
else if MBTEST(ch == close) /* ending delimiter */
count--;
-#if 1
/* handle nested ${...} specially. */
- else if MBTEST(open != close && was_dollar && open == '{' && ch == open) /* } */
+ else if MBTEST(open != close && (tflags & LEX_WASDOL) && open == '{' && ch == open) /* } */
count++;
-#endif
else if MBTEST(((flags & P_FIRSTCLOSE) == 0) && ch == open) /* nested begin */
count++;
RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
ret[retind++] = ch;
+ /* If we just read the ending character, don't bother continuing. */
+ if (count == 0)
+ break;
+
if (open == '\'') /* '' inside grouping construct */
{
if MBTEST((flags & P_ALLOWESC) && ch == '\\')
- pass_next_character++;
+ tflags |= LEX_PASSNEXT;
continue;
}
if MBTEST(ch == '\\') /* backslashes */
- pass_next_character++;
+ tflags |= LEX_PASSNEXT;
+
+#if 0
+ /* The big hammer. Single quotes aren't special in double quotes. The
+ problem is that Posix says the single quotes are semi-special:
+ within a double-quoted ${...} construct "an even number of
+ unescaped double-quotes or single-quotes, if any, shall occur." */
+ if MBTEST(open == '{' && (flags & P_DQUOTE) && ch == '\'') /* } */
+ continue;
+#endif
+ /* Could also check open == '`' if we want to parse grouping constructs
+ inside old-style command substitution. */
if (open != close) /* a grouping construct */
{
if MBTEST(shellquote (ch))
{
/* '', ``, or "" inside $(...) or other grouping construct. */
push_delimiter (dstack, ch);
- if MBTEST(was_dollar && ch == '\'') /* $'...' inside group */
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'') /* $'...' inside group */
nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
else
nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
pop_delimiter (dstack);
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
- }
- if MBTEST(was_dollar && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
+ CHECK_NESTRET_ERROR ();
+
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
/* Translate $'...' here. */
ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
xfree (nestret);
- nestret = sh_single_quote (ttrans);
- free (ttrans);
- nestlen = strlen (nestret);
+
+ if ((rflags & P_DQUOTE) == 0)
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else
+ {
+ nestret = ttrans;
+ nestlen = ttranslen;
+ }
retind -= 2; /* back up before the $' */
}
- else if MBTEST(was_dollar && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
+ else if MBTEST((tflags & LEX_WASDOL) && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
/* Locale expand $"..." here. */
ttrans = localeexpand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
xfree (nestret);
- nestret = (char *)xmalloc (ttranslen + 3);
- nestret[0] = '"';
- strcpy (nestret + 1, ttrans);
- nestret[ttranslen + 1] = '"';
- nestret[ttranslen += 2] = '\0';
+
+ nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
free (ttrans);
- nestlen = ttranslen;
+ nestlen = ttranslen + 2;
retind -= 2; /* back up before the $" */
}
- if (nestlen)
- {
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
- }
+ APPEND_NESTRET ();
FREE (nestret);
}
}
else if MBTEST(open == '"' && ch == '`')
{
nestret = parse_matched_pair (0, '`', '`', &nestlen, rflags);
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
- }
- if (nestlen)
- {
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
- }
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
FREE (nestret);
}
- else if MBTEST(was_dollar && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ else if MBTEST(open != '`' && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
