-/* Yacc grammar for bash. */
+/* parse.y - Yacc grammar for bash. */
-/* Copyright (C) 1989-2005 Free Software Foundation, Inc.
+/* Copyright (C) 1989-2017 Free Software Foundation, Inc.
This file is part of GNU Bash, the Bourne Again SHell.
- Bash is free software; you can redistribute it and/or modify it under
- the terms of the GNU General Public License as published by the Free
- Software Foundation; either version 2, or (at your option) any later
- version.
+ Bash is free software: you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
- Bash is distributed in the hope that it will be useful, but WITHOUT ANY
- WARRANTY; without even the implied warranty of MERCHANTABILITY or
- FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
- for more details.
+ Bash is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
- You should have received a copy of the GNU General Public License along
- with Bash; see the file LICENSE. If not, write to the Free Software
- Foundation, 59 Temple Place, Suite 330, Boston, MA 02111 USA. */
+ You should have received a copy of the GNU General Public License
+ along with Bash. If not, see <http://www.gnu.org/licenses/>.
+*/
%{
#include "config.h"
#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"
#if defined (JOB_CONTROL)
# include "jobs.h"
+#else
+extern int cleanup_dead_jobs __P((void));
#endif /* JOB_CONTROL */
#if defined (ALIAS)
#define RE_READ_TOKEN -99
#define NO_EXPANSION -100
+#define END_ALIAS -2
+
#ifdef DEBUG
# define YYDEBUG 1
#else
extern int extended_glob;
#endif
-extern int eof_encountered;
-extern int no_line_editing, running_under_emacs;
-extern int current_command_number;
-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;
extern int dump_translatable_strings, dump_po_strings;
-extern sh_builtin_func_t *last_shell_builtin, *this_shell_builtin;
-#if defined (BUFFERED_INPUT)
-extern int bash_input_fd_changed;
-#endif
+#if !defined (errno)
extern int errno;
+#endif
+
/* **************************************************************** */
/* */
/* "Forward" declarations */
static int yy_string_get __P((void));
static int yy_string_unget __P((int));
+static void rewind_input_string __P((void));
static int yy_stream_get __P((void));
static int yy_stream_unget __P((int));
static int reserved_word_acceptable __P((int));
static int yylex __P((void));
+
+static void push_heredoc __P((REDIRECT *));
+static char *mk_alexpansion __P((char *));
static int alias_expand_token __P((char *));
static int time_command_acceptable __P((void));
static int special_case_tokens __P((char *));
static int read_token __P((int));
static char *parse_matched_pair __P((int, int, int, int *, int));
+static char *parse_comsub __P((int, int, int, int *, int));
#if defined (ARRAY_VARS)
static char *parse_compound_assignment __P((int *));
#endif
#endif
static void print_prompt __P((void));
-#if defined (HISTORY)
-char *history_delimiting_chars __P((void));
-#endif
-
#if defined (HANDLE_MULTIBYTE)
static void set_line_mbstate __P((void));
static char *shell_input_line_property = NULL;
+static size_t shell_input_line_propsize = 0;
#else
# define set_line_mbstate()
#endif
/* 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;
quotes. */
int extended_quote = 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;
-
/* 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 at which an arithmetic for command starts. */
static int arith_for_lineno;
+/* The decoded prompt string. Used if READLINE is not defined or if
+ editing is turned off. Analogous to current_readline_prompt. */
+static char *current_decoded_prompt;
+
+/* The last read token, or NULL. read_token () uses this for context
+ checking. */
+static int last_read_token;
+
+/* The token read prior to last_read_token. */
+static int token_before_that;
+
+/* The token read prior to token_before_that. */
+static int two_tokens_ago;
+
+static int global_extglob;
+
/* The line number in a script where the word in a `case WORD', `select WORD'
or `for WORD' begins. This is a nested command maximum, since the array
index is decremented after a case, select, or for command is parsed. */
#define MAX_CASE_NEST 128
-static int word_lineno[MAX_CASE_NEST];
+static int word_lineno[MAX_CASE_NEST+1];
static int word_top = -1;
/* If non-zero, it is the token that we want read_token to return
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 ARITH_FOR_EXPRS
%token <command> COND_CMD
%token AND_AND OR_OR GREATER_GREATER LESS_LESS LESS_AND LESS_LESS_LESS
-%token GREATER_AND SEMI_SEMI LESS_LESS_MINUS AND_GREATER LESS_GREATER
-%token GREATER_BAR
+%token GREATER_AND SEMI_SEMI SEMI_AND SEMI_SEMI_AND
+%token LESS_LESS_MINUS AND_GREATER AND_GREATER_GREATER LESS_GREATER
+%token GREATER_BAR BAR_AND
/* The types that the various syntactical units return. */
%type <command> arith_command
%type <command> cond_command
%type <command> arith_for_command
+%type <command> coproc
%type <command> function_def function_body if_command elif_clause subshell
%type <redirect> redirection redirection_list
%type <element> simple_command_element
%left '&' ';' '\n' yacc_EOF
%left AND_AND OR_OR
-%right '|'
+%right '|' BAR_AND
%%
inputunit: simple_list simple_list_terminator
global_command = $1;
eof_encountered = 0;
/* discard_parser_constructs (0); */
+ if (parser_state & PST_CMDSUBST)
+ parser_state |= PST_EOFTOKEN;
YYACCEPT;
}
| '\n'
/* Case of regular command, but not a very
interesting one. Return a NULL command. */
global_command = (COMMAND *)NULL;
+ if (parser_state & PST_CMDSUBST)
+ parser_state |= PST_EOFTOKEN;
YYACCEPT;
}
| error '\n'
global_command = (COMMAND *)NULL;
eof_encountered = 0;
/* discard_parser_constructs (1); */
- if (interactive)
+ if (interactive && parse_and_execute_level == 0)
+ {
+ YYACCEPT;
+ }
+ 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 */
+ handle_eof_input_unit ();
+ if (interactive && parse_and_execute_level == 0)
{
YYACCEPT;
}
redirection: '>' WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_output_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, 0);
}
| '<' WORD
{
+ source.dest = 0;
redir.filename = $2;
- $$ = make_redirection (0, r_input_direction, redir);
+ $$ = make_redirection (source, r_input_direction, redir, 0);
}
| NUMBER '>' WORD
{
+ source.dest = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_output_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, 0);
}
| NUMBER '<' WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_direction, redir, 0);
+ }
+ | REDIR_WORD '>' WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_input_direction, redir);
+ $$ = make_redirection (source, r_output_direction, redir, REDIR_VARASSIGN);
+ }
+ | REDIR_WORD '<' WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_direction, redir, REDIR_VARASSIGN);
}
| GREATER_GREATER WORD
{
+ source.dest = 1;
redir.filename = $2;
- $$ = make_redirection (1, r_appending_to, redir);
+ $$ = make_redirection (source, r_appending_to, redir, 0);
}
| NUMBER GREATER_GREATER WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_appending_to, redir, 0);
+ }
+ | REDIR_WORD GREATER_GREATER WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_appending_to, redir, REDIR_VARASSIGN);
+ }
+ | GREATER_BAR WORD
+ {
+ source.dest = 1;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_output_force, redir, 0);
+ }
+ | NUMBER GREATER_BAR WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_output_force, redir, 0);
+ }
+ | REDIR_WORD GREATER_BAR WORD
+ {
+ source.filename = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_output_force, redir, REDIR_VARASSIGN);
+ }
+ | LESS_GREATER WORD
+ {
+ source.dest = 0;
+ redir.filename = $2;
+ $$ = make_redirection (source, r_input_output, redir, 0);
+ }
+ | NUMBER LESS_GREATER WORD
+ {
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_input_output, redir, 0);
+ }
+ | REDIR_WORD LESS_GREATER WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_appending_to, redir);
+ $$ = make_redirection (source, r_input_output, redir, REDIR_VARASSIGN);
}
| LESS_LESS WORD
{
+ source.dest = 0;
redir.filename = $2;
- $$ = make_redirection (0, r_reading_until, redir);
- 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 ($1, r_reading_until, redir);
- redir_stack[need_here_doc++] = $$;
+ $$ = 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 (0, r_reading_string, redir);
+ $$ = make_redirection (source, r_reading_string, redir, 0);
}
| NUMBER LESS_LESS_LESS WORD
{
+ source.dest = $1;
+ redir.filename = $3;
+ $$ = make_redirection (source, r_reading_string, redir, 0);
+ }
+ | REDIR_WORD LESS_LESS_LESS WORD
+ {
+ source.filename = $1;
redir.filename = $3;
- $$ = make_redirection ($1, r_reading_string, redir);
+ $$ = 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 ($1, r_duplicating_output, redir);
+ $$ = make_redirection (source, r_duplicating_output, redir, 0);
+ }
+ | REDIR_WORD GREATER_AND NUMBER
+ {
+ source.filename = $1;
+ redir.dest = $3;
+ $$ = 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 '-'
{
+ source.dest = 1;
redir.dest = 0;
- $$ = make_redirection (1, r_close_this, redir);
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
| NUMBER GREATER_AND '-'
{
+ source.dest = $1;
redir.dest = 0;
- $$ = make_redirection ($1, r_close_this, redir);
