-/* Yacc grammar for bash. */
+/* parse.y - Yacc grammar for bash. */
-/* Copyright (C) 1989 Free Software Foundation, Inc.
+/* Copyright (C) 1989-2012 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 1, 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, 675 Mass Ave, Cambridge, MA 02139, 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 "bashtypes.h"
#include "bashansi.h"
+#include "filecntl.h"
+
#if defined (HAVE_UNISTD_H)
# include <unistd.h>
#endif
#endif
#include <stdio.h>
+#include "chartypes.h"
#include <signal.h>
#include "memalloc.h"
+#include "bashintl.h"
+
+#define NEED_STRFTIME_DECL /* used in externs.h */
+
#include "shell.h"
+#include "typemax.h" /* SIZE_MAX if needed */
#include "trap.h"
#include "flags.h"
#include "parser.h"
#include "mailcheck.h"
+#include "test.h"
+#include "builtins.h"
#include "builtins/common.h"
#include "builtins/builtext.h"
+#include "shmbutil.h"
+
#if defined (READLINE)
# include "bashline.h"
# include <readline/readline.h>
#if defined (ALIAS)
# include "alias.h"
+#else
+typedef void *alias_t;
#endif /* ALIAS */
#if defined (PROMPT_STRING_DECODE)
# include <sys/param.h>
# endif
# include <time.h>
+# if defined (TM_IN_SYS_TIME)
+# include <sys/types.h>
+# include <sys/time.h>
+# endif /* TM_IN_SYS_TIME */
# include "maxpath.h"
#endif /* PROMPT_STRING_DECODE */
#define RE_READ_TOKEN -99
#define NO_EXPANSION -100
-#define YYDEBUG 0
+#ifdef DEBUG
+# define YYDEBUG 1
+#else
+# define YYDEBUG 0
+#endif
+
+#if defined (HANDLE_MULTIBYTE)
+# define last_shell_getc_is_singlebyte \
+ ((shell_input_line_index > 1) \
+ ? shell_input_line_property[shell_input_line_index - 1] \
+ : 1)
+# define MBTEST(x) ((x) && last_shell_getc_is_singlebyte)
+#else
+# define last_shell_getc_is_singlebyte 1
+# define MBTEST(x) ((x))
+#endif
#if defined (EXTENDED_GLOB)
-#define PATTERN_CHAR(c) \
- ((c) == '@' || (c) == '*' || (c) == '+' || (c) == '?' || (c) == '!')
-
extern int extended_glob;
#endif
extern int eof_encountered;
extern int no_line_editing, running_under_emacs;
extern int current_command_number;
-extern int interactive, interactive_shell, login_shell;
-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 pid_t last_command_subst_pid;
extern char *shell_name, *current_host_name;
extern char *dist_version;
extern int patch_level;
extern int dump_translatable_strings, dump_po_strings;
-extern Function *last_shell_builtin, *this_shell_builtin;
+extern sh_builtin_func_t *last_shell_builtin, *this_shell_builtin;
#if defined (BUFFERED_INPUT)
extern int bash_input_fd_changed;
#endif
+extern int errno;
/* **************************************************************** */
/* */
/* "Forward" declarations */
/* */
/* **************************************************************** */
-static char *ansiexpand ();
-static char *localeexpand ();
-static int reserved_word_acceptable ();
-static int read_token ();
-static int yylex ();
-static int parse_arith_cmd ();
+#ifdef DEBUG
+static void debug_parser __P((int));
+#endif
+
+static int yy_getc __P((void));
+static int yy_ungetc __P((int));
+
+#if defined (READLINE)
+static int yy_readline_get __P((void));
+static int yy_readline_unget __P((int));
+#endif
+
+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 shell_getc __P((int));
+static void shell_ungetc __P((int));
+static void discard_until __P((int));
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+static void push_string __P((char *, int, alias_t *));
+static void pop_string __P((void));
+static void free_string_list __P((void));
+#endif
+
+static char *read_a_line __P((int));
+
+static int reserved_word_acceptable __P((int));
+static int yylex __P((void));
+static int alias_expand_token __P((char *));
+static int time_command_acceptable __P((void));
+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
+#if defined (DPAREN_ARITHMETIC) || defined (ARITH_FOR_COMMAND)
+static int parse_dparen __P((int));
+static int parse_arith_cmd __P((char **, int));
+#endif
#if defined (COND_COMMAND)
-static COMMAND *parse_cond_command ();
+static void cond_error __P((void));
+static COND_COM *cond_expr __P((void));
+static COND_COM *cond_or __P((void));
+static COND_COM *cond_and __P((void));
+static COND_COM *cond_term __P((void));
+static int cond_skip_newlines __P((void));
+static COMMAND *parse_cond_command __P((void));
#endif
-static int read_token_word ();
-static void discard_parser_constructs ();
+#if defined (ARRAY_VARS)
+static int token_is_assignment __P((char *, int));
+static int token_is_ident __P((char *, int));
+#endif
+static int read_token_word __P((int));
+static void discard_parser_constructs __P((int));
-static void report_syntax_error ();
-static void handle_eof_input_unit ();
-static void prompt_again ();
+static char *error_token_from_token __P((int));
+static char *error_token_from_text __P((void));
+static void print_offending_line __P((void));
+static void report_syntax_error __P((char *));
+
+static void handle_eof_input_unit __P((void));
+static void prompt_again __P((void));
#if 0
-static void reset_readline_prompt ();
+static void reset_readline_prompt __P((void));
+#endif
+static void print_prompt __P((void));
+
+#if defined (HANDLE_MULTIBYTE)
+static void set_line_mbstate __P((void));
+static char *shell_input_line_property = NULL;
+#else
+# define set_line_mbstate()
#endif
-static void print_prompt ();
-extern int yyerror ();
+extern int yyerror __P((const char *));
+
+#ifdef DEBUG
+extern int yydebug;
+#endif
/* Default prompt strings */
char *primary_prompt = PPROMPT;
decode_prompt_string. */
int promptvars = 1;
-/* The decoded prompt string. Used if READLINE is not defined or if
- editing is turned off. Analogous to current_readline_prompt. */
-static char *current_decoded_prompt;
+/* If non-zero, $'...' and $"..." are expanded when they appear within
+ a ${...} expansion, even when the expansion appears within double
+ quotes. */
+int extended_quote = 1;
/* The number of lines read from input while creating the current command. */
int current_command_line_count;
+/* The number of lines in a command saved while we run parse_and_execute */
+int saved_command_line_count;
+
+/* The token that currently denotes the end of parse. */
+int shell_eof_token;
+
+/* The token currently being read. */
+int current_token;
+
+/* The current parser state. */
+int parser_state;
+
/* 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];
/* Where shell input comes from. History expansion is performed on each
line when the shell is interactive. */
static char *shell_input_line = (char *)NULL;
-static int shell_input_line_index;
-static int shell_input_line_size; /* Amount allocated for shell_input_line. */
-static int shell_input_line_len; /* strlen (shell_input_line) */
+static size_t shell_input_line_index;
+static size_t shell_input_line_size; /* Amount allocated for shell_input_line. */
+static size_t shell_input_line_len; /* strlen (shell_input_line) */
/* Either zero or EOF. */
static int shell_input_line_terminator;
/* The line number in a script on which a function body starts. */
static int function_bstart;
+/* The line number in a script at which an arithmetic for command starts. */
+static int arith_for_lineno;
+
+/* The decoded prompt string. Used if READLINE is not defined or if
+ editing is turned off. Analogous to current_readline_prompt. */
+static char *current_decoded_prompt;
+
+/* 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;
+
+static int global_extglob;
+
+/* 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. */
+#define MAX_CASE_NEST 128
+static int word_lineno[MAX_CASE_NEST];
+static int word_top = -1;
+
+/* If non-zero, it is the token that we want read_token to return
+ regardless of what text is (or isn't) present to be read. This
+ is reset by read_token. If token_to_read == WORD or
+ ASSIGNMENT_WORD, yylval.word should be set to word_desc_to_read. */
+static int token_to_read;
+static WORD_DESC *word_desc_to_read;
+
+static REDIRECTEE source;
static REDIRECTEE redir;
%}
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 IN BANG TIME TIMEOPT TIMEIGN
/* More general tokens. yylex () knows how to make these. */
-%token <word> WORD ASSIGNMENT_WORD
+%token <word> WORD ASSIGNMENT_WORD REDIR_WORD
%token <number> NUMBER
-%token <word_list> ARITH_CMD
+%token <word_list> ARITH_CMD ARITH_FOR_EXPRS
%token <command> COND_CMD
-%token AND_AND OR_OR GREATER_GREATER LESS_LESS LESS_AND
-%token GREATER_AND SEMI_SEMI LESS_LESS_MINUS AND_GREATER LESS_GREATER
-%token GREATER_BAR
+%token AND_AND OR_OR GREATER_GREATER LESS_LESS LESS_AND LESS_LESS_LESS
+%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> for_command select_command case_command group_command
%type <command> arith_command
%type <command> cond_command
-%type <command> function_def if_command elif_clause subshell
+%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
%type <word_list> word_list pattern
%type <pattern> pattern_list case_clause_sequence case_clause
%type <number> timespec
+%type <number> list_terminator
%start inputunit
%left '&' ';' '\n' yacc_EOF
%left AND_AND OR_OR
-%right '|'
+%right '|' BAR_AND
%%
-inputunit: simple_list '\n'
+inputunit: simple_list simple_list_terminator
{
/* Case of regular command. Discard the error
safety net,and return the command just parsed. */
global_command = $1;
eof_encountered = 0;
- discard_parser_constructs (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'
/* Error during parsing. Return NULL command. */
global_command = (COMMAND *)NULL;
eof_encountered = 0;
- discard_parser_constructs (1);
- if (interactive)
+ /* discard_parser_constructs (1); */
+ if (interactive && parse_and_execute_level == 0)
{
YYACCEPT;
}
redirection: '>' WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_output_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, 0);
}
| '<' WORD
{
+ source.dest = 0;
redir.filename = $2;
- $$ = make_redirection (0, r_input_direction, redir);
+ $$ = make_redirection (source, r_input_direction, redir, 0);
}
| NUMBER '>' WORD
{
+ source.dest = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_output_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, 0);
}
| NUMBER '<' WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_direction, redir, 0);
+ }
+ | REDIR_WORD '>' WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_input_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, REDIR_VARASSIGN);
+ }
+ | REDIR_WORD '<' WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_direction, redir, REDIR_VARASSIGN);
}
| GREATER_GREATER WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_appending_to, redir);
+ $$ = make_redirection (source, r_appending_to, redir, 0);
}
| NUMBER GREATER_GREATER WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_appending_to, redir, 0);
+ }
+ | REDIR_WORD GREATER_GREATER WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_appending_to, redir, REDIR_VARASSIGN);
+ }
+ | GREATER_BAR WORD
+ {
+ source.dest = 1;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_output_force, redir, 0);
+ }
+ | NUMBER GREATER_BAR WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_output_force, redir, 0);
+ }
+ | REDIR_WORD GREATER_BAR WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_output_force, redir, REDIR_VARASSIGN);
+ }
+ | LESS_GREATER WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_input_output, redir, 0);
+ }
+ | NUMBER LESS_GREATER WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_output, redir, 0);
+ }
+ | REDIR_WORD LESS_GREATER WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_appending_to, redir);
+ $$ = make_redirection (source, r_input_output, redir, REDIR_VARASSIGN);
}
| LESS_LESS WORD
{
+ source.dest = 0;
redir.filename = $2;
- $$ = make_redirection (0, r_reading_until, redir);
+ $$ = make_redirection (source, r_reading_until, redir, 0);
redir_stack[need_here_doc++] = $$;
}
| NUMBER LESS_LESS WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_until, redir, 0);
+ redir_stack[need_here_doc++] = $$;
+ }
+ | REDIR_WORD LESS_LESS WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_until, redir, REDIR_VARASSIGN);
+ redir_stack[need_here_doc++] = $$;
+ }
+ | LESS_LESS_MINUS WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_deblank_reading_until, redir, 0);
+ redir_stack[need_here_doc++] = $$;
+ }
+ | NUMBER LESS_LESS_MINUS WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_deblank_reading_until, redir, 0);
+ redir_stack[need_here_doc++] = $$;
+ }
+ | REDIR_WORD LESS_LESS_MINUS WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_reading_until, redir);
+ $$ = make_redirection (source, r_deblank_reading_until, redir, REDIR_VARASSIGN);
redir_stack[need_here_doc++] = $$;
}
+ | LESS_LESS_LESS WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_reading_string, redir, 0);
+ }
+ | NUMBER LESS_LESS_LESS WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_string, redir, 0);
+ }
+ | REDIR_WORD LESS_LESS_LESS WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_string, redir, REDIR_VARASSIGN);
+ }
| LESS_AND NUMBER
{
+ source.dest = 0;
redir.dest = $2;
- $$ = make_redirection (0, r_duplicating_input, redir);
+ $$ = make_redirection (source, r_duplicating_input, redir, 0);
}
| NUMBER LESS_AND NUMBER
{
+ source.dest = $1;
+ redir.dest = $3;
+ $$ = make_redirection (source, r_duplicating_input, redir, 0);
+ }
+ | REDIR_WORD LESS_AND NUMBER
+ {
+ source.filename = $1;
redir.dest = $3;
- $$ = make_redirection ($1, r_duplicating_input, redir);
+ $$ = make_redirection (source, r_duplicating_input, redir, REDIR_VARASSIGN);
}
| GREATER_AND NUMBER
{
+ source.dest = 1;
redir.dest = $2;
- $$ = make_redirection (1, r_duplicating_output, redir);
+ $$ = make_redirection (source, r_duplicating_output, redir, 0);
}
| NUMBER GREATER_AND NUMBER
{
+ source.dest = $1;
+ redir.dest = $3;
+ $$ = make_redirection (source, r_duplicating_output, redir, 0);
+ }
+ | REDIR_WORD GREATER_AND NUMBER
+ {
+ source.filename = $1;
redir.dest = $3;
- $$ = make_redirection ($1, r_duplicating_output, redir);
+ $$ = make_redirection (source, r_duplicating_output, redir, REDIR_VARASSIGN);
}
| LESS_AND WORD
{
+ source.dest = 0;
redir.filename = $2;
- $$ = make_redirection (0, r_duplicating_input_word, redir);
+ $$ = make_redirection (source, r_duplicating_input_word, redir, 0);
}
| NUMBER LESS_AND WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_duplicating_input_word, redir, 0);
+ }
+ | REDIR_WORD LESS_AND WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_duplicating_input_word, redir);
+ $$ = make_redirection (source, r_duplicating_input_word, redir, REDIR_VARASSIGN);
}
| GREATER_AND WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_duplicating_output_word, redir);
+ $$ = make_redirection (source, r_duplicating_output_word, redir, 0);
}
| NUMBER GREATER_AND WORD
{
+ source.dest = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_duplicating_output_word, redir);
+ $$ = make_redirection (source, r_duplicating_output_word, redir, 0);
}
- | LESS_LESS_MINUS WORD
- {
- redir.filename = $2;
- $$ = make_redirection
- (0, r_deblank_reading_until, redir);
- redir_stack[need_here_doc++] = $$;
- }
- | NUMBER LESS_LESS_MINUS WORD
+ | REDIR_WORD GREATER_AND WORD
{
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection
- ($1, r_deblank_reading_until, redir);
- redir_stack[need_here_doc++] = $$;
+ $$ = make_redirection (source, r_duplicating_output_word, redir, REDIR_VARASSIGN);
}
| GREATER_AND '-'
{
- redir.dest = 0L;
- $$ = make_redirection (1, r_close_this, redir);
+ source.dest = 1;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
| NUMBER GREATER_AND '-'
{
- redir.dest = 0L;
- $$ = make_redirection ($1, r_close_this, redir);
+ source.dest = $1;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
- | LESS_AND '-'
+ | REDIR_WORD GREATER_AND '-'
{
- redir.dest = 0L;
- $$ = make_redirection (0, r_close_this, redir);
+ source.filename = $1;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, REDIR_VARASSIGN);
}
- | NUMBER LESS_AND '-'
+ | LESS_AND '-'
{
- redir.dest = 0L;
- $$ = make_redirection ($1, r_close_this, redir);
+ source.dest = 0;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
- | AND_GREATER WORD
+ | NUMBER LESS_AND '-'
{
- redir.filename = $2;
- $$ = make_redirection (1, r_err_and_out, redir);
+ source.dest = $1;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
- | NUMBER LESS_GREATER WORD
+ | REDIR_WORD LESS_AND '-'
{
- redir.filename = $3;
- $$ = make_redirection ($1, r_input_output, redir);
+ source.filename = $1;
+ redir.dest = 0;
+ $$ = make_redirection (source, r_close_this, redir, REDIR_VARASSIGN);
}
- | LESS_GREATER WORD
+ | AND_GREATER WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (0, r_input_output, redir);
+ $$ = make_redirection (source, r_err_and_out, redir, 0);
}
- | GREATER_BAR WORD
+ | AND_GREATER_GREATER WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_output_force, redir);
- }
- | NUMBER GREATER_BAR WORD
- {
- redir.filename = $3;
- $$ = make_redirection ($1, r_output_force, redir);
+ $$ = make_redirection (source, r_append_err_and_out, redir, 0);
}
;
COMMAND *tc;
tc = $1;
- /* According to Posix.2 3.9.5, redirections
- specified after the body of a function should
- be attached to the function and performed when
- the function is executed, not as part of the