/* check for $(), $[], or ${} inside quoted string. */
{
if (open == ch) /* undo previous increment */
count--;
if (ch == '(') /* ) */
- nestret = parse_matched_pair (0, '(', ')', &nestlen, rflags);
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
else if (ch == '{') /* } */
nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|rflags);
else if (ch == '[') /* ] */
nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
- if (nestret == &matched_pair_error)
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ if MBTEST(ch == '$')
+ tflags |= LEX_WASDOL;
+ else
+ tflags &= ~LEX_WASDOL;
+ }
+
+ ret[retind] = '\0';
+ if (lenp)
+ *lenp = retind;
+ return ret;
+}
+
+/* Parse a $(...) command substitution. This is messier than I'd like, and
+ reproduces a lot more of the token-reading code than I'd like. */
+static char *
+parse_comsub (qc, open, close, lenp, flags)
+ int qc; /* `"' if this construct is within double quotes */
+ int open, close;
+ int *lenp, flags;
+{
+ int count, ch, peekc, tflags, lex_rwlen, lex_firstind;
+ int nestlen, ttranslen, start_lineno;
+ char *ret, *nestret, *ttrans, *heredelim;
+ int retind, retsize, rflags, hdlen;
+
+/*itrace("parse_comsub: qc = `%c' open = %c close = %c", qc, open, close);*/
+ count = 1;
+ tflags = LEX_RESWDOK;
+
+ if ((flags & P_COMMAND) && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0)
+ tflags |= LEX_CKCASE;
+ if ((tflags & LEX_CKCASE) && (interactive == 0 || interactive_comments))
+ tflags |= LEX_CKCOMMENT;
+
+ /* RFLAGS is the set of flags we want to pass to recursive calls. */
+ rflags = (flags & P_DQUOTE);
+
+ ret = (char *)xmalloc (retsize = 64);
+ retind = 0;
+
+ start_lineno = line_number;
+ lex_rwlen = 0;
+
+ heredelim = 0;
+ lex_firstind = -1;
+
+ while (count)
+ {
+comsub_readchar:
+ ch = shell_getc (qc != '\'' && (tflags & LEX_PASSNEXT) == 0);
+
+ if (ch == EOF)
+ {
+eof_error:
+ free (ret);
+ FREE (heredelim);
+ parser_error (start_lineno, _("unexpected EOF while looking for matching `%c'"), close);
+ EOF_Reached = 1; /* XXX */
+ return (&matched_pair_error);
+ }
+
+ /* If we hit the end of a line and are reading the contents of a here
+ document, and it's not the same line that the document starts on,
+ check for this line being the here doc delimiter. Otherwise, if
+ we're in a here document, mark the next character as the beginning
+ of a line. */
+ if (ch == '\n')
+ {
+ if ((tflags & LEX_HEREDELIM) && heredelim)
+ {
+ tflags &= ~LEX_HEREDELIM;
+ tflags |= LEX_INHEREDOC;
+ lex_firstind = retind + 1;
+ }
+ else if (tflags & LEX_INHEREDOC)
+ {
+ int tind;
+ tind = lex_firstind;
+ while ((tflags & LEX_STRIPDOC) && ret[tind] == '\t')
+ tind++;
+ if (STREQN (ret + tind, heredelim, hdlen))
+ {
+ tflags &= ~(LEX_STRIPDOC|LEX_INHEREDOC);
+/*itrace("parse_comsub:%d: found here doc end `%s'", line_number, ret + tind);*/
+ lex_firstind = -1;
+ }
+ else
+ lex_firstind = retind + 1;
+ }
+ }
+
+ /* Possible reprompting. */
+ if (ch == '\n' && SHOULD_PROMPT ())
+ prompt_again ();
+
+ /* Don't bother counting parens or doing anything else if in a comment */
+ if (tflags & (LEX_INCOMMENT|LEX_INHEREDOC))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ if ((tflags & LEX_INCOMMENT) && ch == '\n')
+ tflags &= ~LEX_INCOMMENT;
+
+ continue;
+ }
+
+ /* Skip whitespace */
+ if MBTEST(shellblank (ch) && lex_rwlen == 0)
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ continue;
+ }
+
+ /* Either we are looking for the start of the here-doc delimiter
+ (lex_firstind == -1) or we are reading one (lex_firstind >= 0).