+ $$ = make_redirection (source, r_close_this, redir, 0);
}
- | LESS_AND '-'
+ | REDIR_WORD GREATER_AND '-'
{
+ source.filename = $1;
redir.dest = 0;
- $$ = make_redirection (0, r_close_this, redir);
+ $$ = make_redirection (source, r_close_this, redir, REDIR_VARASSIGN);
}
- | NUMBER LESS_AND '-'
+ | LESS_AND '-'
{
+ source.dest = 0;
redir.dest = 0;
- $$ = make_redirection ($1, r_close_this, redir);
+ $$ = 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);
+ $$ = make_redirection (source, r_append_err_and_out, redir, 0);
}
;
COMMAND *tc;
tc = $1;
- if (tc->redirects)
+ if (tc && tc->redirects)
{
register REDIRECT *t;
for (t = tc->redirects; t->next; t = t->next)
;
t->next = $2;
}
- else
+ else if (tc)
tc->redirects = $2;
$$ = $1;
}
| function_def
{ $$ = $1; }
+ | coproc
+ { $$ = $1; }
;
shell_command: for_command
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--;
}
;
$$ = make_select_command ($2, REVERSE_LIST ($5, WORD_LIST *), $9, word_lineno[word_top]);
if (word_top > 0) word_top--;
}
+ | SELECT WORD newline_list IN list_terminator newline_list DO compound_list DONE
+ {
+ $$ = make_select_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
+ | SELECT WORD newline_list IN list_terminator newline_list '{' compound_list '}'
+ {
+ $$ = make_select_command ($2, (WORD_LIST *)NULL, $8, word_lineno[word_top]);
+ if (word_top > 0) word_top--;
+ }
;
case_command: CASE WORD newline_list IN newline_list ESAC
{ $$ = make_function_def ($2, $4, function_dstart, function_bstart); }
;
-
function_body: shell_command
{ $$ = $1; }
| shell_command redirection_list
redirection. The two are semantically equivalent,
though -- the only difference is in how the
command printing code displays the redirections. */
- if (tc->redirects)
+ if (tc && tc->redirects)
{
register REDIRECT *t;
for (t = tc->redirects; t->next; t = t->next)
;
t->next = $2;
}
- else
+ else if (tc)
tc->redirects = $2;
$$ = $1;
}
}
;
+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
{ $$ = make_if_command ($2, $4, (COMMAND *)NULL); }
| IF compound_list THEN compound_list ELSE compound_list FI
;
case_clause_sequence: pattern_list SEMI_SEMI
+ { $$ = $1; }
| case_clause_sequence pattern_list SEMI_SEMI
{ $2->next = $1; $$ = $2; }
+ | pattern_list SEMI_AND
+ { $1->flags |= CASEPAT_FALLTHROUGH; $$ = $1; }
+ | case_clause_sequence pattern_list SEMI_AND
+ { $2->flags |= CASEPAT_FALLTHROUGH; $2->next = $1; $$ = $2; }
+ | pattern_list SEMI_SEMI_AND
+ { $1->flags |= CASEPAT_TESTNEXT; $$ = $1; }
+ | case_clause_sequence pattern_list SEMI_SEMI_AND
+ { $2->flags |= CASEPAT_TESTNEXT; $2->next = $1; $$ = $2; }
;
pattern: WORD
$$ = $1;
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 '&'
{
$$ = command_connect ($1, (COMMAND *)NULL, '&');
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
| simple_list1 ';'
{
$$ = $1;
if (need_here_doc)
gather_here_documents ();
+ if ((parser_state & PST_CMDSUBST) && current_token == shell_eof_token)
+ {
+ global_command = $1;
+ eof_encountered = 0;
+ rewind_input_string ();
+ YYACCEPT;
+ }
}
;
;
pipeline_command: pipeline
- { $$ = $1; }
- | BANG pipeline
+ { $$ = $1; }
+ | BANG pipeline_command
{
if ($2)
- $2->flags |= CMD_INVERT_RETURN;
+ $2->flags ^= CMD_INVERT_RETURN; /* toggle */
$$ = $2;
}
- | timespec pipeline
+ | timespec pipeline_command
{
if ($2)
$2->flags |= $1;
$$ = $2;
}
- | timespec BANG pipeline
- {
- if ($3)
- $3->flags |= $1|CMD_INVERT_RETURN;
- $$ = $3;
- }
- | BANG timespec pipeline
- {
- if ($3)
- $3->flags |= $2|CMD_INVERT_RETURN;
- $$ = $3;
- }
| timespec list_terminator
{
ELEMENT x;
/* 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 list_terminator
+ {
+ 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
{ $$ = CMD_TIME_PIPELINE|CMD_TIME_POSIX; }
+ | TIME TIMEOPT TIMEIGN
+ { $$ = CMD_TIME_PIPELINE|CMD_TIME_POSIX; }
;
%%
-/* Possible states for the parser that require it to do special things. */
-#define PST_CASEPAT 0x0001 /* in a case pattern list */
-#define PST_ALEXPNEXT 0x0002 /* expand next word for aliases */
-#define PST_ALLOWOPNBRC 0x0004 /* allow open brace for function def */
-#define PST_NEEDCLOSBRC 0x0008 /* need close brace */
-#define PST_DBLPAREN 0x0010 /* double-paren parsing */
-#define PST_SUBSHELL 0x0020 /* ( ... ) subshell */
-#define PST_CMDSUBST 0x0040 /* $( ... ) command substitution */
-#define PST_CASESTMT 0x0080 /* parsing a case statement */
-#define PST_CONDCMD 0x0100 /* parsing a [[...]] command */
-#define PST_CONDEXPR 0x0200 /* parsing the guts of [[...]] */
-#define PST_ARITHFOR 0x0400 /* parsing an arithmetic for command */
-#define PST_ALEXPAND 0x0800 /* OK to expand aliases - unused */
-#define PST_CMDTOKEN 0x1000 /* command token OK - unused */
-#define PST_COMPASSIGN 0x2000 /* parsing x=(...) compound assignment */
-#define PST_ASSIGNOK 0x4000 /* assignment statement ok in this context */
-
/* Initial size to allocate for tokens, and the
amount to grow them by. */
#define TOKEN_DEFAULT_INITIAL_SIZE 496
# define expanding_alias() 0
#endif
-/* The token currently being read. */
-static int current_token;
-
-/* The last read token, or NULL. read_token () uses this for context
- checking. */
-static int last_read_token;
-
-/* The token read prior to last_read_token. */
-static int token_before_that;
-
-/* The token read prior to token_before_that. */
-static int two_tokens_ago;
-
-/* The current parser state. */
-static int parser_state;
-
/* Global var is non-zero when end of file has been reached. */
int EOF_Reached = 0;
{
#if YYDEBUG != 0
yydebug = i;
+ yyoutstream = stdout;
+ yyerrstream = stderr;
#endif
}
#endif
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)
give_terminal_to (shell_pgrp, 0);
#endif /* JOB_CONTROL */
- old_sigint = (SigHandler *)NULL;
+ old_sigint = IMPOSSIBLE_TRAP_HANDLER;
if (signal_is_ignored (SIGINT) == 0)
{
+ /* interrupt_immediately++; */
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 : "");
- if (signal_is_ignored (SIGINT) == 0 && old_sigint)
+ CHECK_TERMSIG;
+ if (signal_is_ignored (SIGINT) == 0)
{
- interrupt_immediately--;
- set_signal_handler (SIGINT, old_sigint);
+ /* interrupt_immediately--; */
+ if (old_sigint != IMPOSSIBLE_TRAP_HANDLER)
+ set_signal_handler (SIGINT, old_sigint);
}
#if 0
init_yy_io (yy_string_get, yy_string_unget, st_string, name, location);
}
+/* Count the number of characters we've consumed from bash_input.location.string
+ and read into shell_input_line, but have not returned from shell_getc.
+ That is the true input location. Rewind bash_input.location.string by
+ that number of characters, so it points to the last character actually
+ consumed by the parser. */
+static void
+rewind_input_string ()
+{
+ int xchars;
+
+ /* number of unconsumed characters in the input -- XXX need to take newlines
+ into account, e.g., $(...\n) */
+ xchars = shell_input_line_len - shell_input_line_index;
+ if (bash_input.location.string[-1] == '\n')
+ xchars++;
+
+ /* XXX - how to reflect bash_input.location.string back to string passed to
+ parse_and_execute or xparse_dolparen? xparse_dolparen needs to know how
+ far into the string we parsed. parse_and_execute knows where bash_input.
+ location.string is, and how far from orig_string that is -- that's the
+ number of characters the command consumed. */
+
+ /* bash_input.location.string - xchars should be where we parsed to */
+ /* need to do more validation on xchars value for sanity -- test cases. */
+ bash_input.location.string -= xchars;
+}
+
/* **************************************************************** */
/* */
/* Let input come from STREAM. */
result = EOF;
if (bash_input.location.file)
{
- if (interactive)
- interrupt_immediately++;
+ /* 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);
- if (interactive)
- interrupt_immediately--;
}
return (result);
}
/* 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;
{
int *ret;
- ret = (int *)xmalloc (3 * sizeof (int));
+ 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;
}
last_read_token = ts[0];
token_before_that = ts[1];
two_tokens_ago = ts[2];
+ current_token = ts[3];
}
/*
* implement alias expansion on a per-token basis.
*/
+#define PSH_ALIAS 0x01
+#define PSH_DPAREN 0x02
+#define PSH_SOURCE 0x04
+
typedef struct string_saver {
struct string_saver *next;
int expand_alias; /* Value to set expand_alias to when string is popped. */
#if defined (ALIAS)
alias_t *expander; /* alias that caused this line to be pushed. */
#endif
- int saved_line_size, saved_line_index, saved_line_terminator;
+ size_t saved_line_size, saved_line_index;
+ int saved_line_terminator;
+ int flags;
} STRING_SAVER;
STRING_SAVER *pushed_string_list = (STRING_SAVER *)NULL;
temp->saved_line_size = shell_input_line_size;
temp->saved_line_index = shell_input_line_index;
temp->saved_line_terminator = shell_input_line_terminator;
+ temp->flags = 0;
#if defined (ALIAS)
temp->expander = ap;
+ if (ap)
+ temp->flags = PSH_ALIAS;
#endif
temp->next = pushed_string_list;
pushed_string_list = temp;
#endif
shell_input_line = s;
- shell_input_line_size = strlen (s);
+ shell_input_line_size = STRLEN (s);
shell_input_line_index = 0;
shell_input_line_terminator = '\0';
#if 0
#endif
}
-/* Return a line of text, taken from wherever yylex () reads input.
