-/* Yacc grammar for bash. */
+/* parse.y - Yacc grammar for bash. */
-/* Copyright (C) 1989 Free Software Foundation, Inc.
+/* Copyright (C) 1989-2022 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 "execute_cmd.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 (JOB_CONTROL)
# include "jobs.h"
+#else
+extern int cleanup_dead_jobs PARAMS((void));
#endif /* JOB_CONTROL */
#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
+#define END_ALIAS -2
-#if defined (EXTENDED_GLOB)
-#define PATTERN_CHAR(c) \
- ((c) == '@' || (c) == '*' || (c) == '+' || (c) == '?' || (c) == '!')
+#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
+
+#define EXTEND_SHELL_INPUT_LINE_PROPERTY() \
+do { \
+ if (shell_input_line_len + 2 > shell_input_line_propsize) \
+ { \
+ shell_input_line_propsize = shell_input_line_len + 2; \
+ shell_input_line_property = (char *)xrealloc (shell_input_line_property, \
+ shell_input_line_propsize); \
+ } \
+} while (0)
+#if defined (EXTENDED_GLOB)
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 posixly_correct;
-extern int last_command_exit_value;
-extern int interrupt_immediately;
-extern char *shell_name, *current_host_name;
-extern char *dist_version;
-extern int patch_level;
+#if defined (TRANSLATABLE_STRINGS)
extern int dump_translatable_strings, dump_po_strings;
-extern Function *last_shell_builtin, *this_shell_builtin;
-#if defined (BUFFERED_INPUT)
-extern int bash_input_fd_changed;
+extern int singlequote_translations;
+#endif /* TRANSLATABLE_STRINGS */
+
+#if !defined (errno)
+extern int errno;
#endif
/* **************************************************************** */
/* */
/* **************************************************************** */
-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 PARAMS((int));
+#endif
+
+static int yy_getc PARAMS((void));
+static int yy_ungetc PARAMS((int));
+
+#if defined (READLINE)
+static int yy_readline_get PARAMS((void));
+static int yy_readline_unget PARAMS((int));
+#endif
+
+static int yy_string_get PARAMS((void));
+static int yy_string_unget PARAMS((int));
+static int yy_stream_get PARAMS((void));
+static int yy_stream_unget PARAMS((int));
+
+static int shell_getc PARAMS((int));
+static void shell_ungetc PARAMS((int));
+static void discard_until PARAMS((int));
+
+static void push_string PARAMS((char *, int, alias_t *));
+static void pop_string PARAMS((void));
+static void free_string_list PARAMS((void));
+
+static char *read_a_line PARAMS((int));
+
+static int reserved_word_acceptable PARAMS((int));
+static int yylex PARAMS((void));
+
+static void push_heredoc PARAMS((REDIRECT *));
+static char *mk_alexpansion PARAMS((char *));
+static int alias_expand_token PARAMS((char *));
+static int time_command_acceptable PARAMS((void));
+static int special_case_tokens PARAMS((char *));
+static int read_token PARAMS((int));
+static char *parse_matched_pair PARAMS((int, int, int, int *, int));
+static char *parse_comsub PARAMS((int, int, int, int *, int));
+#if defined (ARRAY_VARS)
+static char *parse_compound_assignment PARAMS((int *));
+#endif
+#if defined (DPAREN_ARITHMETIC) || defined (ARITH_FOR_COMMAND)
+static int parse_dparen PARAMS((int));
+static int parse_arith_cmd PARAMS((char **, int));
+#endif
#if defined (COND_COMMAND)
-static COMMAND *parse_cond_command ();
+static void cond_error PARAMS((void));
+static COND_COM *cond_expr PARAMS((void));
+static COND_COM *cond_or PARAMS((void));
+static COND_COM *cond_and PARAMS((void));
+static COND_COM *cond_term PARAMS((void));
+static int cond_skip_newlines PARAMS((void));
+static COMMAND *parse_cond_command PARAMS((void));
+#endif
+#if defined (ARRAY_VARS)
+static int token_is_assignment PARAMS((char *, int));
+static int token_is_ident PARAMS((char *, int));
#endif
-static int read_token_word ();
-static void discard_parser_constructs ();
+static int read_token_word PARAMS((int));
+static void discard_parser_constructs PARAMS((int));
-static void report_syntax_error ();
-static void handle_eof_input_unit ();
-static void prompt_again ();
+static char *error_token_from_token PARAMS((int));
+static char *error_token_from_text PARAMS((void));
+static void print_offending_line PARAMS((void));
+static void report_syntax_error PARAMS((char *));
+
+static void handle_eof_input_unit PARAMS((void));
+static void prompt_again PARAMS((int));
#if 0
-static void reset_readline_prompt ();
+static void reset_readline_prompt PARAMS((void));
+#endif
+static void print_prompt PARAMS((void));
+
+#if defined (HANDLE_MULTIBYTE)
+static void set_line_mbstate PARAMS((void));
+static char *shell_input_line_property = NULL;
+static size_t shell_input_line_propsize = 0;
+#else
+# define set_line_mbstate()
#endif
-static void print_prompt ();
-extern int yyerror ();
+extern int yyerror PARAMS((const char *));
+
+#ifdef DEBUG
+extern int yydebug;
+#endif
/* Default prompt strings */
char *primary_prompt = PPROMPT;
/* PROMPT_STRING_POINTER points to one of these, never to an actual string. */
char *ps1_prompt, *ps2_prompt;
+/* Displayed after reading a command but before executing it in an interactive shell */
+char *ps0_prompt;
+
/* Handle on the current prompt string. Indirectly points through
ps1_ or ps2_prompt. */
char **prompt_string_pointer = (char **)NULL;
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];
+static REDIRECT *redir_stack[HEREDOC_MAX];
int need_here_doc;
/* 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+1];
+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;
+
+static FILE *yyoutstream;
+static FILE *yyerrstream;
%}
%union {
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
+
+/* Special; never created by yylex; only set by parse_comsub and xparse_dolparen */
+%token DOLPAREN
/* The types that the various syntactical units return. */
%type <command> inputunit command pipeline pipeline_command
-%type <command> list list0 list1 compound_list simple_list simple_list1
+%type <command> list0 list1 compound_list simple_list simple_list1
%type <command> simple_command shell_command
%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> comsub
+%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;
+ }
+ | comsub
+ {
+ /* This is special; look at the production and how
+ parse_comsub sets token_to_read */
+ global_command = $1;
+ eof_encountered = 0;
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;
+ }
+ else
+ {
+ YYABORT;
+ }
+ }
+ | error yacc_EOF
+ {
+ /* EOF after an error. Do ignoreeof or not. Really
+ only interesting in non-interactive shells */
+ global_command = (COMMAND *)NULL;
+ if (last_command_exit_value == 0)
+ last_command_exit_value = EX_BADUSAGE; /* force error return */
+ if (interactive && parse_and_execute_level == 0)
{
+ handle_eof_input_unit ();
YYACCEPT;
}
else
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);
- redir_stack[need_here_doc++] = $$;
+ $$ = make_redirection (source, r_reading_until, redir, 0);
+ push_heredoc ($$);
}
| NUMBER LESS_LESS WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_until, redir, 0);
+ push_heredoc ($$);
+ }
+ | REDIR_WORD LESS_LESS WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_until, redir, REDIR_VARASSIGN);
+ push_heredoc ($$);
+ }
+ | LESS_LESS_MINUS WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_deblank_reading_until, redir, 0);
+ push_heredoc ($$);
+ }
+ | NUMBER LESS_LESS_MINUS WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_deblank_reading_until, redir, 0);
+ push_heredoc ($$);
+ }
+ | REDIR_WORD LESS_LESS_MINUS WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_deblank_reading_until, redir, REDIR_VARASSIGN);
+ push_heredoc ($$);
+ }
+ | LESS_LESS_LESS WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_reading_string, redir, 0);
+ }
+ | NUMBER LESS_LESS_LESS WORD
+ {
+ source.dest = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_reading_until, redir);
- redir_stack[need_here_doc++] = $$;
+ $$ = make_redirection (source, r_reading_string, redir, 0);
+ }
+ | REDIR_WORD LESS_LESS_LESS WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = 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);
- }
- | LESS_LESS_MINUS WORD
- {
- redir.filename = $2;
- $$ = make_redirection
- (0, r_deblank_reading_until, redir);
- redir_stack[need_here_doc++] = $$;
+ $$ = make_redirection (source, r_duplicating_output_word, redir, 0);
}
- | NUMBER LESS_LESS_MINUS WORD
+ | REDIR_WORD GREATER_AND WORD
{
+ source.filename = $1;
redir.filename = $3;
- $$ = 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);
+ $$ = make_redirection (source, r_append_err_and_out, redir, 0);
}
;
COMMAND *tc;
tc = $1;
- /* According to Posix.2 3.9.5, redirections
- specified after the body of a function should
- be attached to the function and performed when
- the function is executed, not as part of the
- function definition command. */
- if (tc->type == cm_function_def)
- {
- tc = tc->value.Function_def->command;
- if (tc->type == cm_group)
- tc = tc->value.Group->command;
- }
- if (tc->redirects)
+ if (tc && tc->redirects)
{
register REDIRECT *t;
for (t = tc->redirects; t->next; t = t->next)
;
t->next = $2;
}
- else
+ else if (tc)
tc->redirects = $2;
$$ = $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 ($$ == 0) YYERROR;
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS list_terminator newline_list '{' compound_list '}'
+ {
+ $$ = make_arith_for_command ($2, $6, arith_for_lineno);
+ if ($$ == 0) YYERROR;
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS DO compound_list DONE
+ {
+ $$ = make_arith_for_command ($2, $4, arith_for_lineno);
+ if ($$ == 0) YYERROR;
+ if (word_top > 0) word_top--;
+ }
+ | FOR ARITH_FOR_EXPRS '{' compound_list '}'
+ {
+ $$ = make_arith_for_command ($2, $4, arith_for_lineno);
+ if ($$ == 0) YYERROR;
+ if (word_top > 0) word_top--;
+ }
;
-select_command: SELECT WORD newline_list DO list DONE
+select_command: SELECT WORD newline_list DO compound_list DONE
+ {
+ $$ = 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 '{' compound_list '}'
+ {
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
+ | SELECT WORD ';' newline_list DO compound_list DONE