+ If this character is a shell break character and we are reading
+ the delimiter, save it and note that we are now reading a here
+ document. If we've found the start of the delimiter, note it by
+ setting lex_firstind. Backslashes can quote shell metacharacters
+ in here-doc delimiters. */
+ if (tflags & LEX_HEREDELIM)
+ {
+ if (lex_firstind == -1 && shellbreak (ch) == 0)
+ lex_firstind = retind;
+ else if (lex_firstind >= 0 && (tflags & LEX_PASSNEXT) == 0 && shellbreak (ch))
+ {
+ nestret = substring (ret, lex_firstind, retind);
+ heredelim = string_quote_removal (nestret, 0);
+ free (nestret);
+ hdlen = STRLEN(heredelim);
+/*itrace("parse_comsub:%d: found here doc delimiter `%s' (%d)", line_number, heredelim, hdlen);*/
+ if (ch == '\n')
+ {
+ tflags |= LEX_INHEREDOC;
+ tflags &= ~LEX_HEREDELIM;
+ lex_firstind = retind + 1;
+ }
+ else
+ lex_firstind = -1;
+ }
+ }
+
+ /* Meta-characters that can introduce a reserved word. Not perfect yet. */
+ if MBTEST((tflags & LEX_RESWDOK) == 0 && (tflags & LEX_CKCASE) && (tflags & LEX_INCOMMENT) == 0 && shellmeta(ch))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ peekc = shell_getc (1);
+ if (ch == peekc && (ch == '&' || ch == '|' || ch == ';')) /* two-character tokens */
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+/*itrace("parse_comsub:%d: set lex_reswordok = 1, ch = `%c'", line_number, ch); */
+ tflags |= LEX_RESWDOK;
+ lex_rwlen = 0;
+ continue;
+ }
+ else if (ch == '\n' || COMSUB_META(ch))
+ {
+ shell_ungetc (peekc);
+ tflags |= LEX_RESWDOK;
+/*itrace("parse_comsub:%d: set lex_reswordok = 1, ch = `%c'", line_number, ch);*/
+ lex_rwlen = 0;
+ continue;
+ }
+ else if (ch == EOF)
+ goto eof_error;
+ else
+ {
+ /* `unget' the character we just added and fall through */
+ retind--;
+ shell_ungetc (peekc);
+ }
+ }
+
+ /* If we can read a reserved word, try to read one. */
+ if (tflags & LEX_RESWDOK)
+ {
+ if MBTEST(islower (ch))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ lex_rwlen++;
+ continue;
+ }
+ else if MBTEST(lex_rwlen == 4 && shellbreak (ch))
+ {
+ if (STREQN (ret + retind - 4, "case", 4))
+{
+ tflags |= LEX_INCASE;
+/*itrace("parse_comsub:%d: found `case', lex_incase -> 1", line_number);*/
+}
+ else if (STREQN (ret + retind - 4, "esac", 4))
+{
+ tflags &= ~LEX_INCASE;
+/*itrace("parse_comsub:%d: found `esac', lex_incase -> 0", line_number);*/
+}
+ tflags &= ~LEX_RESWDOK;
+ }
+ else if (shellbreak (ch) == 0)
+{
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%c', lex_reswordok -> 0", line_number, ch);*/
+}
+ }
+
+ if MBTEST((tflags & LEX_INCOMMENT) == 0 && (tflags & LEX_CKCASE) && ch == '<')
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ peekc = shell_getc (1);
+ if (peekc == EOF)
+ goto eof_error;
+ if (peekc == ch)
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+ peekc = shell_getc (1);
+ if (peekc == EOF)
+ goto eof_error;
+ if (peekc == '-')
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+ tflags |= LEX_STRIPDOC;
+ }
+ else
+ shell_ungetc (peekc);
+ tflags |= LEX_HEREDELIM;
+ lex_firstind = -1;
+ continue;
+ }
+ else
+ ch = peekc; /* fall through and continue XXX - this skips comments if peekc == '#' */
+ }
+ /* Not exactly right yet, should handle shell metacharacters, too. If
+ any changes are made to this test, make analogous changes to subst.c:
+ extract_delimited_string(). */
+ else if MBTEST((tflags & LEX_CKCOMMENT) && (tflags & LEX_INCOMMENT) == 0 && ch == '#' && (retind == 0 || ret[retind-1] == '\n' || shellblank (ret[retind - 1])))
+ tflags |= LEX_INCOMMENT;
+
+ if (tflags & LEX_PASSNEXT) /* last char was backslash */