- function definition command. */
- if (tc->type == cm_function_def)
- {
- tc = tc->value.Function_def->command;
- if (tc->type == cm_group)
- tc = tc->value.Group->command;
- }
if (tc->redirects)
{
register REDIRECT *t;
tc->redirects = $2;
$$ = $1;
}
+ | function_def
+ { $$ = $1; }
+ | coproc
+ { $$ = $1; }
;
shell_command: for_command
{ $$ = $1; }
| cond_command
{ $$ = $1; }
- | function_def
+ | arith_for_command
{ $$ = $1; }
;
for_command: FOR WORD newline_list DO compound_list DONE
- { $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5); }
+ {
+ $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| FOR WORD newline_list '{' compound_list '}'
- { $$ = make_for_command ($2, add_string_to_list ("$@", (WORD_LIST *)NULL), $5); }
+ {
+ $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| FOR WORD ';' newline_list DO compound_list DONE
- { $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6); }
+ {
+ $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| FOR WORD ';' newline_list '{' compound_list '}'
- { $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6); }
+ {
+ $$ = make_for_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| FOR WORD newline_list IN word_list list_terminator newline_list DO compound_list DONE
- { $$ = make_for_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9); }
+ {
+ $$ = make_for_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| FOR WORD newline_list IN word_list list_terminator newline_list '{' compound_list '}'
- { $$ = make_for_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9); }
+ {
+ $$ = make_for_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
+ | FOR WORD newline_list IN list_terminator newline_list DO compound_list DONE
+ {
+ $$ = make_for_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
+ | FOR WORD newline_list IN list_terminator newline_list '{' compound_list '}'
+ {
+ $$ = make_for_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
+ ;
+
+arith_for_command: FOR ARITH_FOR_EXPRS list_terminator newline_list DO compound_list DONE
+ {
+ $$ = make_arith_for_command ($2, $6, arith_for_lineno);
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS list_terminator newline_list '{' compound_list '}'
+ {
+ $$ = make_arith_for_command ($2, $6, arith_for_lineno);
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS DO compound_list DONE
+ {
+ $$ = make_arith_for_command ($2, $4, arith_for_lineno);
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS '{' compound_list '}'
+ {
+ $$ = make_arith_for_command ($2, $4, arith_for_lineno);
+ if (word_top > 0) word_top--;
+ }
;
select_command: SELECT WORD newline_list DO list DONE
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
| SELECT WORD newline_list '{' list '}'
{
- $$ = make_select_command ($2, add_string_to_list ("$@", (WORD_LIST *)NULL), $5);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
| SELECT WORD ';' newline_list DO list DONE
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
| SELECT WORD ';' newline_list '{' list '}'
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
| SELECT WORD newline_list IN word_list list_terminator newline_list DO list DONE
{
- $$ = make_select_command ($2, (WORD_LIST *)reverse_list ($5), $9);
+ $$ = make_select_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
| SELECT WORD newline_list IN word_list list_terminator newline_list '{' list '}'
{
- $$ = make_select_command ($2, (WORD_LIST *)reverse_list ($5), $9);
+ $$ = make_select_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
;
case_command: CASE WORD newline_list IN newline_list ESAC
- { $$ = make_case_command ($2, (PATTERN_LIST *)NULL); }
+ {
+ $$ = make_case_command ($2, (PATTERN_LIST *)NULL, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| CASE WORD newline_list IN case_clause_sequence newline_list ESAC
- { $$ = make_case_command ($2, $5); }
+ {
+ $$ = make_case_command ($2, $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
| CASE WORD newline_list IN case_clause ESAC
- { $$ = make_case_command ($2, $5); }
+ {
+ $$ = make_case_command ($2, $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
;
-function_def: WORD '(' ')' newline_list group_command
+function_def: WORD '(' ')' newline_list function_body
{ $$ = make_function_def ($1, $5, function_dstart, function_bstart); }
-
- | FUNCTION WORD '(' ')' newline_list group_command
+ | FUNCTION WORD '(' ')' newline_list function_body
{ $$ = make_function_def ($2, $6, function_dstart, function_bstart); }
- | FUNCTION WORD newline_list group_command
+ | FUNCTION WORD newline_list function_body
{ $$ = make_function_def ($2, $4, function_dstart, function_bstart); }
;
+function_body: shell_command
+ { $$ = $1; }
+ | shell_command redirection_list
+ {
+ COMMAND *tc;
+
+ tc = $1;
+ /* According to Posix.2 3.9.5, redirections
+ specified after the body of a function should
+ be attached to the function and performed when
+ the function is executed, not as part of the
+ function definition command. */
+ /* XXX - I don't think it matters, but we might
+ want to change this in the future to avoid
+ problems differentiating between a function
+ definition with a redirection and a function
+ definition containing a single command with a
+ redirection. The two are semantically equivalent,
+ though -- the only difference is in how the
+ command printing code displays the redirections. */
+ if (tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $2;
+ }
+ else
+ tc->redirects = $2;
+ $$ = $1;
+ }
+ ;
+
subshell: '(' compound_list ')'
- { $2->flags |= CMD_WANT_SUBSHELL; $$ = $2; }
+ {
+ $$ = make_subshell_command ($2);
+ $$->flags |= CMD_WANT_SUBSHELL;
+ }
+ ;
+
+coproc: COPROC shell_command
+ {
+ $$ = 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));
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
;
if_command: IF compound_list THEN compound_list FI
;
-group_command: '{' list '}'
+group_command: '{' compound_list '}'
{ $$ = make_group_command ($2); }
;
;
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
{ $$ = $1; }
;
+simple_list_terminator: '\n'
+ | yacc_EOF
+ ;
+
list_terminator:'\n'
+ { $$ = '\n'; }
| ';'
+ { $$ = ';'; }
| yacc_EOF
+ { $$ = yacc_EOF; }
;
newline_list:
$$ = $1;
if (need_here_doc)
gather_here_documents ();
- }
- | simple_list1 '&'
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
+ }
+ | simple_list1 '&'
{
if ($1->type == cm_connection)
$$ = connect_async_list ($1, (COMMAND *)NULL, '&');
$$ = command_connect ($1, (COMMAND *)NULL, '&');
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 ';'
{
$$ = $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;
+ }
}
;
;
pipeline_command: pipeline
- { $$ = $1; }
- | BANG pipeline
+ { $$ = $1; }
+ | BANG pipeline_command
{
- $2->flags |= CMD_INVERT_RETURN;
+ if ($2)
+ $2->flags ^= CMD_INVERT_RETURN; /* toggle */
$$ = $2;
}
- | timespec pipeline
+ | timespec pipeline_command
{
- $2->flags |= $1;
+ if ($2)
+ $2->flags |= $1;
$$ = $2;
}
- | timespec BANG pipeline
+ | timespec list_terminator
{
- $3->flags |= $1;
- $$ = $3;
+ ELEMENT x;
+
+ /* Boy, this is unclean. `time' by itself can
+ time a null command. We cheat and push a
+ newline back if the list_terminator was a newline
+ to avoid the double-newline problem (one to
+ terminate this, one to terminate the command) */
+ x.word = 0;
+ x.redirect = 0;
+ $$ = make_simple_command (x, (COMMAND *)NULL);
+ $$->flags |= $1;
+ /* XXX - let's cheat and push a newline back */
+ if ($2 == '\n')
+ token_to_read = '\n';
}
- | BANG timespec pipeline
+ | BANG list_terminator
{
- $3->flags |= $2|CMD_INVERT_RETURN;
- $$ = $3;
+ ELEMENT x;
+
+ /* This is just as unclean. Posix says that `!'
+ by itself should be equivalent to `false'.
+ We cheat and push a
+ newline back if the list_terminator was a newline
+ to avoid the double-newline problem (one to
+ terminate this, one to terminate the command) */
+ x.word = 0;
+ x.redirect = 0;
+ $$ = make_simple_command (x, (COMMAND *)NULL);
+ $$->flags |= CMD_INVERT_RETURN;
+ /* XXX - let's cheat and push a newline back */
+ if ($2 == '\n')
+ token_to_read = '\n';
}
;
-pipeline:
- pipeline '|' newline_list pipeline
+pipeline: pipeline '|' newline_list pipeline
{ $$ = command_connect ($1, $4, '|'); }
+ | pipeline BAR_AND newline_list pipeline
+ {
+ /* Make cmd1 |& cmd2 equivalent to cmd1 2>&1 | cmd2 */
+ COMMAND *tc;
+ REDIRECTEE rd, sd;
+ REDIRECT *r;
+
+ tc = $1->type == cm_simple ? (COMMAND *)$1->value.Simple : $1;
+ sd.dest = 2;
+ rd.dest = 1;
+ r = make_redirection (sd, r_duplicating_output, rd, 0);
+ if (tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = r;
+ }
+ else
+ tc->redirects = r;
+
+ $$ = command_connect ($1, $4, '|');
+ }
| command
{ $$ = $1; }
;
{ $$ = CMD_TIME_PIPELINE; }
| TIME TIMEOPT
{ $$ = CMD_TIME_PIPELINE|CMD_TIME_POSIX; }
+ | TIME TIMEOPT TIMEIGN
+ { $$ = CMD_TIME_PIPELINE|CMD_TIME_POSIX; }
;
%%
-/* Possible states for the parser that require it to do special things. */
-#define PST_CASEPAT 0x001 /* in a case pattern list */
-#define PST_ALEXPNEXT 0x002 /* expand next word for aliases */
-#define PST_ALLOWOPNBRC 0x004 /* allow open brace for function def */
-#define PST_NEEDCLOSBRC 0x008 /* need close brace */
-#define PST_DBLPAREN 0x010 /* double-paren parsing */
-#define PST_SUBSHELL 0x020 /* ( ... ) subshell */
-#define PST_CMDSUBST 0x040 /* $( ... ) command substitution */
-#define PST_CASESTMT 0x080 /* parsing a case statement */
-#define PST_CONDCMD 0x100 /* parsing a [[...]] command */
-#define PST_CONDEXPR 0x200 /* parsing the guts of [[...]] */
-
/* Initial size to allocate for tokens, and the
amount to grow them by. */
#define TOKEN_DEFAULT_INITIAL_SIZE 496
#define TOKEN_DEFAULT_GROW_SIZE 512
-/* Shell meta-characters that, when unquoted, separate words. */
-#define shellmeta(c) (strchr ("()<>;&|", (c)) != 0)
-#define shellbreak(c) (strchr ("()<>;&| \t\n", (c)) != 0)
-#define shellquote(c) ((c) == '"' || (c) == '`' || (c) == '\'')
-#define shellexp(c) ((c) == '$' || (c) == '<' || (c) == '>')
-
-/* 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;
-
-/* If non-zero, it is the token that we want read_token to return
- regardless of what text is (or isn't) present to be read. This
- is reset by read_token. If token_to_read == WORD or
- ASSIGNMENT_WORD, yylval.word should be set to word_desc_to_read. */
-static int token_to_read;
-static WORD_DESC *word_desc_to_read;
+/* Should we call prompt_again? */
+#define SHOULD_PROMPT() \
+ (interactive && (bash_input.type == st_stdin || bash_input.type == st_stream))
-/* The current parser state. */
-static int parser_state;
+#if defined (ALIAS)
+# define expanding_alias() (pushed_string_list && pushed_string_list->expander)
+#else
+# define expanding_alias() 0
+#endif
/* Global var is non-zero when end of file has been reached. */
int EOF_Reached = 0;
-void
+#ifdef DEBUG
+static void
debug_parser (i)
int i;
{
yydebug = i;
#endif
}
+#endif
/* yy_getc () returns the next available character from input or EOF.
yy_ungetc (c) makes `c' the next character to read.
bash_input.name = (char *)NULL;
bash_input.location.file = (FILE *)NULL;
bash_input.location.string = (char *)NULL;
- bash_input.getter = (Function *)NULL;
- bash_input.ungetter = (Function *)NULL;
+ bash_input.getter = (sh_cget_func_t *)NULL;
+ bash_input.ungetter = (sh_cunget_func_t *)NULL;
}
/* Set the contents of the current bash input stream from
GET, UNGET, TYPE, NAME, and LOCATION. */
void
init_yy_io (get, unget, type, name, location)
- Function *get, *unget;
+ sh_cget_func_t *get;
+ sh_cunget_func_t *unget;
enum stream_type type;
- char *name;
+ const char *name;
INPUT_STREAM location;
{
bash_input.type = type;
bash_input.ungetter = unget;
}
+char *
+yy_input_name ()
+{
+ return (bash_input.name ? bash_input.name : "stdin");
+}
+
/* Call this to get the next character of input. */
-int
+static int
yy_getc ()
{
return (*(bash_input.getter)) ();
/* Call this to unget C. That is, to make C the next character
to be read. */
-int
+static int
yy_ungetc (c)
int c;
{
}
#if defined (BUFFERED_INPUT)
+#ifdef INCLUDE_UNUSED
int
input_file_descriptor ()
{
return (fileno (stdin));
}
}
+#endif
#endif /* BUFFERED_INPUT */
/* **************************************************************** */
yy_readline_get ()
{
SigHandler *old_sigint;
- int line_len, c;
+ int line_len;
+ unsigned char c;
if (!current_readline_line)
{
#if defined (JOB_CONTROL)
if (job_control)
- give_terminal_to (shell_pgrp);
+ give_terminal_to (shell_pgrp, 0);
#endif /* JOB_CONTROL */
+ old_sigint = (SigHandler *)IMPOSSIBLE_TRAP_HANDLER;
if (signal_is_ignored (SIGINT) == 0)
{
+ /* interrupt_immediately++; */
old_sigint = (SigHandler *)set_signal_handler (SIGINT, sigint_sighandler);
- interrupt_immediately++;
}
current_readline_line = readline (current_readline_prompt ?
current_readline_prompt : "");
+ CHECK_TERMSIG;
if (signal_is_ignored (SIGINT) == 0)
{
- interrupt_immediately--;
- set_signal_handler (SIGINT, old_sigint);
+ /* interrupt_immediately--; */
+ if (old_sigint != IMPOSSIBLE_TRAP_HANDLER)
+ set_signal_handler (SIGINT, old_sigint);
}
#if 0
current_readline_line_index = 0;
line_len = strlen (current_readline_line);
- current_readline_line = xrealloc (current_readline_line, 2 + line_len);
+ current_readline_line = (char *)xrealloc (current_readline_line, 2 + line_len);
current_readline_line[line_len++] = '\n';
current_readline_line[line_len] = '\0';
}
}
else
{
- c = (unsigned char)current_readline_line[current_readline_line_index++];
+ c = current_readline_line[current_readline_line_index++];
return (c);
}
}
yy_string_get ()
{
register char *string;
- register int c;
+ register unsigned char c;
string = bash_input.location.string;
- c = EOF;
/* If the string doesn't exist, or is empty, EOF found. */
if (string && *string)
{
- c = *(unsigned char *)string++;
+ c = *string++;
bash_input.location.string = string;
+ return (c);
}
- return (c);
+ else
+ return (EOF);
}
static int
void
with_input_from_string (string, name)
- char *string, *name;
+ char *string;
+ const char *name;
{
INPUT_STREAM location;
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. */
/* */
/* **************************************************************** */
+/* These two functions used to test the value of the HAVE_RESTARTABLE_SYSCALLS
+ define, and just use getc/ungetc if it was defined, but since bash
+ installs its signal handlers without the SA_RESTART flag, some signals
+ (like SIGCHLD, SIGWINCH, etc.) received during a read(2) will not cause
+ the read to be restarted. We need to restart it ourselves. */
+
static int
yy_stream_get ()
{
- int result = EOF;
+ int result;
+ result = EOF;
if (bash_input.location.file)
{
-#if !defined (HAVE_RESTARTABLE_SYSCALLS)
+#if 0
+ if (interactive)
+ interrupt_immediately++;
+#endif
+
+ /* XXX - don't need terminate_immediately; getc_with_restart checks
+ for terminating signals itself if read returns < 0 */
result = getc_with_restart (bash_input.location.file);
-#else /* HAVE_RESTARTABLE_SYSCALLS */
- result = getc (bash_input.location.file);
- result = (feof (bash_input.location.file)) ? EOF : (unsigned char)result;
-#endif /* HAVE_RESTARTABLE_SYSCALLS */
+
+#if 0
+ if (interactive)
+ interrupt_immediately--;
+#endif
}
return (result);
}
yy_stream_unget (c)
int c;
{
-#if !defined (HAVE_RESTARTABLE_SYSCALLS)
return (ungetc_with_restart (c, bash_input.location.file));
-#else /* HAVE_RESTARTABLE_SYSCALLS */
- return (ungetc (c, bash_input.location.file));
-#endif /* HAVE_RESTARTABLE_SYSCALLS */
}
void
with_input_from_stream (stream, name)
FILE *stream;
- char *name;
+ const char *name;
{
INPUT_STREAM location;
/* The globally known line number. */
int line_number = 0;
+/* The line number offset set by assigning to LINENO. Not currently used. */
+int line_number_base = 0;
+
#if defined (COND_COMMAND)
static int cond_lineno;
static int cond_token;
save stack, update the buffered fd to the new file descriptor and
re-establish the buffer <-> bash_input fd correspondence. */
if (bash_input.type == st_bstream && bash_input.location.buffered_fd >= 0)
- {
- if (bash_input_fd_changed)
+ {
+ if (bash_input_fd_changed)
{
bash_input_fd_changed = 0;
if (default_buffered_input >= 0)
{
bash_input.location.buffered_fd = default_buffered_input;
saver->bstream->b_fd = default_buffered_input;
+ SET_CLOSE_ON_EXEC (default_buffered_input);
}
}
+ /* XXX could free buffered stream returned as result here. */
set_buffered_stream (bash_input.location.buffered_fd, saver->bstream);
- }
+ }
#endif /* BUFFERED_INPUT */
line_number = saver->line;
return 0;
}
+/* Save the current token state and return it in a malloced array. */
+int *
+save_token_state ()
+{
+ int *ret;
+
+ ret = (int *)xmalloc (4 * sizeof (int));
+ ret[0] = last_read_token;
+ ret[1] = token_before_that;
+ ret[2] = two_tokens_ago;
+ ret[3] = current_token;
+ return ret;
+}
+
+void
+restore_token_state (ts)
+ int *ts;
+{
+ if (ts == 0)
+ return;
+ last_read_token = ts[0];
+ token_before_that = ts[1];
+ two_tokens_ago = ts[2];
+ current_token = ts[3];
+}
+
/*
* This is used to inhibit alias expansion and reserved word recognition
* inside case statement pattern lists. A `case statement pattern list' is:
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
-#if !defined (ALIAS)
-typedef void *alias_t;
-#endif
-
#define END_OF_ALIAS 0
/*
* implement alias expansion on a per-token basis.
