2006-04-24 17:41:27 +00:00
%{
/*
2006-04-26 18:43:29 +00:00
* Asterisk -- An open source telephony toolkit.
2006-04-24 17:41:27 +00:00
*
* Copyright (C) 2006, Digium, Inc.
*
* Steve Murphy <murf@parsetree.com>
*
* See http://www.asterisk.org for more information about
* the Asterisk project. Please do not directly contact
* any of the maintainers of this project for assistance;
* the project provides a web site, mailing lists and IRC
* channels for your use.
*
* This program is free software, distributed under the terms of
* the GNU General Public License Version 2. See the LICENSE file
* at the top of the source tree.
*/
/*! \file
*
* \brief Bison Grammar description of AEL2.
2006-04-26 18:43:29 +00:00
*
2006-04-24 17:41:27 +00:00
*/
2006-06-07 18:54:56 +00:00
#include "asterisk.h"
ASTERISK_FILE_VERSION(__FILE__, "$Revision$")
2006-04-24 17:41:27 +00:00
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
2006-06-07 18:54:56 +00:00
2006-04-24 17:41:27 +00:00
#include "asterisk/logger.h"
Merged revisions 130145 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
Merging this rev from trunk to 1.6.0 was not
simple. Why? Because we've enhanced trunk to
do a [fast] merge-and-delete operation which
also solved problems with contexts having
entries from different registrars.
Fast as in the amount of time the contexts
are locked down. That *is* fast, but traversing
the entire dialplan looking for priorities to
delete takes more time overall.
This particular fix involved pulling in those
enhancements from trunk, along with all the
various fixes and refinements made along the
way.
Merging all this from trunk into 1.6 involved:
a. mergetrunk6 in the stuff from 130145;
b. revert all but the prop changes
c. catalog all revisions to pbx.c since 1.6.0 was forked
(at rev 105596).
d. catalog all revisions to pbx.c in trunk since 1.6.0
was forked, making special note of all revs that
were not merged into 1.6.0.
e. study each rev in trunk not applied to 1.6.0, and
determine if it was involved in the merge_and_delete
enhancements in trunk. 25 commits were done in 1.6.0,
all but one (106306) was a merge from trunk.
Trunk had 22 additional changes, of which 7 were
involved in the merge_and_delete enhancements:
106757
108894
109169
116461
123358
130145
130297
f. Go to trunk and collect patches, one by one,
of the changes made by each rev across the
entire source tree, using svn diff -c <num> > pfile
g. Apply each patch in order to 1.6.0, and
resolve all failures and compilation problems
before proceding to the next patch.
h. test the stuff.
i. profit!
........
r130145 | murf | 2008-07-11 12:24:31 -0600 (Fri, 11 Jul 2008) | 40 lines
(closes issue #13041)
Reported by: eliel
Tested by: murf
(closes issue #12960)
Reported by: mnicholson
In this 'omnibus' fix, I **think** I solved both
the problem in 13041, where unloading pbx_ael.so
caused crashes, or incomplete removal of previous
registrar'ed entries. And I added code to completely
remove all includes, switches, and ignorepats that
had a matching registrar entry, which should
appease 12960.
I also added a lot of seemingly useless brackets
around single statement if's, which helped debug
so much that I'm leaving them there.
I added a routine to check the correlation between
the extension tree lists and the hashtab
tables. It can be amazingly helpful when you have
lots of dialplan stuff, and need to narrow
down where a problem is occurring. It's ifdef'd
out by default.
I cleaned up the code around the new CIDmatch code.
It was leaving hanging extens with bad ptrs, getting confused
over which objects to remove, etc. I tightened
up the code and changed the call to remove_exten
in the merge_and_delete code.
I added more conditions to check for empty context
worthy of deletion. It's not empty if there are
any includes, switches, or ignorepats present.
If I've missed anything, please re-open this bug,
and be prepared to supply example dialplan code.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@130946 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-07-15 13:14:07 +00:00
#include "asterisk/lock.h"
#include "asterisk/hashtab.h"
2006-04-24 17:41:27 +00:00
#include "asterisk/ael_structs.h"
Merged revisions 138815 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
........
r138815 | murf | 2008-08-19 09:59:12 -0600 (Tue, 19 Aug 2008) | 19 lines
These changes are in regards to bug 13249, where users are being surprised by the changes made
to the Set app in trunk/1.6.x, as they come from the 1.4 world. They are only bitten if
they write their AEL dialplan in the 1.4 world, and then carry it over to a trunk/1.6.x
installation where a "make samples" was executed, or where they hand-edited the
asterisk.conf file and added the [compat] category with app_set = 1.6 (or higher).
(this commit does not totally solve 13249, at least not yet)
The change involves issueing a single warning while the AEL file is loading, if:
1. app_set is present in the config file, and set to 1.6 or higher.
2. there are double quotes in an assignment statement (eg x = "hi there";)
3. the warning was not already issued.
The standalone app, aelparse, does not (yet) issue this warning. I'd have to
have it read in the asterisk.conf file, and that's a bit of hassle. I'll add
it if users request it, tho.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@138846 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-19 16:36:22 +00:00
#include "asterisk/utils.h"
2006-04-24 17:41:27 +00:00
2008-08-19 16:38:21 +00:00
extern struct ast_flags ast_compat;
2007-08-15 19:21:27 +00:00
pval * linku1(pval *head, pval *tail);
2006-08-07 12:59:47 +00:00
static void set_dads(pval *dad, pval *child_list);
2006-04-26 22:41:16 +00:00
void reset_parencount(yyscan_t yyscanner);
void reset_semicount(yyscan_t yyscanner);
void reset_argcount(yyscan_t yyscanner );
2007-08-15 19:21:27 +00:00
2006-04-24 17:41:27 +00:00
#define YYLEX_PARAM ((struct parse_io *)parseio)->scanner
#define YYERROR_VERBOSE 1
extern char *my_file;
#ifdef AAL_ARGCHECK
int ael_is_funcname(char *name);
#endif
2007-09-27 00:06:06 +00:00
static char *ael_token_subst(const char *mess);
Merged revisions 138815 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
........
r138815 | murf | 2008-08-19 09:59:12 -0600 (Tue, 19 Aug 2008) | 19 lines
These changes are in regards to bug 13249, where users are being surprised by the changes made
to the Set app in trunk/1.6.x, as they come from the 1.4 world. They are only bitten if
they write their AEL dialplan in the 1.4 world, and then carry it over to a trunk/1.6.x
installation where a "make samples" was executed, or where they hand-edited the
asterisk.conf file and added the [compat] category with app_set = 1.6 (or higher).