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $5);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
- | SELECT WORD newline_list '{' list '}'
+ | SELECT WORD ';' newline_list '{' compound_list '}'
{
- $$ = make_select_command ($2, add_string_to_list ("$@", (WORD_LIST *)NULL), $5);
+ $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
- | SELECT WORD ';' newline_list DO list DONE
+ | SELECT WORD newline_list IN word_list list_terminator newline_list DO compound_list DONE
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6);
+ $$ = make_select_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
- | SELECT WORD ';' newline_list '{' list '}'
+ | SELECT WORD newline_list IN word_list list_terminator newline_list '{' compound_list '}'
{
- $$ = make_select_command ($2, add_string_to_list ("\"$@\"", (WORD_LIST *)NULL), $6);
+ $$ = make_select_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
- | SELECT WORD newline_list IN word_list list_terminator newline_list DO list DONE
+ | SELECT WORD newline_list IN list_terminator newline_list DO compound_list DONE
{
- $$ = make_select_command ($2, (WORD_LIST *)reverse_list ($5), $9);
+ $$ = make_select_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
- | SELECT WORD newline_list IN word_list list_terminator newline_list '{' list '}'
+ | SELECT WORD newline_list IN list_terminator newline_list '{' compound_list '}'
{
- $$ = make_select_command ($2, (WORD_LIST *)reverse_list ($5), $9);
+ $$ = make_select_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
}
;
case_command: CASE WORD newline_list IN newline_list ESAC
- { $$ = 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
{ $$ = make_function_def ($2, $6, function_dstart, function_bstart); }
+ | FUNCTION WORD function_body
+ { $$ = make_function_def ($2, $3, function_dstart, function_bstart); }
+ | FUNCTION WORD '\n' newline_list function_body
+ { $$ = make_function_def ($2, $5, function_dstart, function_bstart); }
+ ;
+
+function_body: shell_command
+ { $$ = $1; }
+ | shell_command redirection_list
+ {
+ COMMAND *tc;
- | FUNCTION WORD newline_list group_command
- { $$ = make_function_def ($2, $4, function_dstart, function_bstart); }
+ tc = $1;
+ /* According to Posix.2 3.9.5, redirections
+ specified after the body of a function should
+ be attached to the function and performed when
+ the function is executed, not as part of the
+ function definition command. */
+ /* XXX - I don't think it matters, but we might
+ want to change this in the future to avoid
+ problems differentiating between a function
+ definition with a redirection and a function
+ definition containing a single command with a
+ redirection. The two are semantically equivalent,
+ though -- the only difference is in how the
+ command printing code displays the redirections. */
+ if (tc && tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $2;
+ }
+ else if (tc)
+ tc->redirects = $2;
+ $$ = $1;
+ }
;
subshell: '(' compound_list ')'
- { $2->flags |= CMD_WANT_SUBSHELL; $$ = $2; }
+ {
+ $$ = make_subshell_command ($2);
+ $$->flags |= CMD_WANT_SUBSHELL;
+ }
;
-if_command: IF compound_list THEN compound_list FI
- { $$ = make_if_command ($2, $4, (COMMAND *)NULL); }
- | IF compound_list THEN compound_list ELSE compound_list FI
- { $$ = make_if_command ($2, $4, $6); }
- | IF compound_list THEN compound_list elif_clause FI
- { $$ = make_if_command ($2, $4, $5); }
+comsub: DOLPAREN compound_list ')'
+ {
+ $$ = $2;
+ }
+ | DOLPAREN newline_list ')'
+ {
+ $$ = (COMMAND *)NULL;
+ }
;
-
-group_command: '{' list '}'
- { $$ = make_group_command ($2); }
- ;
+coproc: COPROC shell_command
+ {
+ $$ = make_coproc_command ("COPROC", $2);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC shell_command redirection_list
+ {
+ COMMAND *tc;
+
+ tc = $2;
+ if (tc && tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $3;
+ }
+ else if (tc)
+ tc->redirects = $3;
+ $$ = make_coproc_command ("COPROC", $2);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC WORD shell_command
+ {
+ $$ = make_coproc_command ($2->word, $3);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC WORD shell_command redirection_list
+ {
+ COMMAND *tc;
+
+ tc = $3;
+ if (tc && tc->redirects)
+ {
+ register REDIRECT *t;
+ for (t = tc->redirects; t->next; t = t->next)
+ ;
+ t->next = $4;
+ }
+ else if (tc)
+ tc->redirects = $4;
+ $$ = make_coproc_command ($2->word, $3);
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ | COPROC simple_command
+ {
+ $$ = make_coproc_command ("COPROC", clean_simple_command ($2));
+ $$->flags |= CMD_WANT_SUBSHELL|CMD_COPROC_SUBSHELL;
+ }
+ ;
+
+if_command: IF compound_list THEN compound_list FI
+ { $$ = make_if_command ($2, $4, (COMMAND *)NULL); }
+ | IF compound_list THEN compound_list ELSE compound_list FI
+ { $$ = make_if_command ($2, $4, $6); }
+ | IF compound_list THEN compound_list elif_clause FI
+ { $$ = make_if_command ($2, $4, $5); }
+ ;
+
+
+group_command: '{' compound_list '}'
+ { $$ = make_group_command ($2); }
+ ;
arith_command: ARITH_CMD
{ $$ = make_arith_command ($1); }
;
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
It must end with a newline or semicolon.
Lists are used within commands such as if, for, while. */
-list: newline_list list0
+compound_list: newline_list list0
{
$$ = $2;
- if (need_here_doc)
+ if (need_here_doc && last_read_token == '\n')
gather_here_documents ();
}
- ;
-
-compound_list: list
| newline_list list1
{
$$ = $2;
| list1 ';' newline_list list1
{ $$ = command_connect ($1, $4, ';'); }
| list1 '\n' newline_list list1
- { $$ = command_connect ($1, $4, ';'); }
+ {
+ if (parser_state & PST_CMDSUBST)
+ $$ = command_connect ($1, $4, '\n');
+ else
+ $$ = command_connect ($1, $4, ';');
+ }
| pipeline_command
{ $$ = $1; }
;
+simple_list_terminator: '\n'
+ | yacc_EOF
+ ;
+
list_terminator:'\n'
+ { $$ = '\n'; }
| ';'
+ { $$ = ';'; }
| yacc_EOF
+ { $$ = yacc_EOF; }
;
newline_list:
{
$$ = $1;
if (need_here_doc)
- gather_here_documents ();
+ gather_here_documents (); /* XXX */
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+INTERNAL_DEBUG (("LEGACY: parser: command substitution simple_list1 -> simple_list"));
+ global_command = $1;
+ eof_encountered = 0;
+ if (bash_input.type == st_string)
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 '&'
{
else
$$ = command_connect ($1, (COMMAND *)NULL, '&');
if (need_here_doc)
- gather_here_documents ();
+ gather_here_documents (); /* XXX */
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+INTERNAL_DEBUG (("LEGACY: parser: command substitution simple_list1 '&' -> simple_list"));
+ global_command = $1;
+ eof_encountered = 0;
+ if (bash_input.type == st_string)
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 ';'
{
$$ = $1;
if (need_here_doc)
- gather_here_documents ();
+ gather_here_documents (); /* XXX */
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+INTERNAL_DEBUG (("LEGACY: parser: command substitution simple_list1 ';' -> simple_list"));
+ global_command = $1;
+ eof_encountered = 0;
+ if (bash_input.type == st_string)
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
;
;
pipeline_command: pipeline
- { $$ = $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';
+ else if ($2 == ';')
+ token_to_read = ';';
+ parser_state &= ~PST_REDIRLIST; /* make_simple_command sets this */
}
- | 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);
+ $$->flags |= CMD_INVERT_RETURN;
+ /* XXX - let's cheat and push a newline back */
+ if ($2 == '\n')
+ token_to_read = '\n';
+ if ($2 == ';')
+ token_to_read = ';';
+ parser_state &= ~PST_REDIRLIST; /* make_simple_command sets this */
}
;
-pipeline:
- pipeline '|' newline_list pipeline
+pipeline: pipeline '|' newline_list pipeline
{ $$ = 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 TIMEIGN
+ { $$ = CMD_TIME_PIPELINE|CMD_TIME_POSIX; }
+ | TIME TIMEOPT TIMEIGN
+ { $$ = 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 /* $( ... ) command substitution */
-#define PST_CASESTMT 0x080 /* parsing a case statement */
-#define PST_CONDCMD 0x100 /* parsing a [[...]] command */
-#define PST_CONDEXPR 0x200 /* parsing the guts of [[...]] */
-
/* Initial size to allocate for tokens, and the
amount to grow them by. */
#define TOKEN_DEFAULT_INITIAL_SIZE 496
#define TOKEN_DEFAULT_GROW_SIZE 512
-/* Shell meta-characters that, when unquoted, separate words. */
-#define shellmeta(c) (strchr (shell_meta_chars, (c)) != 0)
-#define shellbreak(c) (strchr (shell_break_chars, (c)) != 0)
-#define shellquote(c) ((c) == '"' || (c) == '`' || (c) == '\'')
-#define shellexp(c) ((c) == '$' || (c) == '<' || (c) == '>')
-
-char *shell_meta_chars = "()<>;&|";
-char *shell_break_chars = "()<>;&| \t\n";
-
-/* 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;
{
#if YYDEBUG != 0
yydebug = i;
+ yyoutstream = stdout;
+ yyerrstream = stderr;
#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 (current_readline_line == 0)
{
- if (!bash_readline_initialized)
+ if (bash_readline_initialized == 0)
initialize_readline ();
#if defined (JOB_CONTROL)
if (job_control)
- give_terminal_to (shell_pgrp);
+ give_terminal_to (shell_pgrp, 0);
#endif /* JOB_CONTROL */
+ old_sigint = IMPOSSIBLE_TRAP_HANDLER;
if (signal_is_ignored (SIGINT) == 0)
{
old_sigint = (SigHandler *)set_signal_handler (SIGINT, sigint_sighandler);
- interrupt_immediately++;
}
+ sh_unset_nodelay_mode (fileno (rl_instream)); /* just in case */
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);
+ 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);
}
}
}
}
+/* Will we be collecting another input line and printing a prompt? This uses
+ different conditions than SHOULD_PROMPT(), since readline allows a user to
+ embed a newline in the middle of the line it collects, which the parser
+ will interpret as a line break and command delimiter. */
+int
+parser_will_prompt ()
+{
+ return (current_readline_line == 0 || current_readline_line[current_readline_line_index] == 0);
+}
+
#else /* !READLINE */
void
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. */
+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 most of its signal handlers without the SA_RESTART flag, some
+ signals received during a read(2) will not cause the read to be restarted.