- If there is no more input, then we return NULL. If REMOVE_QUOTED_NEWLINE
- is non-zero, we remove unquoted \<newline> pairs. This is used by
- read_secondary_line to read here documents. */
-static char *
-read_a_line (remove_quoted_newline)
- int remove_quoted_newline;
+int
+parser_expanding_alias ()
{
- static char *line_buffer = (char *)NULL;
- static int buffer_size = 0;
- int indx = 0, c, peekc, pass_next;
+ return (expanding_alias ());
+}
-#if defined (READLINE)
- if (no_line_editing && SHOULD_PROMPT ())
+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
- if (SHOULD_PROMPT ())
+ ;
#endif
- print_prompt ();
-
- pass_next = 0;
- while (1)
- {
- /* Allow immediate exit if interrupted during input. */
- QUIT;
-
- c = yy_getc ();
+}
+
+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
+ is non-zero, we remove unquoted \<newline> pairs. This is used by
+ read_secondary_line to read here documents. */
+static char *
+read_a_line (remove_quoted_newline)
+ int remove_quoted_newline;
+{
+ static char *line_buffer = (char *)NULL;
+ static int buffer_size = 0;
+ int indx, c, peekc, pass_next;
+
+#if defined (READLINE)
+ if (no_line_editing && SHOULD_PROMPT ())
+#else
+ if (SHOULD_PROMPT ())
+#endif
+ print_prompt ();
+
+ pass_next = indx = 0;
+ while (1)
+ {
+ /* Allow immediate exit if interrupted during input. */
+ QUIT;
+
+ c = yy_getc ();
/* Ignore null bytes in input. */
if (c == 0)
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
}
else if (c == '\\' && remove_quoted_newline)
{
+ QUIT;
peekc = yy_getc ();
if (peekc == '\n')
{
}
}
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;
if (SHOULD_PROMPT())
prompt_again ();
- return (read_a_line (remove_quoted_newline));
+ 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;
}
/* **************************************************************** */
#if defined (COND_COMMAND)
{ "[[", 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 },
{ "<&", 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 */
{ ">", '>' },
int remove_quoted_newline;
{
register int i;
- int c;
+ int c, truncating, last_was_backslash;
unsigned char uc;
- static int mustpop = 0;
QUIT;
+ last_was_backslash = 0;
if (sigwinch_received)
{
sigwinch_received = 0;
{
line_number++;
+ /* Let's not let one really really long line blow up memory allocation */
+ if (shell_input_line && shell_input_line_size >= 32768)
+ {
+ free (shell_input_line);
+ shell_input_line = 0;
+ shell_input_line_size = 0;
+ }
+
restart_read:
/* Allow immediate exit if interrupted during input. */
QUIT;
- i = 0;
+ i = truncating = 0;
shell_input_line_terminator = 0;
/* If the shell is interatctive, but not currently printing a prompt
#if 0
internal_warning ("shell_getc: ignored null byte in input");
#endif
+ /* 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;
}
- RESIZE_MALLOCED_BUFFER (shell_input_line, i, 2, shell_input_line_size, 256);
+ /* 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;
{
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
else
{
char *hdcs;
- hdcs = history_delimiting_chars ();
+ hdcs = history_delimiting_chars (shell_input_line);
if (hdcs && hdcs[0] == ';')
maybe_add_history (shell_input_line);
}
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
not already end in an EOF character. */
if (shell_input_line_terminator != EOF)
{
- if (shell_input_line_len + 3 > shell_input_line_size)
+ 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';
set_line_mbstate ();
}
}
+next_alias_char:
uc = shell_input_line[shell_input_line_index];
if (uc)
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 (!uc && (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' &&
+ shellmeta (shell_input_line[shell_input_line_index-1]) == 0 &&
+ (current_delimiter (dstack) != '\'' && current_delimiter (dstack) != '"'))
+ {
+ parser_state |= PST_ENDALIAS;
+ return ' '; /* END_ALIAS */
+ }
+#endif
+
+pop_alias:
+ /* This case works for PSH_DPAREN as well */
+ 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)
if (SHOULD_PROMPT ())
prompt_again ();
line_number++;
- goto restart_read;
+ /* 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 (!uc && shell_input_line_terminator == EOF)
+ if (uc == 0 && shell_input_line_terminator == EOF)
return ((shell_input_line_index != 0) ? '\n' : EOF);
+#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);
}
eol_ungetc_lookahead = c;
}
+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
}
void
-execute_prompt_command (command)
- char *command;
+execute_variable_command (command, vname)
+ char *command, *vname;
{
char *last_lastarg;
sh_parser_state_t ps;
if (last_lastarg)
last_lastarg = savestring (last_lastarg);
- parse_and_execute (savestring (command), "PROMPT_COMMAND", SEVAL_NONINT|SEVAL_NOHIST);
+ parse_and_execute (savestring (command), vname, SEVAL_NONINT|SEVAL_NOHIST);
restore_parser_state (&ps);
bind_variable ("_", last_lastarg, 0);
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;
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 ();
token_before_that = last_read_token;
last_read_token = current_token;
current_token = read_token (READ);
+
+ if ((parser_state & PST_EOFTOKEN) && current_token == shell_eof_token)
+ {
+ current_token = yacc_EOF;
+ if (bash_input.type == st_string)
+ rewind_input_string ();
+ }
+ parser_state &= ~PST_EOFTOKEN; /* ??? */
+
return (current_token);
}
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)))
+/* 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))
break; \
if (word_token_alist[i].token == TIME && time_command_acceptable () == 0) \
break; \
+ if ((parser_state & PST_CASEPAT) && last_read_token == '|' && word_token_alist[i].token == ESAC) \
+ break; /* 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) \
l = strlen (s);
r = xmalloc (l + 2);
strcpy (r, s);
- if (r[l -1] != ' ')
+#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;
}
if (ap && (ap->flags & AL_BEINGEXPANDED))
return (NO_EXPANSION);
+#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. 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. */
+ 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)
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 ')': /* 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;
`}' is recognized if there is an unclosed `{' present.
`-p' is returned as TIMEOPT if the last read token was TIME.
+ `--' is returned as TIMEIGN if the last read token was TIMEOPT.
']]' is returned as COND_END if the parser is currently parsing
a conditional expression ((parser_state & PST_CONDEXPR) != 0)
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)) &&
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) &&
(token_before_that == FOR) &&
#endif
(tokstr[0] == 'd' && tokstr[1] == 'o' && tokstr[2] == '\0'))
- return (DO);
+ {
+ 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 (tokstr, "esac"))
+ if (last_read_token == IN && STREQ (tokstr, "esac"))
{
+ esacs_needed_count--;
parser_state &= ~PST_CASEPAT;
return (ESAC);
}
/* Handle -p after `time'. */
if (last_read_token == TIME && tokstr[0] == '-' && tokstr[1] == 'p' && !tokstr[2])
return (TIMEOPT);
-#endif
-
-#if 0
-#if defined (COMMAND_TIMING)
- if (STREQ (token, "time") && ((parser_state & PST_CASEPAT) == 0) && time_command_acceptable ())
- return (TIME);
-#endif /* COMMAND_TIMING */
+ /* Handle -- after `time -p'. */
+ if (last_read_token == TIMEOPT && tokstr[0] == '-' && tokstr[1] == '-' && !tokstr[2])
+ return (TIMEIGN);
#endif
#if defined (COND_COMMAND) /* [[ */
dstack.delimiter_depth = 0; /* No delimiters found so far. */
open_brace_count = 0;
+#if defined (EXTENDED_GLOB)
+ /* Reset to global value of extended glob */
+ if (parser_state & PST_EXTPAT)
+ extended_glob = global_extglob;
+#endif
+
parser_state = 0;
+ 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;
+
+ 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
#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 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)
+ {
+#if defined (DEBUG)
+itrace("shell_getc: bash_input.location.string = `%s'", bash_input.location.string);
+#endif
+ 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. */
}
return (character);
}
+ if (parser_state & PST_REGEXP)
+ goto tokword;
+
/* Shell meta-characters. */
if MBTEST(shellmeta (character) && ((parser_state & PST_DBLPAREN) == 0))
{
parser_state &= ~PST_ASSIGNOK;
- peek_char = shell_getc (1);
+ /* 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 (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 (peek_char == '<')
+ else if MBTEST(peek_char == '<')
return (LESS_LESS_LESS);
else
{
parser_state &= ~PST_ALEXPNEXT;
#endif /* ALIAS */
- return (SEMI_SEMI);
+ peek_char = shell_getc (1);
+ if MBTEST(peek_char == '&')
+ return (SEMI_SEMI_AND);
+ else
+ {
+ shell_ungetc (peek_char);
+ return (SEMI_SEMI);
+ }
case '&':
return (AND_AND);
return (LESS_GREATER);
else if MBTEST(character == '>' && peek_char == '|')
return (GREATER_BAR);
- else if MBTEST(peek_char == '>' && character == '&')
- return (AND_GREATER);
+ else if MBTEST(character == '&' && peek_char == '>')
+ {
+ peek_char = shell_getc (1);
+ if MBTEST(peek_char == '>')
+ return (AND_GREATER_GREATER);
+ else
+ {
+ shell_ungetc (peek_char);
+ return (AND_GREATER);
+ }
+ }
+ else if MBTEST(character == '|' && peek_char == '&')
+ return (BAR_AND);
+ else if MBTEST(character == ';' && peek_char == '&')
+ {
+ parser_state |= PST_CASEPAT;
+#if defined (ALIAS)
+ parser_state &= ~PST_ALEXPNEXT;
+#endif /* ALIAS */
+ return (SEMI_AND);
+ }
shell_ungetc (peek_char);
if MBTEST(character == '-' && (last_read_token == LESS_AND || last_read_token == GREATER_AND))
return (character);
+tokword:
/* Okay, if we got this far, we have to read a word. Read one,
and then check it against the known ones. */
result = read_token_word (character);
* reprompting the user, if necessary, after reading a newline, and returning
* correct error values if it reads EOF.