(this commit does not totally solve 13249, at least not yet)
The change involves issueing a single warning while the AEL file is loading, if:
1. app_set is present in the config file, and set to 1.6 or higher.
2. there are double quotes in an assignment statement (eg x = "hi there";)
3. the warning was not already issued.
The standalone app, aelparse, does not (yet) issue this warning. I'd have to
have it read in the asterisk.conf file, and that's a bit of hassle. I'll add
it if users request it, tho.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@138846 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-19 16:36:22 +00:00
static int only_one_app_set_warning = 0;
2006-04-26 18:43:29 +00:00
2006-04-24 17:41:27 +00:00
%}
%union {
2006-04-30 12:44:54 +00:00
int intval; /* integer value, typically flags */
char *str; /* strings */
struct pval *pval; /* full objects */
2006-04-24 17:41:27 +00:00
}
%{
/* declaring these AFTER the union makes things a lot simpler! */
void yyerror(YYLTYPE *locp, struct parse_io *parseio, char const *s);
int ael_yylex (YYSTYPE * yylval_param, YYLTYPE * yylloc_param , void * yyscanner);
2006-04-26 18:43:29 +00:00
2006-04-30 13:57:08 +00:00
/* create a new object with start-end marker */
2007-08-15 19:21:27 +00:00
pval *npval(pvaltype type, int first_line, int last_line,
2006-04-30 13:57:08 +00:00
int first_column, int last_column);
2006-04-27 18:09:20 +00:00
/* create a new object with start-end marker, simplified interface.
* Must be declared here because YYLTYPE is not known before
*/
static pval *npval2(pvaltype type, YYLTYPE *first, YYLTYPE *last);
2006-04-30 12:12:39 +00:00
2006-04-30 13:57:08 +00:00
/* another frontend for npval, this time for a string */
static pval *nword(char *string, YYLTYPE *pos);
2006-04-30 12:12:39 +00:00
/* update end position of an object, return the object */
static pval *update_last(pval *, YYLTYPE *);
2006-04-24 17:41:27 +00:00
%}
2006-04-26 18:43:29 +00:00
2006-04-24 17:41:27 +00:00
%token KW_CONTEXT LC RC LP RP SEMI EQ COMMA COLON AMPER BAR AT
Merged revisions 87168 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r87168 | murf | 2007-10-26 10:34:02 -0600 (Fri, 26 Oct 2007) | 1 line
closes issue #11086 where a user complains that references to following contexts report a problem; The problem was REALLy that he was referring to empty contexts, which were being ignored. Reporter stated that empty contexts should be OK. I checked it out against extensions.conf, and sure enough, empty contexts ARE ok. So, I removed the restriction from AEL. This, though, highlighted a problem with multiple contexts of the same name. This should be OK, also. So, I added the extend keyword to AEL, and it can preceed the 'context' keyword (mixed with 'abstract', if nec.). This will turn off the warnings in AEL if the same context name is used 2 or more times. Also, I now call ast_context_find_or_create for contexts now, instead of just ast_context_create; I did this because pbx_config does this. The 'extend' keyword thus becomes a statement of intent. AEL can now duplicate the behavior of pbx_config,
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@87187 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-10-26 17:39:39 +00:00
%token KW_MACRO KW_GLOBALS KW_IGNOREPAT KW_SWITCH KW_IF KW_IFTIME KW_ELSE KW_RANDOM KW_ABSTRACT KW_EXTEND
2006-04-24 17:41:27 +00:00
%token EXTENMARK KW_GOTO KW_JUMP KW_RETURN KW_BREAK KW_CONTINUE KW_REGEXTEN KW_HINT
%token KW_FOR KW_WHILE KW_CASE KW_PATTERN KW_DEFAULT KW_CATCH KW_SWITCHES KW_ESWITCHES
2007-06-20 20:10:19 +00:00
%token KW_INCLUDES KW_LOCAL
2006-04-24 17:41:27 +00:00
2006-04-30 08:21:46 +00:00
%right BAR COMMA
2006-04-24 17:41:27 +00:00
%token <str> word
%type <pval>includes
%type <pval>includeslist
%type <pval>switchlist
%type <pval>eswitches
%type <pval>switches
%type <pval>macro_statement
%type <pval>macro_statements
%type <pval>case_statement
%type <pval>case_statements
%type <pval>eval_arglist
%type <pval>application_call
%type <pval>application_call_head
%type <pval>macro_call
%type <pval>target jumptarget
%type <pval>statement
2006-05-03 16:08:35 +00:00
%type <pval>switch_statement
2006-04-27 19:51:59 +00:00
2006-05-02 18:08:27 +00:00
%type <pval>if_like_head
2006-04-24 17:41:27 +00:00
%type <pval>statements
%type <pval>extension
%type <pval>ignorepat
%type <pval>element
%type <pval>elements
%type <pval>arglist
2006-05-02 20:11:24 +00:00
%type <pval>assignment
2007-06-20 20:10:19 +00:00
%type <pval>local_assignment
2006-04-24 17:41:27 +00:00
%type <pval>global_statements
%type <pval>globals
%type <pval>macro
%type <pval>context
%type <pval>object
2006-04-26 18:43:29 +00:00
%type <pval>objects
%type <pval>file
2006-04-30 08:21:46 +00:00
/* XXX lr changes */
%type <pval>opt_else
2006-05-02 18:33:15 +00:00
%type <pval>timespec
2006-05-02 18:41:57 +00:00
%type <pval>included_entry
2006-04-30 08:21:46 +00:00
%type <str>opt_word
2006-05-02 18:45:18 +00:00
%type <str>context_name
2006-05-02 18:23:41 +00:00
%type <str>timerange
2006-04-24 17:41:27 +00:00
2006-04-27 19:51:59 +00:00