*/
+#define PSH_ALIAS 0x01
+#define PSH_DPAREN 0x02
+#define PSH_SOURCE 0x04
+
typedef struct string_saver {
struct string_saver *next;
int expand_alias; /* Value to set expand_alias to when string is popped. */
#if defined (ALIAS)
alias_t *expander; /* alias that caused this line to be pushed. */
#endif
- int saved_line_size, saved_line_index, saved_line_terminator;
+ size_t saved_line_size, saved_line_index;
+ int saved_line_terminator;
+ int flags;
} STRING_SAVER;
STRING_SAVER *pushed_string_list = (STRING_SAVER *)NULL;
int expand;
alias_t *ap;
{
- STRING_SAVER *temp = (STRING_SAVER *) xmalloc (sizeof (STRING_SAVER));
+ STRING_SAVER *temp = (STRING_SAVER *)xmalloc (sizeof (STRING_SAVER));
temp->expand_alias = expand;
temp->saved_line = shell_input_line;
temp->saved_line_size = shell_input_line_size;
temp->saved_line_index = shell_input_line_index;
temp->saved_line_terminator = shell_input_line_terminator;
+ temp->flags = 0;
#if defined (ALIAS)
temp->expander = ap;
+ if (ap)
+ temp->flags = PSH_ALIAS;
#endif
temp->next = pushed_string_list;
pushed_string_list = temp;
#endif
shell_input_line = s;
- shell_input_line_size = strlen (s);
+ shell_input_line_size = STRLEN (s);
shell_input_line_index = 0;
shell_input_line_terminator = '\0';
- parser_state &= ~PST_ALEXPNEXT;
+#if 0
+ parser_state &= ~PST_ALEXPNEXT; /* XXX */
+#endif
+
+ set_line_mbstate ();
}
/*
#endif
free ((char *)t);
+
+ set_line_mbstate ();
}
static void
#endif /* ALIAS || DPAREN_ARITHMETIC */
+void
+free_pushed_string_input ()
+{
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ free_string_list ();
+#endif
+}
+
+int
+parser_expanding_alias ()
+{
+ return (expanding_alias ());
+}
+
+void
+parser_save_alias ()
+{
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ push_string ((char *)NULL, 0, (alias_t *)NULL);
+ pushed_string_list->flags = PSH_SOURCE; /* XXX - for now */
+#else
+ ;
+#endif
+}
+
+void
+parser_restore_alias ()
+{
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ if (pushed_string_list)
+ pop_string ();
+#else
+ ;
+#endif
+}
+
/* Return a line of text, taken from wherever yylex () reads input.
If there is no more input, then we return NULL. If REMOVE_QUOTED_NEWLINE
is non-zero, we remove unquoted \<newline> pairs. This is used by
{
static char *line_buffer = (char *)NULL;
static int buffer_size = 0;
- int indx = 0, c, peekc, pass_next;
+ int indx, c, peekc, pass_next;
#if defined (READLINE)
- if (interactive && bash_input.type != st_string && no_line_editing)
+ if (no_line_editing && SHOULD_PROMPT ())
#else
- if (interactive && bash_input.type != st_string)
+ if (SHOULD_PROMPT ())
#endif
print_prompt ();
- pass_next = 0;
+ pass_next = indx = 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)
- continue;
+ {
+#if 0
+ internal_warning ("read_a_line: ignored null byte in input");
+#endif
+ continue;
+ }
/* If there is no more input, then we return NULL. */
if (c == EOF)
need to treat the backslash specially only if a backslash
quoting a backslash-newline pair appears in the line. */
if (pass_next)
- {
+ {
line_buffer[indx++] = c;
pass_next = 0;
- }
+ }
else if (c == '\\' && remove_quoted_newline)
{
+ QUIT;
peekc = yy_getc ();
if (peekc == '\n')
- continue; /* Make the unquoted \<newline> pair disappear. */
+ {
+ line_number++;
+ continue; /* Make the unquoted \<newline> pair disappear. */
+ }
else
{
yy_ungetc (peekc);
read_secondary_line (remove_quoted_newline)
int remove_quoted_newline;
{
+ char *ret;
+ int n, c;
+
prompt_string_pointer = &ps2_prompt;
- prompt_again ();
- return (read_a_line (remove_quoted_newline));
+ if (SHOULD_PROMPT())
+ prompt_again ();
+ ret = read_a_line (remove_quoted_newline);
+#if defined (HISTORY)
+ if (ret && 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}
+};
+
+/* other tokens that can be returned by read_token() */
+STRING_INT_ALIST other_token_alist[] = {
+ /* Multiple-character tokens with special values */
+ { "--", TIMEIGN },
+ { "-p", TIMEOPT },
+ { "&&", AND_AND },
+ { "||", OR_OR },
+ { ">>", GREATER_GREATER },
+ { "<<", LESS_LESS },
+ { "<&", 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 */
+ { ">", '>' },
+ { "<", '<' },
+ { "-", '-' },
+ { "{", '{' },
+ { "}", '}' },
+ { ";", ';' },
+ { "(", '(' },
+ { ")", ')' },
+ { "|", '|' },
+ { "&", '&' },
+ { "newline", '\n' },
{ (char *)NULL, 0}
};
+/* others not listed here:
+ WORD look at yylval.word
+ ASSIGNMENT_WORD look at yylval.word
+ NUMBER look at yylval.number
+ ARITH_CMD look at yylval.word_list
+ ARITH_FOR_EXPRS look at yylval.word_list
+ COND_CMD look at yylval.command
+*/
+
/* These are used by read_token_word, but appear up here so that shell_getc
can use them to decide when to add otherwise blank lines to the history. */
do \
{ \
if (ds.delimiter_depth + 2 > ds.delimiter_space) \
- ds.delimiters = xrealloc \
+ ds.delimiters = (char *)xrealloc \
(ds.delimiters, (ds.delimiter_space += 10) * sizeof (char)); \
ds.delimiters[ds.delimiter_depth] = character; \
ds.delimiter_depth++; \
read the next line. This is called by read_token when the shell is
processing normal command input. */
+/* This implements one-character lookahead/lookbehind across physical input
+ lines, to avoid something being lost because it's pushed back with
+ shell_ungetc when we're at the start of a line. */
+static int eol_ungetc_lookahead = 0;
+
static int
shell_getc (remove_quoted_newline)
int remove_quoted_newline;
{
register int i;
- int c;
- static int mustpop = 0;
+ int c, truncating;
+ unsigned char uc;
QUIT;
+ if (sigwinch_received)
+ {
+ sigwinch_received = 0;
+ get_new_window_size (0, (int *)0, (int *)0);
+ }
+
+ if (eol_ungetc_lookahead)
+ {
+ c = eol_ungetc_lookahead;
+ eol_ungetc_lookahead = 0;
+ return (c);
+ }
+
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
/* If shell_input_line[shell_input_line_index] == 0, but there is
something on the pushed list of strings, then we don't want to go
{
line_number++;
+ /* Let's not let one really really long line blow up memory allocation */
+ if (shell_input_line && shell_input_line_size >= 32768)
+ {
+ free (shell_input_line);
+ shell_input_line = 0;
+ shell_input_line_size = 0;
+ }
+
restart_read:
/* Allow immediate exit if interrupted during input. */
QUIT;
- i = 0;
+ i = truncating = 0;
shell_input_line_terminator = 0;
+ /* If the shell is interatctive, but not currently printing a prompt
+ (interactive_shell && interactive == 0), we don't want to print
+ notifies or cleanup the jobs -- we want to defer it until we do
+ print the next prompt. */
+ if (interactive_shell == 0 || SHOULD_PROMPT())
+ {
#if defined (JOB_CONTROL)
/* This can cause a problem when reading a command as the result
of a trap, when the trap is called from flush_child. This call
had better not cause jobs to disappear from the job table in
that case, or we will have big trouble. */
- notify_and_cleanup ();
+ notify_and_cleanup ();
#else /* !JOB_CONTROL */
- cleanup_dead_jobs ();
+ cleanup_dead_jobs ();
#endif /* !JOB_CONTROL */
+ }
#if defined (READLINE)
- if (interactive && bash_input.type != st_string && no_line_editing)
+ if (no_line_editing && SHOULD_PROMPT())
#else
- if (interactive && bash_input.type != st_string)
+ if (SHOULD_PROMPT())
#endif
print_prompt ();
if (bash_input.type == st_stream)
clearerr (stdin);
- while (c = yy_getc ())
+ while (1)
{
+ c = yy_getc ();
+
/* Allow immediate exit if interrupted during input. */
QUIT;
- RESIZE_MALLOCED_BUFFER (shell_input_line, i, 2, shell_input_line_size, 256);
+ if (c == '\0')
+ {
+#if 0
+ internal_warning ("shell_getc: ignored null byte in input");
+#endif
+ continue;
+ }
+
+ /* Theoretical overflow */
+ /* If we can't put 256 bytes more into the buffer, allocate
+ everything we can and fill it as full as we can. */
+ /* XXX - we ignore rest of line using `truncating' flag */
+ if (shell_input_line_size > (SIZE_MAX - 256))
+ {
+ size_t n;
+
+ n = SIZE_MAX - i; /* how much more can we put into the buffer? */
+ if (n <= 2) /* we have to save 1 for the newline added below */
+ {
+ if (truncating == 0)
+ internal_warning("shell_getc: shell_input_line_size (%zu) exceeds SIZE_MAX (%llu): line truncated", shell_input_line_size, SIZE_MAX);
+ shell_input_line[i] = '\0';
+ truncating = 1;
+ }
+ if (shell_input_line_size < SIZE_MAX)
+ {
+ shell_input_line_size = SIZE_MAX;
+ shell_input_line = xrealloc (shell_input_line, shell_input_line_size);
+ }
+ }
+ else
+ RESIZE_MALLOCED_BUFFER (shell_input_line, i, 2, shell_input_line_size, 256);
if (c == EOF)
{
break;
}
- shell_input_line[i++] = c;
+ if (truncating == 0 || c == '\n')
+ shell_input_line[i++] = c;
if (c == '\n')
{
break;
}
}
+
shell_input_line_index = 0;
shell_input_line_len = i; /* == strlen (shell_input_line) */
+ set_line_mbstate ();
+
#if defined (HISTORY)
if (remember_on_history && shell_input_line && shell_input_line[0])
{
shell_input_line = expansions;
shell_input_line_len = shell_input_line ?
strlen (shell_input_line) : 0;
- if (!shell_input_line_len)
+ if (shell_input_line_len == 0)
current_command_line_count--;
/* We have to force the xrealloc below because we don't know
- the true allocated size of shell_input_line anymore. */
+ the true allocated size of shell_input_line anymore. */
shell_input_line_size = shell_input_line_len;
+
+ set_line_mbstate ();
}
}
- /* XXX - this is grotesque */
+ /* Try to do something intelligent with blank lines encountered while
+ entering multi-line commands. XXX - this is grotesque */
else if (remember_on_history && shell_input_line &&
shell_input_line[0] == '\0' &&
- current_command_line_count > 1 && current_delimiter (dstack))
+ current_command_line_count > 1)
{
- /* We know shell_input_line[0] == 0 and we're reading some sort of
- quoted string. This means we've got a line consisting of only
- a newline in a quoted string. We want to make sure this line
- gets added to the history. */
- maybe_add_history (shell_input_line);
+ if (current_delimiter (dstack))
+ /* We know shell_input_line[0] == 0 and we're reading some sort of
+ quoted string. This means we've got a line consisting of only
+ a newline in a quoted string. We want to make sure this line
+ gets added to the history. */
+ maybe_add_history (shell_input_line);
+ else
+ {
+ char *hdcs;
+ hdcs = history_delimiting_chars (shell_input_line);
+ if (hdcs && hdcs[0] == ';')
+ maybe_add_history (shell_input_line);
+ }
}
#endif /* HISTORY */
if (shell_input_line)
{
/* Lines that signify the end of the shell's input should not be
- echoed. */
+ echoed. We should not echo lines while parsing command
+ substitutions with recursive calls into the parsing engine; those
+ should only be echoed once when we read the word. That is the
+ reason for the test against shell_eof_token, which is set to a
+ right paren when parsing the contents of command substitutions. */
if (echo_input_at_read && (shell_input_line[0] ||
- shell_input_line_terminator != EOF))
+ shell_input_line_terminator != EOF) &&
+ shell_eof_token == 0)
fprintf (stderr, "%s\n", shell_input_line);
}
else
{
shell_input_line_size = 0;
prompt_string_pointer = ¤t_prompt_string;
- prompt_again ();
+ if (SHOULD_PROMPT ())
+ prompt_again ();
goto restart_read;
}
not already end in an EOF character. */
if (shell_input_line_terminator != EOF)
{
- if (shell_input_line_len + 3 > shell_input_line_size)
- shell_input_line = xrealloc (shell_input_line,
+ if (shell_input_line_size < SIZE_MAX && shell_input_line_len > shell_input_line_size - 3)
+ shell_input_line = (char *)xrealloc (shell_input_line,
1 + (shell_input_line_size += 2));
shell_input_line[shell_input_line_len] = '\n';
shell_input_line[shell_input_line_len + 1] = '\0';
+
+ set_line_mbstate ();
}
}
- c = shell_input_line[shell_input_line_index];
+next_alias_char:
+ uc = shell_input_line[shell_input_line_index];
- if (c)
+ if (uc)
shell_input_line_index++;
- if (c == '\\' && remove_quoted_newline &&
- shell_input_line[shell_input_line_index] == '\n')
- {
- prompt_again ();
- line_number++;
- goto restart_read;
- }
-
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
- /* If C is NULL, we have reached the end of the current input string. If
+ /* 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
because we have fully consumed the result of the last alias expansion.
Do it transparently; just return the next character of the string popped
to. */
- if (!c && (pushed_string_list != (STRING_SAVER *)NULL))
+ /* If pushed_string_list != 0 but pushed_string_list->expander == 0 (not
+ currently tested) and the flags value is not PSH_SOURCE, we are not
+ parsing an alias, we have just saved one (push_string, when called by
+ the parse_dparen code) In this case, just go on as well. The PSH_SOURCE
+ case is handled below. */
+pop_alias:
+ if (uc == 0 && pushed_string_list && pushed_string_list->flags != PSH_SOURCE)
{
- if (mustpop)
- {
- pop_string ();
- c = shell_input_line[shell_input_line_index];
- if (c)
- shell_input_line_index++;
- mustpop--;
- }
- else
- {
- mustpop++;
- c = ' ';
- }
+ pop_string ();
+ uc = shell_input_line[shell_input_line_index];
+ if (uc)
+ shell_input_line_index++;
}
#endif /* ALIAS || DPAREN_ARITHMETIC */
- if (!c && shell_input_line_terminator == EOF)
+ if MBTEST(uc == '\\' && remove_quoted_newline && shell_input_line[shell_input_line_index] == '\n')
+ {
+ if (SHOULD_PROMPT ())
+ prompt_again ();
+ line_number++;
+ /* What do we do here if we're expanding an alias whose definition
+ includes an escaped newline? If that's the last character in the
+ alias expansion, we just pop the pushed string list (recall that
+ we inhibit the appending of a space in mk_alexpansion() if newline
+ is the last character). If it's not the last character, we need
+ to consume the quoted newline and move to the next character in
+ the expansion. */
+#if defined (ALIAS)
+ if (expanding_alias () && shell_input_line[shell_input_line_index+1] == '\0')
+ {
+ uc = 0;
+ goto pop_alias;
+ }
+ else if (expanding_alias () && shell_input_line[shell_input_line_index+1] != '\0')
+ {
+ shell_input_line_index++; /* skip newline */
+ goto next_alias_char; /* and get next character */
+ }
+ else
+#endif
+ goto restart_read;
+ }
+
+ if (uc == 0 && shell_input_line_terminator == EOF)
return ((shell_input_line_index != 0) ? '\n' : EOF);
- return ((unsigned char)c);
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ /* We already know that we are not parsing an alias expansion because of the
+ check for expanding_alias() above. This knows how parse_and_execute
+ handles switching to st_string input while an alias is being expanded,
+ hence the check for pushed_string_list without pushed_string_list->expander
+ and the check for PSH_SOURCE as pushed_string_list->flags.
+ parse_and_execute and parse_string both change the input type to st_string
+ and place the string to be parsed and executed into location.string, so
+ we should not stop reading that until the pointer is '\0'.
+ The check for shell_input_line_terminator may be superfluous.