+ We will 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)
+ /* 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 */
}
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:
* everything between a `;;' and the next `)' or `esac'
*/
-#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
+#define PSH_ARRAY 0x08
+
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, saved_line_len;
+ 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_len = shell_input_line_len;
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 = shell_input_line_len = 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 ();
}
/*
shell_input_line = pushed_string_list->saved_line;
shell_input_line_index = pushed_string_list->saved_line_index;
shell_input_line_size = pushed_string_list->saved_line_size;
+ shell_input_line_len = pushed_string_list->saved_line_len;
shell_input_line_terminator = pushed_string_list->saved_line_terminator;
+#if defined (ALIAS)
if (pushed_string_list->expand_alias)
parser_state |= PST_ALEXPNEXT;
else
parser_state &= ~PST_ALEXPNEXT;
+#endif
t = pushed_string_list;
pushed_string_list = pushed_string_list->next;
#endif
free ((char *)t);
+
+ set_line_mbstate ();
}
static void
pushed_string_list = (STRING_SAVER *)NULL;
}
-#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
+}
+
+#if defined (ALIAS)
+/* Before freeing AP, make sure that there aren't any cases of pointer
+ aliasing that could cause us to reference freed memory later on. */
+void
+clear_string_list_expander (ap)
+ alias_t *ap;
+{
+ register STRING_SAVER *t;
+
+ for (t = pushed_string_list; t; t = t->next)
+ {
+ if (t->expander && t->expander == ap)
+ t->expander = 0;
+ }
+}
+#endif
+
+void
+clear_shell_input_line ()
+{
+ if (shell_input_line)
+ shell_input_line[shell_input_line_index = 0] = '\0';
+}
/* 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
{
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;
c = '\n';
}
- /* `+2' in case the final character in the buffer is a newline. */
+ /* `+2' in case the final character in the buffer is a newline or we
+ have to handle CTLESC or CTLNUL. */
RESIZE_MALLOCED_BUFFER (line_buffer, indx, 2, buffer_size, 128);
/* IF REMOVE_QUOTED_NEWLINES is non-zero, we are reading a
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);
}
}
else
- line_buffer[indx++] = c;
+ {
+ /* remove_quoted_newline is non-zero if the here-document delimiter
+ is unquoted. In this case, we will be expanding the lines and
+ need to make sure CTLESC and CTLNUL in the input are quoted. */
+ if (remove_quoted_newline && (c == CTLESC || c == CTLNUL))
+ line_buffer[indx++] = CTLESC;
+ line_buffer[indx++] = c;
+ }
if (c == '\n')
{
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 (0);
+ ret = read_a_line (remove_quoted_newline);
+#if defined (HISTORY)
+ if (ret && remember_on_history && (parser_state & PST_HEREDOC))
+ {
+ /* To make adding 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}
};
-/* XXX - we should also have an alist with strings for other tokens, so we
- can give more descriptive error messages. Look at y.tab.h for the
- other tokens. */
+/* 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 unquoted_backslash = 0;
+
static int
shell_getc (remove_quoted_newline)
int remove_quoted_newline;
{
register int i;
- int c;
- static int mustpop = 0;
+ int c, truncating, last_was_backslash;
+ unsigned char uc;
QUIT;
+ last_was_backslash = 0;
+ 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 interactive, 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 we get EOS while parsing a string, treat it as EOF so we
+ don't just keep looping. Happens very rarely */
+ if (bash_input.type == st_string)
+ {
+ if (i == 0)
+ shell_input_line_terminator = EOF;
+ shell_input_line[i] = '\0';
+ c = EOF;
+ break;
+ }
+ 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 (%lu): line truncated"), shell_input_line_size, (unsigned long)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')
{
current_command_line_count++;
break;
}
+
+ last_was_backslash = last_was_backslash == 0 && c == '\\';
}
+
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])
{
char *expansions;
# if defined (BANG_HISTORY)
- int old_hist;
-
/* If the current delimiter is a single quote, we should not be
performing history expansion, even if we're on a different
line from the original single quote. */
- old_hist = history_expansion_inhibited;
if (current_delimiter (dstack) == '\'')
- history_expansion_inhibited = 1;
+ history_quoting_state = '\'';
+ else if (current_delimiter (dstack) == '"')
+ history_quoting_state = '"';
+ else
+ history_quoting_state = 0;
# endif
+ /* Calling with a third argument of 1 allows remember_on_history to
+ determine whether or not the line is saved to the history list */
expansions = pre_process_line (shell_input_line, 1, 1);
# if defined (BANG_HISTORY)
- history_expansion_inhibited = old_hist;
+ history_quoting_state = 0;
# endif
if (expansions != shell_input_line)
{
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 (0);
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-3 && (shell_input_line_len+3 > shell_input_line_size))
+ shell_input_line = (char *)xrealloc (shell_input_line,
1 + (shell_input_line_size += 2));
- shell_input_line[shell_input_line_len] = '\n';
+ /* Don't add a newline to a string that ends with a backslash if we're
+ going to be removing quoted newlines, since that will eat the
+ backslash. Add another backslash instead (will be removed by
+ word expansion). */
+ if (bash_input.type == st_string && expanding_alias() == 0 && last_was_backslash && c == EOF && remove_quoted_newline)
+ shell_input_line[shell_input_line_len] = '\\';
+ else
+ shell_input_line[shell_input_line_len] = '\n';
shell_input_line[shell_input_line_len + 1] = '\0';
+
+#if defined (HANDLE_MULTIBYTE)
+ /* This is kind of an abstraction violation, but there's no need to
+ go through the entire shell_input_line again with a call to
+ set_line_mbstate(). */
+ EXTEND_SHELL_INPUT_LINE_PROPERTY();
+ shell_input_line_property[shell_input_line_len] = 1;
+#endif
}
}
- c = shell_input_line[shell_input_line_index];
+next_alias_char:
+ if (shell_input_line_index == 0)
+ unquoted_backslash = 0;
- if (c)
- shell_input_line_index++;
+ uc = shell_input_line[shell_input_line_index];
- if (c == '\\' && remove_quoted_newline &&
- shell_input_line[shell_input_line_index] == '\n')
+ if (uc)
{
- prompt_again ();
- line_number++;
- goto restart_read;
+ unquoted_backslash = unquoted_backslash == 0 && uc == '\\';
+ shell_input_line_index++;
}
#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. */
+
+ /* If we're at the end of an alias expansion add a space to make sure that
+ the alias remains marked as being in use while we expand its last word.
+ This makes sure that pop_string doesn't mark the alias as not in use
+ before the string resulting from the alias expansion is tokenized and
+ checked for alias expansion, preventing recursion. At this point, the
+ last character in shell_input_line is the last character of the alias
+ expansion. We test that last character to determine whether or not to
+ return the space that will delimit the token and postpone the pop_string.
+ This set of conditions duplicates what used to be in mk_alexpansion ()
+ below, with the addition that we don't add a space if we're currently
+ reading a quoted string or in a shell comment. */
+#ifndef OLD_ALIAS_HACK
+ if (uc == 0 && pushed_string_list && pushed_string_list->flags != PSH_SOURCE &&
+ pushed_string_list->flags != PSH_DPAREN &&
+ (parser_state & PST_COMMENT) == 0 &&
+ (parser_state & PST_ENDALIAS) == 0 && /* only once */
+ shell_input_line_index > 0 &&
+ shellblank (shell_input_line[shell_input_line_index-1]) == 0 &&
+ shell_input_line[shell_input_line_index-1] != '\n' &&
+ unquoted_backslash == 0 &&
+ shellmeta (shell_input_line[shell_input_line_index-1]) == 0 &&
+ (current_delimiter (dstack) != '\'' && current_delimiter (dstack) != '"'))
{
- if (mustpop)
- {
- pop_string ();
- c = shell_input_line[shell_input_line_index];
- if (c)
- shell_input_line_index++;
- mustpop--;
- }
- else
- {
- mustpop++;
- c = ' ';
- }
+ parser_state |= PST_ENDALIAS;
+ /* We need to do this to make sure last_shell_getc_is_singlebyte returns
+ true, since we are returning a single-byte space. */
+ if (shell_input_line_index == shell_input_line_len && last_shell_getc_is_singlebyte == 0)
+ {
+#if 0
+ EXTEND_SHELL_INPUT_LINE_PROPERTY();
+ shell_input_line_property[shell_input_line_len++] = 1;
+ /* extend shell_input_line to accommodate the shell_ungetc that
+ read_token_word() will perform, since we're extending the index */
+ RESIZE_MALLOCED_BUFFER (shell_input_line, shell_input_line_index, 2, shell_input_line_size, 16);
+ shell_input_line[++shell_input_line_index] = '\0'; /* XXX */
+#else
+ shell_input_line_property[shell_input_line_index - 1] = 1;
+#endif
+ }
+ return ' '; /* END_ALIAS */
}
+#endif
+
+pop_alias:
#endif /* ALIAS || DPAREN_ARITHMETIC */
+ /* This case works for PSH_DPAREN as well as the shell_ungets() case that uses
+ push_string */
+ if (uc == 0 && pushed_string_list && pushed_string_list->flags != PSH_SOURCE)
+ {
+ parser_state &= ~PST_ENDALIAS;
+ pop_string ();
+ uc = shell_input_line[shell_input_line_index];
+ if (uc)
+ shell_input_line_index++;
+ }
+
+ if MBTEST(uc == '\\' && remove_quoted_newline && shell_input_line[shell_input_line_index] == '\n')
+ {
+ if (SHOULD_PROMPT ())
+ prompt_again (0);
+ 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 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 (!c && shell_input_line_terminator == EOF)
+ 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;
+}
+
+/* Push S back into shell_input_line; updating shell_input_line_index */
+void
+shell_ungets (s)
+ char *s;
+{
+ size_t slen, chars_left;
+
+ slen = strlen (s);
+
+ if (shell_input_line[shell_input_line_index] == '\0')
+ {
+ /* Easy, just overwrite shell_input_line. This is preferred because it
+ saves on set_line_mbstate () and other overhead like push_string */
+ if (shell_input_line_size <= slen)
+ RESIZE_MALLOCED_BUFFER (shell_input_line, shell_input_line_index, slen + 1, shell_input_line_size, 64);
+ strcpy (shell_input_line, s);
+ shell_input_line_index = 0;
+ shell_input_line_len = slen;
+ shell_input_line_terminator = 0;
+ }
+ else if (shell_input_line_index >= slen)
+ {
+ /* Just as easy, just back up shell_input_line_index, but it means we
+ will re-process some characters in set_line_mbstate(). Need to
+ watch pushing back newlines here. */
+ while (slen > 0)
+ shell_input_line[--shell_input_line_index] = s[--slen];
+ }
+ else if (s[slen - 1] == '\n')
+ {
+ push_string (savestring (s), 0, (alias_t *)NULL);
+ /* push_string does set_line_mbstate () */
+ return;
+ }
+ else
+ {
+ /* Harder case: pushing back input string that's longer than what we've
+ consumed from shell_input_line so far. */
+ INTERNAL_DEBUG (("shell_ungets: not at end of shell_input_line"));
+
+ chars_left = shell_input_line_len - shell_input_line_index;
+ if (shell_input_line_size <= (slen + chars_left))
+ RESIZE_MALLOCED_BUFFER (shell_input_line, shell_input_line_index, chars_left + slen + 1, shell_input_line_size, 64);
+ memmove (shell_input_line + slen, shell_input_line + shell_input_line_index, shell_input_line_len - shell_input_line_index);
+ strcpy (shell_input_line, s);
+ shell_input_line_index = 0;
+ shell_input_line_len = strlen (shell_input_line); /* chars_left + slen? */
+ }
+
+#if defined (HANDLE_MULTIBYTE)
+ set_line_mbstate (); /* XXX */
+#endif
+}
+
+char *
+parser_remaining_input ()
+{
+ if (shell_input_line == 0)
+ return 0;
+ if ((int)shell_input_line_index < 0 || shell_input_line_index >= shell_input_line_len)
+ return ""; /* XXX */
+ return (shell_input_line + shell_input_line_index);
}
+#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);
-
- last_shell_builtin = temp_last;
- this_shell_builtin = temp_this;
- last_command_exit_value = temp_exit_value;
- eof_encountered = temp_eof_encountered;
+ parse_and_execute (savestring (command), vname, SEVAL_NONINT|SEVAL_NOHIST);
- 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 */
token_to_read = 0;
}
+void
+push_token (x)
+ int x;
+{
+ two_tokens_ago = token_before_that;
+ token_before_that = last_read_token;
+ last_read_token = current_token;
+
+ current_token = x;
+}
+
/* Place to remember the token. We try to keep the buffer
at a reasonable size, but it can grow. */
static char *token = (char *)NULL;
/* Current size of the token buffer. */
-static int token_buffer_size;
+static size_t token_buffer_size;
/* Command to read_token () explaining what we want it to do. */
#define READ 0
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)
- prompt_again ();
+ if (token_to_read == 0 && SHOULD_PROMPT ())
+ prompt_again (0);
}
two_tokens_ago = token_before_that;
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)
+ {
+ /* placeholder for any special handling. */
+ return (current_token);
+ }
+
+ if (current_token < 0)
+#if defined (YYERRCODE) && !defined (YYUNDEF)
+ current_token = YYERRCODE;
+#else
+ current_token = YYerror;
+#endif
+
return (current_token);
}
which allow ESAC to be the next one read. */
static int esacs_needed_count;
+/* When non-zero, we can read IN as an acceptable token, regardless of how
+ many newlines we read. */
+static int expecting_in_token;
+
+static void
+push_heredoc (r)
+ REDIRECT *r;
+{
+ if (need_here_doc >= HEREDOC_MAX)
+ {
+ last_command_exit_value = EX_BADUSAGE;
+ need_here_doc = 0;
+ report_syntax_error (_("maximum here-document count exceeded"));
+ reset_parser ();
+ exit_shell (last_command_exit_value);
+ }
+ redir_stack[need_here_doc++] = r;
+}
+
void
gather_here_documents ()
{
- int r = 0;
- while (need_here_doc)
+ int r;
+
+ r = 0;
+ here_doc_first_line = 1;
+ while (need_here_doc > 0)
{
- 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--;
+ redir_stack[r - 1] = 0; /* XXX */
}
+ here_doc_first_line = 0; /* just in case */
}
/* When non-zero, an open-brace used to create a group is awaiting a close
brace partner. */
static int open_brace_count;
+/* In the following three macros, `token' is always last_read_token */
+
+/* Are we in the middle of parsing a redirection where we are about to read
+ a word? This is used to make sure alias expansion doesn't happen in the
+ middle of a redirection, even though we're parsing a simple command. */
+#define parsing_redirection(token) \
+ (token == '<' || token == '>' || \
+ token == GREATER_GREATER || token == GREATER_BAR || \
+ token == LESS_GREATER || token == LESS_LESS_MINUS || \
+ token == LESS_LESS || token == LESS_LESS_LESS || \
+ token == LESS_AND || token == GREATER_AND || token == AND_GREATER)
+
+/* Is `token' one that will allow a WORD to be read in a command position?