%type <str>goto_word
%type <str>word_list
2006-08-12 19:28:33 +00:00
%type <str>word3_list hint_word
2006-05-02 19:17:49 +00:00
%type <str>test_expr
2006-05-03 17:07:56 +00:00
%type <str>opt_pri
2006-04-27 19:51:59 +00:00
2006-04-30 12:44:54 +00:00
%type <intval>opt_abstract
2006-04-27 19:51:59 +00:00
/*
* OPTIONS
*/
%locations /* track source location using @n variables (yylloc in flex) */
%pure-parser /* pass yylval and yylloc as arguments to yylex(). */
2006-04-24 17:41:27 +00:00
%name-prefix="ael_yy"
2006-04-27 19:51:59 +00:00
/*
* add an additional argument, parseio, to yyparse(),
* which is then accessible in the grammar actions
*/
2006-04-24 17:41:27 +00:00
%parse-param {struct parse_io *parseio}
2006-04-27 19:51:59 +00:00
2006-04-26 18:43:29 +00:00
/* there will be two shift/reduce conflicts, they involve the if statement, where a single statement occurs not wrapped in curlies in the "true" section
2006-04-24 17:41:27 +00:00
the default action to shift will attach the else to the preceeding if. */
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
%expect 30
2006-04-24 17:41:27 +00:00
%error-verbose
2006-04-27 19:51:59 +00:00
/*
* declare destructors for objects.
* The former is for pval, the latter for strings.
2006-04-27 21:09:52 +00:00
* NOTE: we must not have a destructor for a 'file' object.
2006-04-27 19:51:59 +00:00
*/
%destructor {
2006-04-27 21:09:52 +00:00
destroy_pval($$);
prev_word=0;
2006-04-27 19:51:59 +00:00
} includes includeslist switchlist eswitches switches
macro_statement macro_statements case_statement case_statements
eval_arglist application_call application_call_head
2006-05-03 16:08:35 +00:00
macro_call target jumptarget statement switch_statement
2006-05-02 18:08:27 +00:00
if_like_head statements extension
2007-06-20 20:10:19 +00:00
ignorepat element elements arglist assignment local_assignment
2006-04-27 19:51:59 +00:00
global_statements globals macro context object objects
2006-04-30 08:21:46 +00:00
opt_else
2006-05-02 18:41:57 +00:00
timespec included_entry
2006-04-27 19:51:59 +00:00
2006-05-02 18:45:18 +00:00
%destructor { free($$);} word word_list goto_word word3_list opt_word context_name
2006-05-02 18:23:41 +00:00
timerange
2006-05-02 19:17:49 +00:00
test_expr
2006-05-03 17:07:56 +00:00
opt_pri
2006-04-24 17:41:27 +00:00
%%
file : objects { $$ = parseio->pval = $1; }
;
objects : object {$$=$1;}
2006-05-02 14:12:01 +00:00
| objects object { $$ = linku1($1, $2); }
2006-04-24 17:41:27 +00:00
| objects error {$$=$1;}
;
object : context {$$=$1;}
| macro {$$=$1;}
| globals {$$=$1;}
| SEMI {$$=0;/* allow older docs to be read */}
;
2006-05-02 18:45:18 +00:00
context_name : word { $$ = $1; }
2006-04-30 08:21:46 +00:00
| KW_DEFAULT { $$ = strdup("default"); }
;
2006-05-03 16:46:55 +00:00
context : opt_abstract KW_CONTEXT context_name LC elements RC {
Merged revisions 87168 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r87168 | murf | 2007-10-26 10:34:02 -0600 (Fri, 26 Oct 2007) | 1 line
closes issue #11086 where a user complains that references to following contexts report a problem; The problem was REALLy that he was referring to empty contexts, which were being ignored. Reporter stated that empty contexts should be OK. I checked it out against extensions.conf, and sure enough, empty contexts ARE ok. So, I removed the restriction from AEL. This, though, highlighted a problem with multiple contexts of the same name. This should be OK, also. So, I added the extend keyword to AEL, and it can preceed the 'context' keyword (mixed with 'abstract', if nec.). This will turn off the warnings in AEL if the same context name is used 2 or more times. Also, I now call ast_context_find_or_create for contexts now, instead of just ast_context_create; I did this because pbx_config does this. The 'extend' keyword thus becomes a statement of intent. AEL can now duplicate the behavior of pbx_config,
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@87187 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-10-26 17:39:39 +00:00
$$ = npval2(PV_CONTEXT, &@1, &@6);
$$->u1.str = $3;
$$->u2.statements = $5;
set_dads($$,$5);
$$->u3.abstract = $1;}
2006-04-30 12:44:54 +00:00
;
2006-05-02 17:58:57 +00:00
/* optional "abstract" keyword XXX there is no regression test for this */
2006-04-30 12:44:54 +00:00
opt_abstract: KW_ABSTRACT { $$ = 1; }
| /* nothing */ { $$ = 0; }
Merged revisions 87168 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r87168 | murf | 2007-10-26 10:34:02 -0600 (Fri, 26 Oct 2007) | 1 line
closes issue #11086 where a user complains that references to following contexts report a problem; The problem was REALLy that he was referring to empty contexts, which were being ignored. Reporter stated that empty contexts should be OK. I checked it out against extensions.conf, and sure enough, empty contexts ARE ok. So, I removed the restriction from AEL. This, though, highlighted a problem with multiple contexts of the same name. This should be OK, also. So, I added the extend keyword to AEL, and it can preceed the 'context' keyword (mixed with 'abstract', if nec.). This will turn off the warnings in AEL if the same context name is used 2 or more times. Also, I now call ast_context_find_or_create for contexts now, instead of just ast_context_create; I did this because pbx_config does this. The 'extend' keyword thus becomes a statement of intent. AEL can now duplicate the behavior of pbx_config,