*/
-#define P_FIRSTCLOSE 0x01
-#define P_ALLOWESC 0x02
-#define P_DQUOTE 0x04
-#define P_COMMAND 0x08 /* parsing a command, so look for comments */
+#define P_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 COMSUB_META(ch) ((ch) == ';' || (ch) == '&' || (ch) == '|')
+
+#define CHECK_NESTRET_ERROR() \
+ do { \
+ if (nestret == &matched_pair_error) \
+ { \
+ free (ret); \
+ return &matched_pair_error; \
+ } \
+ } while (0)
+
+#define APPEND_NESTRET() \
+ do { \
+ if (nestlen) \
+ { \
+ RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64); \
+ strcpy (ret + retind, nestret); \
+ retind += nestlen; \
+ } \
+ } while (0)
static char matched_pair_error;
+
static char *
parse_matched_pair (qc, open, close, lenp, flags)
int qc; /* `"' if this construct is within double quotes */
int open, close;
int *lenp, flags;
{
- int count, ch, was_dollar, in_comment, check_comment;
- int pass_next_character, nestlen, ttranslen, start_lineno;
+ 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 = in_comment = 0;
- check_comment = (flags & P_COMMAND) && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0;
+ tflags = 0;
+
+ if ((flags & P_COMMAND) && qc != '`' && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0)
+ tflags |= LEX_CKCOMMENT;
/* RFLAGS is the set of flags we want to pass to recursive calls. */
rflags = (qc == '"') ? P_DQUOTE : (flags & P_DQUOTE);
retind = 0;
start_lineno = line_number;
+ ch = EOF; /* just in case */
while (count)
{
-#if 0
- ch = shell_getc ((qc != '\'' || (flags & P_ALLOWESC)) && pass_next_character == 0);
-#else
- ch = shell_getc (qc != '\'' && pass_next_character == 0);
-#endif
+ prevch = ch;
+ ch = shell_getc (qc != '\'' && (tflags & (LEX_PASSNEXT)) == 0);
+
if (ch == EOF)
{
free (ret);
if (ch == '\n' && SHOULD_PROMPT ())
prompt_again ();
- if (in_comment)
+ /* Don't bother counting parens or doing anything else if in a comment
+ or part of a case statement */
+ if (tflags & LEX_INCOMMENT)
{
/* Add this character. */
RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
ret[retind++] = ch;
if (ch == '\n')
- in_comment = 0;
+ tflags &= ~LEX_INCOMMENT;
continue;
}
- /* Not exactly right yet */
- else if (check_comment && in_comment == 0 && ch == '#' && (retind == 0 || ret[retind-1] == '\n' || whitespace (ret[retind -1])))
- in_comment = 1;
- if (pass_next_character) /* last char was backslash */
+ /* 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 */
{
- pass_next_character = 0;
+ tflags &= ~LEX_PASSNEXT;
if (qc != '\'' && ch == '\n') /* double-quoted \<newline> disappears. */
{
- if (retind > 0) retind--; /* swallow previously-added backslash */
+ if (retind > 0)
+ retind--; /* swallow previously-added backslash */
continue;
}
RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
- if MBTEST(ch == CTLESC || ch == CTLNUL)
+ if MBTEST(ch == CTLESC)
ret[retind++] = CTLESC;
ret[retind++] = ch;
continue;
}
+ /* If we're reparsing the input (e.g., from parse_string_to_word_list),
+ we've already prepended CTLESC to single-quoted results of $'...'.
+ We may want to do this for other CTLESC-quoted characters in
+ reparse, too. */
+ else if MBTEST((parser_state & PST_REPARSE) && open == '\'' && (ch == CTLESC || ch == CTLNUL))
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ continue;
+ }
else if MBTEST(ch == CTLESC || ch == CTLNUL) /* special shell escapes */
{
RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
}
else if MBTEST(ch == close) /* ending delimiter */
count--;
-#if 1
/* handle nested ${...} specially. */
- else if MBTEST(open != close && was_dollar && open == '{' && ch == open) /* } */
+ else if MBTEST(open != close && (tflags & LEX_WASDOL) && open == '{' && ch == open) /* } */
count++;
-#endif
else if MBTEST(((flags & P_FIRSTCLOSE) == 0) && ch == open) /* nested begin */
count++;
RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
ret[retind++] = ch;
- if (open == '\'') /* '' inside grouping construct */
- {
- if MBTEST((flags & P_ALLOWESC) && ch == '\\')
- pass_next_character++;
- continue;
- }
+ /* If we just read the ending character, don't bother continuing. */
+ if (count == 0)
+ break;
+
+ if (open == '\'') /* '' inside grouping construct */
+ {
+ if MBTEST((flags & P_ALLOWESC) && ch == '\\')
+ tflags |= LEX_PASSNEXT;
+ continue;
+ }
+
+ if MBTEST(ch == '\\') /* backslashes */
+ tflags |= LEX_PASSNEXT;
+
+ /* Based on which dolstate is currently in (param, op, or word),
+ decide what the op is. We're really only concerned if it's % or
+ #, so we can turn on a flag that says whether or not we should
+ treat single quotes as special when inside a double-quoted
+ ${...}. This logic must agree with subst.c:extract_dollar_brace_string
+ since they share the same defines. */
+ /* FLAG POSIX INTERP 221 */
+ if (flags & P_DOLBRACE)
+ {
+ /* ${param%[%]word} */
+ if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '%' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param#[#]word} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '#' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param/[/]pat/rep} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '/' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE2; /* XXX */
+ /* ${param^[^]pat} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == '^' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ /* ${param,[,]pat} */
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && ch == ',' && retind > 1)
+ dolbrace_state = DOLBRACE_QUOTE;
+ else if MBTEST(dolbrace_state == DOLBRACE_PARAM && strchr ("#%^,~:-=?+/", ch) != 0)
+ dolbrace_state = DOLBRACE_OP;
+ else if MBTEST(dolbrace_state == DOLBRACE_OP && strchr ("#%^,~:-=?+/", ch) == 0)
+ dolbrace_state = DOLBRACE_WORD;
+ }
+
+ /* The big hammer. Single quotes aren't special in double quotes. The
+ problem is that Posix used to say the single quotes are semi-special:
+ within a double-quoted ${...} construct "an even number of
+ unescaped double-quotes or single-quotes, if any, shall occur." */
+ /* This was changed in Austin Group Interp 221 */
+ if MBTEST(posixly_correct && shell_compatibility_level > 41 && dolbrace_state != DOLBRACE_QUOTE && dolbrace_state != DOLBRACE_QUOTE2 && (flags & P_DQUOTE) && (flags & P_DOLBRACE) && ch == '\'')
+ continue;
+
+ /* Could also check open == '`' if we want to parse grouping constructs
+ inside old-style command substitution. */
+ if (open != close) /* a grouping construct */
+ {
+ if MBTEST(shellquote (ch))
+ {
+ /* '', ``, or "" inside $(...) or other grouping construct. */
+ push_delimiter (dstack, ch);
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'') /* $'...' inside group */
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
+ else
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
+ pop_delimiter (dstack);
+ CHECK_NESTRET_ERROR ();
+
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
+ {
+ /* Translate $'...' here. */
+ ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
+ 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) && (flags & P_DOLBRACE))
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else if ((rflags & P_DQUOTE) == 0)
+ {
+ nestret = sh_single_quote (ttrans);
+ free (ttrans);
+ nestlen = strlen (nestret);
+ }
+ else
+ {
+ nestret = ttrans;
+ nestlen = ttranslen;
+ }
+ retind -= 2; /* back up before the $' */
+ }
+ else if MBTEST((tflags & LEX_WASDOL) && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
+ {
+ /* Locale expand $"..." here. */
+ ttrans = localeexpand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
+ free (nestret);
+
+ nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = ttranslen + 2;
+ retind -= 2; /* back up before the $" */
+ }
+
+ APPEND_NESTRET ();
+ FREE (nestret);
+ }
+ else if ((flags & (P_ARRAYSUB|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 MBTEST(open == '"' && ch == '`')
+ {
+ nestret = parse_matched_pair (0, '`', '`', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+ else if MBTEST(open != '`' && (tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ /* check for $(), $[], or ${} inside quoted string. */
+ {
+parse_dollar_word:
+ if (open == ch) /* undo previous increment */
+ count--;
+ if (ch == '(') /* ) */
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
+ else if (ch == '{') /* } */
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|P_DOLBRACE|rflags);
+ else if (ch == '[') /* ] */
+ nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
+
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
+ }
+#if defined (PROCESS_SUBSTITUTION)
+ if MBTEST((ch == '<' || ch == '>') && (tflags & LEX_GTLT) == 0)
+ tflags |= LEX_GTLT;
+ else
+ tflags &= ~LEX_GTLT;
+#endif
+ if MBTEST(ch == '$' && (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 (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_INCASE)
+ {
+ f &= ~LEX_INCASE;
+ fprintf (stderr, "LEX_INCASE%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 is messier than I'd like, and
+ reproduces a lot more of the token-reading code than I'd like. */
+static char *
+parse_comsub (qc, open, close, lenp, flags)
+ int qc; /* `"' if this construct is within double quotes */
+ int open, close;
+ int *lenp, flags;
+{
+ int count, ch, peekc, tflags, lex_rwlen, lex_wlen, lex_firstind;