+ We can read a simple command name on which we should attempt alias expansion
+ or we can read an assignment statement. */
#define command_token_position(token) \
(((token) == ASSIGNMENT_WORD) || \
- ((token) != SEMI_SEMI && reserved_word_acceptable(token)))
+ ((parser_state&PST_REDIRLIST) && parsing_redirection(token) == 0) || \
+ ((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)
+/* Are we in a position where we can read an assignment statement? */
+#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) \
+ if ((parser_state & PST_CASEPAT) && last_read_token == '|' && word_token_alist[i].token == ESAC) \
+ break; /* Posix grammar rule 4 */ \
+ if ((parser_state & PST_CASEPAT) && last_read_token == '(' && word_token_alist[i].token == ESAC) /*)*/ \
+ break; /* phantom Posix grammar rule 4 */ \
+ if (word_token_alist[i].token == ESAC) { \
parser_state &= ~(PST_CASEPAT|PST_CASESTMT); \
- else if (word_token_alist[i].token == CASE) \
+ esacs_needed_count--; \
+ } else if (word_token_alist[i].token == CASE) \
parser_state |= PST_CASESTMT; \
else if (word_token_alist[i].token == COND_END) \
parser_state &= ~(PST_CONDCMD|PST_CONDEXPR); \
/* 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);
+#ifdef OLD_ALIAS_HACK
+ /* If the last character in the alias is a newline, don't add a trailing
+ space to the expansion. Works with shell_getc above. */
+ /* Need to do something about the case where the alias expansion contains
+ an unmatched quoted string, since appending this space affects the
+ subsequent output. */
+ if (l > 0 && r[l - 1] != ' ' && r[l - 1] != '\n' && shellmeta(r[l - 1]) == 0)
+ r[l++] = ' ';
+#endif
+ 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 0
if (((parser_state & PST_ALEXPNEXT) || command_token_position (last_read_token)) &&
(parser_state & PST_CASEPAT) == 0)
+#else
+ if ((parser_state & PST_ALEXPNEXT) || assignment_acceptable (last_read_token))
+#endif
{
- 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;
+#ifdef OLD_ALIAS_HACK
+ /* mk_alexpansion puts an extra space on the end of the alias expansion,
+ so the lookahead by the parser works right (the alias needs to remain
+ `in use' while parsing its last word to avoid alias recursion for
+ something like "alias echo=echo"). 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. */
+#endif
+ 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 ';':
case '\n':
+ if (token_before_that == '|')
+ return (0);
+ /* FALLTHROUGH */
case AND_AND:
case OR_OR:
case '&':
+ case WHILE:
case DO:
+ case UNTIL:
+ case IF:
case THEN:
+ case ELIF:
case ELSE:
- case '{':
- case '(':
+ case '{': /* } */
+ case '(': /* )( */
+ case ')': /* only valid in case statement */
+ 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 TIME or 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;
{
+ /* Posix grammar rule 6 */
if ((last_read_token == WORD) &&
#if defined (SELECT_COMMAND)
((token_before_that == FOR) || (token_before_that == CASE) || (token_before_that == SELECT)) &&
#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)
{
parser_state |= PST_CASEPAT;
esacs_needed_count++;
}
+ if (expecting_in_token)
+ expecting_in_token--;
+ return (IN);
+ }
+
+ /* XXX - leaving above code intact for now, but it should eventually be
+ removed in favor of this clause. */
+ /* Posix grammar rule 6 */
+ if (expecting_in_token && (last_read_token == WORD || last_read_token == '\n') &&
+ (tokstr[0] == 'i' && tokstr[1] == 'n' && tokstr[2] == 0))
+ {
+ if (parser_state & PST_CASESTMT)
+ {
+ parser_state |= PST_CASEPAT;
+ esacs_needed_count++;
+ }
+ expecting_in_token--;
return (IN);
}
+ /* Posix grammar rule 6, third word in FOR: for i; do command-list; done */
+ else if (expecting_in_token && (last_read_token == '\n' || last_read_token == ';') &&
+ (tokstr[0] == 'd' && tokstr[1] == 'o' && tokstr[2] == '\0'))
+ {
+ expecting_in_token--;
+ return (DO);
+ }
+ /* for i do; command-list; done */
if (last_read_token == WORD &&
#if defined (SELECT_COMMAND)
(token_before_that == FOR || token_before_that == SELECT) &&
#else
(token_before_that == FOR) &&
#endif
- (token[0] == 'd' && token[1] == 'o' && token[2] == '\0'))
- return (DO);
+ (tokstr[0] == 'd' && tokstr[1] == 'o' && tokstr[2] == '\0'))
+ {
+ if (expecting_in_token)
+ expecting_in_token--;
+ return (DO);
+ }
/* Ditto for ESAC in the CASE case.