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@87187 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2007-10-26 17:39:39 +00:00
| KW_EXTEND { $$ = 2; }
| KW_EXTEND KW_ABSTRACT { $$=3; }
| KW_ABSTRACT KW_EXTEND { $$=3; }
2006-04-24 17:41:27 +00:00
;
2006-04-27 02:29:32 +00:00
macro : KW_MACRO word LP arglist RP LC macro_statements RC {
2006-04-27 18:18:12 +00:00
$$ = npval2(PV_MACRO, &@1, &@8);
2006-08-07 12:59:47 +00:00
$$->u1.str = $2; $$->u2.arglist = $4; $$->u3.macro_statements = $7;
set_dads($$,$7);}
2006-04-24 17:41:27 +00:00
;
2006-04-27 02:29:32 +00:00
globals : KW_GLOBALS LC global_statements RC {
2006-04-28 10:06:10 +00:00
$$ = npval2(PV_GLOBALS, &@1, &@4);
2006-08-07 12:59:47 +00:00
$$->u1.statements = $3;
set_dads($$,$3);}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:28:48 +00:00
global_statements : { $$ = NULL; }
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| global_statements assignment {$$ = linku1($1, $2); }
2007-08-23 23:37:33 +00:00
| error global_statements {$$=$2;}
2006-04-24 17:41:27 +00:00
;
2006-05-02 20:11:24 +00:00
assignment : word EQ { reset_semicount(parseio->scanner); } word SEMI {
2006-04-28 10:06:10 +00:00
$$ = npval2(PV_VARDEC, &@1, &@5);
Merged revisions 138815 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
........
r138815 | murf | 2008-08-19 09:59:12 -0600 (Tue, 19 Aug 2008) | 19 lines
These changes are in regards to bug 13249, where users are being surprised by the changes made
to the Set app in trunk/1.6.x, as they come from the 1.4 world. They are only bitten if
they write their AEL dialplan in the 1.4 world, and then carry it over to a trunk/1.6.x
installation where a "make samples" was executed, or where they hand-edited the
asterisk.conf file and added the [compat] category with app_set = 1.6 (or higher).
(this commit does not totally solve 13249, at least not yet)
The change involves issueing a single warning while the AEL file is loading, if:
1. app_set is present in the config file, and set to 1.6 or higher.
2. there are double quotes in an assignment statement (eg x = "hi there";)
3. the warning was not already issued.
The standalone app, aelparse, does not (yet) issue this warning. I'd have to
have it read in the asterisk.conf file, and that's a bit of hassle. I'll add
it if users request it, tho.
........
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@138846 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2008-08-19 16:36:22 +00:00
if (!ast_compat_app_set && !only_one_app_set_warning && strchr($4,'"')) {
ast_log(LOG_NOTICE,"Note: In asterisk.conf, in the [compat] section, the app_set is set to 1.6 or greater. The Set() function no longer removes double quotes from the value. If this is a surprise to you, you can set app_set to 1.4.\n");
only_one_app_set_warning = 1;
}
2006-04-27 02:29:32 +00:00
$$->u1.str = $1;
$$->u2.val = $4; }
2006-04-24 17:41:27 +00:00
;
2007-06-20 20:10:19 +00:00
local_assignment : KW_LOCAL word EQ { reset_semicount(parseio->scanner); } word SEMI {
$$ = npval2(PV_LOCALVARDEC, &@1, &@6);
$$->u1.str = $2;
$$->u2.val = $5; }
;
2006-05-03 16:58:16 +00:00
/* XXX this matches missing arguments, is this desired ? */
2006-05-02 18:48:47 +00:00
arglist : /* empty */ { $$ = NULL; }
| word { $$ = nword($1, &@1); }
2006-05-02 14:27:19 +00:00
| arglist COMMA word { $$ = linku1($1, nword($3, &@3)); }
2006-04-24 17:41:27 +00:00
| arglist error {$$=$1;}
;
2006-05-03 16:28:48 +00:00
elements : {$$=0;}
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| elements element { $$ = linku1($1, $2); }
2007-08-23 23:37:33 +00:00
| error elements { $$=$2;}
2006-04-24 17:41:27 +00:00
;
element : extension {$$=$1;}
| includes {$$=$1;}
| switches {$$=$1;}
| eswitches {$$=$1;}
| ignorepat {$$=$1;}
2006-05-02 20:11:24 +00:00
| assignment {$$=$1;}
2007-06-20 20:10:19 +00:00
| local_assignment {$$=$1;}
2006-04-26 18:43:29 +00:00
| word error {free($1); $$=0;}
2006-04-24 17:41:27 +00:00
| SEMI {$$=0;/* allow older docs to be read */}
;
2006-04-27 02:29:32 +00:00
ignorepat : KW_IGNOREPAT EXTENMARK word SEMI {
2006-04-27 18:26:37 +00:00
$$ = npval2(PV_IGNOREPAT, &@1, &@4);
2006-04-27 02:29:32 +00:00
$$->u1.str = $3;}
2006-04-24 17:41:27 +00:00
;
2006-04-27 02:29:32 +00:00
extension : word EXTENMARK statement {
2006-04-28 11:20:21 +00:00
$$ = npval2(PV_EXTENSION, &@1, &@3);
2006-04-27 02:29:32 +00:00
$$->u1.str = $1;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $3; set_dads($$,$3);}
2007-11-27 18:50:44 +00:00
| word AT word EXTENMARK statement {
$$ = npval2(PV_EXTENSION, &@1, &@3);
$$->u1.str = malloc(strlen($1)+strlen($3)+2);
strcpy($$->u1.str,$1);
strcat($$->u1.str,"@");
strcat($$->u1.str,$3);
free($1);
$$->u2.statements = $5; set_dads($$,$5);}