+ {
+ tflags &= ~LEX_PASSNEXT;
+ if (qc != '\'' && ch == '\n') /* double-quoted \<newline> disappears. */
+ {
+ if (retind > 0)
+ retind--; /* swallow previously-added backslash */
+ continue;
+ }
+
+ RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
+ if MBTEST(ch == CTLESC || ch == CTLNUL)
+ ret[retind++] = CTLESC;
+ ret[retind++] = ch;
+ continue;
+ }
+ else if MBTEST(ch == CTLESC || ch == CTLNUL) /* special shell escapes */
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
+ ret[retind++] = CTLESC;
+ ret[retind++] = ch;
+ continue;
+ }
+#if 0
+ else if MBTEST((tflags & LEX_INCASE) && ch == close && close == ')')
+ tflags &= ~LEX_INCASE; /* XXX */
+#endif
+ else if MBTEST(ch == close && (tflags & LEX_INCASE) == 0) /* ending delimiter */
+{
+ count--;
+/*itrace("parse_comsub:%d: found close: count = %d", line_number, count);*/
+}
+ else if MBTEST(((flags & P_FIRSTCLOSE) == 0) && (tflags & LEX_INCASE) == 0 && ch == open) /* nested begin */
+ count++;
+
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ /* If we just read the ending character, don't bother continuing. */
+ if (count == 0)
+ break;
+
+ if MBTEST(ch == '\\') /* backslashes */
+ tflags |= LEX_PASSNEXT;
+
+ if MBTEST(shellquote (ch))
+ {
+ /* '', ``, or "" inside $(...). */
+ push_delimiter (dstack, ch);
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'') /* $'...' inside group */
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
+ else
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
+ pop_delimiter (dstack);
+ CHECK_NESTRET_ERROR ();
+
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
- free (ret);
- return &matched_pair_error;
+ /* Translate $'...' here. */
+ ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
+ xfree (nestret);
+
+ if ((rflags & P_DQUOTE) == 0)
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else
+ {
+ nestret = ttrans;
+ nestlen = ttranslen;
+ }
+ retind -= 2; /* back up before the $' */
}
- if (nestlen)
+ else if MBTEST((tflags & LEX_WASDOL) && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
+ /* Locale expand $"..." here. */
+ ttrans = localeexpand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
+ xfree (nestret);
+
+ nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = ttranslen + 2;
+ retind -= 2; /* back up before the $" */
}
+
+ APPEND_NESTRET ();
FREE (nestret);
}
- was_dollar = MBTEST(ch == '$');
+ else if MBTEST((tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ /* check for $(), $[], or ${} inside command substitution. */
+ {
+ if ((tflags & LEX_INCASE) == 0 && open == ch) /* undo previous increment */
+ count--;
+ if (ch == '(') /* ) */
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
+ else if (ch == '{') /* } */
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|rflags);
+ else if (ch == '[') /* ] */
+ nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ if MBTEST(ch == '$')
+ tflags |= LEX_WASDOL;
+ else
+ tflags &= ~LEX_WASDOL;
}
+ FREE (heredelim);
ret[retind] = '\0';
if (lenp)
*lenp = retind;
+/*itrace("parse_comsub:%d: returning `%s'", line_number, ret);*/
+ return ret;
+}
+
+/* XXX - this needs to handle functionality like subst.c:no_longjmp_on_fatal_error;
+ maybe extract_command_subst should handle it. */
+char *
+xparse_dolparen (base, string, indp, flags)
+ char *base;
+ char *string;
+ int *indp;
+ int flags;
+{
+ sh_parser_state_t ps;
+ int orig_ind, nc, sflags;
+ char *ret, *s, *ep, *ostring;
+
+ /*yydebug = 1;*/
+ orig_ind = *indp;
+ ostring = string;
+
+ sflags = SEVAL_NONINT|SEVAL_NOHIST|SEVAL_NOFREE;
+ if (flags & SX_NOLONGJMP)