Specifically, this handles "case word in esac", which is a legal
the designers disagree. */
if (esacs_needed_count)
{
- esacs_needed_count--;
- if (STREQ (token, "esac"))
+ if (last_read_token == IN && STREQ (tokstr, "esac"))
{
+ esacs_needed_count--;
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'. */
+ if (last_read_token == TIME && tokstr[0] == '-' && tokstr[1] == '-' && !tokstr[2])
+ return (TIMEIGN);
+ /* 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") && ((parser_state & PST_CASEPAT) == 0) && 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|PST_CMDSUBST))
+ extended_glob = global_extglob;
+#endif
+
parser_state = 0;
+ here_doc_first_line = 0;
#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
if (pushed_string_list)
free_string_list ();
#endif /* ALIAS || DPAREN_ARITHMETIC */
+ /* This is where we resynchronize to the next newline on error/reset */
if (shell_input_line)
{
free (shell_input_line);
FREE (word_desc_to_read);
word_desc_to_read = (WORD_DESC *)NULL;
+ eol_ungetc_lookahead = 0;
+
+ /* added post-bash-5.1 */
+ need_here_doc = 0;
+ redir_stack[0] = 0;
+ esacs_needed_count = expecting_in_token = 0;
+
+ current_token = '\n'; /* XXX */
last_read_token = '\n';
token_to_read = '\n';
}
+void
+reset_readahead_token ()
+{
+ if (token_to_read == '\n')
+ token_to_read = 0;
+}
+
/* Read the next token. Command can be READ (normal operation) or
RESET (to normalize state). */
static int
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 we hit the end of the string and we're not expanding an alias (e.g.,
+ we are eval'ing a string that is an incomplete command), return EOF */
+ if (character == '\0' && bash_input.type == st_string && expanding_alias() == 0)
+ {
+ INTERNAL_DEBUG (("shell_getc: bash_input.location.string = `%s'", bash_input.location.string));
+ EOF_Reached = 1;
+ return (yacc_EOF);
+ }
+
+ if MBTEST(character == '#' && (!interactive || interactive_comments))
{
/* A comment. Discard until EOL or EOF, and then return a newline. */
+ parser_state |= PST_COMMENT;
discard_until ('\n');
shell_getc (0);
+ parser_state &= ~PST_COMMENT;
character = '\n'; /* this will take the next if statement and return. */
}
- if (character == '\n')
+ if MBTEST(character == '\n')
{
/* If we're about to return an unquoted newline, we can go and collect
the text of any pending here document. */
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))
{
#if defined (ALIAS)
/* Turn off alias tokenization iff this character sequence would
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
- peek_char = shell_getc (1);
- if (character == peek_char)
+ parser_state &= ~PST_ASSIGNOK;
+
+ /* If we are parsing a command substitution and we have read a character
+ that marks the end of it, don't bother to skip over quoted newlines
+ when we read the next token. We're just interested in a character
+ that will turn this into a two-character token, so we let the higher
+ layers deal with quoted newlines following the command substitution. */
+ if ((parser_state & PST_CMDSUBST) && character == shell_eof_token)
+ peek_char = shell_getc (0);
+ else
+ peek_char = shell_getc (1);
+
+ if MBTEST(character == peek_char)
{
switch (character)
{
/* 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((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 0x0001
+#define LEX_CKCOMMENT 0x0002
+#define LEX_INCOMMENT 0x0004
+#define LEX_PASSNEXT 0x0008
+#define LEX_RESWDOK 0x0010
+#define LEX_CKCASE 0x0020
+#define LEX_INCASE 0x0040
+#define LEX_INHEREDOC 0x0080
+#define LEX_HEREDELIM 0x0100 /* reading here-doc delimiter */
+#define LEX_STRIPDOC 0x0200 /* <<- strip tabs from here doc delim */
+#define LEX_QUOTEDDOC 0x0400 /* here doc with quoted delim */
+#define LEX_INWORD 0x0800
+#define LEX_GTLT 0x1000
+#define LEX_CKESAC 0x2000 /* check esac after in -- for later */
+#define LEX_CASEWD 0x4000 /* word after case */
+#define LEX_PATLIST 0x8000 /* case statement pattern list */
+
+#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 P_FIRSTCLOSE 0x01
+#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, prevch, 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;
+
+ if ((flags & P_COMMAND) && qc != '`' && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0)
+ tflags |= LEX_CKCOMMENT;
- ret = xmalloc (retsize = 64);
+ /* 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;
+ ch = EOF; /* just in case */
while (count)
{
- ch = shell_getc (qc != '\'' && pass_next_character == 0);
+ prevch = ch;
+ 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))
- prompt_again ();
+ if MBTEST(ch == '\n' && SHOULD_PROMPT ())
+ prompt_again (0);
- 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;
- if (qc != '\'' && ch == '\n') /* double-quoted \<newline> disappears. */
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ if MBTEST(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;
+ /* XXX - PST_NOEXPAND? */
+ if MBTEST(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 (((flags & P_FIRSTCLOSE) == 0) && 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 we just read the ending character, don't bother continuing. */
+ if (count == 0)
+ break;
+
if (open == '\'') /* '' inside grouping construct */
- continue;
+ {
+ 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;
+ }
- if (ch == '\\') /* backslashes */
- pass_next_character++;
+ /* 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 (shellquote (ch))
+ if MBTEST(shellquote (ch))
{
/* '', ``, or "" inside $(...) or other grouping construct. */
push_delimiter (dstack, ch);
- nestret = parse_matched_pair (ch, ch, ch, &nestlen, 0);
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'') /* $'...' inside group */
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
+ else
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
pop_delimiter (dstack);
- if (nestret == &matched_pair_error)
+ CHECK_NESTRET_ERROR ();
+
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0 || dolbrace_state == DOLBRACE_QUOTE || dolbrace_state == DOLBRACE_QUOTE2))
{
- free (ret);
- return &matched_pair_error;
+ /* Translate $'...' here. */
+ /* PST_NOEXPAND */
+ ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
+ free (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 || dolbrace_state == DOLBRACE_QUOTE) && (flags & P_DOLBRACE))
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+#if 0 /* TAG:bash-5.3 */
+ /* This single-quotes PARAM in ${PARAM OP WORD} when PARAM
+ contains a $'...' even when extended_quote is set. */
+ else if ((rflags & P_DQUOTE) && (dolbrace_state == DOLBRACE_PARAM) && (flags & P_DOLBRACE))
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+#endif
+ else if ((rflags & P_DQUOTE) == 0)
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else
+ {
+ /* Should we quote CTLESC here? */
+ nestret = ttrans;
+ nestlen = ttranslen;
+ }
+ retind -= 2; /* back up before the $' */
}
- if (nestlen)
+#if defined (TRANSLATABLE_STRINGS)
+ else if MBTEST((tflags & LEX_WASDOL) && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
+ /* Locale expand $"..." here. */
+ /* PST_NOEXPAND */
+ ttrans = locale_expand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
+ free (nestret);
+
+ /* If we're supposed to single-quote translated strings,
+ check whether the translated result is different from
+ the original and single-quote the string if it is. */
+ if (singlequote_translations &&
+ ((nestlen - 1) != ttranslen || STREQN (nestret, ttrans, ttranslen) == 0))
+ {
+ if ((rflags & P_DQUOTE) == 0)
+ nestret = sh_single_quote (ttrans);
+ else if ((rflags & P_DQUOTE) && (dolbrace_state == DOLBRACE_QUOTE2) && (flags & P_DOLBRACE))
+ nestret = sh_single_quote (ttrans);
+ else
+ /* single quotes aren't special, use backslash instead */
+ nestret = sh_backslash_quote_for_double_quotes (ttrans, 0);
+ }
+ else
+ nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ retind -= 2; /* back up before the $" */
}
+#endif /* TRANSLATABLE_STRINGS */
+
+ APPEND_NESTRET ();
FREE (nestret);
}
+ else if ((flags & (P_ARRAYSUB|P_DOLBRACE)) && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ goto parse_dollar_word;
+#if defined (PROCESS_SUBSTITUTION)
+ /* XXX - technically this should only be recognized at the start of
+ a word */
+ else if ((flags & (P_ARRAYSUB|P_DOLBRACE)) && (tflags & LEX_GTLT) && (ch == '(')) /* ) */
+ goto parse_dollar_word;
+#endif
}
/* 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 if MBTEST(open == '"' && 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;
- }
+ nestret = parse_matched_pair (0, '`', '`', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
FREE (nestret);
}
- else if (was_dollar && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ else if MBTEST(open != '`' && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
/* check for $(), $[], or ${} inside quoted string. */
{
+parse_dollar_word:
if (open == ch) /* undo previous increment */
count--;
if (ch == '(') /* ) */
- nestret = parse_matched_pair (0, '(', ')', &nestlen, 0);
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
else if (ch == '{') /* } */
- nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE);
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|P_DOLBRACE|rflags);
else if (ch == '[') /* ] */
- nestret = parse_matched_pair (0, '[', ']', &nestlen, 0);
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
- }
- if (nestlen)
- {
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
- }
+ nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
FREE (nestret);
}
- was_dollar = (ch == '$');
+#if defined (PROCESS_SUBSTITUTION)
+ if MBTEST((ch == '<' || ch == '>') && (tflags & LEX_GTLT) == 0)
+ tflags |= LEX_GTLT;
+ else
+ tflags &= ~LEX_GTLT;
+#endif
+ if MBTEST(ch == '$' && (tflags & LEX_WASDOL) == 0)
+ 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;
}
-#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)
- char **ep;
+#if defined (DEBUG)
+static void
+dump_tflags (flags)
+ int flags;
+{
+ int f;
+
+ f = flags;
+ fprintf (stderr, "%d -> ", f);
+ if (f & LEX_WASDOL)
+ {
+ f &= ~LEX_WASDOL;
+ fprintf (stderr, "LEX_WASDOL%s", f ? "|" : "");
+ }
+ if (f & LEX_CKCOMMENT)
+ {
+ f &= ~LEX_CKCOMMENT;
+ fprintf (stderr, "LEX_CKCOMMENT%s", f ? "|" : "");
+ }
+ if (f & LEX_INCOMMENT)
+ {
+ f &= ~LEX_INCOMMENT;
+ fprintf (stderr, "LEX_INCOMMENT%s", f ? "|" : "");
+ }
+ if (f & LEX_PASSNEXT)
+ {
+ f &= ~LEX_PASSNEXT;
+ fprintf (stderr, "LEX_PASSNEXT%s", f ? "|" : "");
+ }
+ if (f & LEX_RESWDOK)
+ {
+ f &= ~LEX_RESWDOK;
+ fprintf (stderr, "LEX_RESWDOK%s", f ? "|" : "");
+ }
+ if (f & LEX_CKCASE)
+ {
+ f &= ~LEX_CKCASE;
+ fprintf (stderr, "LEX_CKCASE%s", f ? "|" : "");
+ }
+ if (f & LEX_CKESAC)
+ {
+ f &= ~LEX_CKESAC;
+ fprintf (stderr, "LEX_CKESAC%s", f ? "|" : "");
+ }
+ if (f & LEX_INCASE)
+ {
+ f &= ~LEX_INCASE;
+ fprintf (stderr, "LEX_INCASE%s", f ? "|" : "");
+ }
+ if (f & LEX_CASEWD)
+ {
+ f &= ~LEX_CASEWD;
+ fprintf (stderr, "LEX_CASEWD%s", f ? "|" : "");
+ }
+ if (f & LEX_PATLIST)
+ {
+ f &= ~LEX_PATLIST;
+ fprintf (stderr, "LEX_PATLIST%s", f ? "|" : "");
+ }
+ if (f & LEX_INHEREDOC)
+ {
+ f &= ~LEX_INHEREDOC;
+ fprintf (stderr, "LEX_INHEREDOC%s", f ? "|" : "");
+ }
+ if (f & LEX_HEREDELIM)
+ {
+ f &= ~LEX_HEREDELIM;
+ fprintf (stderr, "LEX_HEREDELIM%s", f ? "|" : "");
+ }
+ if (f & LEX_STRIPDOC)
+ {
+ f &= ~LEX_STRIPDOC;
+ fprintf (stderr, "LEX_WASDOL%s", f ? "|" : "");
+ }
+ if (f & LEX_QUOTEDDOC)
+ {
+ f &= ~LEX_QUOTEDDOC;
+ fprintf (stderr, "LEX_QUOTEDDOC%s", f ? "|" : "");
+ }
+ if (f & LEX_INWORD)
+ {
+ f &= ~LEX_INWORD;
+ fprintf (stderr, "LEX_INWORD%s", f ? "|" : "");
+ }
+
+ fprintf (stderr, "\n");
+ fflush (stderr);
+}
+#endif
+
+/* Parse a $(...) command substitution. This reads input from the current
+ input stream. */
+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 peekc, r;
+ int start_lineno, local_extglob, was_extpat;
+ char *ret, *tcmd;
+ int retlen;
+ sh_parser_state_t ps;
+ STRING_SAVER *saved_strings;
+ COMMAND *saved_global, *parsed_command;
+
+ /* Posix interp 217 says arithmetic expressions have precedence, so
+ assume $(( introduces arithmetic expansion and parse accordingly. */
+ if (open == '(') /*)*/
+ {
+ peekc = shell_getc (1);
+ 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);*/
+
+ /*debug_parser(1);*/
+ start_lineno = line_number;
+
+ save_parser_state (&ps);
+
+ was_extpat = (parser_state & PST_EXTPAT);
+
+ /* State flags we don't want to persist into command substitutions. */
+ parser_state &= ~(PST_REGEXP|PST_EXTPAT|PST_CONDCMD|PST_CONDEXPR|PST_COMPASSIGN);
+ /* Could do PST_CASESTMT too, but that also affects history. Setting
+ expecting_in_token below should take care of the parsing requirements.