2006-04-27 02:29:32 +00:00
| KW_REGEXTEN word EXTENMARK statement {
2006-04-28 11:20:21 +00:00
$$ = npval2(PV_EXTENSION, &@1, &@4);
2006-04-27 02:29:32 +00:00
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $4; set_dads($$,$4);
2006-04-27 02:29:32 +00:00
$$->u4.regexten=1;}
2006-08-12 19:28:33 +00:00
| KW_HINT LP hint_word RP word EXTENMARK statement {
2006-04-28 11:20:21 +00:00
$$ = npval2(PV_EXTENSION, &@1, &@7);
2006-04-27 02:29:32 +00:00
$$->u1.str = $5;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $7; set_dads($$,$7);
2006-04-27 02:29:32 +00:00
$$->u3.hints = $3;}
2006-08-12 19:28:33 +00:00
| KW_REGEXTEN KW_HINT LP hint_word RP word EXTENMARK statement {
2006-04-28 11:20:21 +00:00
$$ = npval2(PV_EXTENSION, &@1, &@8);
2006-04-27 02:29:32 +00:00
$$->u1.str = $6;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $8; set_dads($$,$8);
2006-04-27 02:29:32 +00:00
$$->u4.regexten=1;
$$->u3.hints = $4;}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:33:00 +00:00
/* list of statements in a block or after a case label - can be empty */
statements : /* empty */ { $$ = NULL; }
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| statements statement { $$ = linku1($1, $2); }
2007-08-23 23:37:33 +00:00
| error statements {$$=$2;}
2006-04-24 17:41:27 +00:00
;
2006-05-02 18:23:41 +00:00
/* hh:mm-hh:mm, due to the way the parser works we do not
* detect the '-' but only the ':' as separator
*/
timerange: word3_list COLON word3_list COLON word3_list {
2008-11-03 00:39:04 +00:00
if (asprintf(&$$, "%s:%s:%s", $1, $3, $5) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($3);
free($5);
}
}
2006-05-02 18:23:41 +00:00
| word { $$ = $1; }
;
2006-05-02 18:33:15 +00:00
/* full time specification range|dow|*|* */
timespec : timerange BAR word3_list BAR word3_list BAR word3_list {
$$ = nword($1, &@1);
2006-05-02 20:44:56 +00:00
$$->next = nword($3, &@3);
$$->next->next = nword($5, &@5);
$$->next->next->next = nword($7, &@7); }
2006-05-02 18:33:15 +00:00
;
2006-05-02 19:17:49 +00:00
/* expression used in if, random, while, switch */
2006-05-03 17:07:56 +00:00
test_expr : LP { reset_parencount(parseio->scanner); } word_list RP { $$ = $3; }
2006-05-02 19:17:49 +00:00
;
2006-05-02 18:08:27 +00:00
/* 'if' like statements: if, iftime, random */
2006-05-02 19:17:49 +00:00
if_like_head : KW_IF test_expr {
$$= npval2(PV_IF, &@1, &@2);
$$->u1.str = $2; }
| KW_RANDOM test_expr {
$$ = npval2(PV_RANDOM, &@1, &@2);
$$->u1.str=$2;}
2006-05-02 18:33:15 +00:00
| KW_IFTIME LP timespec RP {
$$ = npval2(PV_IFTIME, &@1, &@4);
$$->u1.list = $3;
prev_word = 0; }
2006-04-24 17:41:27 +00:00
;
/* word_list is a hack to fix a problem with context switching between bison and flex;
2006-04-26 18:43:29 +00:00
by the time you register a new context with flex, you've already got a look-ahead token
2006-04-24 17:41:27 +00:00
from the old context, with no way to put it back and start afresh. So, we kludge this
and merge the words back together. */
2006-04-26 18:43:29 +00:00
word_list : word { $$ = $1;}
2006-04-27 08:24:00 +00:00
| word word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s%s", $1, $2) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($2);
prev_word = $$;
}
}
2006-04-24 17:41:27 +00:00
;
2006-04-27 18:26:37 +00:00
2006-08-12 19:28:33 +00:00
hint_word : word { $$ = $1; }
| hint_word word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s %s", $1, $2) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($2);
}
}
2007-10-24 04:44:27 +00:00
| hint_word COLON word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s:%s", $1, $3) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($3);
}
}
2006-08-12 19:28:33 +00:00
| hint_word AMPER word { /* there are often '&' in hints */
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s&%s", $1, $3) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($3);
}
}
2009-04-20 21:31:36 +00:00
| hint_word AT word {
if (asprintf(&($$), "%s@%s", $1, $3) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($3);
}
}
2008-11-03 00:39:04 +00:00
;
2006-08-12 19:28:33 +00:00
2006-04-26 18:43:29 +00:00
word3_list : word { $$ = $1;}
2006-04-27 08:24:00 +00:00
| word word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s%s", $1, $2) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($2);
prev_word = $$;
}
}
2006-04-27 08:24:00 +00:00
| word word word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s%s%s", $1, $2, $3) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($2);
free($3);
prev_word=$$;
}
}
2006-04-24 17:41:27 +00:00
;
goto_word : word { $$ = $1;}
2006-04-27 08:24:00 +00:00
| word word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s%s", $1, $2) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($2);
}
}
2006-07-19 02:55:24 +00:00
| goto_word COLON word {
2008-11-03 00:39:04 +00:00
if (asprintf(&($$), "%s:%s", $1, $3) < 0) {
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
free($1);
free($3);
}
}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:08:35 +00:00
switch_statement : KW_SWITCH test_expr LC case_statements RC {