+ int nestlen, ttranslen, start_lineno;
+ char *ret, *nestret, *ttrans, *heredelim;
+ int retind, retsize, rflags, hdlen;
+
+ /* Posix interp 217 says arithmetic expressions have precedence, so
+ assume $(( introduces arithmetic expansion and parse accordingly. */
+ peekc = shell_getc (0);
+ shell_ungetc (peekc);
+ if (peekc == '(')
+ return (parse_matched_pair (qc, open, close, lenp, 0));
+
+/*itrace("parse_comsub: qc = `%c' open = %c close = %c", qc, open, close);*/
+ count = 1;
+ tflags = LEX_RESWDOK;
+
+ if ((flags & P_COMMAND) && qc != '\'' && qc != '"' && (flags & P_DQUOTE) == 0)
+ tflags |= LEX_CKCASE;
+ if ((tflags & LEX_CKCASE) && (interactive == 0 || interactive_comments))
+ tflags |= LEX_CKCOMMENT;
+
+ /* RFLAGS is the set of flags we want to pass to recursive calls. */
+ rflags = (flags & P_DQUOTE);
+
+ ret = (char *)xmalloc (retsize = 64);
+ retind = 0;
+
+ start_lineno = line_number;
+ lex_rwlen = lex_wlen = 0;
+
+ heredelim = 0;
+ lex_firstind = -1;
+
+ while (count)
+ {
+comsub_readchar:
+ ch = shell_getc (qc != '\'' && (tflags & (LEX_INCOMMENT|LEX_PASSNEXT|LEX_QUOTEDDOC)) == 0);
+
+ if (ch == EOF)
+ {
+eof_error:
+ free (ret);
+ FREE (heredelim);
+ parser_error (start_lineno, _("unexpected EOF while looking for matching `%c'"), close);
+ EOF_Reached = 1; /* XXX */
+ return (&matched_pair_error);
+ }
+
+ /* If we hit the end of a line and are reading the contents of a here
+ document, and it's not the same line that the document starts on,
+ check for this line being the here doc delimiter. Otherwise, if
+ we're in a here document, mark the next character as the beginning
+ of a line. */
+ if (ch == '\n')
+ {
+ if ((tflags & LEX_HEREDELIM) && heredelim)
+ {
+ tflags &= ~LEX_HEREDELIM;
+ tflags |= LEX_INHEREDOC;
+ lex_firstind = retind + 1;
+ }
+ else if (tflags & LEX_INHEREDOC)
+ {
+ int tind;
+ tind = lex_firstind;
+ while ((tflags & LEX_STRIPDOC) && ret[tind] == '\t')
+ tind++;
+ if (retind-tind == hdlen && STREQN (ret + tind, heredelim, hdlen))
+ {
+ tflags &= ~(LEX_STRIPDOC|LEX_INHEREDOC|LEX_QUOTEDDOC);
+/*itrace("parse_comsub:%d: found here doc end `%s'", line_number, ret + tind);*/
+ free (heredelim);
+ heredelim = 0;
+ lex_firstind = -1;
+ }
+ else
+ lex_firstind = retind + 1;
+ }
+ }
+
+ /* Possible reprompting. */
+ if (ch == '\n' && SHOULD_PROMPT ())
+ prompt_again ();
+
+ /* XXX -- possibly allow here doc to be delimited by ending right
+ paren. */
+ if ((tflags & LEX_INHEREDOC) && ch == close && count == 1)
+ {
+ int tind;
+/*itrace("parse_comsub:%d: in here doc, ch == close, retind - firstind = %d hdlen = %d retind = %d", line_number, retind-lex_firstind, hdlen, retind);*/
+ tind = lex_firstind;
+ while ((tflags & LEX_STRIPDOC) && ret[tind] == '\t')
+ tind++;
+ if (retind-tind == hdlen && STREQN (ret + tind, heredelim, hdlen))
+ {
+ tflags &= ~(LEX_STRIPDOC|LEX_INHEREDOC|LEX_QUOTEDDOC);
+/*itrace("parse_comsub:%d: found here doc end `%*s'", line_number, hdlen, ret + tind);*/
+ free (heredelim);
+ heredelim = 0;
+ lex_firstind = -1;
+ }
+ }
+
+ /* Don't bother counting parens or doing anything else if in a comment or
+ here document (not exactly right for here-docs -- if we want to allow
+ recursive calls to parse_comsub to have their own here documents,
+ change the LEX_INHEREDOC to LEX_QUOTEDDOC here and uncomment the next
+ clause below. Note that to make this work completely, we need to make
+ additional changes to allow xparse_dolparen to work right when the
+ command substitution is parsed, because read_secondary_line doesn't know
+ to recursively parse through command substitutions embedded in here-
+ documents */
+ if (tflags & (LEX_INCOMMENT|LEX_INHEREDOC))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ if ((tflags & LEX_INCOMMENT) && ch == '\n')
+ {
+/*itrace("parse_comsub:%d: lex_incomment -> 0 ch = `%c'", line_number, ch);*/
+ tflags &= ~LEX_INCOMMENT;
+ }
+
+ continue;
+ }
+#if 0
+ /* If we're going to recursively parse a command substitution inside a
+ here-document, make sure we call parse_comsub recursively below. See
+ above for additional caveats. */
+ if ((tflags & LEX_INHEREDOC) && ((tflags & LEX_WASDOL) == 0 || ch != '(')) /*)*/
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ if MBTEST(ch == '$')
+ tflags |= LEX_WASDOL;
+ else
+ tflags &= ~LEX_WASDOL;
+ }
+#endif
+
+ if (tflags & LEX_PASSNEXT) /* last char was backslash */
+ {
+/*itrace("parse_comsub:%d: lex_passnext -> 0 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ tflags &= ~LEX_PASSNEXT;
+ if (qc != '\'' && ch == '\n') /* double-quoted \<newline> disappears. */
+ {
+ if (retind > 0)
+ retind--; /* swallow previously-added backslash */
+ continue;
+ }
+
+ RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
+ if MBTEST(ch == CTLESC)
+ ret[retind++] = CTLESC;
+ ret[retind++] = ch;
+ continue;
+ }
+
+ /* If this is a shell break character, we are not in a word. If not,
+ we either start or continue a word. */
+ if MBTEST(shellbreak (ch))
+ {
+ tflags &= ~LEX_INWORD;
+/*itrace("parse_comsub:%d: lex_inword -> 0 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ }
+ else
+ {
+ if (tflags & LEX_INWORD)
+ {
+ lex_wlen++;
+/*itrace("parse_comsub:%d: lex_inword == 1 ch = `%c' lex_wlen = %d (%d)", line_number, ch, lex_wlen, __LINE__);*/
+ }
+ else
+ {
+/*itrace("parse_comsub:%d: lex_inword -> 1 ch = `%c' (%d)", line_number, ch, __LINE__);*/
+ tflags |= LEX_INWORD;
+ lex_wlen = 0;
+ if (tflags & LEX_RESWDOK)
+ lex_rwlen = 0;
+ }
+ }
+
+ /* Skip whitespace */
+ if MBTEST(shellblank (ch) && (tflags & LEX_HEREDELIM) == 0 && lex_rwlen == 0)
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ continue;
+ }
+
+ /* Either we are looking for the start of the here-doc delimiter
+ (lex_firstind == -1) or we are reading one (lex_firstind >= 0).
+ If this character is a shell break character and we are reading
+ the delimiter, save it and note that we are now reading a here
+ document. If we've found the start of the delimiter, note it by
+ setting lex_firstind. Backslashes can quote shell metacharacters
+ in here-doc delimiters. */
+ if (tflags & LEX_HEREDELIM)
+ {
+ if (lex_firstind == -1 && shellbreak (ch) == 0)
+ lex_firstind = retind;
+#if 0
+ else if (heredelim && (tflags & LEX_PASSNEXT) == 0 && ch == '\n')
+ {
+ tflags |= LEX_INHEREDOC;
+ tflags &= ~LEX_HEREDELIM;
+ lex_firstind = retind + 1;
+ }
+#endif
+ else if (lex_firstind >= 0 && (tflags & LEX_PASSNEXT) == 0 && shellbreak (ch))
+ {
+ if (heredelim == 0)
+ {
+ nestret = substring (ret, lex_firstind, retind);
+ heredelim = string_quote_removal (nestret, 0);
+ hdlen = STRLEN(heredelim);
+/*itrace("parse_comsub:%d: found here doc delimiter `%s' (%d)", line_number, heredelim, hdlen);*/
+ if (STREQ (heredelim, nestret) == 0)
+ tflags |= LEX_QUOTEDDOC;
+ free (nestret);
+ }
+ if (ch == '\n')
+ {
+ tflags |= LEX_INHEREDOC;
+ tflags &= ~LEX_HEREDELIM;
+ lex_firstind = retind + 1;
+ }
+ else
+ lex_firstind = -1;
+ }
+ }
+
+ /* Meta-characters that can introduce a reserved word. Not perfect yet. */
+ if MBTEST((tflags & LEX_RESWDOK) == 0 && (tflags & LEX_CKCASE) && (tflags & LEX_INCOMMENT) == 0 && (shellmeta(ch) || ch == '\n'))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ peekc = shell_getc (1);
+ if (ch == peekc && (ch == '&' || ch == '|' || ch == ';')) /* two-character tokens */
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+/*itrace("parse_comsub:%d: set lex_reswordok = 1, ch = `%c'", line_number, ch);*/
+ tflags |= LEX_RESWDOK;
+ lex_rwlen = 0;
+ continue;
+ }
+ else if (ch == '\n' || COMSUB_META(ch))
+ {
+ shell_ungetc (peekc);
+/*itrace("parse_comsub:%d: set lex_reswordok = 1, ch = `%c'", line_number, ch);*/
+ tflags |= LEX_RESWDOK;
+ lex_rwlen = 0;
+ continue;
+ }
+ else if (ch == EOF)
+ goto eof_error;
+ else
+ {
+ /* `unget' the character we just added and fall through */
+ retind--;
+ shell_ungetc (peekc);
+ }
+ }
+
+ /* If we can read a reserved word, try to read one. */
+ if (tflags & LEX_RESWDOK)
+ {
+ if MBTEST(islower ((unsigned char)ch))
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ lex_rwlen++;
+ continue;
+ }
+ else if MBTEST(lex_rwlen == 4 && shellbreak (ch))
+ {
+ if (STREQN (ret + retind - 4, "case", 4))
+ {
+ tflags |= LEX_INCASE;
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `case', lex_incase -> 1 lex_reswdok -> 0", line_number);*/
+ }
+ else if (STREQN (ret + retind - 4, "esac", 4))
+ {
+ tflags &= ~LEX_INCASE;
+/*itrace("parse_comsub:%d: found `esac', lex_incase -> 0 lex_reswdok -> 1", line_number);*/
+ tflags |= LEX_RESWDOK;
+ lex_rwlen = 0;
+ }
+ else if (STREQN (ret + retind - 4, "done", 4) ||
+ STREQN (ret + retind - 4, "then", 4) ||
+ STREQN (ret + retind - 4, "else", 4) ||
+ STREQN (ret + retind - 4, "elif", 4) ||
+ STREQN (ret + retind - 4, "time", 4))
+ {
+ /* these are four-character reserved words that can be
+ followed by a reserved word; anything else turns off
+ the reserved-word-ok flag */
+/*itrace("parse_comsub:%d: found `%.4s', lex_reswdok -> 1", line_number, ret+retind-4);*/
+ tflags |= LEX_RESWDOK;
+ lex_rwlen = 0;
+ }
+ else if (shellmeta (ch) == 0)
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%.4s', lex_reswdok -> 0", line_number, ret+retind-4);*/
+ }