+
+ This solves the problem of `.' inside a multi-line alias with embedded
+ newlines executing things out of order. */
+ if (uc == 0 && bash_input.type == st_string && *bash_input.location.string &&
+ pushed_string_list && pushed_string_list->flags == PSH_SOURCE &&
+ shell_input_line_terminator == 0)
+ {
+ shell_input_line_index = 0;
+ goto restart_read;
+ }
+#endif
+
+ return (uc);
}
-/* Put C back into the input for the shell. */
+/* Put C back into the input for the shell. This might need changes for
+ HANDLE_MULTIBYTE around EOLs. Since we (currently) never push back a
+ character different than we read, shell_input_line_property doesn't need
+ to change when manipulating shell_input_line. The define for
+ last_shell_getc_is_singlebyte should take care of it, though. */
static void
shell_ungetc (c)
int c;
{
if (shell_input_line && shell_input_line_index)
shell_input_line[--shell_input_line_index] = c;
+ else
+ eol_ungetc_lookahead = c;
}
+#ifdef INCLUDE_UNUSED
+/* Back the input pointer up by one, effectively `ungetting' a character. */
static void
shell_ungetchar ()
{
if (shell_input_line && shell_input_line_index)
shell_input_line_index--;
}
+#endif
/* Discard input until CHARACTER is seen, then push that character back
onto the input stream. */
}
void
-execute_prompt_command (command)
- char *command;
+execute_variable_command (command, vname)
+ char *command, *vname;
{
- Function *temp_last, *temp_this;
char *last_lastarg;
- int temp_exit_value, temp_eof_encountered;
+ sh_parser_state_t ps;
- temp_last = last_shell_builtin;
- temp_this = this_shell_builtin;
- temp_exit_value = last_command_exit_value;
- temp_eof_encountered = eof_encountered;
+ save_parser_state (&ps);
last_lastarg = get_string_value ("_");
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);
- last_shell_builtin = temp_last;
- this_shell_builtin = temp_this;
- last_command_exit_value = temp_exit_value;
- eof_encountered = temp_eof_encountered;
-
- bind_variable ("_", last_lastarg);
+ restore_parser_state (&ps);
+ bind_variable ("_", last_lastarg, 0);
FREE (last_lastarg);
if (token_to_read == '\n') /* reset_parser was called */
We do this only if it is time to do so. Notice that only here
is the mail alarm reset; nothing takes place in check_mail ()
except the checking of mail. Please don't change this. */
- if (prompt_is_ps1 && time_to_check_mail ())
+ if (prompt_is_ps1 && parse_and_execute_level == 0 && time_to_check_mail ())
{
check_mail ();
reset_mail_timer ();
/* Avoid printing a prompt if we're not going to read anything, e.g.
after resetting the parser with read_token (RESET). */
- if (token_to_read == 0 && interactive)
+ if (token_to_read == 0 && SHOULD_PROMPT ())
prompt_again ();
}
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--;
}
}
static int open_brace_count;
#define command_token_position(token) \
- (((token) == ASSIGNMENT_WORD) || \
- ((token) != SEMI_SEMI && reserved_word_acceptable(token)))
+ (((token) == ASSIGNMENT_WORD) || (parser_state&PST_REDIRLIST) || \
+ ((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)
+#define assignment_acceptable(token) \
+ (command_token_position(token) && ((parser_state & PST_CASEPAT) == 0))
/* Check to see if TOKEN is a reserved word and return the token
value if it is. */
{ \
if ((parser_state & PST_CASEPAT) && (word_token_alist[i].token != ESAC)) \
break; \
- if (word_token_alist[i].token == TIME) \
+ if (word_token_alist[i].token == TIME && time_command_acceptable () == 0) \
break; \
if (word_token_alist[i].token == ESAC) \
parser_state &= ~(PST_CASEPAT|PST_CASESTMT); \
/* OK, we have a token. Let's try to alias expand it, if (and only if)
it's eligible.
- It is eligible for expansion if the shell is in interactive mode, and
+ It is eligible for expansion if EXPAND_ALIASES is set, and
the token is unquoted and the last token read was a command
separator (or expand_next_token is set), and we are currently
processing an alias (pushed_string_list is non-empty) and this
Special cases that disqualify:
In a pattern list in a case statement (parser_state & PST_CASEPAT). */
+
+static char *
+mk_alexpansion (s)
+ char *s;
+{
+ int l;
+ char *r;
+
+ l = strlen (s);
+ r = xmalloc (l + 2);
+ strcpy (r, s);
+ /* If the last character in the alias is a newline, don't add a trailing
+ space to the expansion. Works with shell_getc above. */
+ if (r[l - 1] != ' ' && r[l - 1] != '\n')
+ r[l++] = ' ';
+ r[l] = '\0';
+ return r;
+}
+
static int
-alias_expand_token (token)
- char *token;
+alias_expand_token (tokstr)
+ char *tokstr;
{
char *expanded;
alias_t *ap;
if (((parser_state & PST_ALEXPNEXT) || command_token_position (last_read_token)) &&
(parser_state & PST_CASEPAT) == 0)
{
- ap = find_alias (token);
+ ap = find_alias (tokstr);
/* Currently expanding this token. */
if (ap && (ap->flags & AL_BEINGEXPANDED))
return (NO_EXPANSION);
- expanded = ap ? savestring (ap->value) : (char *)NULL;
+ /* mk_alexpansion puts an extra space on the end of the alias expansion,
+ so the lookahead by the parser works right. If this gets changed,
+ make sure the code in shell_getc that deals with reaching the end of
+ an expanded alias is changed with it. */
+ expanded = ap ? mk_alexpansion (ap->value) : (char *)NULL;
+
if (expanded)
{
push_string (expanded, ap->flags & AL_EXPANDNEXT, ap);
time_command_acceptable ()
{
#if defined (COMMAND_TIMING)
+ int i;
+
+ if (posixly_correct && shell_compatibility_level > 41)
+ {
+ /* Quick check of the rest of the line to find the next token. If it
+ begins with a `-', Posix says to not return `time' as the token.
+ This was interp 267. */
+ i = shell_input_line_index;
+ while (i < shell_input_line_len && (shell_input_line[i] == ' ' || shell_input_line[i] == '\t'))
+ i++;
+ if (shell_input_line[i] == '-')
+ return 0;
+ }
+
switch (last_read_token)
{
case 0:
case AND_AND:
case OR_OR:
case '&':
+ case DO:
+ case THEN:
+ case ELSE:
+ case '{': /* } */
+ case '(': /* ) */
+ case BANG: /* ! time pipeline */
+ case TIME: /* time time pipeline */
+ case TIMEOPT: /* time -p time pipeline */
+ case TIMEIGN: /* time -p -- ... */
return 1;
default:
return 0;
to be set
`{' is recognized if the last token as WORD and the token
- before that was FUNCTION.
+ before that was FUNCTION, or if we just parsed an arithmetic
+ `for' command.
- `}' is recognized if there is an unclosed `{' prsent.
+ `}' is recognized if there is an unclosed `{' present.
`-p' is returned as TIMEOPT if the last read token was TIME.
+ `--' is returned as TIMEIGN if the last read token was TIMEOPT.
']]' is returned as COND_END if the parser is currently parsing
a conditional expression ((parser_state & PST_CONDEXPR) != 0)
*/
static int
-special_case_tokens (token)
- char *token;
+special_case_tokens (tokstr)
+ char *tokstr;
{
if ((last_read_token == WORD) &&
#if defined (SELECT_COMMAND)
#else
((token_before_that == FOR) || (token_before_that == CASE)) &&
#endif
- (token[0] == 'i' && token[1] == 'n' && token[2] == 0))
+ (tokstr[0] == 'i' && tokstr[1] == 'n' && tokstr[2] == 0))
{
if (token_before_that == CASE)
{
#else
(token_before_that == FOR) &&
#endif
- (token[0] == 'd' && token[1] == 'o' && token[2] == '\0'))
+ (tokstr[0] == 'd' && tokstr[1] == 'o' && tokstr[2] == '\0'))
return (DO);
/* Ditto for ESAC in the CASE case.
if (esacs_needed_count)
{
esacs_needed_count--;
- if (STREQ (token, "esac"))
+ if (STREQ (tokstr, "esac"))
{
parser_state &= ~PST_CASEPAT;
return (ESAC);
if (parser_state & PST_ALLOWOPNBRC)
{
parser_state &= ~PST_ALLOWOPNBRC;
- if (token[0] == '{' && token[1] == '\0') /* } */
+ if (tokstr[0] == '{' && tokstr[1] == '\0') /* } */
{
open_brace_count++;
function_bstart = line_number;
}
}
- if (open_brace_count && reserved_word_acceptable (last_read_token) && token[0] == '}' && !token[1])
+ /* We allow a `do' after a for ((...)) without an intervening
+ list_terminator */
+ if (last_read_token == ARITH_FOR_EXPRS && tokstr[0] == 'd' && tokstr[1] == 'o' && !tokstr[2])
+ return (DO);
+ if (last_read_token == ARITH_FOR_EXPRS && tokstr[0] == '{' && tokstr[1] == '\0') /* } */
+ {
+ open_brace_count++;
+ return ('{'); /* } */
+ }
+
+ if (open_brace_count && reserved_word_acceptable (last_read_token) && tokstr[0] == '}' && !tokstr[1])
{
open_brace_count--; /* { */
return ('}');
#if defined (COMMAND_TIMING)
/* Handle -p after `time'. */
- if (last_read_token == TIME && token[0] == '-' && token[1] == 'p' && !token[2])
+ if (last_read_token == TIME && tokstr[0] == '-' && tokstr[1] == 'p' && !tokstr[2])
return (TIMEOPT);
+ /* Handle -- after `time -p'. */
+ if (last_read_token == TIMEOPT && tokstr[0] == '-' && tokstr[1] == '-' && !tokstr[2])
+ return (TIMEIGN);
#endif
-#if defined (COMMAND_TIMING)
- if (STREQ (token, "time") && time_command_acceptable ())
- return (TIME);
-#endif /* COMMAND_TIMING */
-
#if defined (COND_COMMAND) /* [[ */
- if ((parser_state & PST_CONDEXPR) && token[0] == ']' && token[1] == ']' && token[2] == '\0')
+ if ((parser_state & PST_CONDEXPR) && tokstr[0] == ']' && tokstr[1] == ']' && tokstr[2] == '\0')
return (COND_END);
#endif
dstack.delimiter_depth = 0; /* No delimiters found so far. */
open_brace_count = 0;
+#if defined (EXTENDED_GLOB)
+ /* Reset to global value of extended glob */
+ if (parser_state & PST_EXTPAT)
+ extended_glob = global_extglob;
+#endif
+
parser_state = 0;
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
FREE (word_desc_to_read);
word_desc_to_read = (WORD_DESC *)NULL;
+ current_token = '\n'; /* XXX */
last_read_token = '\n';
token_to_read = '\n';
}
yylval.command = parse_cond_command ();
if (cond_token != COND_END)
{
- if (EOF_Reached && cond_token != COND_ERROR) /* [[ */
- parser_error (cond_lineno, "unexpected EOF while looking for `]]'");
- else if (cond_token != COND_ERROR)
- parser_error (cond_lineno, "syntax error in conditional expression");
+ cond_error ();
return (-1);
}
token_to_read = COND_END;
#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)
return (yacc_EOF);
}
- if (character == '#' && (!interactive || interactive_comments))
+ if MBTEST(character == '#' && (!interactive || interactive_comments))
{
/* A comment. Discard until EOL or EOF, and then return a newline. */
discard_until ('\n');
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
+ parser_state &= ~PST_ASSIGNOK;
+
return (character);
}
+ if (parser_state & PST_REGEXP)
+ goto tokword;
+
/* Shell meta-characters. */
- if (shellmeta (character) && ((parser_state & PST_DBLPAREN) == 0))
+ if MBTEST(shellmeta (character) && ((parser_state & PST_DBLPAREN) == 0))
{
#if defined (ALIAS)
/* Turn off alias tokenization iff this character sequence would
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 MBTEST(peek_char == '<')
+ return (LESS_LESS_LESS);
else
{
shell_ungetc (peek_char);
#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);
case '|':
return (OR_OR);
-#if defined (DPAREN_ARITHMETIC)
+#if defined (DPAREN_ARITHMETIC) || defined (ARITH_FOR_COMMAND)
case '(': /* ) */
- if (reserved_word_acceptable (last_read_token))
- {
- int cmdtyp, sline;
- char *wval;
- WORD_DESC *wd;
-
- sline = line_number;
- cmdtyp = parse_arith_cmd (&wval);
- if (cmdtyp == 1) /* arithmetic command */
- {
- wd = make_word (wval);
- wd->flags = W_QUOTED;
- 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 */
- {
- push_string (wval, 0, (alias_t *)NULL);
- if ((parser_state & PST_CASEPAT) == 0)
- parser_state |= PST_SUBSHELL;
- return (character);
- }
- else /* ERROR */
- return -1;
- }
- break;
+ result = parse_dparen (character);
+ if (result == -2)
+ break;
+ else
+ return result;
#endif
}
}
- else if (character == '<' && peek_char == '&')
+ else if MBTEST(character == '<' && peek_char == '&')
return (LESS_AND);
- else if (character == '>' && peek_char == '&')
+ else if MBTEST(character == '>' && peek_char == '&')
return (GREATER_AND);
- else if (character == '<' && peek_char == '>')
+ else if MBTEST(character == '<' && peek_char == '>')
return (LESS_GREATER);
- else if (character == '>' && peek_char == '|')
+ else if MBTEST(character == '>' && peek_char == '|')
return (GREATER_BAR);
- else if (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 we look like we are reading the start of a function
definition, then let the reader know about it so that
we will do the right thing with `{'. */
- if (character == ')' && last_read_token == '(' && token_before_that == WORD)
+ if MBTEST(character == ')' && last_read_token == '(' && token_before_that == WORD)
{
parser_state |= PST_ALLOWOPNBRC;
#if defined (ALIAS)
/* case pattern lists may be preceded by an optional left paren. If
we're not trying to parse a case pattern list, the left paren
indicates a subshell. */
- if (character == '(' && (parser_state & PST_CASEPAT) == 0) /* ) */
+ if MBTEST(character == '(' && (parser_state & PST_CASEPAT) == 0) /* ) */
parser_state |= PST_SUBSHELL;
/*(*/
- else if ((parser_state & PST_CASEPAT) && character == ')')
- parser_state &= ~PST_CASEPAT;
+ else if MBTEST((parser_state & PST_CASEPAT) && character == ')')
+ parser_state &= ~PST_CASEPAT;
/*(*/
- else if ((parser_state & PST_SUBSHELL) && character == ')')
+ else if MBTEST((parser_state & PST_SUBSHELL) && character == ')')
parser_state &= ~PST_SUBSHELL;
#if defined (PROCESS_SUBSTITUTION)
/* Check for the constructs which introduce process substitution.
Shells running in `posix mode' don't do process substitution. */
- if (posixly_correct ||
- ((character != '>' && character != '<') || peek_char != '('))
+ if MBTEST(posixly_correct || ((character != '>' && character != '<') || peek_char != '(')) /*)*/
#endif /* PROCESS_SUBSTITUTION */
return (character);
}
- /* Hack <&- (close stdin) case. */
- if (character == '-' && (last_read_token == LESS_AND || last_read_token == GREATER_AND))
+ /* Hack <&- (close stdin) case. Also <&N- (dup and close). */
+ 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);
return result;
}
-/* 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, and returning
- correct error values if it reads EOF. */
+/*
+ * 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, and returning
+ * correct error values if it reads EOF.