$$ = npval2(PV_SWITCH, &@1, &@5);
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $4; set_dads($$,$4);}
2006-04-24 17:41:27 +00:00
;
2006-04-27 18:09:20 +00:00
/*
* Definition of a statememt in our language
*/
2006-04-27 16:40:25 +00:00
statement : LC statements RC {
2006-04-27 18:09:20 +00:00
$$ = npval2(PV_STATEMENTBLOCK, &@1, &@3);
2006-08-07 12:59:47 +00:00
$$->u1.list = $2; set_dads($$,$2);}
2006-05-02 20:13:58 +00:00
| assignment { $$ = $1; }
2007-06-20 20:10:19 +00:00
| local_assignment { $$ = $1; }
2006-04-27 16:40:25 +00:00
| KW_GOTO target SEMI {
2006-04-27 19:14:07 +00:00
$$ = npval2(PV_GOTO, &@1, &@3);
2006-04-27 16:40:25 +00:00
$$->u1.list = $2;}
| KW_JUMP jumptarget SEMI {
2006-04-28 06:26:27 +00:00
$$ = npval2(PV_GOTO, &@1, &@3);
2006-04-27 16:40:25 +00:00
$$->u1.list = $2;}
| word COLON {
2006-04-28 06:26:27 +00:00
$$ = npval2(PV_LABEL, &@1, &@2);
2006-04-27 16:40:25 +00:00
$$->u1.str = $1; }
2006-04-24 17:41:27 +00:00
| KW_FOR LP {reset_semicount(parseio->scanner);} word SEMI
{reset_semicount(parseio->scanner);} word SEMI
2006-05-02 19:17:49 +00:00
{reset_parencount(parseio->scanner);} word RP statement { /* XXX word_list maybe ? */
2006-04-28 06:26:27 +00:00
$$ = npval2(PV_FOR, &@1, &@12);
2006-04-27 16:40:25 +00:00
$$->u1.for_init = $4;
$$->u2.for_test=$7;
$$->u3.for_inc = $10;
2006-08-07 12:59:47 +00:00
$$->u4.for_statements = $12; set_dads($$,$12);}
2006-05-02 19:17:49 +00:00
| KW_WHILE test_expr statement {
$$ = npval2(PV_WHILE, &@1, &@3);
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $3; set_dads($$,$3);}
2006-05-03 16:08:35 +00:00
| switch_statement { $$ = $1; }
2006-05-03 17:07:56 +00:00
| AMPER macro_call SEMI { $$ = update_last($2, &@2); }
| application_call SEMI { $$ = update_last($1, &@2); }
2006-04-27 08:37:40 +00:00
| word SEMI {
2006-04-28 11:20:21 +00:00
$$= npval2(PV_APPLICATION_CALL, &@1, &@2);
2006-04-27 08:37:40 +00:00
$$->u1.str = $1;}
2006-04-24 17:41:27 +00:00
| application_call EQ {reset_semicount(parseio->scanner);} word SEMI {
2006-04-27 08:37:40 +00:00
char *bufx;
int tot=0;
pval *pptr;
2006-04-28 11:20:21 +00:00
$$ = npval2(PV_VARDEC, &@1, &@5);
2006-04-27 08:37:40 +00:00
$$->u2.val=$4;
/* rebuild the original string-- this is not an app call, it's an unwrapped vardec, with a func call on the LHS */
/* string to big to fit in the buffer? */
tot+=strlen($1->u1.str);
for(pptr=$1->u2.arglist;pptr;pptr=pptr->next) {
tot+=strlen(pptr->u1.str);
tot++; /* for a sep like a comma */
}
tot+=4; /* for safety */
2006-05-02 03:03:27 +00:00
bufx = calloc(1, tot);
2006-04-27 08:37:40 +00:00
strcpy(bufx,$1->u1.str);
strcat(bufx,"(");
/* XXX need to advance the pointer or the loop is very inefficient */
for (pptr=$1->u2.arglist;pptr;pptr=pptr->next) {
if ( pptr != $1->u2.arglist )
strcat(bufx,",");
strcat(bufx,pptr->u1.str);
}
strcat(bufx,")");
2006-04-24 17:41:27 +00:00
#ifdef AAL_ARGCHECK
2006-04-27 08:37:40 +00:00
if ( !ael_is_funcname($1->u1.str) )
ast_log(LOG_WARNING, "==== File: %s, Line %d, Cols: %d-%d: Function call? The name %s is not in my internal list of function names\n",
my_file, @1.first_line, @1.first_column, @1.last_column, $1->u1.str);
2006-04-24 17:41:27 +00:00
#endif
2006-04-27 08:37:40 +00:00
$$->u1.str = bufx;
destroy_pval($1); /* the app call it is not, get rid of that chain */
prev_word = 0;
}
2006-04-28 11:20:21 +00:00
| KW_BREAK SEMI { $$ = npval2(PV_BREAK, &@1, &@2); }
| KW_RETURN SEMI { $$ = npval2(PV_RETURN, &@1, &@2); }
| KW_CONTINUE SEMI { $$ = npval2(PV_CONTINUE, &@1, &@2); }
2006-05-02 18:08:27 +00:00
| if_like_head statement opt_else {
$$ = update_last($1, &@2);
2006-08-07 12:59:47 +00:00
$$->u2.statements = $2; set_dads($$,$2);
$$->u3.else_statements = $3;set_dads($$,$3);}
2006-04-24 17:41:27 +00:00
| SEMI { $$=0; }
;
2006-04-30 08:21:46 +00:00
opt_else : KW_ELSE statement { $$ = $2; }
| { $$ = NULL ; }
2007-08-23 23:37:33 +00:00
2006-04-30 13:57:08 +00:00
target : goto_word { $$ = nword($1, &@1); }
2006-04-27 06:44:38 +00:00
| goto_word BAR goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword($1, &@1);
$$->next = nword($3, &@3); }
2006-04-27 06:44:38 +00:00
| goto_word COMMA goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword($1, &@1);
$$->next = nword($3, &@3); }
2006-04-27 17:59:09 +00:00
| goto_word BAR goto_word BAR goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword($1, &@1);
$$->next = nword($3, &@3);
$$->next->next = nword($5, &@5); }
2006-04-27 17:59:09 +00:00
| goto_word COMMA goto_word COMMA goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword($1, &@1);
$$->next = nword($3, &@3);
$$->next->next = nword($5, &@5); }
2006-04-27 17:59:09 +00:00
| KW_DEFAULT BAR goto_word BAR goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword(strdup("default"), &@1);
$$->next = nword($3, &@3);
$$->next->next = nword($5, &@5); }
2006-04-27 17:59:09 +00:00
| KW_DEFAULT COMMA goto_word COMMA goto_word {
2006-04-30 13:57:08 +00:00
$$ = nword(strdup("default"), &@1);
$$->next = nword($3, &@3);
$$->next->next = nword($5, &@5); }
2006-04-24 17:41:27 +00:00
;