+ else /* can't be in a reserved word any more */
+ lex_rwlen = 0;
+ }
+ else if MBTEST((tflags & LEX_CKCOMMENT) && ch == '#' && (lex_rwlen == 0 || ((tflags & LEX_INWORD) && lex_wlen == 0)))
+ ; /* don't modify LEX_RESWDOK if we're starting a comment */
+ /* Allow `do' followed by space, tab, or newline to preserve the
+ RESWDOK flag, but reset the reserved word length counter so we
+ can read another one. */
+ else if MBTEST(((tflags & LEX_INCASE) == 0) &&
+ (isblank((unsigned char)ch) || ch == '\n') &&
+ lex_rwlen == 2 &&
+ STREQN (ret + retind - 2, "do", 2))
+ {
+/*itrace("parse_comsub:%d: lex_incase == 0 found `%c', found \"do\"", line_number, ch);*/
+ lex_rwlen = 0;
+ }
+ else if MBTEST((tflags & LEX_INCASE) && ch != '\n')
+ /* If we can read a reserved word and we're in case, we're at the
+ point where we can read a new pattern list or an esac. We
+ handle the esac case above. If we read a newline, we want to
+ leave LEX_RESWDOK alone. If we read anything else, we want to
+ turn off LEX_RESWDOK, since we're going to read a pattern list. */
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: lex_incase == 1 found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+ else if MBTEST(shellbreak (ch) == 0)
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+#if 0
+ /* If we find a space or tab but have read something and it's not
+ `do', turn off the reserved-word-ok flag */
+ else if MBTEST(isblank ((unsigned char)ch) && lex_rwlen > 0)
+ {
+ tflags &= ~LEX_RESWDOK;
+/*itrace("parse_comsub:%d: found `%c', lex_reswordok -> 0", line_number, ch);*/
+ }
+#endif
+ }
+
+ /* Might be the start of a here-doc delimiter */
+ if MBTEST((tflags & LEX_INCOMMENT) == 0 && (tflags & LEX_CKCASE) && ch == '<')
+ {
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+ peekc = shell_getc (1);
+ if (peekc == EOF)
+ goto eof_error;
+ if (peekc == ch)
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+ peekc = shell_getc (1);
+ if (peekc == EOF)
+ goto eof_error;
+ if (peekc == '-')
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = peekc;
+ tflags |= LEX_STRIPDOC;
+ }
+ else
+ shell_ungetc (peekc);
+ if (peekc != '<')
+ {
+ tflags |= LEX_HEREDELIM;
+ lex_firstind = -1;
+ }
+ continue;
+ }
+ else
+ {
+ shell_ungetc (peekc); /* not a here-doc, start over */
+ continue;
+ }
+ }
+ else if MBTEST((tflags & LEX_CKCOMMENT) && (tflags & LEX_INCOMMENT) == 0 && ch == '#' && (((tflags & LEX_RESWDOK) && lex_rwlen == 0) || ((tflags & LEX_INWORD) && lex_wlen == 0)))
+ {
+/*itrace("parse_comsub:%d: lex_incomment -> 1 (%d)", line_number, __LINE__);*/
+ tflags |= LEX_INCOMMENT;
+ }
+
+ if MBTEST(ch == CTLESC || ch == CTLNUL) /* special shell escapes */
+ {
+ RESIZE_MALLOCED_BUFFER (ret, retind, 2, retsize, 64);
+ ret[retind++] = CTLESC;
+ ret[retind++] = ch;
+ continue;
+ }
+#if 0
+ else if MBTEST((tflags & LEX_INCASE) && ch == close && close == ')')
+ tflags &= ~LEX_INCASE; /* XXX */
+#endif
+ else if MBTEST(ch == close && (tflags & LEX_INCASE) == 0) /* ending delimiter */
+ {
+ count--;
+/*itrace("parse_comsub:%d: found close: count = %d", line_number, count);*/
+ }
+ else if MBTEST(((flags & P_FIRSTCLOSE) == 0) && (tflags & LEX_INCASE) == 0 && ch == open) /* nested begin */
+ {
+ count++;
+/*itrace("parse_comsub:%d: found open: count = %d", line_number, count);*/
+ }
+
+ /* Add this character. */
+ RESIZE_MALLOCED_BUFFER (ret, retind, 1, retsize, 64);
+ ret[retind++] = ch;
+
+ /* If we just read the ending character, don't bother continuing. */
+ if (count == 0)
+ break;
if MBTEST(ch == '\\') /* backslashes */
- pass_next_character++;
+ tflags |= LEX_PASSNEXT;
- if (open != close) /* a grouping construct */
- {
- if MBTEST(shellquote (ch))
+ if MBTEST(shellquote (ch))
+ {
+ /* '', ``, or "" inside $(...). */
+ push_delimiter (dstack, ch);
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'') /* $'...' inside group */
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
+ else
+ nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
+ pop_delimiter (dstack);
+ CHECK_NESTRET_ERROR ();
+
+ if MBTEST((tflags & LEX_WASDOL) && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
- /* '', ``, or "" inside $(...) or other grouping construct. */
- push_delimiter (dstack, ch);
- if MBTEST(was_dollar && ch == '\'') /* $'...' inside group */
- nestret = parse_matched_pair (ch, ch, ch, &nestlen, P_ALLOWESC|rflags);
- else
- nestret = parse_matched_pair (ch, ch, ch, &nestlen, rflags);
- pop_delimiter (dstack);
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
- }
- if MBTEST(was_dollar && ch == '\'' && (extended_quote || (rflags & P_DQUOTE) == 0))
- {
- /* Translate $'...' here. */
- ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
- xfree (nestret);
+ /* Translate $'...' here. */
+ ttrans = ansiexpand (nestret, 0, nestlen - 1, &ttranslen);
+ free (nestret);
- if ((rflags & P_DQUOTE) == 0)
- {
- nestret = sh_single_quote (ttrans);
- free (ttrans);
- nestlen = strlen (nestret);
- }
- else
- {
- nestret = ttrans;
- nestlen = ttranslen;
- }
- retind -= 2; /* back up before the $' */
- }
- else if MBTEST(was_dollar && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
+ if ((rflags & P_DQUOTE) == 0)
{
- /* Locale expand $"..." here. */
- ttrans = localeexpand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
- xfree (nestret);
-
- nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ nestret = sh_single_quote (ttrans);
free (ttrans);
- nestlen = ttranslen + 2;
- retind -= 2; /* back up before the $" */
+ nestlen = strlen (nestret);
}
-
- if (nestlen)
+ else
{
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
+ nestret = ttrans;
+ nestlen = ttranslen;
}
- FREE (nestret);
- }
- }
- /* Parse an old-style command substitution within double quotes as a
- single word. */
- /* XXX - sh and ksh93 don't do this - XXX */
- else if MBTEST(open == '"' && ch == '`')
- {
- nestret = parse_matched_pair (0, '`', '`', &nestlen, rflags);
-add_nestret:
- if (nestret == &matched_pair_error)
- {
- free (ret);
- return &matched_pair_error;
+ retind -= 2; /* back up before the $' */
}
- if (nestlen)
+ else if MBTEST((tflags & LEX_WASDOL) && ch == '"' && (extended_quote || (rflags & P_DQUOTE) == 0))
{
- RESIZE_MALLOCED_BUFFER (ret, retind, nestlen, retsize, 64);
- strcpy (ret + retind, nestret);
- retind += nestlen;
+ /* Locale expand $"..." here. */
+ ttrans = localeexpand (nestret, 0, nestlen - 1, start_lineno, &ttranslen);
+ free (nestret);
+
+ nestret = sh_mkdoublequoted (ttrans, ttranslen, 0);
+ free (ttrans);
+ nestlen = ttranslen + 2;
+ retind -= 2; /* back up before the $" */
}
+
+ APPEND_NESTRET ();
FREE (nestret);
}
- else if MBTEST(qc == '`' && (ch == '"' || ch == '\'') && in_comment == 0)
+ else if MBTEST((tflags & LEX_WASDOL) && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
+ /* check for $(), $[], or ${} inside command substitution. */
{
- nestret = parse_matched_pair (0, ch, ch, &nestlen, rflags);
- goto add_nestret;
- }
- else if MBTEST(was_dollar && (ch == '(' || ch == '{' || ch == '[')) /* ) } ] */
- /* check for $(), $[], or ${} inside quoted string. */
- {
- if (open == ch) /* undo previous increment */
+ if ((tflags & LEX_INCASE) == 0 && open == ch) /* undo previous increment */
count--;
if (ch == '(') /* ) */
- nestret = parse_matched_pair (0, '(', ')', &nestlen, rflags);
+ nestret = parse_comsub (0, '(', ')', &nestlen, (rflags|P_COMMAND) & ~P_DQUOTE);
else if (ch == '{') /* } */
- nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|rflags);
+ nestret = parse_matched_pair (0, '{', '}', &nestlen, P_FIRSTCLOSE|P_DOLBRACE|rflags);
else if (ch == '[') /* ] */
nestret = parse_matched_pair (0, '[', ']', &nestlen, rflags);
- goto add_nestret;
+ CHECK_NESTRET_ERROR ();
+ APPEND_NESTRET ();
+
+ FREE (nestret);
}
- was_dollar = MBTEST(ch == '$');
+ if MBTEST(ch == '$' && (tflags & LEX_WASDOL) == 0)
+ tflags |= LEX_WASDOL;
+ else
+ tflags &= ~LEX_WASDOL;
}
+ FREE (heredelim);
ret[retind] = '\0';
if (lenp)
*lenp = retind;
+/*itrace("parse_comsub:%d: returning `%s'", line_number, ret);*/
+ return ret;
+}
+
+/* Recursively call the parser to parse a $(...) command substitution. */
+char *
+xparse_dolparen (base, string, indp, flags)
+ char *base;
+ char *string;
+ int *indp;
+ int flags;
+{
+ sh_parser_state_t ps;
+ sh_input_line_state_t ls;
+ int orig_ind, nc, sflags, orig_eof_token;
+ char *ret, *ep, *ostring;
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ STRING_SAVER *saved_pushed_strings;
+#endif
+
+/*debug_parser(1);*/
+ orig_ind = *indp;
+ ostring = string;
+
+ 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'", shell_input_line_size, shell_input_line);*/
+ sflags = SEVAL_NONINT|SEVAL_NOHIST|SEVAL_NOFREE;
+ if (flags & SX_NOLONGJMP)
+ sflags |= SEVAL_NOLONGJMP;
+ save_parser_state (&ps);
+ save_input_line_state (&ls);
+ orig_eof_token = shell_eof_token;
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ saved_pushed_strings = pushed_string_list; /* separate parsing context */
+ pushed_string_list = (STRING_SAVER *)NULL;
+#endif
+
+ /*(*/
+ parser_state |= PST_CMDSUBST|PST_EOFTOKEN; /* allow instant ')' */ /*(*/
+ shell_eof_token = ')';
+
+ /* Should we save and restore the bison/yacc lookahead token (yychar) here?