+ Unsetting PST_REDIRLIST isn't strictly necessary because of how we set
+ token_to_read below, but we do it anyway. */
+ parser_state &= ~(PST_CASEPAT|PST_ALEXPNEXT|PST_SUBSHELL|PST_REDIRLIST);
+ /* State flags we want to set for this run through the parser. */
+ parser_state |= PST_CMDSUBST|PST_EOFTOKEN|PST_NOEXPAND;
+
+ /* leave pushed_string_list alone, since we might need to consume characters
+ from it to satisfy this command substitution (in some perverse case). */
+ shell_eof_token = close;
+
+ saved_global = global_command; /* might not be necessary */
+ global_command = (COMMAND *)NULL;
+
+ /* These are reset by reset_parser() */
+ need_here_doc = 0;
+ esacs_needed_count = expecting_in_token = 0;
+
+ /* We want to expand aliases on this pass if we're in posix mode, since the
+ standard says you have to take aliases into account when looking for the
+ terminating right paren. Otherwise, we defer until execution time for
+ backwards compatibility. */
+ if (expand_aliases)
+ expand_aliases = posixly_correct != 0;
+#if defined (EXTENDED_GLOB)
+ /* If (parser_state & PST_EXTPAT), we're parsing an extended pattern for a
+ conditional command and have already set global_extglob appropriately. */
+ if (shell_compatibility_level <= 51 && was_extpat == 0)
+ {
+ local_extglob = global_extglob = extended_glob;
+ extended_glob = 1;
+ }
+#endif
+
+ current_token = '\n'; /* XXX */
+ token_to_read = DOLPAREN; /* let's trick the parser */
+
+ r = yyparse ();
+
+ if (need_here_doc > 0)
+ {
+ internal_warning ("command substitution: %d unterminated here-document%s", need_here_doc, (need_here_doc == 1) ? "" : "s");
+ gather_here_documents (); /* XXX check compatibility level? */
+ }
+
+#if defined (EXTENDED_GLOB)
+ if (shell_compatibility_level <= 51 && was_extpat == 0)
+ extended_glob = local_extglob;
+#endif
+
+ parsed_command = global_command;
+
+ if (EOF_Reached)
+ {
+ shell_eof_token = ps.eof_token;
+ expand_aliases = ps.expand_aliases;
+
+ /* yyparse() has already called yyerror() and reset_parser() */
+ return (&matched_pair_error);
+ }
+ else if (r != 0)
+ {
+ /* parser_error (start_lineno, _("could not parse command substitution")); */
+ /* Non-interactive shells exit on parse error in a command substitution. */
+ if (last_command_exit_value == 0)
+ last_command_exit_value = EXECUTION_FAILURE;
+ set_exit_status (last_command_exit_value);
+ if (interactive_shell == 0)
+ jump_to_top_level (FORCE_EOF); /* This is like reader_loop() */
+ else
+ {
+ shell_eof_token = ps.eof_token;
+ expand_aliases = ps.expand_aliases;
+
+ jump_to_top_level (DISCARD); /* XXX - return (&matched_pair_error)? */
+ }
+ }
+
+ if (current_token != shell_eof_token)
+ {
+INTERNAL_DEBUG(("current_token (%d) != shell_eof_token (%c)", current_token, shell_eof_token));
+ token_to_read = current_token;
+
+ /* If we get here we can check eof_encountered and if it's 1 but the
+ previous EOF_Reached test didn't succeed, we can assume that the shell
+ is interactive and ignoreeof is set. We might want to restore the
+ parser state in this case. */
+ shell_eof_token = ps.eof_token;
+ expand_aliases = ps.expand_aliases;
+
+ return (&matched_pair_error);
+ }
+
+ /* We don't want to restore the old pushed string list, since we might have
+ used it to consume additional input from an alias while parsing this
+ command substitution. */
+ saved_strings = pushed_string_list;
+ restore_parser_state (&ps);
+ pushed_string_list = saved_strings;
+
+ tcmd = print_comsub (parsed_command); /* returns static memory */
+ retlen = strlen (tcmd);
+ if (tcmd[0] == '(') /* ) need a space to prevent arithmetic expansion */
+ retlen++;
+ ret = xmalloc (retlen + 2);
+ if (tcmd[0] == '(') /* ) */
+ {
+ ret[0] = ' ';
+ strcpy (ret + 1, tcmd);
+ }
+ else
+ strcpy (ret, tcmd);
+ ret[retlen++] = ')';
+ ret[retlen] = '\0';
+
+ dispose_command (parsed_command);
+ global_command = saved_global;
+
+ if (lenp)
+ *lenp = retlen;
+
+/*itrace("parse_comsub:%d: returning `%s'", line_number, ret);*/
+ return ret;
+}
+
+/* Recursively call the parser to parse a $(...) command substitution. This is
+ called by the word expansion code and so does not have to reset as much
+ parser state before calling yyparse(). */
+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, start_lineno;
+ char *ret, *ep, *ostring;
+
+/*debug_parser(1);*/
+ orig_ind = *indp;
+ ostring = string;
+ start_lineno = line_number;
+
+ if (*string == 0)
+ {
+ if (flags & SX_NOALLOC)
+ return (char *)NULL;
+
+ ret = xmalloc (1);
+ ret[0] = '\0';
+ return ret;
+ }
+
+/*itrace("xparse_dolparen: size = %d shell_input_line = `%s' string=`%s'", shell_input_line_size, shell_input_line, string);*/
+
+ sflags = SEVAL_NONINT|SEVAL_NOHIST|SEVAL_NOFREE;
+ if (flags & SX_NOLONGJMP)
+ sflags |= SEVAL_NOLONGJMP;
+
+ save_parser_state (&ps);
+ save_input_line_state (&ls);
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ pushed_string_list = (STRING_SAVER *)NULL;
+#endif
+ /*(*/
+ parser_state |= PST_CMDSUBST|PST_EOFTOKEN; /* allow instant ')' */ /*(*/
+ shell_eof_token = ')';
+ if (flags & SX_COMPLETE)
+ parser_state |= PST_NOERROR;
+
+ /* Don't expand aliases on this pass at all. Either parse_comsub() does it
+ at parse time, in which case this string already has aliases expanded,
+ or command_substitute() does it in the child process executing the
+ command substitution and we want to defer it completely until then. The
+ old value will be restored by restore_parser_state(). */
+ expand_aliases = 0;
+#if defined (EXTENDED_GLOB)
+ global_extglob = extended_glob; /* for reset_parser() */
+#endif
+
+ token_to_read = DOLPAREN; /* let's trick the parser */
+
+ nc = parse_string (string, "command substitution", sflags, (COMMAND **)NULL, &ep);
+
+ /* Should we save and restore the bison/yacc lookahead token (yychar) here?
+ Or only if it's not YYEMPTY? */
+ if (current_token == shell_eof_token)
+ yyclearin; /* might want to clear lookahead token unconditionally */
+
+ reset_parser (); /* resets extended_glob too */
+ /* reset_parser() clears shell_input_line and associated variables, including
+ parser_state, so we want to reset things, then restore what we need. */
+ restore_input_line_state (&ls);
+ restore_parser_state (&ps);
+
+ token_to_read = 0;
+
+ /* If parse_string returns < 0, we need to jump to top level with the
+ negative of the return value. We abandon the rest of this input line
+ first */
+ if (nc < 0)
+ {
+ clear_shell_input_line (); /* XXX */
+ if (bash_input.type != st_string) /* paranoia */
+ parser_state &= ~(PST_CMDSUBST|PST_EOFTOKEN);
+ if ((flags & SX_NOLONGJMP) == 0)
+ jump_to_top_level (-nc); /* XXX */
+ }
+
+ /* Need to find how many characters parse_string() consumed, update
+ *indp, if flags != 0, copy the portion of the string parsed into RET
+ and return it. If flags & 1 (SX_NOALLOC) we can return NULL. */
+
+ /*(*/
+ if (ep[-1] != ')')
+ {
+#if 0
+ 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 0
+ if (base[*indp] != ')')
+ itrace("xparse_dolparen:%d: base[%d] != RPAREN (%d), base = `%s'", line_number, *indp, base[*indp], base);
+ if (*indp < orig_ind)
+ itrace("xparse_dolparen:%d: *indp (%d) < orig_ind (%d), orig_string = `%s'", line_number, *indp, orig_ind, ostring);
+#endif
+
+ if (base[*indp] != ')' && (flags & SX_NOLONGJMP) == 0)
+ {
+ /*(*/
+ if ((flags & SX_NOERROR) == 0)
+ parser_error (start_lineno, _("unexpected EOF while looking for matching `%c'"), ')');
+ jump_to_top_level (DISCARD);
+ }
+
+ if (flags & SX_NOALLOC)
+ return (char *)NULL;
+
+ if (nc == 0)
+ {
+ ret = xmalloc (1);
+ ret[0] = '\0';
+ }
+ else
+ ret = substring (ostring, 0, nc - 1);
+
+ return ret;
+}
+
+/* Recursively call the parser to parse the string from a $(...) command
+ substitution to a COMMAND *. This is called from command_substitute() and
+ has the same parser state constraints as xparse_dolparen(). */
+COMMAND *
+parse_string_to_command (string, flags)
+ char *string;
+ int flags;
+{
+ sh_parser_state_t ps;
+ sh_input_line_state_t ls;
+ int nc, sflags;
+ size_t slen;
+ char *ret, *ep;
+ COMMAND *cmd;
+
+ if (*string == 0)
+ return (COMMAND *)NULL;
+
+ ep = string;
+ slen = STRLEN (string);
+
+/*itrace("parse_string_to_command: size = %d shell_input_line = `%s' string=`%s'", shell_input_line_size, shell_input_line, string);*/
+
+ sflags = SEVAL_NONINT|SEVAL_NOHIST|SEVAL_NOFREE;
+ if (flags & SX_NOLONGJMP)
+ sflags |= SEVAL_NOLONGJMP;
+
+ save_parser_state (&ps);
+ save_input_line_state (&ls);
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ pushed_string_list = (STRING_SAVER *)NULL;
+#endif
+ if (flags & SX_COMPLETE)
+ parser_state |= PST_NOERROR;
+
+ expand_aliases = 0;
+
+ cmd = 0;
+ nc = parse_string (string, "command substitution", sflags, &cmd, &ep);
+
+ reset_parser ();
+ /* reset_parser() clears shell_input_line and associated variables, including
+ parser_state, so we want to reset things, then restore what we need. */
+ restore_input_line_state (&ls);
+ restore_parser_state (&ps);
+
+ /* If parse_string returns < 0, we need to jump to top level with the
+ negative of the return value. We abandon the rest of this input line
+ first */
+ if (nc < 0)
+ {
+ clear_shell_input_line (); /* XXX */
+ if ((flags & SX_NOLONGJMP) == 0)
+ jump_to_top_level (-nc); /* XXX */
+ }
+
+ /* Need to check how many characters parse_string() consumed, make sure it's
+ the entire string. */
+ if (nc < slen)
+ {
+ dispose_command (cmd);
+ return (COMMAND *)NULL;
+ }
+
+ return cmd;
+}
+
+#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)
+ {
+ if (word_top < MAX_CASE_NEST)
+ word_top++;
+ arith_for_lineno = word_lineno[word_top] = line_number;
+ cmdtyp = parse_arith_cmd (&wval, 0);
+ if (cmdtyp == 1)
+ {
+ wd = alloc_word_desc ();
+ wd->word = wval;
+ yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
+ return (ARITH_FOR_EXPRS);
+ }
+ 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 */
+ {
+ wd = alloc_word_desc ();
+ wd->word = wval;
+ wd->flags = W_QUOTED|W_NOSPLIT|W_NOGLOB|W_NOTILDE|W_NOPROCSUB;
+ yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
+ return (ARITH_CMD);
+ }
+ else if (cmdtyp == 0) /* nested subshell */
+ {
+ 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);
+ }
+ else /* ERROR */
+ return -1;
+ }
+#endif
+
+ return -2; /* XXX */
+}
+
+/* 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, 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 */
{
- token[ttoklen] = ')';
- token[ttoklen+1] = c;
- token[ttoklen+2] = '\0';
+ strncpy (tokstr, ttok, ttoklen - 1);
+ tokstr[ttoklen-1] = '\0';
}
- *ep = token;
+ else /* nested subshell */
+ {
+ tokstr[0] = '(';
+ strncpy (tokstr + 1, ttok, ttoklen - 1);
+ tokstr[ttoklen] = ')';
+ tokstr[ttoklen+1] = c;
+ tokstr[ttoklen+2] = '\0';
+ }
+
+ *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))
- prompt_again ();
+ if (SHOULD_PROMPT ())