2006-05-03 17:07:56 +00:00
opt_pri : /* empty */ { $$ = strdup("1"); }
| COMMA word { $$ = $2; }
;
2006-04-30 14:06:04 +00:00
/* XXX please document the form of jumptarget */
2006-05-03 17:07:56 +00:00
jumptarget : goto_word opt_pri { /* ext[, pri] default 1 */
2006-04-30 13:57:08 +00:00
$$ = nword($1, &@1);
2006-05-03 17:07:56 +00:00
$$->next = nword($2, &@2); } /* jump extension[,priority][@context] */
| goto_word opt_pri AT context_name { /* context, ext, pri */
$$ = nword($4, &@4);
2006-05-02 14:25:01 +00:00
$$->next = nword($1, &@1);
2006-05-03 17:07:56 +00:00
$$->next->next = nword($2, &@2); }
2006-04-27 06:44:38 +00:00
;
2006-04-24 17:41:27 +00:00
2006-04-27 17:59:09 +00:00
macro_call : word LP {reset_argcount(parseio->scanner);} eval_arglist RP {
2006-04-27 18:18:12 +00:00
/* XXX original code had @2 but i think we need @5 */
$$ = npval2(PV_MACRO_CALL, &@1, &@5);
2006-04-27 17:59:09 +00:00
$$->u1.str = $1;
$$->u2.arglist = $4;}
| word LP RP {
2006-04-27 18:18:12 +00:00
$$= npval2(PV_MACRO_CALL, &@1, &@3);
2006-04-27 17:59:09 +00:00
$$->u1.str = $1; }
2006-04-24 17:41:27 +00:00
;
2006-04-30 23:53:22 +00:00
/* XXX application_call_head must be revised. Having 'word LP { ...'
* just as above should work fine, however it gives a different result.
*/
application_call_head: word LP {reset_argcount(parseio->scanner);} {
2006-04-27 08:37:40 +00:00
if (strcasecmp($1,"goto") == 0) {
2006-05-01 00:02:12 +00:00
$$ = npval2(PV_GOTO, &@1, &@2);
2006-04-27 06:44:38 +00:00
free($1); /* won't be using this */
ast_log(LOG_WARNING, "==== File: %s, Line %d, Cols: %d-%d: Suggestion: Use the goto statement instead of the Goto() application call in AEL.\n", my_file, @1.first_line, @1.first_column, @1.last_column );
2006-05-01 00:02:12 +00:00
} else {
$$= npval2(PV_APPLICATION_CALL, &@1, &@2);
$$->u1.str = $1;
} }
2006-04-24 17:41:27 +00:00
;
2006-04-30 12:12:39 +00:00
application_call : application_call_head eval_arglist RP {
$$ = update_last($1, &@3);
2006-04-24 17:41:27 +00:00
if( $$->type == PV_GOTO )
$$->u1.list = $2;
else
$$->u2.arglist = $2;
2006-04-30 12:12:39 +00:00
}
| application_call_head RP { $$ = update_last($1, &@2); }
2006-04-24 17:41:27 +00:00
;
2006-04-30 08:21:46 +00:00
opt_word : word { $$ = $1 }
| { $$ = strdup(""); }
;
2006-04-30 14:06:04 +00:00
eval_arglist : word_list { $$ = nword($1, &@1); }
2006-04-27 17:59:09 +00:00
| /*nothing! */ {
$$= npval(PV_WORD,0/*@1.first_line*/,0/*@1.last_line*/,0/* @1.first_column*/, 0/*@1.last_column*/);
$$->u1.str = strdup(""); }
2006-05-02 14:27:19 +00:00
| eval_arglist COMMA opt_word { $$ = linku1($1, nword($3, &@3)); }
2006-04-24 17:41:27 +00:00
;
2006-05-02 20:18:02 +00:00
case_statements: /* empty */ { $$ = NULL; }
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| case_statements case_statement { $$ = linku1($1, $2); }
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:33:00 +00:00
case_statement: KW_CASE word COLON statements {
2006-04-27 19:14:07 +00:00
$$ = npval2(PV_CASE, &@1, &@3); /* XXX 3 or 4 ? */
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $4; set_dads($$,$4);}
2006-05-03 16:33:00 +00:00
| KW_DEFAULT COLON statements {
2006-04-27 19:14:07 +00:00
$$ = npval2(PV_DEFAULT, &@1, &@3);
$$->u1.str = NULL;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $3;set_dads($$,$3);}
2006-05-03 16:33:00 +00:00
| KW_PATTERN word COLON statements {
2006-04-27 19:14:07 +00:00
$$ = npval2(PV_PATTERN, &@1, &@4); /* XXX@3 or @4 ? */
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $4;set_dads($$,$4);}
2006-04-24 17:41:27 +00:00
;
2006-05-02 18:51:33 +00:00
macro_statements: /* empty */ { $$ = NULL; }
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| macro_statements macro_statement { $$ = linku1($1, $2); }
2006-04-24 17:41:27 +00:00
;
macro_statement : statement {$$=$1;}
2007-06-05 22:04:22 +00:00
| includes { $$=$1;}
2006-04-28 06:26:27 +00:00
| KW_CATCH word LC statements RC {
$$ = npval2(PV_CATCH, &@1, &@5);
$$->u1.str = $2;
2006-08-07 12:59:47 +00:00
$$->u2.statements = $4; set_dads($$,$4);}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:38:50 +00:00
switches : KW_SWITCHES LC switchlist RC {
2006-04-30 12:46:15 +00:00
$$ = npval2(PV_SWITCHES, &@1, &@2);
2006-08-07 12:59:47 +00:00
$$->u1.list = $3; set_dads($$,$3);}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:38:50 +00:00
eswitches : KW_ESWITCHES LC switchlist RC {
2006-04-30 12:44:54 +00:00
$$ = npval2(PV_ESWITCHES, &@1, &@2);
2006-08-07 12:59:47 +00:00
$$->u1.list = $3; set_dads($$,$3);}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:37:14 +00:00
switchlist : /* empty */ { $$ = NULL; }
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
| switchlist word SEMI { $$ = linku1($1,nword($2, &@2)); }
| switchlist word AT word SEMI {
2008-11-03 00:39:04 +00:00
char *x;
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
if (asprintf(&x,"%s@%s", $2, $4) < 0) {
2008-11-03 00:39:04 +00:00
ast_log(LOG_WARNING, "asprintf() failed\n");
$$ = NULL;
} else {
Merged revisions 177286 via svnmerge from
https://origsvn.digium.com/svn/asterisk/trunk
................