+ */
+#define P_FIRSTCLOSE 0x0001
+#define P_ALLOWESC 0x0002
+#define P_DQUOTE 0x0004
+#define P_COMMAND 0x0008 /* parsing a command, so look for comments */
+#define P_BACKQUOTE 0x0010 /* parsing a backquoted command substitution */
+#define P_ARRAYSUB 0x0020 /* parsing a [...] array subscript for assignment */
+#define P_DOLBRACE 0x0040 /* parsing a ${...} construct */
+
+/* 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 LEX_INWORD 0x400
+
+#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, start_lineno;
- char *ret, *nestret;
- int retind, retsize;
+ int count, ch, tflags;
+ int nestlen, ttranslen, start_lineno;
+ char *ret, *nestret, *ttrans;
+ int retind, retsize, rflags;
+ int dolbrace_state;
+
+ dolbrace_state = (flags & P_DOLBRACE) ? DOLBRACE_PARAM : 0;
+/*itrace("parse_matched_pair[%d]: open = %c close = %c flags = %d", line_number, open, close, flags);*/
count = 1;
- pass_next_character = was_dollar = 0;
+ tflags = 0;
- ret = xmalloc (retsize = 64);
+ 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);
+
+ ret = (char *)xmalloc (retsize = 64);
retind = 0;
start_lineno = line_number;
while (count)
{
- ch = shell_getc (qc != '\'' && pass_next_character == 0);
+ ch = shell_getc (qc != '\'' && (tflags & (LEX_PASSNEXT)) == 0);
+
if (ch == EOF)
{
free (ret);
- parser_error (start_lineno, "unexpected EOF while looking for matching `%c'", close);
+ parser_error (start_lineno, _("unexpected EOF while looking for matching `%c'"), close);
EOF_Reached = 1; /* XXX */
return (&matched_pair_error);
}
/* Possible reprompting. */
- if (ch == '\n' && interactive &&
- (bash_input.type == st_stdin || bash_input.type == st_stream))
+ 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;
}
RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
- if (ch == CTLESC || ch == CTLNUL)
+ if MBTEST(ch == CTLESC)
ret[retind++] = CTLESC;
ret[retind++] = ch;
continue;
}
- else if (ch == CTLESC || ch == CTLNUL) /* special shell escapes */
+ /* 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);
ret[retind++] = CTLESC;
ret[retind++] = ch;
continue;
}
- else if (ch == close) /* ending delimiter */
+ else if MBTEST(ch == close) /* ending delimiter */
count--;
- else if (ch == open) /* nested begin */
+ /* handle nested ${...} specially. */
+ else if MBTEST(open != close && (tflags & LEX_WASDOL) && open == '{' && ch == open) /* } */
+ count++;
+ else if MBTEST(((flags & P_FIRSTCLOSE) == 0) && ch == open) /* nested begin */
count++;
/* Add this character. */
RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
ret[retind++] = ch;
- if (open == '\'') /* '' inside grouping construct */
- continue;
+ /* 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 == '\\')
+ tflags |= LEX_PASSNEXT;
+ continue;
+ }
+
+ if MBTEST(ch == '\\') /* backslashes */
+ tflags |= LEX_PASSNEXT;
+
+ /* Based on which dolstate is currently in (param, op, or word),
+ decide what the op is. We're really only concerned if it's % or
+ #, so we can turn on a flag that says whether or not we should
+ treat single quotes as special when inside a double-quoted
+ ${...}. This logic must agree with subst.c:extract_dollar_brace_string
+ since they share the same defines. */
+ /* FLAG POSIX INTERP 221 */
+ if (flags & P_DOLBRACE)
+ {
+ /* ${param%[%]word} */
+ if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '%' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param#[#]word} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '#' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param/[/]pat/rep} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '/' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE2; /* XXX */
+ /* ${param^[^]pat} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '^' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param,[,]pat} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == ',' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && strchr ("#%^,~:-=?+/", ch) != 0)
+ dolbrace_state = DOLBRACE_OP;
+ else if MBTEST(dolbrace_state == DOLBRACE_OP && strchr ("#%^,~:-=?+/", ch) == 0)
+ dolbrace_state = DOLBRACE_WORD;
+ }
+
+ /* The big hammer. Single quotes aren't special in double quotes. The
+ problem is that Posix used to say 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." */
+ /* This was changed in Austin Group Interp 221 */
+ if MBTEST(posixly_correct && shell_compatibility_level > 41 && dolbrace_state != DOLBRACE_QUOTE && dolbrace_state != DOLBRACE_QUOTE2 && (flags & P_DQUOTE) && (flags & P_DOLBRACE) && ch == '\'')
+ continue;
+
+ /* 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((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))
+ {
+ /* Translate $'...' here. */
+ ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
+ xfree (nestret);
+
+ /* If we're parsing a double-quoted brace expansion and we are
+ not in a place where single quotes are treated specially,
+ make sure we single-quote the results of the ansi
+ expansion because quote removal should remove them later */
+ /* FLAG POSIX INTERP 221 */
+ if ((shell_compatibility_level > 42) && (rflags & P_DQUOTE) && (dolbrace_state == DOLBRACE_QUOTE2) && (flags & P_DOLBRACE))
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else 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((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 = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = ttranslen + 2;
+ retind -= 2; /* back up before the $" */
+ }
+
+ APPEND_NESTRET ();
+ FREE (nestret);
+ }
+ else if ((flags & P_ARRAYSUB) && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ goto parse_dollar_word;
+ }
+ /* Parse an old-style command substitution within double quotes as a
+ single word. */
+ /* XXX - sh and ksh93 don't do this - XXX */
+ else if MBTEST(open == '"' && ch == '`')
+ {
+ nestret = parse_matched_pair (0, '`', '`', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ else if MBTEST(open != '`' && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ /* check for $(), $[], or ${} inside quoted string. */
+ {
+parse_dollar_word:
+ if (open == ch) /* undo previous increment */
+ count--;
+ if (ch == '(') /* ) */
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
+ else if (ch == '{') /* } */
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|P_DOLBRACE|rflags);
+ else if (ch == '[') /* ] */
+ nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ if MBTEST(ch == '$')
+ tflags |= LEX_WASDOL;
+ else
+ tflags &= ~LEX_WASDOL;
+ }
+
+ ret[retind] = '\0';
+ if (lenp)
+ *lenp = retind;
+/*itrace("parse_matched_pair[%d]: returning %s", line_number, ret);*/
+ 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_wlen, lex_firstind;
+ int nestlen, ttranslen, start_lineno;
+ char *ret, *nestret, *ttrans, *heredelim;
+ int retind, retsize, rflags, hdlen;
+
+ /* Posix interp 217 says arithmetic expressions have precedence, so
+ assume $(( introduces arithmetic expansion and parse accordingly. */
+ peekc = shell_getc (0);
+ shell_ungetc (peekc);
+ if (peekc == '(')
+ return (parse_matched_pair (qc, open, close, lenp, 0));
+
+/*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 = lex_wlen = 0;
+
+ heredelim = 0;
+ lex_firstind = -1;
+
+ while (count)
+ {
+comsub_readchar:
+ ch = shell_getc (qc != '\'' && (tflags & (LEX_INCOMMENT|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);*/
+ free (heredelim);
+ heredelim = 0;
+ lex_firstind = -1;
+ }
+ else
+ lex_firstind = retind + 1;
+ }
+ }
+
+ /* Possible reprompting. */
+ if (ch == '\n' && SHOULD_PROMPT ())
+ prompt_again ();
+
+ /* XXX -- possibly allow here doc to be delimited by ending right
+ paren. */
+ if ((tflags & LEX_INHEREDOC) && ch == close && count == 1)
+ {
+ int tind;
+/*itrace("parse_comsub: in here doc, ch == close, retind - firstind = %d hdlen = %d retind = %d", retind-lex_firstind, hdlen, retind);*/
+ tind = lex_firstind;
+ while ((tflags & LEX_STRIPDOC) && ret[tind] == '\t')
+ tind++;
+ if (retind-tind == hdlen && STREQN (ret + tind, heredelim, hdlen))
+ {
+ tflags &= ~(LEX_STRIPDOC|LEX_INHEREDOC);
+/*itrace("parse_comsub:%d: found here doc end `%s'", line_number, ret + tind);*/
+ free (heredelim);
+ heredelim = 0;
+ lex_firstind = -1;
+ }
+ }
+
+ /* 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')
+ {
+/*itrace("parse_comsub:%d: lex_incomment -> 0 ch = `%c'", line_number, ch);*/
+ tflags &= ~LEX_INCOMMENT;
+ }
+
+ continue;
+ }
+
+ if (tflags & LEX_PASSNEXT) /* last char was backslash */
+ {
+/*itrace("parse_comsub:%d: lex_passnext -> 0 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ 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)
+ ret[retind++] = CTLESC;
+ ret[retind++] = ch;
+ continue;
+ }
+
+ /* If this is a shell break character, we are not in a word. If not,
+ we either start or continue a word. */
+ if MBTEST(shellbreak (ch))
+ {
+ tflags &= ~LEX_INWORD;
+/*itrace("parse_comsub:%d: lex_inword -> 0 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ }
+ else
+ {
+ if (tflags & LEX_INWORD)
+ {
+ lex_wlen++;
+/*itrace("parse_comsub:%d: lex_inword == 1 ch = `%c' lex_wlen = %d (%d)", line_number, ch, lex_wlen, __LINE__);*/
+ }
+ else
+ {
+/*itrace("parse_comsub:%d: lex_inword -> 1 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ tflags |= LEX_INWORD;
+ lex_wlen = 0;
+ }
+ }
+
+ /* Skip whitespace */
+ if MBTEST(shellblank (ch) && (tflags & LEX_HEREDELIM) == 0 && 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;
+#if 0
+ else if (heredelim && (tflags & LEX_PASSNEXT) == 0 && ch == '\n')
+ {
+ tflags |= LEX_INHEREDOC;
+ tflags &= ~LEX_HEREDELIM;
+ lex_firstind = retind + 1;
+ }
+#endif
+ else if (lex_firstind >= 0 && (tflags & LEX_PASSNEXT) == 0 && shellbreak (ch))
+ {
+ if (heredelim == 0)
+ {
+ 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) || ch == '\n'))
+ {
+ /* 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);
+/*itrace("parse_comsub:%d: set lex_reswordok = 1, ch = `%c'", line_number, ch);*/
+ tflags |= LEX_RESWDOK;
+ 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 lex_reswdok -> 0", line_number);*/
+ }
+ else if (STREQN (ret + retind - 4, "esac", 4))
+ {
+ tflags &= ~LEX_INCASE;
+/*itrace("parse_comsub:%d: found `esac', lex_incase -> 0 lex_reswdok -> 0", line_number);*/
+ }
+ tflags &= ~LEX_RESWDOK;
+ }
+ else if MBTEST((tflags & LEX_CKCOMMENT) && ch == '#' && (lex_rwlen == 0 || ((tflags & LEX_INWORD) && lex_wlen == 0)))
+ ; /* don't modify LEX_RESWDOK if we're starting a comment */
+ /* Allow `do' followed by space, tab, or newline to preserve the
+ RESWDOK flag, but reset the reserved word length counter so we
+ can read another one. */
+ else if MBTEST(((tflags & LEX_INCASE) == 0) &&
+ (isblank(ch) || ch == '\n') &&
+ lex_rwlen == 2 &&
+ STREQN (ret + retind - 2, "do", 2))
+ {
+/*itrace("parse_comsub:%d: lex_incase == 1 found `%c', found \"do\"", line_number, ch);*/
+ lex_rwlen = 0;
+ }
+ else if MBTEST((tflags & LEX_INCASE) && ch != '\n')
+ /* If we can read a reserved word and we're in case, we're at the
+ point where we can read a new pattern list or an esac. We
+ handle the esac case above. If we read a newline, we want to
+ leave LEX_RESWDOK alone. If we read anything else, we want to
+ turn off LEX_RESWDOK, since we're going to read a pattern list. */
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: lex_incase == 1 found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+ else if MBTEST(shellbreak (ch) == 0)
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+#if 0
+ /* If we find a space or tab but have read something and it's not
+ `do', turn off the reserved-word-ok flag */
+ else if MBTEST(isblank (ch) && lex_rwlen > 0)
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+#endif
+ }
+
+ /* Might be the start of a here-doc delimiter */
+ 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);
+ if (peekc != '<')
+ {
+ tflags |= LEX_HEREDELIM;
+ lex_firstind = -1;
+ }
+ continue;
+ }
+ else
+ ch = peekc; /* fall through and continue XXX */
+ }
+ else if MBTEST((tflags & LEX_CKCOMMENT) && (tflags & LEX_INCOMMENT) == 0 && ch == '#' && (((tflags & LEX_RESWDOK) && lex_rwlen == 0) || ((tflags & LEX_INWORD) && lex_wlen == 0)))
+ {
+/*itrace("parse_comsub:%d: lex_incomment -> 1 (%d)", line_number, __LINE__);*/
+ tflags |= LEX_INCOMMENT;
+ }
+
+ 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++;
+/*itrace("parse_comsub:%d: found open: count = %d", line_number, 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))
+ {
+ /* 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 $' */
+ }
+ 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 = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = ttranslen + 2;
+ retind -= 2; /* back up before the $" */
+ }
+
+ APPEND_NESTRET ();
+ FREE (nestret);
+ }
+ else if MBTEST((tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ /* check for $(), $[], or ${} inside command substitution. */
+ {
+ if ((tflags & LEX_INCASE) == 0 && open == ch) /* undo previous increment */
+ count--;
+ if (ch == '(') /* ) */
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
+ else if (ch == '{') /* } */
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|P_DOLBRACE|rflags);
+ else if (ch == '[') /* ] */
+ nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ if MBTEST(ch == '$')
+ 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;
+}
+
+/* Recursively call the parser to parse a $(...) command substitution. */
+char *
+xparse_dolparen (base, string, indp, flags)
+ char *base;
+ char *string;
+ int *indp;
+ int flags;
+{
+ sh_parser_state_t ps;
+ sh_input_line_state_t ls;
+ int orig_ind, nc, sflags, orig_eof_token;
+ char *ret, *s, *ep, *ostring;
+
+ /*yydebug = 1;*/
+ orig_ind = *indp;
+ ostring = string;
+
+/*itrace("xparse_dolparen: size = %d shell_input_line = `%s'", shell_input_line_size, shell_input_line);*/
+ sflags = SEVAL_NONINT|SEVAL_NOHIST|SEVAL_NOFREE;
+ if (flags & SX_NOLONGJMP)
+ sflags |= SEVAL_NOLONGJMP;
+ save_parser_state (&ps);
+ save_input_line_state (&ls);
+ orig_eof_token = shell_eof_token;
+
+ /*(*/
+ parser_state |= PST_CMDSUBST|PST_EOFTOKEN; /* allow instant ')' */ /*(*/
+ shell_eof_token = ')';
+ parse_string (string, "command substitution", sflags, &ep);
+
+ shell_eof_token = orig_eof_token;
+ restore_parser_state (&ps);
+ reset_parser ();
+ /* reset_parser clears shell_input_line and associated variables */
+ restore_input_line_state (&ls);
+ 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 (ch == '\\') /* backslashes */
- pass_next_character++;
+ if (nc == 0)
+ {
+ ret = xmalloc (1);
+ ret[0] = '\0';
+ }
+ else
+ ret = substring (ostring, 0, nc - 1);
- if (open != close) /* a grouping construct */
+ return ret;
+}
+
+#if defined (DPAREN_ARITHMETIC) || defined (ARITH_FOR_COMMAND)
+/* Parse a double-paren construct. It can be either an arithmetic
+ command, an arithmetic `for' command, or a nested subshell. Returns
+ the parsed token, -1 on error, or -2 if we didn't do anything and
+ should just go on. */
+static int
+parse_dparen (c)
+ int c;
+{
+ int cmdtyp, sline;
+ char *wval;
+ WORD_DESC *wd;
+
+#if defined (ARITH_FOR_COMMAND)
+ if (last_read_token == FOR)
+ {
+ arith_for_lineno = line_number;
+ cmdtyp = parse_arith_cmd (&wval, 0);
+ if (cmdtyp == 1)
{
- if (shellquote (ch))
- {
- /* '', ``, or "" inside $(...) or other grouping construct. */
- push_delimiter (dstack, ch);
- nestret = parse_matched_pair (ch, ch, ch, &nestlen, 0);
- pop_delimiter (dstack);
- 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;
- }
- FREE (nestret);
- }
+ wd = alloc_word_desc ();
+ wd->word = wval;
+ yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
+ return (ARITH_FOR_EXPRS);
}
- /* Parse an old-style command substitution within double quotes as a
- single word. */
- /* XXX - sh and ksh93 don't do this - XXX */
- else if (open == '"' && ch == '`')
+ else
+ return -1; /* ERROR */
+ }
+#endif
+
+#if defined (DPAREN_ARITHMETIC)
+ if (reserved_word_acceptable (last_read_token))
+ {
+ sline = line_number;
+
+ cmdtyp = parse_arith_cmd (&wval, 0);
+ if (cmdtyp == 1) /* arithmetic command */
{
- nestret = parse_matched_pair (0, '`', '`', &nestlen, 0);
- 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;
- }
- FREE (nestret);
+ 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);
+ return (ARITH_CMD);
}
- else if (was_dollar && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
- /* check for $(), $[], or ${} inside quoted string. */
+ else if (cmdtyp == 0) /* nested subshell */
{
- if (open == ch) /* undo previous increment */
- count--;
- if (ch == '(') /* ) */
- nestret = parse_matched_pair (0, '(', ')', &nestlen, 0);
- else if (ch == '{') /* } */
- nestret = parse_matched_pair (0, '{', '}', &nestlen, 0);
- else if (ch == '[') /* ] */
- nestret = parse_matched_pair (0, '[', ']', &nestlen, 0);
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
- }
- if (nestlen)
- {
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
- }
- FREE (nestret);
+ push_string (wval, 0, (alias_t *)NULL);
+ pushed_string_list->flags = PSH_DPAREN;
+ if ((parser_state & PST_CASEPAT) == 0)
+ parser_state |= PST_SUBSHELL;
+ return (c);
}
- was_dollar = (ch == '$');
+ else /* ERROR */
+ return -1;
}
+#endif
- ret[retind] = '\0';
- if (lenp)
- *lenp = retind;
- return ret;
+ return -2; /* XXX */
}
-#if defined (DPAREN_ARITHMETIC)
/* We've seen a `(('. Look for the matching `))'. If we get it, return 1.