+ prompt_again (0);
}
return (cond_token);
}
{
WORD_DESC *op;
COND_COM *term, *tleft, *tright;
- int tok, lineno;
+ int tok, lineno, local_extglob;
+ 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);
dispose_word (yylval.word); /* not needed */
term = cond_term ();
if (term)
- term->flags |= CMD_INVERT_RETURN;
+ 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[1] && 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 = cond_skip_newlines (); ? */
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 */
+ local_extglob = extended_glob;
+ if (parser_state & PST_EXTPAT)
+ extended_glob = 1;
tok = read_token (READ);
+ if (parser_state & PST_EXTPAT)
+ extended_glob = local_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 use a buffer with room for the characters we add just
+ in case assignment() ends up doing something like parsing a command
+ substitution that will reallocate atoken. We don't want to write beyond
+ the end of an allocated buffer. */
+static int
+token_is_assignment (t, i)
+ char *t;
+ int i;
+{
+ int r;
+ char *atoken;
+
+ atoken = xmalloc (i + 3);
+ memcpy (atoken, t, i);
+ atoken[i] = '=';
+ atoken[i+1] = '\0';
+
+ r = assignment (atoken, (parser_state & PST_COMPASSIGN) != 0);
+
+ free (atoken);
+
+ /* XXX - check that r == i to avoid returning false positive for
+ t containing `=' before t[i]. */
+ return (r > 0 && r == i);
+}
+
+/* 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;
/* Non-zero means to ignore the value of the next character, and just
to add it no matter what. */
- int pass_next_character;
+ int pass_next_character;
/* The current delimiting character. */
int cd;
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 == '\\')
{
+ if (parser_state & PST_NOEXPAND)
+ {
+ pass_next_character++;
+ quoted = 1;
+ goto got_character;
+ }
+
peek_char = shell_getc (0);
/* Backslash-newline is ignored in all cases except
when quoted with single quotes. */
- if (peek_char == '\n')
+ if MBTEST(peek_char == '\n')
{
character = '\n';
goto next_character;
shell_ungetc (peek_char);
/* If the next character is to be quoted, note it now. */
- if (cd == 0 || cd == '`' ||
- (cd == '"' && member (peek_char, slashify_in_quotes)))
+ if MBTEST(cd == 0 || cd == '`' ||
+ (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;
- quoted = 1;
+ all_digit_token = 0;
+ if (character != '`')
+ quoted = 1;
dollar_present |= (character == '"' && strchr (ttok, '$') != 0);
FREE (ttok);
goto next_character;
}
-#ifdef EXTENDED_GLOB
- /* Parse a ksh-style extended pattern matching specification. */
- if (extended_glob && PATTERN_CHAR(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 == '|')) /*)*/
{
- peek_char = shell_getc (1);
- if (peek_char == '(') /* ) */
- {
- push_delimiter (dstack, peek_char);
- ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
+ 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 MBTEST(extended_glob && PATTERN_CHAR (character))
+ {
+ peek_char = shell_getc (1);
+ 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 MBTEST(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, P_FIRSTCLOSE);
+ 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 == '"'))
+#if defined (TRANSLATABLE_STRINGS)
+ else if MBTEST(character == '$' && (peek_char == '\'' || peek_char == '"'))
+#else
+ else if MBTEST(character == '$' && peek_char == '\'')
+#endif
{
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);
+ {
+ /* PST_NOEXPAND */
+ 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;
+ }
+#if defined (TRANSLATABLE_STRINGS)
else
- ttrans = localeexpand (ttok, 0, ttoklen - 1, first_line, &ttranslen);
- free (ttok);
- RESIZE_MALLOCED_BUFFER (token, token_index, ttranslen + 2,
+ {
+ /* PST_NOEXPAND */
+ /* Try to locale-expand the converted string. */
+ ttrans = locale_expand (ttok, 0, ttoklen - 1, first_line, &ttranslen);
+ free (ttok);
+
+ /* Add the double quotes back (or single quotes if the user
+ has set that option). */
+ if (singlequote_translations &&
+ ((ttoklen - 1) != ttranslen || STREQN (ttok, ttrans, ttranslen) == 0))
+ ttok = sh_single_quote (ttrans);
+ else
+ ttok = sh_mkdoublequoted (ttrans, ttranslen, 0);
+
+ free (ttrans);
+ ttrans = ttok;
+ ttranslen = strlen (ttrans);
+ }
+#endif /* TRANSLATABLE_STRINGS */
+
+ 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 (character == '$' && peek_char == '$')
+ else if MBTEST(character == '$' && peek_char == '$')
{
- ttok = xmalloc (3);
- ttok[0] = ttok[1] = '$';
- ttok[2] = '\0';
RESIZE_MALLOCED_BUFFER (token, token_index, 3,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
- strcpy (token + token_index, ttok);
- token_index += 2;
+ token[token_index++] = '$';
+ token[token_index++] = peek_char;
dollar_present = 1;
- all_digits = 0;
- FREE (ttok);
+ 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. */
- if (ttok[0] == '(') /* ) */
- {
- FREE (ttok);
- return -1;
- }
- 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 == '$';
-
- if (character == CTLESC || character == CTLNUL)
- token[token_index++] = CTLESC;
+got_character:
+ if MBTEST(character == CTLESC || character == CTLNUL)
+ {
+ RESIZE_MALLOCED_BUFFER (token, token_index, 2, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = CTLESC;
+ }
+ else
+got_escaped_character:
+ RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
token[token_index++] = character;
- RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
- TOKEN_DEFAULT_GROW_SIZE);
+ all_digit_token &= DIGIT (character);
+ dollar_present |= character == '$';
next_character:
- if (character == '\n' && interactive &&
- (bash_input.type == st_stdin || bash_input.type == st_stream))
- prompt_again ();
+ if (character == '\n' && SHOULD_PROMPT ())
+ prompt_again (0);
/* We want to remove quoted newlines (that is, a \<newline> pair)
unless we are within single quotes or pass_next_character is
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 = alloc_word_desc ();
+ 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;
+ /* should we check that quoted == 0 as well? */
+ if MBTEST(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, 0))
+#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); */
+ yylval.word = the_word; /* accommodate recursive call */
+ return (REDIR_WORD);
+ }
+ else
+ /* valid_array_reference can call the parser recursively; need to
+ make sure that yylval.word doesn't change if we are going to
+ return WORD or ASSIGNMENT_WORD */
+ yylval.word = the_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;
+ expecting_in_token++;
+ 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 ARITH_CMD:
+ case BANG:
+ case BAR_AND:
+ case COND_END:
+ 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:
+ case DOLPAREN:
+ 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 (here_doc_first_line ? "\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 " ";
- else
- return "; "; /* (...) subshell */
+ return " ";
+ else
+ 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 ((parser_state & PST_HEREDOC) == 0 && current_command_line_count > 1 && need_here_doc > 0)
+ 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])
return (" ");
}
+ /* Assume that by this point we are reading lines in a multi-line command.
+ If we have multiple consecutive blank lines we want to return only one
+ semicolon. */
+ if (line_isblank (line))
+ return (current_command_line_count > 1 && last_read_token == '\n' && token_before_that != '\n') ? "; " : "";
+
return ("; ");
}
#endif /* HISTORY */
/* Issue a prompt, or prepare to issue a prompt when the next character
is read. */
static void
-prompt_again ()
+prompt_again (force)
+ int force;
{
char *temp_prompt;
- if (!interactive) /* XXX */
+ if (interactive == 0 || expanding_alias ()) /* XXX */
return;
ps1_prompt = get_string_value ("PS1");
ps2_prompt = get_string_value ("PS2");
+ ps0_prompt = get_string_value ("PS0");
+
if (!prompt_string_pointer)
prompt_string_pointer = &ps1_prompt;
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 ()
{
fflush (stderr);
}
+#if defined (HISTORY)
+ /* The history library increments the history offset as soon as it stores
+ the first line of a potentially multi-line command, so we compensate
+ here by returning one fewer when appropriate. */
+static int
+prompt_history_number (pmt)
+ char *pmt;
+{
+ int ret;
+
+ ret = history_number ();
+ if (ret == 1)
+ return ret;
+
+ if (pmt == ps1_prompt) /* are we expanding $PS1? */
+ return ret;
+ else if (pmt == ps2_prompt && command_oriented_history == 0)
+ return ret; /* not command oriented history */
+ else if (pmt == ps2_prompt && command_oriented_history && current_command_first_line_saved)
+ return ret - 1;
+ else
+ return ret - 1; /* PS0, PS4, ${var@P}, PS2 other cases */
+}
+#endif
+
/* Return a string which will be printed as a prompt. The string
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
char *string;
{
WORD_LIST *list;
- char *result, *t;
+ char *result, *t, *orig_string;
struct dstack save_dstack;
+ int last_exit_value, last_comsub_pid;
#if defined (PROMPT_STRING_DECODE)
- int result_size, result_index;
- int c, n;
- char *temp, octal_string[4];
+ size_t result_size;
+ size_t result_index;
+ int c, n, i;
+ char *temp, *t_host, 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;
+ orig_string = string;
while (c = *string++)
{
#if !defined (HISTORY)
temp = savestring ("1");
#else /* HISTORY */
- temp = itos (history_number ());
+ temp = itos (prompt_history_number (orig_string));
#endif /* HISTORY */
string--; /* add_string increments string again. */
goto add_string;
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 = savestring (timebuf);
+ goto add_string;
- temp = (c != 'd') ? savestring (temp + 11) : savestring (temp);
- temp[(c != 'd') ? 8 : 10] = '\0';
+ case 'D': /* strftime format */
+ if (string[1] != '{') /* } */
+ goto not_escape;
- /* quick and dirty conversion to 12-hour time */
- if (c == 'T' || c == '@')
+ (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, 0);
+ 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 's':
temp = base_pathname (shell_name);
- temp = savestring (temp);
+ /* Try to quote anything the user can set in the file system */
+ if (promptvars || posixly_correct)
+ {
+ char *t;
+ t = sh_strvis (temp);
+ temp = sh_backslash_quote_for_double_quotes (t, 0);
+ free (t);
+ }
+ else
+ temp = sh_strvis (temp);
goto add_string;
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. */
+ {
+ char *t;
+ t = sh_strvis (t_string);
+ temp = sh_backslash_quote_for_double_quotes (t, 0);
+ free (t);
+ }
else
- temp = savestring (t_string);
+ temp = sh_strvis (t_string);
goto add_string;
}
case 'u':
+ if (current_user.user_name == 0)
+ get_current_user_info ();
temp = savestring (current_user.user_name);
goto add_string;
case 'h':
case 'H':
- temp = savestring (current_host_name);
- if (c == 'h' && (t = (char *)strchr (temp, '.')))