+ Or only if it's not YYEMPTY? */
+
+ nc = parse_string (string, "command substitution", sflags, &ep);
+
+ if (current_token == shell_eof_token)
+ yyclearin; /* might want to clear lookahead token unconditionally */
+
+ shell_eof_token = orig_eof_token;
+ restore_parser_state (&ps);
+ reset_parser ();
+ /* reset_parser clears shell_input_line and associated variables */
+ restore_input_line_state (&ls);
+
+#if defined (ALIAS) || defined (DPAREN_ARITHMETIC)
+ pushed_string_list = saved_pushed_strings;
+#endif
+
+ 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 */
+ jump_to_top_level (-nc); /* XXX */
+ }
+
+ /* Need to find how many characters parse_and_execute consumed, update
+ *indp, if flags != 0, copy the portion of the string parsed into RET
+ and return it. If flags & 1 (SX_NOALLOC) we can return NULL. */
+
+ /*(*/
+ if (ep[-1] != ')')
+ {
+#if DEBUG
+ if (ep[-1] != '\n')
+ itrace("xparse_dolparen:%d: ep[-1] != RPAREN (%d), ep = `%s'", line_number, ep[-1], ep);
+#endif
+ while (ep > ostring && ep[-1] == '\n') ep--;
+ }
+
+ nc = ep - ostring;
+ *indp = ep - base - 1;
+
+ /*(*/
+#if DEBUG
+ if (base[*indp] != ')')
+ itrace("xparse_dolparen:%d: base[%d] != RPAREN (%d), base = `%s'", line_number, *indp, base[*indp], base);
+ if (*indp < orig_ind)
+ itrace("xparse_dolparen:%d: *indp (%d) < orig_ind (%d), orig_string = `%s'", line_number, *indp, orig_ind, ostring);
+#endif
+
+ if (flags & SX_NOALLOC)
+ return (char *)NULL;
+
+ if (nc == 0)
+ {
+ ret = xmalloc (1);
+ ret[0] = '\0';
+ }
+ else
+ ret = substring (ostring, 0, nc - 1);
+
return ret;
}
parse_dparen (c)
int c;
{
- int cmdtyp, len, sline;
- char *wval, *wv2;
+ int cmdtyp, sline;
+ char *wval;
WORD_DESC *wd;
#if defined (ARITH_FOR_COMMAND)
{
wd = alloc_word_desc ();
wd->word = wval;
- wd = make_word (wval);
yylval.word_list = make_word_list (wd, (WORD_LIST *)NULL);
return (ARITH_FOR_EXPRS);
}
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);
if (term)
term->flags |= CMD_INVERT_RETURN;
}
- else if (tok == WORD && test_unop (yylval.word->word))
+ else if (tok == WORD && yylval.word->word[0] == '-' && yylval.word->word[1] && yylval.word->word[2] == 0 && test_unop (yylval.word->word))
{
op = yylval.word;
tok = read_token (READ);
/* binop */
tok = read_token (READ);
if (tok == WORD && test_binop (yylval.word->word))
- op = yylval.word;
+ {
+ op = yylval.word;
+ if (op->word[0] == '=' && (op->word[1] == '\0' || (op->word[1] == '=' && op->word[2] == '\0')))
+ parser_state |= PST_EXTPAT;
+ else if (op->word[0] == '!' && op->word[1] == '=' && op->word[2] == '\0')
+ parser_state |= PST_EXTPAT;
+ }
#if defined (COND_REGEXP)
- else if (tok == WORD && STREQ (yylval.word->word,"=~"))
- op = yylval.word;
+ else if (tok == WORD && STREQ (yylval.word->word, "=~"))
+ {
+ op = yylval.word;
+ parser_state |= PST_REGEXP;
+ }
#endif
else if (tok == '<' || tok == '>')
op = make_word_from_token (tok); /* ( */
}
/* rhs */
+ if (parser_state & PST_EXTPAT)
+ extended_glob = 1;
tok = read_token (READ);
+ if (parser_state & PST_EXTPAT)
+ extended_glob = global_extglob;
+ parser_state &= ~(PST_REGEXP|PST_EXTPAT);
+
if (tok == WORD)
{
tright = make_cond_node (COND_TERM, yylval.word, (COND_COM *)NULL, (COND_COM *)NULL);
{
COND_COM *cexp;
+ global_extglob = extended_glob;
cexp = cond_expr ();
return (make_cond_command (cexp));
}
#if defined (ARRAY_VARS)
/* When this is called, it's guaranteed that we don't care about anything
- in t beyond i. We do save and restore the chars, though. */
+ 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;
{
- unsigned char c, c1;
int r;
+ char *atoken;
- c = t[i]; c1 = t[i+1];
- t[i] = '='; t[i+1] = '\0';
- r = assignment (t, (parser_state & PST_COMPASSIGN) != 0);
- t[i] = c; t[i+1] = c1;
- return r;
+ 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 `+=' */
/* 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;
if (pass_next_character)
{
pass_next_character = 0;
- goto got_character;
+ goto got_escaped_character;
}
cd = current_delimiter (dstack);
strcpy (token + token_index, ttok);
token_index += ttoklen;
all_digit_token = 0;
- quoted = 1;
+ if (character != '`')
+ quoted = 1;
dollar_present |= (character == '"' && strchr (ttok, '$') != 0);
FREE (ttok);
goto next_character;
}
+#ifdef COND_REGEXP
+ /* When parsing a regexp as a single word inside a conditional command,
+ we need to special-case characters special to both the shell and
+ regular expressions. Right now, that is only '(' and '|'. */ /*)*/
+ if MBTEST((parser_state & PST_REGEXP) && (character == '(' || character == '|')) /*)*/
+ {
+ if (character == '|')
+ goto got_character;
+
+ push_delimiter (dstack, character);
+ ttok = parse_matched_pair (cd, '(', ')', &ttoklen, 0);
+ pop_delimiter (dstack);
+ if (ttok == &matched_pair_error)
+ return -1; /* Bail immediately. */
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttoklen + 2,
+ token_buffer_size, TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = character;
+ strcpy (token + token_index, ttok);
+ token_index += ttoklen;
+ FREE (ttok);
+ dollar_present = all_digit_token = 0;
+ goto next_character;
+ }
+#endif /* COND_REGEXP */
+
#ifdef EXTENDED_GLOB
/* Parse a ksh-style extended pattern matching specification. */
- if (extended_glob && PATTERN_CHAR (character))
+ if MBTEST(extended_glob && PATTERN_CHAR (character))
{
peek_char = shell_getc (1);
if MBTEST(peek_char == '(') /* ) */
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;
{
peek_char = shell_getc (1);
/* $(...), <(...), >(...), $((...)), ${...}, and $[...] constructs */
- if MBTEST(peek_char == '(' || \
+ if MBTEST(peek_char == '(' ||
((peek_char == '{' || peek_char == '[') && character == '$')) /* ) ] } */
{
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
history literally rather than causing a possibly-
incorrect `;' to be added. ) */
push_delimiter (dstack, peek_char);
- ttok = parse_matched_pair (cd, '(', ')', &ttoklen, P_COMMAND);
+ 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;
}
else
{
- /* Try to locale)-expand the converted string. */
+ /* Try to locale-expand the converted string. */
ttrans = localeexpand (ttok, 0, ttoklen - 1, first_line, &ttranslen);
free (ttok);
ttrans = ttok;
}
- RESIZE_MALLOCED_BUFFER (token, token_index, ttranslen + 2,
+ RESIZE_MALLOCED_BUFFER (token, token_index, ttranslen + 1,
token_buffer_size,
TOKEN_DEFAULT_GROW_SIZE);
strcpy (token + token_index, ttrans);
shell's single-character parameter expansions, and set flags.*/
else if MBTEST(character == '$' && peek_char == '$')
{
- ttok = (char *)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_digit_token = 0;
- FREE (ttok);
goto next_character;
}
else
}
#if defined (ARRAY_VARS)
- /* Identify possible array subscript assignment; match [...] */
- else if MBTEST(character == '[' && token_index > 0 && assignment_acceptable (last_read_token) && token_is_ident (token, token_index)) /* ] */
+ /* Identify possible array subscript assignment; match [...]. If
+ parser_state&PST_COMPASSIGN, we need to parse [sub]=words treating
+ `sub' as if it were enclosed in double quotes. */
+ else if MBTEST(character == '[' && /* ] */
+ ((token_index > 0 && assignment_acceptable (last_read_token) && token_is_ident (token, token_index)) ||
+ (token_index == 0 && (parser_state&PST_COMPASSIGN))))
{
- ttok = parse_matched_pair (cd, '[', ']', &ttoklen, 0);
+ 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,
FREE (ttok);
all_digit_token = 0;
compound_assignment = 1;
-#if 0
+#if 1
goto next_character;
#else
goto got_token; /* ksh93 seems to do this */
goto got_token;
}
- got_character:
-
- all_digit_token &= DIGIT (character);
- dollar_present |= character == '$';
-
+got_character:
if (character == CTLESC || character == CTLNUL)
- token[token_index++] = CTLESC;
+ {
+ RESIZE_MALLOCED_BUFFER (token, token_index, 2, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
+ token[token_index++] = CTLESC;
+ }
+ else
+got_escaped_character:
+ RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
+ TOKEN_DEFAULT_GROW_SIZE);
token[token_index++] = character;
- RESIZE_MALLOCED_BUFFER (token, token_index, 1, token_buffer_size,
- TOKEN_DEFAULT_GROW_SIZE);
+ all_digit_token &= DIGIT (character);
+ dollar_present |= character == '$';
next_character:
if (character == '\n' && SHOULD_PROMPT ())
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 MBTEST(all_digit_token && (character == '<' || character == '>' || \
- last_read_token == LESS_AND || \
+ if MBTEST(all_digit_token && (character == '<' || character == '>' ||
+ last_read_token == LESS_AND ||
last_read_token == GREATER_AND))
{
if (legal_number (token, &lvalue) && (int)lvalue == lvalue)
- yylval.number = lvalue;
- else
- yylval.number = -1;
- return (NUMBER);
+ {
+ yylval.number = lvalue;
+ return (NUMBER);
+ }
}
/* Check for special case tokens. */
#endif
CHECK_FOR_RESERVED_WORD (token);
- the_word = (WORD_DESC *)xmalloc (sizeof (WORD_DESC));
+ 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;
- if (compound_assignment)
+ 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
the_word->flags |= W_ASSIGNMENT;
/* Don't perform word splitting on assignment statements. */
if (assignment_acceptable (last_read_token) || (parser_state & PST_COMPASSIGN) != 0)
- the_word->flags |= W_NOSPLIT;
+ {
+ 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;
+ 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 (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 (word_top < MAX_CASE_NEST)
word_top++;