If not, assume it's a nested subshell for backwards compatibility and
return 0. In any case, put the characters we've consumed into a locally-
allocated buffer and make *ep point to that buffer. Return -1 on an
error, for example EOF. */
static int
-parse_arith_cmd (ep)
+parse_arith_cmd (ep, adddq)
char **ep;
+ int adddq;
{
int exp_lineno, rval, c;
- char *ttok, *token;
+ char *ttok, *tokstr;
int ttoklen;
exp_lineno = line_number;
return -1;
/* Check that the next character is the closing right paren. If
not, this is a syntax error. ( */
- if ((c = shell_getc (0)) != ')')
+ c = shell_getc (0);
+ if MBTEST(c != ')')
rval = 0;
- token = xmalloc(ttoklen + 4);
+ tokstr = (char *)xmalloc (ttoklen + 4);
- /* (( ... )) -> "..." */
- token[0] = (rval == 1) ? '"' : '(';
- strncpy (token + 1, ttok, ttoklen - 1); /* don't copy the final `)' */
- if (rval == 1)
+ /* if ADDDQ != 0 then (( ... )) -> "..." */
+ if (rval == 1 && adddq) /* arith cmd, add double quotes */
{
- token[ttoklen] = '"';
- token[ttoklen+1] = '\0';
+ tokstr[0] = '"';
+ strncpy (tokstr + 1, ttok, ttoklen - 1);
+ tokstr[ttoklen] = '"';
+ tokstr[ttoklen+1] = '\0';
}
- else
+ else if (rval == 1) /* arith cmd, don't add double quotes */
+ {
+ strncpy (tokstr, ttok, ttoklen - 1);
+ tokstr[ttoklen-1] = '\0';
+ }
+ else /* nested subshell */
{
- token[ttoklen] = ')';
- token[ttoklen+1] = c;
- token[ttoklen+2] = '\0';
+ tokstr[0] = '(';
+ strncpy (tokstr + 1, ttok, ttoklen - 1);
+ tokstr[ttoklen] = ')';
+ tokstr[ttoklen+1] = c;
+ tokstr[ttoklen+2] = '\0';
}
- *ep = token;
+
+ *ep = tokstr;
FREE (ttok);
return rval;
}
-#endif /* DPAREN_ARITHMETIC */
+#endif /* DPAREN_ARITHMETIC || ARITH_FOR_COMMAND */
#if defined (COND_COMMAND)
-static COND_COM *cond_term ();
-static COND_COM *cond_and ();
-static COND_COM *cond_or ();
-static COND_COM *cond_expr ();
+static void
+cond_error ()
+{
+ char *etext;
+
+ if (EOF_Reached && cond_token != COND_ERROR) /* [[ */
+ parser_error (cond_lineno, _("unexpected EOF while looking for `]]'"));
+ else if (cond_token != COND_ERROR)
+ {
+ if (etext = error_token_from_token (cond_token))
+ {
+ parser_error (cond_lineno, _("syntax error in conditional expression: unexpected token `%s'"), etext);
+ free (etext);
+ }
+ else
+ parser_error (cond_lineno, _("syntax error in conditional expression"));
+ }
+}
static COND_COM *
cond_expr ()
{
while ((cond_token = read_token (READ)) == '\n')
{
- if (interactive && (bash_input.type == st_stdin || bash_input.type == st_stream))
+ if (SHOULD_PROMPT ())
prompt_again ();
}
return (cond_token);
WORD_DESC *op;
COND_COM *term, *tleft, *tright;
int tok, lineno;
+ char *etext;
/* Read a token. It can be a left paren, a `!', a unary operator, or a
word that should be the first argument of a binary operator. Start by
{
if (term)
dispose_cond_node (term); /* ( */
- parser_error (lineno, "expected `)'");
+ if (etext = error_token_from_token (cond_token))
+ {
+ parser_error (lineno, _("unexpected token `%s', expected `)'"), etext);
+ free (etext);
+ }
+ else
+ parser_error (lineno, _("expected `)'"));
COND_RETURN_ERROR ();
}
term = make_cond_node (COND_EXPR, (WORD_DESC *)NULL, term, (COND_COM *)NULL);
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);
else
{
dispose_word (op);
- parser_error (line_number, "unexpected argument to conditional unary operator");
+ if (etext = error_token_from_token (tok))
+ {
+ parser_error (line_number, _("unexpected argument `%s' to conditional unary operator"), etext);
+ free (etext);
+ }
+ else
+ parser_error (line_number, _("unexpected argument to conditional unary operator"));
COND_RETURN_ERROR ();
}
(void)cond_skip_newlines ();
}
- else /* left argument to binary operator */
+ else if (tok == WORD) /* left argument to binary operator */
{
/* lhs */
tleft = make_cond_node (COND_TERM, yylval.word, (COND_COM *)NULL, (COND_COM *)NULL);
/* binop */
tok = read_token (READ);
if (tok == WORD && test_binop (yylval.word->word))
- op = yylval.word;
+ {
+ op = yylval.word;
+ if (op->word[0] == '=' && (op->word[1] == '\0' || (op->word[1] == '=' && op->word[2] == '\0')))
+ parser_state |= PST_EXTPAT;
+ else if (op->word[0] == '!' && op->word[1] == '=' && op->word[2] == '\0')
+ parser_state |= PST_EXTPAT;
+ }
+#if defined (COND_REGEXP)
+ 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);
- else if (tok == COND_END || tok == AND_AND || tok == OR_OR)
+ op = make_word_from_token (tok); /* ( */
+ /* There should be a check before blindly accepting the `)' that we have
+ seen the opening `('. */
+ else if (tok == COND_END || tok == AND_AND || tok == OR_OR || tok == ')')
{
/* Special case. [[ x ]] is equivalent to [[ -n x ]], just like
the test command. Similarly for [[ x && expr ]] or
- [[ x || expr ]] */
+ [[ x || expr ]] or [[ (x) ]]. */
op = make_word ("-n");
term = make_cond_node (COND_UNARY, op, tleft, (COND_COM *)NULL);
cond_token = tok;
}
else
{
- parser_error (line_number, "conditional binary operator expected");
+ if (etext = error_token_from_token (tok))
+ {
+ parser_error (line_number, _("unexpected token `%s', conditional binary operator expected"), etext);
+ free (etext);
+ }
+ else
+ parser_error (line_number, _("conditional binary operator expected"));
dispose_cond_node (tleft);
COND_RETURN_ERROR ();
}
/* rhs */
+ if (parser_state & PST_EXTPAT)
+ extended_glob = 1;
tok = read_token (READ);
+ if (parser_state & PST_EXTPAT)
+ extended_glob = global_extglob;
+ parser_state &= ~(PST_REGEXP|PST_EXTPAT);
+
if (tok == WORD)
{
tright = make_cond_node (COND_TERM, yylval.word, (COND_COM *)NULL, (COND_COM *)NULL);
}
else
{
- parser_error (line_number, "unexpected argument to conditional binary operator");
+ if (etext = error_token_from_token (tok))
+ {
+ parser_error (line_number, _("unexpected argument `%s' to conditional binary operator"), etext);
+ free (etext);
+ }
+ else
+ parser_error (line_number, _("unexpected argument to conditional binary operator"));
dispose_cond_node (tleft);
dispose_word (op);
COND_RETURN_ERROR ();
(void)cond_skip_newlines ();
}
+ else
+ {
+ if (tok < 256)
+ parser_error (line_number, _("unexpected token `%c' in conditional command"), tok);
+ else if (etext = error_token_from_token (tok))
+ {
+ parser_error (line_number, _("unexpected token `%s' in conditional command"), etext);
+ free (etext);
+ }
+ else
+ parser_error (line_number, _("unexpected token %d in conditional command"), tok);
+ COND_RETURN_ERROR ();
+ }
return (term);
}
{
COND_COM *cexp;
+ global_extglob = extended_glob;
cexp = cond_expr ();
return (make_cond_command (cexp));
}
#endif
+#if defined (ARRAY_VARS)
+/* When this is called, it's guaranteed that we don't care about anything
+ in t beyond i. We do save and restore the chars, though. */
+static int
+token_is_assignment (t, i)
+ char *t;
+ int i;
+{
+ unsigned char c, c1;
+ int r;
+
+ c = t[i]; c1 = t[i+1];
+ t[i] = '='; t[i+1] = '\0';
+ r = assignment (t, (parser_state & PST_COMPASSIGN) != 0);
+ t[i] = c; t[i+1] = c1;
+ return r;
+}
+
+/* XXX - possible changes here for `+=' */
+static int
+token_is_ident (t, i)
+ char *t;
+ int i;
+{
+ unsigned char c;
+ int r;
+
+ c = t[i];
+ t[i] = '\0';
+ r = legal_identifier (t);
+ t[i] = c;
+ return r;
+}
+#endif
+
static int
read_token_word (character)
int character;
int token_index;
/* ALL_DIGITS becomes zero when we see a non-digit. */
- int all_digits;
+ int all_digit_token;
/* 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;
int result, peek_char;
char *ttok, *ttrans;
int ttoklen, ttranslen;
+ intmax_t lvalue;
if (token_buffer_size < TOKEN_DEFAULT_INITIAL_SIZE)
- token = xrealloc (token, token_buffer_size = TOKEN_DEFAULT_INITIAL_SIZE);
+ token = (char *)xrealloc (token, token_buffer_size = TOKEN_DEFAULT_INITIAL_SIZE);
token_index = 0;
- all_digits = digit (character);
- dollar_present = quoted = pass_next_character = 0;
+ all_digit_token = DIGIT (character);
+ 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);
/* Handle backslashes. Quote lots of things when not inside of
double-quotes, quote some things inside of double-quotes. */
- if (character == '\\')
+ if MBTEST(character == '\\')
{
peek_char = shell_getc (0);
/* If the next character is to be quoted, note it now. */
if (cd == 0 || cd == '`' ||
- (cd == '"' && member (peek_char, slashify_in_quotes)))
+ (cd == '"' && peek_char >= 0 && (sh_syntaxtab[peek_char] & CBSDQUOTE)))
pass_next_character++;
quoted = 1;
}
/* Parse a matched pair of quote characters. */
- if (shellquote (character))
+ 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. */
token[token_index++] = character;
strcpy (token + token_index, ttok);
token_index += ttoklen;
- all_digits = 0;
+ all_digit_token = 0;
quoted = 1;
dollar_present |= (character == '"' && strchr (ttok, '$') != 0);
FREE (ttok);
goto next_character;
}
+#ifdef COND_REGEXP
+ /* When parsing a regexp as a single word inside a conditional command,
+ we need to special-case characters special to both the shell and
+ regular expressions. Right now, that is only '(' and '|'. */ /*)*/
+ if MBTEST((parser_state & PST_REGEXP) && (character == '(' || character == '|')) /*)*/
+ {
+ if (character == '|')
+ goto got_character;
+
+ push_delimiter (dstack, character);
+ ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
+ pop_delimiter (dstack);
+ if (ttok == &matched_pair_error)
+ return -1; /* Bail immediately. */
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ token_buffer_size, TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = character;
+ strcpy (token + token_index, ttok);
+ token_index += ttoklen;
+ FREE (ttok);
+ dollar_present = all_digit_token = 0;
+ goto next_character;
+ }
+#endif /* COND_REGEXP */
+
#ifdef EXTENDED_GLOB
/* Parse a ksh-style extended pattern matching specification. */
- if (extended_glob && PATTERN_CHAR(character))
+ if MBTEST(extended_glob && PATTERN_CHAR (character))
{
peek_char = shell_getc (1);
- if (peek_char == '(') /* ) */
+ if MBTEST(peek_char == '(') /* ) */
{
push_delimiter (dstack, peek_char);
ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
pop_delimiter (dstack);
if (ttok == &matched_pair_error)
return -1; /* Bail immediately. */
- RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 3,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
token[token_index++] = character;
strcpy (token + token_index, ttok);
token_index += ttoklen;
FREE (ttok);
- dollar_present = all_digits = 0;
+ dollar_present = all_digit_token = 0;
goto next_character;
}
else
/* If the delimiter character is not single quote, parse some of
the shell expansions that must be read as a single word. */
-#if defined (PROCESS_SUBSTITUTION)
- if (character == '$' || character == '<' || character == '>')
-#else
- if (character == '$')
-#endif /* !PROCESS_SUBSTITUTION */
+ if (shellexp (character))
{
peek_char = shell_getc (1);
/* $(...), <(...), >(...), $((...)), ${...}, and $[...] constructs */
- if (peek_char == '(' ||
+ if MBTEST(peek_char == '(' ||
((peek_char == '{' || peek_char == '[') && character == '$')) /* ) ] } */
{
if (peek_char == '{') /* } */
- ttok = parse_matched_pair (cd, '{', '}', &ttoklen, 0);
+ ttok = parse_matched_pair (cd, '{', '}', &ttoklen, P_FIRSTCLOSE|P_DOLBRACE);
else if (peek_char == '(') /* ) */
{
/* XXX - push and pop the `(' as a delimiter for use by
the command-oriented-history code. This way newlines
appearing in the $(...) string get added to the
history literally rather than causing a possibly-
- incorrect `;' to be added. */
+ 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
ttok = parse_matched_pair (cd, '[', ']', &ttoklen, 0);
if (ttok == &matched_pair_error)
return -1; /* Bail immediately. */
- RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 3,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
token[token_index++] = character;
token_index += ttoklen;
FREE (ttok);
dollar_present = 1;
- all_digits = 0;
+ all_digit_token = 0;
goto next_character;
}
/* This handles $'...' and $"..." new-style quoted strings. */
- else if (character == '$' && (peek_char == '\'' || peek_char == '"'))
+ else if MBTEST(character == '$' && (peek_char == '\'' || peek_char == '"'))
{
int first_line;
first_line = line_number;
- ttok = parse_matched_pair (peek_char, peek_char, peek_char, &ttoklen, 0);
+ push_delimiter (dstack, peek_char);
+ ttok = parse_matched_pair (peek_char, peek_char, peek_char,
+ &ttoklen,
+ (peek_char == '\'') ? P_ALLOWESC : 0);
+ pop_delimiter (dstack);
if (ttok == &matched_pair_error)
return -1;
if (peek_char == '\'')
- ttrans = ansiexpand (ttok, 0, ttoklen - 1, &ttranslen);
+ {
+ ttrans = ansiexpand (ttok, 0, ttoklen - 1, &ttranslen);
+ free (ttok);
+
+ /* Insert the single quotes and correctly quote any
+ embedded single quotes (allowed because P_ALLOWESC was
+ passed to parse_matched_pair). */
+ ttok = sh_single_quote (ttrans);
+ free (ttrans);
+ ttranslen = strlen (ttok);
+ ttrans = ttok;
+ }
else
- ttrans = localeexpand (ttok, 0, ttoklen - 1, first_line, &ttranslen);
- free (ttok);
- RESIZE_MALLOCED_BUFFER (token, token_index, ttranslen + 2,
+ {
+ /* Try to locale-expand the converted string. */
+ ttrans = localeexpand (ttok, 0, ttoklen - 1, first_line, &ttranslen);
+ free (ttok);
+
+ /* Add the double quotes back */
+ ttok = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ ttranslen += 2;
+ ttrans = ttok;
+ }
+
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttranslen + 1,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
- token[token_index++] = peek_char;
strcpy (token + token_index, ttrans);
token_index += ttranslen;
- token[token_index++] = peek_char;
FREE (ttrans);
quoted = 1;
- all_digits = 0;
+ all_digit_token = 0;
+ goto next_character;
+ }
+ /* This could eventually be extended to recognize all of the
+ shell's single-character parameter expansions, and set flags.*/
+ else if MBTEST(character == '$' && peek_char == '$')
+ {
+ RESIZE_MALLOCED_BUFFER (token, token_index, 3,
+ token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = '$';
+ token[token_index++] = peek_char;
+ dollar_present = 1;
+ all_digit_token = 0;
goto next_character;
}
else
}
#if defined (ARRAY_VARS)
+ /* 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, P_ARRAYSUB);
+ 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);
+ all_digit_token = 0;
+ goto next_character;
+ }
/* Identify possible compound array variable assignment. */
- else if (character == '=' && token_index > 0)
+ 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 (peek_char == '(') /* ) */
+ if MBTEST(peek_char == '(') /* ) */
{
- ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
- if (ttok == &matched_pair_error)
- return -1; /* Bail immediately. */
- RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ ttok = parse_compound_assignment (&ttoklen);
+
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 4,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
- token[token_index++] = character;
- token[token_index++] = peek_char;
- strcpy (token + token_index, ttok);
- token_index += ttoklen;
+
+ token[token_index++] = '=';
+ token[token_index++] = '(';
+ if (ttok)
+ {
+ strcpy (token + token_index, ttok);
+ token_index += ttoklen;
+ }
+ token[token_index++] = ')';
FREE (ttok);
- all_digits = 0;
+ 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);
/* When not parsing a multi-character word construct, shell meta-
characters break words. */
- if (shellbreak (character))
+ if MBTEST(shellbreak (character))
{
shell_ungetc (character);
goto got_token;
}
- got_character:
-
- all_digits &= digit (character);
- dollar_present |= character == '$';
+got_character:
if (character == CTLESC || character == CTLNUL)
- token[token_index++] = CTLESC;
+ {
+ RESIZE_MALLOCED_BUFFER (token, token_index, 2, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = CTLESC;
+ }
+ else
+got_escaped_character:
+ RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
token[token_index++] = character;
- RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
- TOKEN_DEFAULT_GROW_SIZE);
+ all_digit_token &= DIGIT (character);
+ dollar_present |= character == '$';
next_character:
- if (character == '\n' && interactive &&
- (bash_input.type == st_stdin || bash_input.type == st_stream))
+ if (character == '\n' && SHOULD_PROMPT ())
prompt_again ();
/* We want to remove quoted newlines (that is, a \<newline> pair)
got_token:
+ /* Calls to RESIZE_MALLOCED_BUFFER ensure there is sufficient room. */
token[token_index] = '\0';
/* Check to see what thing we should return. If the last_read_token
is a `<', or a `&', or the character which ended this token is
a '>' or '<', then, and ONLY then, is this input token a NUMBER.
Otherwise, it is just a word, and should be returned as such. */
- if (all_digits && (character == '<' || character == '>' ||
+ if MBTEST(all_digit_token && (character == '<' || character == '>' ||
last_read_token == LESS_AND ||
last_read_token == GREATER_AND))
{
- yylval.number = atoi (token);
- return (NUMBER);
+ if (legal_number (token, &lvalue) && (int)lvalue == lvalue)
+ {
+ yylval.number = lvalue;
+ return (NUMBER);
+ }
}
/* Check for special case tokens. */
- result = special_case_tokens (token);
+ result = (last_shell_getc_is_singlebyte) ? special_case_tokens (token) : -1;
if (result >= 0)
return result;
/* Posix.2 does not allow reserved words to be aliased, so check for all
of them, including special cases, before expanding the current token
as an alias. */
- if (posixly_correct)
+ if MBTEST(posixly_correct)
CHECK_FOR_RESERVED_WORD (token);
/* Aliases are expanded iff EXPAND_ALIASES is non-zero, and quoting
/* If not in Posix.2 mode, check for reserved words after alias
expansion. */
- if (posixly_correct == 0)
+ if MBTEST(posixly_correct == 0)
#endif
CHECK_FOR_RESERVED_WORD (token);
the_word = (WORD_DESC *)xmalloc (sizeof (WORD_DESC));
- the_word->word = xmalloc (1 + token_index);
+ the_word->word = (char *)xmalloc (1 + token_index);
the_word->flags = 0;
strcpy (the_word->word, token);
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. */
- if (assignment (token))
+ if (assignment (token, (parser_state & PST_COMPASSIGN) != 0))
{
the_word->flags |= W_ASSIGNMENT;
/* Don't perform word splitting on assignment statements. */
- if (assignment_acceptable (last_read_token))
- the_word->flags |= W_NOSPLIT;
+ if (assignment_acceptable (last_read_token) || (parser_state & PST_COMPASSIGN) != 0)
+ {
+ the_word->flags |= W_NOSPLIT;
+ if (parser_state & PST_COMPASSIGN)
+ the_word->flags |= W_NOGLOB; /* XXX - W_NOBRACE? */
+ }
+ }
+
+ 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;
+ if (token[0] == '{' && token[token_index-1] == '}' &&
+ (character == '<' || character == '>'))
+ {
+ /* can use token; already copied to the_word */
+ token[token_index-1] = '\0';
+#if defined (ARRAY_VARS)
+ if (legal_identifier (token+1) || valid_array_reference (token+1))
+#else
+ if (legal_identifier (token+1))
+#endif
+ {
+ strcpy (the_word->word, token+1);
+/*itrace("read_token_word: returning REDIR_WORD for %s", the_word->word);*/
+ return (REDIR_WORD);
+ }
+ }
+
result = ((the_word->flags & (W_ASSIGNMENT|W_NOSPLIT)) == (W_ASSIGNMENT|W_NOSPLIT))
? ASSIGNMENT_WORD : WORD;
- if (last_read_token == FUNCTION)
+ switch (last_read_token)
{
+ case FUNCTION:
parser_state |= PST_ALLOWOPNBRC;
function_dstart = line_number;
+ break;
+ case CASE:
+ case SELECT:
+ case FOR:
+ if (word_top < MAX_CASE_NEST)
+ word_top++;
+ word_lineno[word_top] = line_number;
+ break;
}
return (result);
}
-/* $'...' ANSI-C expand the portion of STRING between START and END and
- return the result. The result cannot be longer than the input string. */
-static char *
-ansiexpand (string, start, end, lenp)
- char *string;
- int start, end, *lenp;
-{
- char *temp, *t;
- int len, tlen;
-
- temp = xmalloc (end - start + 1);
- for (tlen = 0, len = start; len < end; )
- temp[tlen++] = string[len++];
- temp[tlen] = '\0';
-
- if (*temp)
- {
- t = ansicstr (temp, tlen, (int *)NULL, lenp);
- free (temp);
- return (t);
- }
- else
- {
- if (lenp)
- *lenp = 0;
- return (temp);
- }
-}
-
-/* $"..." -- Translate the portion of STRING between START and END
- according to current locale using gettext (if available) and return
- the result. The caller will take care of leaving the quotes intact.