+ sflags |= SEVAL_NOLONGJMP;
+ save_parser_state (&ps);
+
+ /*(*/
+ parser_state |= PST_CMDSUBST|PST_EOFTOKEN; /* allow instant ')' */ /*(*/
+ shell_eof_token = ')';
+ parse_string (string, "command substitution", sflags, &ep);
+
+ restore_parser_state (&ps);
+ reset_parser ();
+ if (interactive)
+ token_to_read = 0;
+
+ /* Need to find how many characters parse_and_execute consumed, update
+ *indp, if flags != 0, copy the portion of the string parsed into RET
+ and return it. If flags & 1 (EX_NOALLOC) we can return NULL. */
+
+ /*(*/
+ if (ep[-1] != ')')
+ {
+#if DEBUG
+ if (ep[-1] != '\n')
+ itrace("xparse_dolparen:%d: ep[-1] != RPAREN (%d), ep = `%s'", line_number, ep[-1], ep);
+#endif
+ while (ep > ostring && ep[-1] == '\n') ep--;
+ }
+
+ nc = ep - ostring;
+ *indp = ep - base - 1;
+
+ /*(*/
+#if DEBUG
+ if (base[*indp] != ')')
+ itrace("xparse_dolparen:%d: base[%d] != RPAREN (%d), base = `%s'", line_number, *indp, base[*indp], base);
+#endif
+
+ if (flags & SX_NOALLOC)
+ return (char *)NULL;
+
+ if (nc == 0)
+ {
+ ret = xmalloc (1);
+ ret[0] = '\0';
+ }
+ else
+ ret = substring (ostring, 0, nc - 1);
+
return ret;
}
parse_dparen (c)
int c;
{
- int cmdtyp, len, sline;
- char *wval, *wv2;
+ int cmdtyp, sline;
+ char *wval;
WORD_DESC *wd;
#if defined (ARITH_FOR_COMMAND)
cmdtyp = parse_arith_cmd (&wval, 0);
if (cmdtyp == 1)
{
- wd = make_word (wval);
+ wd = alloc_word_desc ();
+ wd->word = wval;
yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
- free (wval);
return (ARITH_FOR_EXPRS);
}
else
if (reserved_word_acceptable (last_read_token))
{
sline = line_number;
-#if 0
- cmdtyp = parse_arith_cmd (&wval, 1);
-#else
+
cmdtyp = parse_arith_cmd (&wval, 0);
-#endif
if (cmdtyp == 1) /* arithmetic command */
{
- wd = make_word (wval);
-#if 0
- wd->flags = W_QUOTED;
-#else
- wd->flags = W_QUOTED|W_NOSPLIT|W_NOGLOB;
-#endif
+ wd = alloc_word_desc ();
+ wd->word = wval;
+ wd->flags = W_QUOTED|W_NOSPLIT|W_NOGLOB|W_DQUOTE;
yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
- free (wval); /* make_word copies it */
return (ARITH_CMD);
}
else if (cmdtyp == 0) /* nested subshell */
if (term)
term->flags |= CMD_INVERT_RETURN;
}
- else if (tok == WORD && test_unop (yylval.word->word))
+ else if (tok == WORD && yylval.word->word[0] == '-' && yylval.word->word[2] == 0 && test_unop (yylval.word->word))
{
op = yylval.word;
tok = read_token (READ);
if (tok == WORD && test_binop (yylval.word->word))
op = yylval.word;
#if defined (COND_REGEXP)
- else if (tok == WORD && STREQ (yylval.word->word,"=~"))
- op = yylval.word;
+ else if (tok == WORD && STREQ (yylval.word->word, "=~"))
+ {
+ op = yylval.word;
+ parser_state |= PST_REGEXP;
+ }
#endif
else if (tok == '<' || tok == '>')
op = make_word_from_token (tok); /* ( */
/* rhs */
tok = read_token (READ);
+ parser_state &= ~PST_REGEXP;
if (tok == WORD)
{
tright = make_cond_node (COND_TERM, yylval.word, (COND_COM *)NULL, (COND_COM *)NULL);
return r;
}
+/* XXX - possible changes here for `+=' */
static int
token_is_ident (t, i)
char *t;
/* DOLLAR_PRESENT becomes non-zero if we see a `$'. */
int dollar_present;
+ /* COMPOUND_ASSIGNMENT becomes non-zero if we are parsing a compound
+ assignment. */
+ int compound_assignment;
+
/* QUOTED becomes non-zero if we see one of ("), ('), (`), or (\). */
int quoted;
token_index = 0;
all_digit_token = DIGIT (character);
- dollar_present = quoted = pass_next_character = 0;
+ dollar_present = quoted = pass_next_character = compound_assignment = 0;
for (;;)
{
if (pass_next_character)
{
pass_next_character = 0;
- goto got_character;
+ goto got_escaped_character;