+ t_host = savestring (current_host_name);
+ if (c == 'h' && (t = (char *)strchr (t_host, '.')))
*t = '\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 (t_host, 0);
+ else
+ temp = savestring (t_host);
+ free (t_host);
goto add_string;
case '#':
- temp = itos (current_command_number);
+ n = current_command_number;
+ /* If we have already incremented current_command_number (PS4,
+ ${var@P}), compensate */
+ if (orig_string != ps0_prompt && orig_string != ps1_prompt && orig_string != ps2_prompt)
+ n--;
+ temp = itos (n);
goto add_string;
case '!':
#if !defined (HISTORY)
temp = savestring ("1");
#else /* HISTORY */
- temp = itos (history_number ());
+ temp = itos (prompt_history_number (orig_string));
#endif /* HISTORY */
goto add_string;
case '$':
- t = temp = xmalloc (3);
+ 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';
else
{
RESIZE_MALLOCED_BUFFER (result, result_index, 3, result_size, PROMPT_GROWTH);
+ /* dequote_string should take care of removing this if we are not
+ performing the rest of the word expansions. */
+ if (c == CTLESC || c == CTLNUL)
+ result[result_index++] = CTLESC;
result[result_index++] = c;
result[result_index] = '\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);
+ if ((parser_state & PST_NOERROR) == 0)
+ 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 = (executing_builtin && parse_and_execute_level) ? EX_BADSYNTAX : EX_BADUSAGE;
+ set_pipestatus_from_exit (last_command_exit_value);
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 = (executing_builtin && parse_and_execute_level) ? EX_BADSYNTAX : EX_BADUSAGE;
+ set_pipestatus_from_exit (last_command_exit_value);
+ 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";
- parser_error (line_number, "%s", msg);
+ if (EOF_Reached && shell_eof_token && current_token != shell_eof_token)
+ parser_error (line_number, _("unexpected EOF while looking for matching `%c'"), shell_eof_token);
+ else
+ {
+ 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
decide whether or not to exit. */
if (interactive && EOF_Reached)
EOF_Reached = 0;
}
- last_command_exit_value = EX_USAGE;
+
+ last_command_exit_value = (executing_builtin && parse_and_execute_level) ? EX_BADSYNTAX : EX_BADUSAGE;
+ set_pipestatus_from_exit (last_command_exit_value);
}
/* ??? 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';
+ prompt_again (0);
return;
}
}
/* In this case EOF should exit the shell. Do it now. */
reset_parser ();
+
+ last_shell_builtin = this_shell_builtin;
+ this_shell_builtin = exit_builtin;
exit_builtin ((WORD_LIST *)NULL);
}
else
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;
+ int orig_parser_state;
+ sh_parser_state_t ps;
+ int ea;
+
+ orig_line_number = line_number;
+ save_parser_state (&ps);
+
+#if defined (HISTORY)
+ bash_history_disable ();
+#endif
+
+ push_stream (1);
+ if (ea = expanding_alias ())
+ parser_save_alias ();
+
+ /* WORD to avoid parsing reserved words as themselves and just parse them as
+ WORDs. */
+ last_read_token = WORD;
+
+ 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)
+ {
+ orig_parser_state = parser_state; /* XXX - not needed? */
+ /* State flags we don't want to persist into compound assignments. */
+ parser_state &= ~PST_NOEXPAND; /* parse_comsub sentinel */
+ /* State flags we want to set for this run through the tokenizer. */
+ 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 (ea)
+ parser_restore_alias ();
+
+ restore_parser_state (&ps);
+
+ if (flags & 1)
+ parser_state = orig_parser_state; /* XXX - not needed? */
+
+ if (wl == &parse_string_error)
+ {
+ set_exit_status (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, assignok;
+ sh_parser_state_t ps;
+ char *ret;
+
+ orig_line_number = line_number;
+ save_parser_state (&ps);
+
+ /* WORD to avoid parsing reserved words as themselves and just parse them as
+ WORDs. Plus it means we won't be in a command position and so alias
+ expansion won't happen. */
+ last_read_token = WORD;
+
+ token = (char *)NULL;
+ token_buffer_size = 0;
+ wl = (WORD_LIST *)NULL; /* ( */
+
+ assignok = parser_state&PST_ASSIGNOK; /* XXX */
+
+ /* State flags we don't want to persist into compound assignments. */
+ parser_state &= ~(PST_NOEXPAND|PST_CONDCMD|PST_CONDEXPR|PST_REGEXP|PST_EXTPAT);
+ /* State flags we want to set for this run through the tokenizer. */
+ parser_state |= PST_COMPASSIGN;
+
+ esacs_needed_count = expecting_in_token = 0;
+
+ while ((tok = read_token (READ)) != ')')
+ {
+ if (tok == '\n') /* Allow newlines in compound assignments */
+ {
+ if (SHOULD_PROMPT ())
+ prompt_again (0);
+ 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);
+ }
+
+ restore_parser_state (&ps);
+
+ if (wl == &parse_string_error)
+ {
+ set_exit_status (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);
+ }
+
+ 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->eol_lookahead = eol_ungetc_lookahead;
+
+ 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->need_here_doc = need_here_doc;
+ ps->here_doc_first_line = here_doc_first_line;
+
+ ps->esacs_needed = esacs_needed_count;
+ ps->expecting_in = expecting_in_token;
+
+ if (need_here_doc == 0)
+ ps->redir_stack[0] = 0;
+ else
+ memcpy (ps->redir_stack, redir_stack, sizeof (redir_stack[0]) * HEREDOC_MAX);
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ ps->pushed_strings = pushed_string_list;
+#endif
+
+ ps->eof_token = shell_eof_token;
+ 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;
+{
+ int i;
+
+ 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;
+ eol_ungetc_lookahead = ps->eol_lookahead;
+
+ 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;
+ need_here_doc = ps->need_here_doc;
+ here_doc_first_line = ps->here_doc_first_line;
+
+ esacs_needed_count = ps->esacs_needed;
+ expecting_in_token = ps->expecting_in;
+
+#if 0
+ for (i = 0; i < HEREDOC_MAX; i++)
+ redir_stack[i] = ps->redir_stack[i];
+#else
+ if (need_here_doc == 0)
+ redir_stack[0] = 0;
+ else
+ memcpy (redir_stack, ps->redir_stack, sizeof (redir_stack[0]) * HEREDOC_MAX);
+#endif
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ pushed_string_list = (STRING_SAVER *)ps->pushed_strings;
+#endif
+
+ FREE (token);
+ token = ps->token;
+ token_buffer_size = ps->token_buffer_size;
+ shell_eof_token = ps->eof_token;
+}
+
+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;
+
+#if defined (HANDLE_MULTIBYTE)
+ ls->input_property = shell_input_line_property;
+ ls->input_propsize = shell_input_line_propsize;
+#endif
+
+ /* force reallocation */
+ shell_input_line = 0;
+ shell_input_line_size = shell_input_line_len = shell_input_line_index = 0;
+
+#if defined (HANDLE_MULTIBYTE)
+ shell_input_line_property = 0;
+ shell_input_line_propsize = 0;
+#endif
+
+ 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;
+
+#if defined (HANDLE_MULTIBYTE)
+ FREE (shell_input_line_property);
+ shell_input_line_property = ls->input_property;
+ shell_input_line_propsize = ls->input_propsize;
+#endif
+
+#if 0
+ set_line_mbstate ();
+#endif
+}
+
+/************************************************
+ * *
+ * MULTIBYTE CHARACTER HANDLING *
+ * *
+ ************************************************/
+
+#if defined (HANDLE_MULTIBYTE)
+
+/* We don't let the property buffer get larger than this unless the line is */
+#define MAX_PROPSIZE 32768
+
+static void
+set_line_mbstate ()
+{
+ int c;
+ size_t i, previ, len;
+ mbstate_t mbs, prevs;
+ size_t mbclen;
+ int ilen;
+
+ if (shell_input_line == NULL)
+ return;
+ len = STRLEN (shell_input_line); /* XXX - shell_input_line_len ? */
+ if (len == 0)
+ return;
+ if (shell_input_line_propsize >= MAX_PROPSIZE && len < MAX_PROPSIZE>>1)
+ {
+ free (shell_input_line_property);
+ shell_input_line_property = 0;
+ shell_input_line_propsize = 0;
+ }
+ if (len+1 > shell_input_line_propsize)
+ {
+ shell_input_line_propsize = len + 1;
+ shell_input_line_property = (char *)xrealloc (shell_input_line_property, shell_input_line_propsize);
+ }
+
+ if (locale_mb_cur_max == 1)
+ {
+ memset (shell_input_line_property, 1, len);
+ return;
+ }
+
+ /* XXX - use whether or not we are in a UTF-8 locale to avoid calls to
+ mbrlen */
+ if (locale_utf8locale == 0)
+ memset (&prevs, '\0', sizeof (mbstate_t));
+
+ for (i = previ = 0; i < len; i++)
+ {
+ if (locale_utf8locale == 0)
+ 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;
+ }
+
+ if (locale_utf8locale)
+ {
+ if ((unsigned char)shell_input_line[previ] < 128) /* i != previ */
+ mbclen = 1;
+ else
+ {
+ ilen = utf8_mblen (shell_input_line + previ, i - previ + 1);
+ mbclen = (ilen == -1) ? (size_t)-1
+ : ((ilen == -2) ? (size_t)-2 : (size_t)ilen);
+ }
+ }
+ else
+ 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;
+ if (locale_utf8locale == 0)
+ prevs = mbs;
+ }
+ else
+ {
+ 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