- The string will be left without the leading `$' by the caller.
- If translation is performed, the translated string will be double-quoted
- by the caller. The length of the translated string is returned in LENP,
- if non-null. */
-static char *
-localeexpand (string, start, end, lineno, lenp)
- char *string;
- int start, end, lineno, *lenp;
+/* Return 1 if TOKSYM is a token that after being read would allow
+ a reserved word to be seen, else 0. */
+static int
+reserved_word_acceptable (toksym)
+ int toksym;
{
- int len, tlen;
- char *temp, *t;
-
- temp = xmalloc (end - start + 1);
- for (tlen = 0, len = start; len < end; )
- temp[tlen++] = string[len++];
- temp[tlen] = '\0';
-
- /* If we're just dumping translatable strings, don't do anything. */
- if (dump_translatable_strings)
- {
- if (dump_po_strings)
- printf ("#: %s:%d\nmsgid \"%s\"\nmsgstr \"\"\n",
- (bash_input.name ? bash_input.name : "stdin"), lineno, temp);
- else
- printf ("\"%s\"\n", temp);
- if (lenp)
- *lenp = tlen;
- return (temp);
- }
- else if (*temp)
+ switch (toksym)
{
- t = localetrans (temp, tlen, &len);
- free (temp);
- if (lenp)
- *lenp = len;
- return (t);
- }
- else
- {
- if (lenp)
- *lenp = 0;
- return (temp);
+ case '\n':
+ case ';':
+ case '(':
+ case ')':
+ case '|':
+ case '&':
+ case '{':
+ case '}': /* XXX */
+ case AND_AND:
+ case BANG:
+ case BAR_AND:
+ case DO:
+ case DONE:
+ case ELIF:
+ case ELSE:
+ case ESAC:
+ case FI:
+ case IF:
+ case OR_OR:
+ case SEMI_SEMI:
+ case SEMI_AND:
+ case SEMI_SEMI_AND:
+ case THEN:
+ case TIME:
+ case TIMEOPT:
+ case TIMEIGN:
+ 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
+ if (last_read_token == WORD && token_before_that == FUNCTION)
+ return 1;
+ return 0;
}
}
-
-/* Return 1 if TOKEN is a token that after being read would allow
- a reserved word to be seen, else 0. */
-static int
-reserved_word_acceptable (token)
- int token;
-{
- if (token == '\n' || token == ';' || token == '(' || token == ')' ||
- token == '|' || token == '&' || token == '{' ||
- token == '}' || /* XXX */
- token == AND_AND ||
- token == BANG ||
- token == TIME || token == TIMEOPT ||
- token == DO ||
- token == ELIF ||
- token == ELSE ||
- token == FI ||
- token == IF ||
- token == OR_OR ||
- token == SEMI_SEMI ||
- token == THEN ||
- token == UNTIL ||
- token == WHILE ||
- token == DONE || /* XXX these two are experimental */
- token == ESAC ||
- token == 0)
- return (1);
- else
- return (0);
-}
-
+
/* Return the index of TOKEN in the alist of reserved words, or -1 if
TOKEN is not a shell reserved word. */
int
-find_reserved_word (token)
- char *token;
+find_reserved_word (tokstr)
+ char *tokstr;
{
int i;
for (i = 0; word_token_alist[i].word; i++)
- if (STREQ (token, word_token_alist[i].word))
+ if (STREQ (tokstr, word_token_alist[i].word))
return i;
return -1;
}
+/* An interface to let the rest of the shell (primarily the completion
+ system) know what the parser is expecting. */
+int
+parser_in_command_position ()
+{
+ return (command_token_position (last_read_token));
+}
+
#if 0
#if defined (READLINE)
/* Called after each time readline is called. This insures that whatever
if (temp_prompt == 0)
{
- temp_prompt = xmalloc (1);
+ temp_prompt = (char *)xmalloc (1);
temp_prompt[0] = '\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 we are not within a delimited expression, try to be smart
about which separators can be semi-colons and which must be
newlines. Returns the string that should be added into the
- history entry. */
+ history entry. LINE is the line we're about to add; it helps
+ make some more intelligent decisions in certain cases. */
char *
-history_delimiting_chars ()
+history_delimiting_chars (line)
+ const char *line;
{
+ static int last_was_heredoc = 0; /* was the last entry the start of a here document? */
register int i;
+ if ((parser_state & PST_HEREDOC) == 0)
+ last_was_heredoc = 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). We also keep LAST_WAS_HEREDOC as a private sentinel
+ variable to note when we think we added the first line of a here doc
+ (the one with a "<<" somewhere in it) */
+ if (parser_state & PST_HEREDOC)
+ {
+ if (last_was_heredoc)
+ {
+ last_was_heredoc = 0;
+ return "\n";
+ }
+ return (current_command_line_count == 2 ? "\n" : "");
+ }
+
+ if (parser_state & PST_COMPASSIGN)
+ return (" ");
+
/* First, handle some special cases. */
/*(*/
/* If we just read `()', assume it's a function definition, and don't
/* This does not work for subshells inside case statement
command lists. It's a suboptimal solution. */
else if (parser_state & PST_CASESTMT) /* case statement pattern */
- return " ";
+ return " ";
else
- return "; "; /* (...) subshell */
+ return "; "; /* (...) subshell */
}
else if (token_before_that == WORD && two_tokens_ago == FUNCTION)
return " "; /* function def using `function name' without `()' */
+ /* If we're not in a here document, but we think we're about to parse one,
+ and we would otherwise return a `;', return a newline to delimit the
+ line with the here-doc delimiter */
+ else if ((parser_state & PST_HEREDOC) == 0 && current_command_line_count > 1 && last_read_token == '\n' && strstr (line, "<<"))
+ {
+ last_was_heredoc = 1;
+ return "\n";
+ }
+
+ else if (token_before_that == WORD && two_tokens_ago == FOR)
+ {
+ /* 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++)
+ ;
+ 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++)
{
if (token_before_that == no_semi_successors[i])
{
char *temp_prompt;
- if (!interactive) /* XXX */
+ if (interactive == 0 || expanding_alias ()) /* XXX */
return;
ps1_prompt = get_string_value ("PS1");
if (temp_prompt == 0)
{
- temp_prompt = xmalloc (1);
+ temp_prompt = (char *)xmalloc (1);
temp_prompt[0] = '\0';
}
}
}
+int
+get_current_prompt_level ()
+{
+ return ((current_prompt_string && current_prompt_string == ps2_prompt) ? 2 : 1);
+}
+
+void
+set_current_prompt_level (x)
+ int x;
+{
+ prompt_string_pointer = (x == 2) ? &ps2_prompt : &ps1_prompt;
+ current_prompt_string = *prompt_string_pointer;
+}
+
static void
print_prompt ()
{
may contain special characters which are decoded as follows:
\a bell (ascii 07)
- \e escape (ascii 033)
\d the date in Day Mon Date format
+ \e escape (ascii 033)
\h the hostname up to the first `.'
\H the hostname
+ \j the number of active jobs
+ \l the basename of the shell's tty device name
\n CRLF
+ \r CR
\s the name of the shell
\t the time in 24-hour hh:mm:ss format
\T the time in 12-hour hh:mm:ss format
- \@ the time in 12-hour am/pm format
+ \@ the time in 12-hour hh:mm am/pm format
+ \A the time in 24-hour hh:mm format
+ \D{fmt} the result of passing FMT to strftime(3)
+ \u your username
\v the version of bash (e.g., 2.00)
\V the release of bash, version + patchlevel (e.g., 2.00.0)
\w the current working directory
\W the last element of $PWD
- \u your username
- \# the command number of this command
\! the history number of this command
+ \# the command number of this command
\$ a $ or a # if you are root
\nnn character code nnn in octal
\\ a backslash
WORD_LIST *list;
char *result, *t;
struct dstack save_dstack;
+ int last_exit_value, last_comsub_pid;
#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;
+ char timebuf[128];
+ char *timefmt;
- result = xmalloc (result_size = PROMPT_GROWTH);
+ result = (char *)xmalloc (result_size = PROMPT_GROWTH);
result[result_index = 0] = 0;
temp = (char *)NULL;
octal_string[3] = '\0';
n = read_octal (octal_string);
- temp = xmalloc (3);
+ temp = (char *)xmalloc (3);
if (n == CTLESC || n == CTLNUL)
{
- string += 3;
temp[0] = CTLESC;
temp[1] = n;
temp[2] = '\0';
}
else
{
- string += 3;
temp[0] = n;
temp[1] = '\0';
}
- c = 0;
+ for (c = 0; n != -1 && c < 3 && ISOCTAL (*string); c++)
+ string++;
+
+ c = 0; /* tested at add_string: */
goto add_string;
- case 't':
case 'd':
+ case 't':
case 'T':
case '@':
+ case 'A':
/* Make the current time/date into a string. */
- the_time = time (0);
- temp = ctime (&the_time);
+ (void) time (&the_time);
+#if defined (HAVE_TZSET)
+ sv_tz ("TZ"); /* XXX -- just make sure */
+#endif
+ tm = localtime (&the_time);
+
+ if (c == 'd')
+ n = strftime (timebuf, sizeof (timebuf), "%a %b %d", tm);
+ else if (c == 't')
+ n = strftime (timebuf, sizeof (timebuf), "%H:%M:%S", tm);
+ else if (c == 'T')
+ n = strftime (timebuf, sizeof (timebuf), "%I:%M:%S", tm);
+ else if (c == '@')
+ n = strftime (timebuf, sizeof (timebuf), "%I:%M %p", tm);
+ else if (c == 'A')
+ n = strftime (timebuf, sizeof (timebuf), "%H:%M", tm);
+
+ if (n == 0)
+ timebuf[0] = '\0';
+ else
+ timebuf[sizeof(timebuf) - 1] = '\0';
- temp = (c != 'd') ? savestring (temp + 11) : savestring (temp);
- temp[(c != 'd') ? 8 : 10] = '\0';
+ temp = savestring (timebuf);
+ goto add_string;
- /* quick and dirty conversion to 12-hour time */
- if (c == 'T' || c == '@')
+ case 'D': /* strftime format */
+ if (string[1] != '{') /* } */
+ goto not_escape;
+
+ (void) time (&the_time);
+ tm = localtime (&the_time);
+ string += 2; /* skip { */
+ timefmt = xmalloc (strlen (string) + 3);
+ for (t = timefmt; *string && *string != '}'; )
+ *t++ = *string++;
+ *t = '\0';
+ c = *string; /* tested at add_string */
+ if (timefmt[0] == '\0')
{
- if (c == '@')
- {
- temp[5] = 'a'; /* am/pm format */
- temp[6] = 'm';
- temp[7] = '\0';
- }
- c = temp[2];
- temp[2] = '\0';
- n = atoi (temp);
- temp[2] = c;
- n -= 12;
- if (n > 0)
- {
- temp[0] = (n / 10) + '0';
- temp[1] = (n % 10) + '0';
- }
- if (n >= 0 && temp[5] == 'a')
- temp[5] = 'p';
+ timefmt[0] = '%';
+ timefmt[1] = 'X'; /* locale-specific current time */
+ timefmt[2] = '\0';
}
- goto add_string;
+ n = strftime (timebuf, sizeof (timebuf), timefmt, tm);
+ free (timefmt);
- case 'r':
- temp = xmalloc (2);
- temp[0] = '\r';
- temp[1] = '\0';
- goto add_string;
+ if (n == 0)
+ timebuf[0] = '\0';
+ else
+ timebuf[sizeof(timebuf) - 1] = '\0';
+ if (promptvars || posixly_correct)
+ /* Make sure that expand_prompt_string is called with a
+ second argument of Q_DOUBLE_QUOTES if we use this
+ function here. */
+ temp = sh_backslash_quote_for_double_quotes (timebuf);
+ else
+ temp = savestring (timebuf);
+ goto add_string;
+
case 'n':
- temp = xmalloc (3);
+ temp = (char *)xmalloc (3);
temp[0] = no_line_editing ? '\n' : '\r';
temp[1] = no_line_editing ? '\0' : '\n';
temp[2] = '\0';
case 'v':
case 'V':
- temp = xmalloc (8);
+ temp = (char *)xmalloc (16);
if (c == 'v')
strcpy (temp, dist_version);
else
{
if (getcwd (t_string, sizeof(t_string)) == 0)
{
- t_string[0] = '.';
+ t_string[0] = '.';
tlen = 1;
}
else
}
t_string[tlen] = '\0';
- if (c == 'W')
+#if defined (MACOSX)
+ /* Convert from "fs" format to "input" format */
+ temp = fnx_fromfs (t_string, strlen (t_string));
+ if (temp != t_string)
+ strcpy (t_string, temp);
+#endif
+
+#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 == $HOME */
+ if (c == 'W' && (((t = get_string_value ("HOME")) == 0) || STREQ (t, t_string) == 0))
{
- t = strrchr (t_string, '/');
- if (t && t != t_string)
- strcpy (t_string, t + 1);
+ if (ROOT_PATH (t_string) == 0 && DOUBLE_SLASH_ROOT (t_string) == 0)
+ {
+ t = strrchr (t_string, '/');
+ if (t)
+ memmove (t_string, t + 1, strlen (t)); /* strlen(t) to copy NULL */
+ }
}
+#undef ROOT_PATH
+#undef DOUBLE_SLASH_ROOT
else
- /* polite_directory_format is guaranteed to return a string
- no longer than PATH_MAX - 1 characters. */
- strcpy (t_string, polite_directory_format (t_string));
+ {
+ /* polite_directory_format is guaranteed to return a string
+ no longer than PATH_MAX - 1 characters. */
+ temp = polite_directory_format (t_string);
+ if (temp != t_string)
+ strcpy (t_string, temp);
+ }
+ temp = trim_pathname (t_string, PATH_MAX - 1);
/* If we're going to be expanding the prompt string later,
quote the directory name. */
if (promptvars || posixly_correct)
- temp = backslash_quote (t_string);
+ /* Make sure that expand_prompt_string is called with a
+ second argument of Q_DOUBLE_QUOTES if we use this
+ function here. */
+ temp = sh_backslash_quote_for_double_quotes (t_string);
else
temp = savestring (t_string);
}
case 'u':
+ if (current_user.user_name == 0)
+ get_current_user_info ();
temp = savestring (current_user.user_name);
goto add_string;
goto add_string;
case '$':
- temp = xmalloc (2);
- temp[0] = current_user.euid == 0 ? '#' : '$';
- temp[1] = '\0';
+ t = temp = (char *)xmalloc (3);
+ if ((promptvars || posixly_correct) && (current_user.euid != 0))
+ *t++ = '\\';
+ *t++ = current_user.euid == 0 ? '#' : '$';
+ *t = '\0';
+ goto add_string;
+
+ case 'j':
+ temp = itos (count_all_jobs ());
+ goto add_string;
+
+ case 'l':
+#if defined (HAVE_TTYNAME)
+ temp = (char *)ttyname (fileno (stdin));
+ t = temp ? base_pathname (temp) : "tty";
+ temp = savestring (t);
+#else
+ temp = savestring ("tty");
+#endif /* !HAVE_TTYNAME */
goto add_string;
#if defined (READLINE)
case '[':
case ']':
- temp = xmalloc (3);
- temp[0] = '\001';
- temp[1] = (c == '[') ? RL_PROMPT_START_IGNORE : RL_PROMPT_END_IGNORE;
- temp[2] = '\0';
+ if (no_line_editing)
+ {
+ string++;
+ break;
+ }
+ temp = (char *)xmalloc (3);
+ 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 */
case '\\':
- temp = xmalloc (2);
- temp[0] = c;
- temp[1] = '\0';
- goto add_string;
-
case 'a':
case 'e':
- temp = xmalloc (2);
- temp[0] = (c == 'a') ? '\07' : '\033';
+ case 'r':
+ temp = (char *)xmalloc (2);
+ if (c == 'a')
+ temp[0] = '\07';
+ else if (c == 'e')
+ temp[0] = '\033';
+ else if (c == 'r')
+ temp[0] = '\r';
+ else /* (c == '\\') */
+ temp[0] = c;
temp[1] = '\0';
goto add_string;
default:
- temp = xmalloc (3);
+not_escape:
+ temp = (char *)xmalloc (3);
temp[0] = '\\';
temp[1] = c;
temp[2] = '\0';
the prompt string. */
if (promptvars || posixly_correct)
{
- list = expand_string_unsplit (result, Q_DOUBLE_QUOTES);
+ last_exit_value = last_command_exit_value;
+ last_comsub_pid = last_command_subst_pid;
+ list = expand_prompt_string (result, Q_DOUBLE_QUOTES, 0);
free (result);
result = string_list (list);
dispose_words (list);
+ last_command_exit_value = last_exit_value;
+ last_command_subst_pid = last_comsub_pid;
}
else
{
return (result);
}
+/************************************************
+ * *
+ * ERROR HANDLING *
+ * *
+ ************************************************/
+
/* Report a syntax error, and restart the parser. Call here for fatal
errors. */
int
-yyerror ()
+yyerror (msg)
+ const char *msg;
{
report_syntax_error ((char *)NULL);
reset_parser ();
return (0);
}
+static char *
+error_token_from_token (tok)
+ int tok;
+{
+ char *t;
+
+ if (t = find_token_in_alist (tok, word_token_alist, 0))
+ return t;
+
+ if (t = find_token_in_alist (tok, other_token_alist, 0))
+ return t;
+
+ t = (char *)NULL;
+ /* This stuff is dicy and needs closer inspection */
+ switch (current_token)
+ {
+ case WORD:
+ case ASSIGNMENT_WORD:
+ if (yylval.word)
+ t = savestring (yylval.word->word);
+ break;
+ case NUMBER:
+ t = itos (yylval.number);
+ break;
+ case ARITH_CMD:
+ if (yylval.word_list)
+ t = string_list (yylval.word_list);
+ break;
+ case ARITH_FOR_EXPRS:
+ if (yylval.word_list)
+ t = string_list_internal (yylval.word_list, " ; ");
+ break;
+ case COND_CMD:
+ t = (char *)NULL; /* punt */
+ break;
+ }
+
+ return t;
+}
+
+static char *
+error_token_from_text ()
+{
+ char *msg, *t;
+ int token_end, i;
+
+ t = shell_input_line;
+ i = shell_input_line_index;
+ token_end = 0;
+ msg = (char *)NULL;
+
+ if (i && t[i] == '\0')
+ i--;
+
+ while (i && (whitespace (t[i]) || t[i] == '\n'))
+ i--;
+
+ if (i)
+ token_end = i + 1;
+
+ while (i && (member (t[i], " \n\t;|&") == 0))
+ i--;
+
+ while (i != token_end && (whitespace (t[i]) || t[i] == '\n'))
+ i++;
+
+ /* Return our idea of the offending token. */
+ if (token_end || (i == 0 && token_end == 0))
+ {
+ if (token_end)
+ msg = substring (t, i, token_end);
+ else /* one-character token */
+ {
+ msg = (char *)xmalloc (2);
+ msg[0] = t[i];
+ msg[1] = '\0';
+ }
+ }
+
+ return (msg);
+}
+
+static void
+print_offending_line ()
+{
+ char *msg;
+ int token_end;
+
+ msg = savestring (shell_input_line);
+ token_end = strlen (msg);
+ while (token_end && msg[token_end - 1] == '\n')
+ msg[--token_end] = '\0';
+
+ parser_error (line_number, "`%s'", msg);
+ free (msg);
+}
+
/* Report a syntax error with line numbers, etc.