}
cd = current_delimiter (dstack);
if MBTEST(shellquote (character))
{
push_delimiter (dstack, character);
- ttok = parse_matched_pair (character, character, character, &ttoklen, 0);
+ ttok = parse_matched_pair (character, character, character, &ttoklen, (character == '`') ? P_COMMAND : 0);
pop_delimiter (dstack);
if (ttok == &matched_pair_error)
return -1; /* Bail immediately. */
goto next_character;
}
+#ifdef COND_REGEXP
+ /* When parsing a regexp as a single word inside a conditional command,
+ we need to special-case characters special to both the shell and
+ regular expressions. Right now, that is only '(' and '|'. */ /*)*/
+ if MBTEST((parser_state & PST_REGEXP) && (character == '(' || character == '|')) /*)*/
+ {
+ if (character == '|')
+ goto got_character;
+
+ push_delimiter (dstack, character);
+ ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
+ pop_delimiter (dstack);
+ if (ttok == &matched_pair_error)
+ return -1; /* Bail immediately. */
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ token_buffer_size, TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = character;
+ strcpy (token + token_index, ttok);
+ token_index += ttoklen;
+ FREE (ttok);
+ dollar_present = all_digit_token = 0;
+ goto next_character;
+ }
+#endif /* COND_REGEXP */
+
#ifdef EXTENDED_GLOB
/* Parse a ksh-style extended pattern matching specification. */
- if (extended_glob && PATTERN_CHAR (character))
+ if MBTEST(extended_glob && PATTERN_CHAR (character))
{
peek_char = shell_getc (1);
if MBTEST(peek_char == '(') /* ) */
history literally rather than causing a possibly-
incorrect `;' to be added. ) */
push_delimiter (dstack, peek_char);
- ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
+ ttok = parse_comsub (cd, '(', ')', &ttoklen, P_COMMAND);
pop_delimiter (dstack);
}
else
{
ttrans = ansiexpand (ttok, 0, ttoklen - 1, &ttranslen);
free (ttok);
+
/* Insert the single quotes and correctly quote any
embedded single quotes (allowed because P_ALLOWESC was
passed to parse_matched_pair). */
ttok = sh_single_quote (ttrans);
free (ttrans);
+ ttranslen = strlen (ttok);
ttrans = ttok;
- ttranslen = strlen (ttrans);
}
else
{
free (ttok);
/* Add the double quotes back */
- ttok = (char *)xmalloc (ttranslen + 3);
- ttok[0] = '"';
- strcpy (ttok + 1, ttrans);
- ttok[ttranslen + 1] = '"';
- ttok[ttranslen += 2] = '\0';
+ ttok = sh_mkdoublequoted (ttrans, ttranslen, 0);
free (ttrans);
+ ttranslen += 2;
ttrans = ttok;
}
}
#if defined (ARRAY_VARS)
- /* Identify possible array subscript assignment; match [...] */
- else if MBTEST(character == '[' && token_index > 0 && assignment_acceptable (last_read_token) && token_is_ident (token, token_index)) /* ] */
+ /* Identify possible array subscript assignment; match [...]. If
+ parser_state&PST_COMPASSIGN, we need to parse [sub]=words treating
+ `sub' as if it were enclosed in double quotes. */
+ else if MBTEST(character == '[' && /* ] */
+ ((token_index > 0 && assignment_acceptable (last_read_token) && token_is_ident (token, token_index)) ||
+ (token_index == 0 && (parser_state&PST_COMPASSIGN))))
{
ttok = parse_matched_pair (cd, '[', ']', &ttoklen, 0);
if (ttok == &matched_pair_error)
goto next_character;
}
/* Identify possible compound array variable assignment. */
- else if MBTEST(character == '=' && token_index > 0 && token_is_assignment (token, token_index))
+ else if MBTEST(character == '=' && token_index > 0 && (assignment_acceptable (last_read_token) || (parser_state & PST_ASSIGNOK)) && token_is_assignment (token, token_index))
{
peek_char = shell_getc (1);
if MBTEST(peek_char == '(') /* ) */