word_lineno[word_top] = line_number;
+ expecting_in_token++;
break;
}
case '}': /* XXX */
case AND_AND:
case BANG:
+ case BAR_AND:
case DO:
case DONE:
case ELIF:
case IF:
case OR_OR:
case SEMI_SEMI:
+ case SEMI_AND:
+ case SEMI_SEMI_AND:
case THEN:
case TIME:
case TIMEOPT:
+ case TIMEIGN:
+ case COPROC:
case UNTIL:
case WHILE:
case 0:
return 1;
default:
+#if defined (COPROCESS_SUPPORT)
+ if (last_read_token == WORD && token_before_that == COPROC)
+ return 1;
+#endif
+ if (last_read_token == WORD && token_before_that == FUNCTION)
+ return 1;
return 0;
}
}
return -1;
}
+/* 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
/* 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
command lists. It's a suboptimal solution. */
else if (parser_state & PST_CASESTMT) /* case statement pattern */
return " ";
- else
+ 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++)
+ 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 (" ");
}
+ if (line_isblank (line))
+ return ("");
+
return ("; ");
}
#endif /* HISTORY */
{
char *temp_prompt;
- if (interactive == 0 || expanding_alias()) /* 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;
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:
char *string;
{
WORD_LIST *list;
- char *result, *t;
+ char *result, *t, *orig_string;
struct dstack save_dstack;
- int last_exit_value;
+ 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;
+ int result_index;
+ int c, n, i;
+ char *temp, *t_host, octal_string[4];
struct tm *tm;
time_t the_time;
char timebuf[128];
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;
case 'A':
/* Make the current time/date into a string. */
(void) time (&the_time);
+#if defined (HAVE_TZSET)
+ sv_tz ("TZ"); /* XXX -- just make sure */
+#endif
tm = localtime (&the_time);
if (c == 'd')
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)
+ temp = sh_backslash_quote_for_double_quotes (temp);
+ else
+ temp = savestring (temp);
goto add_string;
case 'v':
case 'W':
{
/* Use the value of PWD because it is much more efficient. */
- char t_string[PATH_MAX], *t;
+ char t_string[PATH_MAX];
int tlen;
temp = get_string_value ("PWD");
}
t_string[tlen] = '\0';
+#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 */
{
t = strrchr (t_string, '/');
if (t)
- strcpy (t_string, t + 1);
+ 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)
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);
+ 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;
break;
}
temp = (char *)xmalloc (3);
- temp[0] = '\001';
- temp[1] = (c == '[') ? RL_PROMPT_START_IGNORE : RL_PROMPT_END_IGNORE;
- temp[2] = '\0';
+ n = (c == '[') ? RL_PROMPT_START_IGNORE : RL_PROMPT_END_IGNORE;
+ i = 0;
+ if (n == CTLESC || n == CTLNUL)
+ temp[i++] = CTLESC;
+ temp[i++] = n;
+ temp[i] = '\0';
goto add_string;
#endif /* READLINE */
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';
}
if (promptvars || posixly_correct)
{
last_exit_value = last_command_exit_value;
- list = expand_prompt_string (result, Q_DOUBLE_QUOTES);
+ 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
{
}
static char *
-error_token_from_token (token)
- int token;
+error_token_from_token (tok)
+ int tok;
{
char *t;
- if (t = find_token_in_alist (token, word_token_alist, 0))
+ if (t = find_token_in_alist (tok, word_token_alist, 0))
return t;
- if (t = find_token_in_alist (token, other_token_alist, 0))
+ if (t = find_token_in_alist (tok, other_token_alist, 0))
return t;
t = (char *)NULL;
report_syntax_error (message)
char *message;
{
- char *msg;
+ char *msg, *p;
if (message)
{
parser_error (line_number, "%s", message);
if (interactive && EOF_Reached)
EOF_Reached = 0;
- last_command_exit_value = EX_USAGE;
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
return;
}
parser's complaining about by looking at current_token. */
if (current_token != 0 && EOF_Reached == 0 && (msg = error_token_from_token (current_token)))
{
+ 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);
if (interactive == 0)
print_offending_line ();
- last_command_exit_value = EX_USAGE;
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
return;
}
EOF_Reached = 0;
}
- last_command_exit_value = EX_USAGE;
+ last_command_exit_value = parse_and_execute_level ? EX_BADSYNTAX : EX_BADUSAGE;
}
/* ??? Needed function. ??? We have to be able to discard the constructs
wl = (WORD_LIST *)NULL;
if (flags & 1)
- parser_state |= PST_COMPASSIGN;
+ parser_state |= PST_COMPASSIGN|PST_REPARSE;
while ((tok = read_token (READ)) != yacc_EOF)
{
shell_input_line_terminator = orig_input_terminator;
if (flags & 1)
- parser_state &= ~PST_COMPASSIGN;
+ parser_state &= ~(PST_COMPASSIGN|PST_REPARSE);
if (wl == &parse_string_error)
{
int *retlenp;
{
WORD_LIST *wl, *rl;
- int tok, orig_line_number, orig_token_size;
+ int tok, orig_line_number, orig_token_size, orig_last_token, assignok;
char *saved_token, *ret;
saved_token = token;
orig_token_size = token_buffer_size;
orig_line_number = line_number;
+ orig_last_token = last_read_token;
last_read_token = WORD; /* WORD to allow reserved words here */
token = (char *)NULL;
token_buffer_size = 0;
+ assignok = parser_state&PST_ASSIGNOK; /* XXX */
+
wl = (WORD_LIST *)NULL; /* ( */
parser_state |= PST_COMPASSIGN;
jump_to_top_level (DISCARD);
}
- last_read_token = WORD;
+ last_read_token = orig_last_token; /* XXX - was WORD? */
+
if (wl)
{
rl = REVERSE_LIST (wl, WORD_LIST *);
if (retlenp)
*retlenp = (ret && *ret) ? strlen (ret) : 0;
+
+ if (assignok)
+ parser_state |= PST_ASSIGNOK;
+
return ret;
}
save_parser_state (ps)
sh_parser_state_t *ps;
{
-#if defined (ARRAY_VARS)
- SHELL_VAR *v;
-#endif
-
if (ps == 0)
ps = (sh_parser_state_t *)xmalloc (sizeof (sh_parser_state_t));
if (ps == 0)
ps->input_line_terminator = shell_input_line_terminator;
ps->eof_encountered = eof_encountered;
+ ps->prompt_string_pointer = prompt_string_pointer;
+
ps->current_command_line_count = current_command_line_count;
#if defined (HISTORY)
ps->last_command_exit_value = last_command_exit_value;
#if defined (ARRAY_VARS)
- v = find_variable ("PIPESTATUS");
- if (v && array_p (v) && array_cell (v))
- ps->pipestatus = array_copy (array_cell (v));
- else
- ps->pipestatus = (ARRAY *)NULL;
+ ps->pipestatus = save_pipestatus_array ();
#endif
ps->last_shell_builtin = last_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;
+
+ if (need_here_doc == 0)
+ ps->redir_stack[0] = 0;
+ else
+ memcpy (ps->redir_stack, redir_stack, sizeof (redir_stack[0]) * HEREDOC_MAX);
+
+ 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);
}
restore_parser_state (ps)
sh_parser_state_t *ps;
{
-#if defined (ARRAY_VARS)
- SHELL_VAR *v;
-#endif
+ int i;
if (ps == 0)
return;
shell_input_line_terminator = ps->input_line_terminator;
eof_encountered = ps->eof_encountered;
+ prompt_string_pointer = ps->prompt_string_pointer;
+
current_command_line_count = ps->current_command_line_count;
#if defined (HISTORY)
last_command_exit_value = ps->last_command_exit_value;
#if defined (ARRAY_VARS)
- v = find_variable ("PIPESTATUS");
- if (v && array_p (v) && array_cell (v))
- {
- array_dispose (array_cell (v));
- var_setarray (v, ps->pipestatus);
- }
+ restore_pipestatus_array (ps->pipestatus);
#endif
last_shell_builtin = ps->last_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;
+
+#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
+
+ FREE (token);
+ token = ps->token;
+ token_buffer_size = ps->token_buffer_size;
+}
+
+sh_input_line_state_t *
+save_input_line_state (ls)
+ sh_input_line_state_t *ls;
+{
+ if (ls == 0)
+ ls = (sh_input_line_state_t *)xmalloc (sizeof (sh_input_line_state_t));
+ if (ls == 0)
+ return ((sh_input_line_state_t *)NULL);
+
+ ls->input_line = shell_input_line;
+ ls->input_line_size = shell_input_line_size;
+ ls->input_line_len = shell_input_line_len;
+ ls->input_line_index = shell_input_line_index;
+
+ /* force reallocation */
+ shell_input_line = 0;
+ shell_input_line_size = shell_input_line_len = shell_input_line_index = 0;
+
+ return ls;
+}
+
+void
+restore_input_line_state (ls)
+ sh_input_line_state_t *ls;
+{
+ FREE (shell_input_line);
+ shell_input_line = ls->input_line;
+ shell_input_line_size = ls->input_line_size;
+ shell_input_line_len = ls->input_line_len;
+ shell_input_line_index = ls->input_line_index;
+
+ set_line_mbstate ();
}
/************************************************
************************************************/
#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 i, previ, len, c;
+ int c;
+ size_t i, previ, len;
mbstate_t mbs, prevs;
size_t mbclen;
if (shell_input_line == NULL)
return;
len = strlen (shell_input_line); /* XXX - shell_input_line_len ? */
- FREE (shell_input_line_property);
- shell_input_line_property = (char *)xmalloc (len + 1);
+ 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);
+ }
+ /* XXX - use whether or not we are in a UTF-8 locale to avoid calls to
+ mbrlen */
memset (&prevs, '\0', sizeof (mbstate_t));
for (i = previ = 0; i < len; i++)
{
c = shell_input_line[i];
if (c == EOF)
{
- int j;
+ size_t j;
for (j = i; j < len; j++)
shell_input_line_property[j] = 1;
break;
}
+ /* I'd love to take more advantage of UTF-8's properties in a UTF-8
+ locale, but mbrlen changes the mbstate_t on every call even when
+ presented with single-byte characters. */
mbclen = mbrlen (shell_input_line + previ, i - previ + 1, &mbs);
if (mbclen == 1 || mbclen == (size_t)-1)
{
else
{
/* XXX - what to do if mbrlen returns 0? (null wide character) */
- int j;
+ size_t j;
for (j = i; j < len; j++)
shell_input_line_property[j] = 1;
break;