Call here for recoverable errors. If you have a message to print,
then place it in MESSAGE, otherwise pass NULL and this will figure
report_syntax_error (message)
char *message;
{
- char *msg, *t;
- int token_end, i;
- char msg2[2];
+ char *msg, *p;
if (message)
{
parser_error (line_number, "%s", message);
if (interactive && EOF_Reached)
EOF_Reached = 0;
- last_command_exit_value = EX_USAGE;
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
return;
}
/* If the line of input we're reading is not null, try to find the
- objectionable token. */
- if (shell_input_line && *shell_input_line)
+ objectionable token. First, try to figure out what token the
+ parser's complaining about by looking at current_token. */
+ if (current_token != 0 && EOF_Reached == 0 && (msg = error_token_from_token (current_token)))
{
- t = shell_input_line;
- i = shell_input_line_index;
- token_end = 0;
-
- if (i && t[i] == '\0')
- i--;
-
- while (i && (whitespace (t[i]) || t[i] == '\n'))
- i--;
-
- if (i)
- token_end = i + 1;
+ if (ansic_shouldquote (msg))
+ {
+ p = ansic_quote (msg, 0, NULL);
+ free (msg);
+ msg = p;
+ }
+ parser_error (line_number, _("syntax error near unexpected token `%s'"), msg);
+ free (msg);
- while (i && (member (t[i], " \n\t;|&") == 0))
- i--;
+ if (interactive == 0)
+ print_offending_line ();
- while (i != token_end && (whitespace (t[i]) || t[i] == '\n'))
- i++;
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
+ return;
+ }
- /* Print the offending token. */
- if (token_end || (i == 0 && token_end == 0))
+ /* If looking at the current token doesn't prove fruitful, try to find the
+ offending token by analyzing the text of the input line near the current
+ input line index and report what we find. */
+ if (shell_input_line && *shell_input_line)
+ {
+ msg = error_token_from_text ();
+ if (msg)
{
- if (token_end)
- {
- msg = xmalloc (1 + (token_end - i));
- strncpy (msg, t + i, token_end - i);
- msg[token_end - i] = '\0';
- }
- else /* one-character token */
- {
- msg2[0] = t[i];
- msg2[1] = '\0';
- msg = msg2;
- }
-
- parser_error (line_number, "syntax error near unexpected token `%s'", msg);
-
- if (msg != msg2)
- free (msg);
+ parser_error (line_number, _("syntax error near `%s'"), msg);
+ free (msg);
}
/* If not interactive, print the line containing the error. */
if (interactive == 0)
- {
- msg = savestring (shell_input_line);
- token_end = strlen (msg);
- while (token_end && msg[token_end - 1] == '\n')
- msg[--token_end] = '\0';
-
- parser_error (line_number, "`%s'", msg);
- free (msg);
- }
+ print_offending_line ();
}
else
{
- msg = EOF_Reached ? "syntax error: unexpected end of file" : "syntax error";
+ msg = EOF_Reached ? _("syntax error: unexpected end of file") : _("syntax error");
parser_error (line_number, "%s", msg);
/* When the shell is interactive, this file uses EOF_Reached
only for error reporting. Other mechanisms are used to
if (interactive && EOF_Reached)
EOF_Reached = 0;
}
- last_command_exit_value = EX_USAGE;
+
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
}
/* ??? Needed function. ??? We have to be able to discard the constructs
created during parsing. In the case of error, we want to return
allocated objects to the memory pool. In the case of no error, we want
to throw away the information about where the allocated objects live.
- (dispose_command () will actually free the command. */
+ (dispose_command () will actually free the command.) */
static void
discard_parser_constructs (error_p)
int error_p;
{
}
+/************************************************
+ * *
+ * EOF HANDLING *
+ * *
+ ************************************************/
+
/* Do that silly `type "bye" to exit' stuff. You know, "ignoreeof". */
/* A flag denoting whether or not ignoreeof is set. */
{
if (eof_encountered < eof_encountered_limit)
{
- fprintf (stderr, "Use \"%s\" to leave the shell.\n",
+ fprintf (stderr, _("Use \"%s\" to leave the shell.\n"),
login_shell ? "logout" : "exit");
eof_encountered++;
+ /* Reset the parsing state. */
+ last_read_token = current_token = '\n';
/* Reset the prompt string to be $PS1. */
prompt_string_pointer = (char **)NULL;
prompt_again ();
- last_read_token = current_token = '\n';
return;
}
}
EOF_Reached = 1;
}
}
+
+/************************************************
+ * *
+ * STRING PARSING FUNCTIONS *
+ * *
+ ************************************************/
+
+/* It's very important that these two functions treat the characters
+ between ( and ) identically. */
+
+static WORD_LIST parse_string_error;
+
+/* Take a string and run it through the shell parser, returning the
+ resultant word list. Used by compound array assignment. */
+WORD_LIST *
+parse_string_to_word_list (s, flags, whom)
+ char *s;
+ int flags;
+ const char *whom;
+{
+ WORD_LIST *wl;
+ int tok, orig_current_token, orig_line_number, orig_input_terminator;
+ int orig_line_count;
+ int old_echo_input, old_expand_aliases;
+#if defined (HISTORY)
+ int old_remember_on_history, old_history_expansion_inhibited;
+#endif
+
+#if defined (HISTORY)
+ old_remember_on_history = remember_on_history;
+# if defined (BANG_HISTORY)
+ old_history_expansion_inhibited = history_expansion_inhibited;
+# endif
+ bash_history_disable ();
+#endif
+
+ orig_line_number = line_number;
+ orig_line_count = current_command_line_count;
+ orig_input_terminator = shell_input_line_terminator;
+ old_echo_input = echo_input_at_read;
+ old_expand_aliases = expand_aliases;
+
+ push_stream (1);
+ last_read_token = WORD; /* WORD to allow reserved words here */
+ current_command_line_count = 0;
+ echo_input_at_read = expand_aliases = 0;
+
+ with_input_from_string (s, whom);
+ wl = (WORD_LIST *)NULL;
+
+ if (flags & 1)
+ parser_state |= PST_COMPASSIGN|PST_REPARSE;
+
+ while ((tok = read_token (READ)) != yacc_EOF)
+ {
+ if (tok == '\n' && *bash_input.location.string == '\0')
+ break;
+ if (tok == '\n') /* Allow newlines in compound assignments */
+ continue;
+ if (tok != WORD && tok != ASSIGNMENT_WORD)
+ {
+ line_number = orig_line_number + line_number - 1;
+ orig_current_token = current_token;
+ current_token = tok;
+ yyerror (NULL); /* does the right thing */
+ current_token = orig_current_token;
+ if (wl)
+ dispose_words (wl);
+ wl = &parse_string_error;
+ break;
+ }
+ wl = make_word_list (yylval.word, wl);
+ }
+
+ last_read_token = '\n';
+ pop_stream ();
+
+#if defined (HISTORY)
+ remember_on_history = old_remember_on_history;
+# if defined (BANG_HISTORY)
+ history_expansion_inhibited = old_history_expansion_inhibited;
+# endif /* BANG_HISTORY */
+#endif /* HISTORY */
+
+ echo_input_at_read = old_echo_input;
+ expand_aliases = old_expand_aliases;
+
+ current_command_line_count = orig_line_count;
+ shell_input_line_terminator = orig_input_terminator;
+
+ if (flags & 1)
+ parser_state &= ~(PST_COMPASSIGN|PST_REPARSE);
+
+ if (wl == &parse_string_error)
+ {
+ last_command_exit_value = EXECUTION_FAILURE;
+ if (interactive_shell == 0 && posixly_correct)
+ jump_to_top_level (FORCE_EOF);
+ else
+ jump_to_top_level (DISCARD);
+ }
+
+ return (REVERSE_LIST (wl, WORD_LIST *));
+}
+
+static char *
+parse_compound_assignment (retlenp)
+ int *retlenp;
+{
+ WORD_LIST *wl, *rl;
+ 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;
+
+ while ((tok = read_token (READ)) != ')')
+ {
+ if (tok == '\n') /* Allow newlines in compound assignments */
+ {
+ if (SHOULD_PROMPT ())
+ prompt_again ();
+ continue;
+ }
+ if (tok != WORD && tok != ASSIGNMENT_WORD)
+ {
+ current_token = tok; /* for error reporting */
+ if (tok == yacc_EOF) /* ( */
+ parser_error (orig_line_number, _("unexpected EOF while looking for matching `)'"));
+ else
+ yyerror(NULL); /* does the right thing */
+ if (wl)
+ dispose_words (wl);
+ wl = &parse_string_error;
+ break;
+ }
+ wl = make_word_list (yylval.word, wl);
+ }
+
+ FREE (token);
+ token = saved_token;
+ token_buffer_size = orig_token_size;
+
+ parser_state &= ~PST_COMPASSIGN;
+
+ if (wl == &parse_string_error)
+ {
+ last_command_exit_value = EXECUTION_FAILURE;
+ last_read_token = '\n'; /* XXX */
+ if (interactive_shell == 0 && posixly_correct)
+ jump_to_top_level (FORCE_EOF);
+ else
+ jump_to_top_level (DISCARD);
+ }
+
+ last_read_token = orig_last_token; /* XXX - was WORD? */
+
+ if (wl)
+ {
+ rl = REVERSE_LIST (wl, WORD_LIST *);
+ ret = string_list (rl);
+ dispose_words (rl);
+ }
+ else
+ ret = (char *)NULL;
+
+ if (retlenp)
+ *retlenp = (ret && *ret) ? strlen (ret) : 0;
+
+ if (assignok)
+ parser_state |= PST_ASSIGNOK;
+
+ return ret;
+}
+
+/************************************************
+ * *
+ * SAVING AND RESTORING PARTIAL PARSE STATE *
+ * *
+ ************************************************/
+
+sh_parser_state_t *
+save_parser_state (ps)
+ sh_parser_state_t *ps;
+{
+ if (ps == 0)
+ ps = (sh_parser_state_t *)xmalloc (sizeof (sh_parser_state_t));
+ if (ps == 0)
+ return ((sh_parser_state_t *)NULL);
+
+ ps->parser_state = parser_state;
+ ps->token_state = save_token_state ();
+
+ ps->input_line_terminator = shell_input_line_terminator;
+ ps->eof_encountered = eof_encountered;
+
+ ps->prompt_string_pointer = prompt_string_pointer;
+
+ ps->current_command_line_count = current_command_line_count;
+
+#if defined (HISTORY)
+ ps->remember_on_history = remember_on_history;
+# if defined (BANG_HISTORY)
+ ps->history_expansion_inhibited = history_expansion_inhibited;
+# endif
+#endif
+
+ ps->last_command_exit_value = last_command_exit_value;
+#if defined (ARRAY_VARS)
+ ps->pipestatus = save_pipestatus_array ();
+#endif
+
+ ps->last_shell_builtin = last_shell_builtin;
+ ps->this_shell_builtin = this_shell_builtin;
+
+ ps->expand_aliases = expand_aliases;
+ ps->echo_input_at_read = echo_input_at_read;
+
+ ps->token = token;
+ ps->token_buffer_size = token_buffer_size;
+ /* Force reallocation on next call to read_token_word */
+ token = 0;
+ token_buffer_size = 0;
+
+ return (ps);
+}
+
+void
+restore_parser_state (ps)
+ sh_parser_state_t *ps;
+{
+ if (ps == 0)
+ return;
+
+ parser_state = ps->parser_state;
+ if (ps->token_state)
+ {
+ restore_token_state (ps->token_state);
+ free (ps->token_state);
+ }
+
+ shell_input_line_terminator = ps->input_line_terminator;
+ eof_encountered = ps->eof_encountered;
+
+ prompt_string_pointer = ps->prompt_string_pointer;
+
+ current_command_line_count = ps->current_command_line_count;
+
+#if defined (HISTORY)
+ remember_on_history = ps->remember_on_history;
+# if defined (BANG_HISTORY)
+ history_expansion_inhibited = ps->history_expansion_inhibited;
+# endif
+#endif
+
+ last_command_exit_value = ps->last_command_exit_value;
+#if defined (ARRAY_VARS)
+ restore_pipestatus_array (ps->pipestatus);
+#endif
+
+ last_shell_builtin = ps->last_shell_builtin;
+ this_shell_builtin = ps->this_shell_builtin;
+
+ expand_aliases = ps->expand_aliases;
+ echo_input_at_read = ps->echo_input_at_read;
+
+ FREE (token);
+ token = ps->token;
+ token_buffer_size = ps->token_buffer_size;
+}
+
+sh_input_line_state_t *
+save_input_line_state (ls)
+ sh_input_line_state_t *ls;
+{
+ if (ls == 0)
+ ls = (sh_input_line_state_t *)xmalloc (sizeof (sh_input_line_state_t));
+ if (ls == 0)
+ return ((sh_input_line_state_t *)NULL);
+
+ ls->input_line = shell_input_line;
+ ls->input_line_size = shell_input_line_size;
+ ls->input_line_len = shell_input_line_len;
+ ls->input_line_index = shell_input_line_index;
+
+ /* force reallocation */
+ shell_input_line = 0;
+ shell_input_line_size = shell_input_line_len = shell_input_line_index = 0;
+
+ return ls;
+}
+
+void
+restore_input_line_state (ls)
+ sh_input_line_state_t *ls;
+{
+ FREE (shell_input_line);
+ shell_input_line = ls->input_line;
+ shell_input_line_size = ls->input_line_size;
+ shell_input_line_len = ls->input_line_len;
+ shell_input_line_index = ls->input_line_index;
+
+ set_line_mbstate ();
+}
+
+/************************************************
+ * *
+ * MULTIBYTE CHARACTER HANDLING *
+ * *
+ ************************************************/
+
+#if defined (HANDLE_MULTIBYTE)
+static void
+set_line_mbstate ()
+{
+ int c;
+ size_t i, previ, len;
+ mbstate_t mbs, prevs;
+ size_t mbclen;
+
+ if (shell_input_line == NULL)
+ return;
+ len = strlen (shell_input_line); /* XXX - shell_input_line_len ? */
+ FREE (shell_input_line_property);
+ shell_input_line_property = (char *)xmalloc (len + 1);
+
+ memset (&prevs, '\0', sizeof (mbstate_t));
+ for (i = previ = 0; i < len; i++)
+ {
+ mbs = prevs;
+
+ c = shell_input_line[i];
+ if (c == EOF)
+ {
+ size_t j;
+ for (j = i; j < len; j++)
+ shell_input_line_property[j] = 1;
+ break;
+ }
+
+ mbclen = mbrlen (shell_input_line + previ, i - previ + 1, &mbs);
+ if (mbclen == 1 || mbclen == (size_t)-1)
+ {
+ mbclen = 1;
+ previ = i + 1;
+ }
+ else if (mbclen == (size_t)-2)
+ mbclen = 0;
+ else if (mbclen > 1)
+ {
+ mbclen = 0;
+ previ = i + 1;
+ prevs = mbs;
+ }
+ else
+ {
+ /* XXX - what to do if mbrlen returns 0? (null wide character) */
+ size_t j;
+ for (j = i; j < len; j++)
+ shell_input_line_property[j] = 1;
+ break;
+ }
+
+ shell_input_line_property[i] = mbclen;
+ }
+}
+#endif /* HANDLE_MULTIBYTE */
projects / thirdparty / bash.git / blobdiff