token[token_index++] = ')';
FREE (ttok);
all_digit_token = 0;
+ compound_assignment = 1;
+#if 1
goto next_character;
+#else
+ goto got_token; /* ksh93 seems to do this */
+#endif
}
else
shell_ungetc (peek_char);
got_character:
- all_digit_token &= DIGIT (character);
- dollar_present |= character == '$';
-
if (character == CTLESC || character == CTLNUL)
token[token_index++] = CTLESC;
+ got_escaped_character:
+
+ all_digit_token &= DIGIT (character);
+ dollar_present |= character == '$';
+
token[token_index++] = character;
RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
if (dollar_present)
the_word->flags |= W_HASDOLLAR;
if (quoted)
- the_word->flags |= W_QUOTED;
+ the_word->flags |= W_QUOTED; /*(*/
+ if (compound_assignment && token[token_index-1] == ')')
+ the_word->flags |= W_COMPASSIGN;
/* A word is an assignment if it appears at the beginning of a
simple command, or after another assignment word. This is
context-dependent, so it cannot be handled in the grammar. */
the_word->flags |= W_NOSPLIT;
}
+ if (command_token_position (last_read_token))
+ {
+ struct builtin *b;
+ b = builtin_address_internal (token, 0);
+ if (b && (b->flags & ASSIGNMENT_BUILTIN))
+ parser_state |= PST_ASSIGNOK;
+ else if (STREQ (token, "eval") || STREQ (token, "let"))
+ parser_state |= PST_ASSIGNOK;
+ }
+
yylval.word = the_word;
result = ((the_word->flags & (W_ASSIGNMENT|W_NOSPLIT)) == (W_ASSIGNMENT|W_NOSPLIT))
case IF:
case OR_OR:
case SEMI_SEMI:
+ case SEMI_AND:
+ case SEMI_SEMI_AND:
case THEN:
case TIME:
case TIMEOPT:
+ case COPROC:
case UNTIL:
case WHILE:
case 0:
return 1;
default:
+#if defined (COPROCESS_SUPPORT)
+ if (last_read_token == WORD && token_before_that == COPROC)
+ return 1;
+#endif
return 0;
}
}
/* A list of tokens which can be followed by newlines, but not by
semi-colons. When concatenating multiple lines of history, the
newline separator for such tokens is replaced with a space. */
-static int no_semi_successors[] = {
+static const int no_semi_successors[] = {
'\n', '{', '(', ')', ';', '&', '|',
- CASE, DO, ELSE, IF, SEMI_SEMI, THEN, UNTIL, WHILE, AND_AND, OR_OR, IN,
+ CASE, DO, ELSE, IF, SEMI_SEMI, SEMI_AND, SEMI_SEMI_AND, THEN, UNTIL,
+ WHILE, AND_AND, OR_OR, IN,
0
};
if (dstack.delimiter_depth != 0)
return ("\n");
-
+
+ /* We look for current_command_line_count == 2 because we are looking to
+ add the first line of the body of the here document (the second line
+ of the command). */
+ if (parser_state & PST_HEREDOC)
+ return (current_command_line_count == 2 ? "\n" : "");
+
/* First, handle some special cases. */
/*(*/
/* If we just read `()', assume it's a function definition, and don't
{
/* Tricky. `for i\nin ...' should not have a semicolon, but
`for i\ndo ...' should. We do what we can. */
- for (i = shell_input_line_index; whitespace(shell_input_line[i]); i++)
+ for (i = shell_input_line_index; whitespace (shell_input_line[i]); i++)
;
if (shell_input_line[i] && shell_input_line[i] == 'i' && shell_input_line[i+1] == 'n')
return " ";
return ";";
}
+ else if (two_tokens_ago == CASE && token_before_that == WORD && (parser_state & PST_CASESTMT))
+ return " ";
for (i = 0; no_semi_successors[i]; i++)
{
int last_exit_value;
#if defined (PROMPT_STRING_DECODE)
int result_size, result_index;
- int c, n;
+ int c, n, i;
char *temp, octal_string[4];
struct tm *tm;
time_t the_time;
case 'W':
{
/* Use the value of PWD because it is much more efficient. */
- char t_string[PATH_MAX], *t;
+ char t_string[PATH_MAX];
int tlen;
temp = get_string_value ("PWD");
#define ROOT_PATH(x) ((x)[0] == '/' && (x)[1] == 0)
#define DOUBLE_SLASH_ROOT(x) ((x)[0] == '/' && (x)[1] == '/' && (x)[2] == 0)
/* Abbreviate \W as ~ if $PWD == $HOME */
- if (c == 'W' && (((t = get_string_value ("HOME")) == 0) || STREQ (t, temp) == 0))
+ if (c == 'W' && (((t = get_string_value ("HOME")) == 0) || STREQ (t, t_string) == 0))
{
if (ROOT_PATH (t_string) == 0 && DOUBLE_SLASH_ROOT (t_string) == 0)
{
no longer than PATH_MAX - 1 characters. */
strcpy (t_string, polite_directory_format (t_string));
+ temp = trim_pathname (t_string, PATH_MAX - 1);
/* If we're going to be expanding the prompt string later,
quote the directory name. */
if (promptvars || posixly_correct)
break;
}
temp = (char *)xmalloc (3);
- temp[0] = '\001';
- temp[1] = (c == '[') ? RL_PROMPT_START_IGNORE : RL_PROMPT_END_IGNORE;
- temp[2] = '\0';
+ n = (c == '[') ? RL_PROMPT_START_IGNORE : RL_PROMPT_END_IGNORE;
+ i = 0;
+ if (n == CTLESC || n == CTLNUL)
+ temp[i++] = CTLESC;
+ temp[i++] = n;
+ temp[i] = '\0';
goto add_string;
#endif /* READLINE */
if (promptvars || posixly_correct)
{
last_exit_value = last_command_exit_value;
- list = expand_prompt_string (result, Q_DOUBLE_QUOTES);
+ list = expand_prompt_string (result, Q_DOUBLE_QUOTES, 0);
free (result);
result = string_list (list);
dispose_words (list);
}
static char *
-error_token_from_token (token)
- int token;
+error_token_from_token (tok)
+ int tok;
{
char *t;
- if (t = find_token_in_alist (token, word_token_alist, 0))
+ if (t = find_token_in_alist (tok, word_token_alist, 0))
return t;
- if (t = find_token_in_alist (token, other_token_alist, 0))
+ if (t = find_token_in_alist (tok, other_token_alist, 0))
return t;
t = (char *)NULL;
wl = (WORD_LIST *)NULL;
if (flags & 1)
- parser_state |= PST_COMPASSIGN;
+ parser_state |= PST_COMPASSIGN|PST_REPARSE;
while ((tok = read_token (READ)) != yacc_EOF)
{
line_number = orig_line_number + line_number - 1;
orig_current_token = current_token;
current_token = tok;
- yyerror ((char *)NULL); /* does the right thing */
+ yyerror (NULL); /* does the right thing */
current_token = orig_current_token;
if (wl)
dispose_words (wl);
shell_input_line_terminator = orig_input_terminator;
if (flags & 1)
- parser_state &= ~PST_COMPASSIGN;
+ parser_state &= ~(PST_COMPASSIGN|PST_REPARSE);
if (wl == &parse_string_error)
{
int *retlenp;
{
WORD_LIST *wl, *rl;
- int tok, orig_line_number, orig_token_size;
+ int tok, orig_line_number, orig_token_size, orig_last_token, assignok;
char *saved_token, *ret;
saved_token = token;
orig_token_size = token_buffer_size;
orig_line_number = line_number;
+ orig_last_token = last_read_token;
last_read_token = WORD; /* WORD to allow reserved words here */
token = (char *)NULL;
token_buffer_size = 0;
+ assignok = parser_state&PST_ASSIGNOK; /* XXX */
+
wl = (WORD_LIST *)NULL; /* ( */
parser_state |= PST_COMPASSIGN;
if (tok == yacc_EOF) /* ( */
parser_error (orig_line_number, _("unexpected EOF while looking for matching `)'"));
else
- yyerror ((char *)NULL); /* does the right thing */
+ yyerror(NULL); /* does the right thing */
if (wl)
dispose_words (wl);
wl = &parse_string_error;
jump_to_top_level (DISCARD);
}
- last_read_token = WORD;
+ last_read_token = orig_last_token; /* XXX - was WORD? */
+
if (wl)
{
rl = REVERSE_LIST (wl, WORD_LIST *);
if (retlenp)
*retlenp = (ret && *ret) ? strlen (ret) : 0;
+
+ if (assignok)
+ parser_state |= PST_ASSIGNOK;
+
return ret;
}
#endif
if (ps == 0)
- ps = xmalloc (sizeof (sh_parser_state_t));
+ ps = (sh_parser_state_t *)xmalloc (sizeof (sh_parser_state_t));
if (ps == 0)
return ((sh_parser_state_t *)NULL);
projects / platform / upstream / bash.git / blobdiff