r177286 | murf | 2009-02-18 16:50:57 -0700 (Wed, 18 Feb 2009) | 39 lines
Merged revisions 177225 via svnmerge from
https://origsvn.digium.com/svn/asterisk/branches/1.4
........
r177225 | murf | 2009-02-18 15:43:14 -0700 (Wed, 18 Feb 2009) | 34 lines
This patch fixes a regression of sorts that was introduced in
rev 24425.
It basically fixes AST-190/ABE-1782.
What was wrong: the user has 6000 extensions in one context; and
then 6000 contexts, one per extension. The parser could only handle
about 4893 of the 6000 extens in the single context.
This was due to the regression I mentioned. To get rid of
shift/reduce conflicts, Luigi set up right-recursive lists
for globals, context elements, switch lists, and statements.
Right recursive lists got rid of the warnings, but instead, they
use up a tremendous amount of stack space when the lists are long.
I saw this a few years back, and resolved not to fix it until
someone complained. That day has arrived!
After the changes were made, I ran the regression test suite,
and there were no problems.
I took the test case the user provided, and added 100,000
extensions to the single context, that already had 6,000 extens
in it. (I'll see your 6, and raise you 100!) It takes a few minutes
to read it all in, check it and generate code for it, but no
problems.
So, I think I can say that fundamentally, there are no longer
any limits on the number of items you can place in contexts,
statement blocks, switches, or globals, beyond your virt mem
constraints.
........
................
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.6.0@177289 65c4cc65-6c06-0410-ace0-fbb531ad65f3
2009-02-19 00:05:07 +00:00
free($2);
free($4);
$$ = linku1($1,nword(x, &@2));
2008-11-03 00:39:04 +00:00
}
}
2007-08-23 23:37:33 +00:00
| error switchlist {$$=$2;}
2006-04-24 17:41:27 +00:00
;
2006-05-03 16:28:48 +00:00
included_entry : context_name { $$ = nword($1, &@1); }
| context_name BAR timespec {
2006-05-02 14:25:01 +00:00
$$ = nword($1, &@1);
2006-05-02 18:33:15 +00:00
$$->u2.arglist = $3;
2006-05-02 18:41:57 +00:00
prev_word=0; /* XXX sure ? */ }
;
/* list of ';' separated context names followed by optional timespec */
2006-05-03 16:28:48 +00:00
includeslist : included_entry SEMI { $$ = $1; }
| includeslist included_entry SEMI { $$ = linku1($1, $2); }
2006-04-24 17:41:27 +00:00
| includeslist error {$$=$1;}
;
2006-04-27 06:44:38 +00:00
includes : KW_INCLUDES LC includeslist RC {
2006-04-27 18:18:12 +00:00
$$ = npval2(PV_INCLUDES, &@1, &@4);
2006-08-07 12:59:47 +00:00
$$->u1.list = $3;set_dads($$,$3);}
2006-04-27 06:44:38 +00:00
| KW_INCLUDES LC RC {
2006-04-27 18:18:12 +00:00
$$ = npval2(PV_INCLUDES, &@1, &@3);}
2006-04-24 17:41:27 +00:00
;
%%
2006-04-26 18:43:29 +00:00
static char *token_equivs1[] =
2006-04-24 17:41:27 +00:00
{
"AMPER",
"AT",
"BAR",
"COLON",
"COMMA",
"EQ",
"EXTENMARK",
2006-04-26 18:43:29 +00:00
"KW_BREAK",
"KW_CASE",
"KW_CATCH",
2006-04-24 17:41:27 +00:00
"KW_CONTEXT",
2006-04-26 18:43:29 +00:00
"KW_CONTINUE",
"KW_DEFAULT",
2006-04-24 17:41:27 +00:00
"KW_ELSE",
"KW_ESWITCHES",
2006-04-26 18:43:29 +00:00
"KW_FOR",
2006-04-24 17:41:27 +00:00
"KW_GLOBALS",
"KW_GOTO",
"KW_HINT",
"KW_IFTIME",
"KW_IF",
"KW_IGNOREPAT",
"KW_INCLUDES"
"KW_JUMP",
"KW_MACRO",
2006-04-26 18:43:29 +00:00
"KW_PATTERN",
"KW_REGEXTEN",
"KW_RETURN",
"KW_SWITCHES",
2006-04-24 17:41:27 +00:00
"KW_SWITCH",
2006-04-26 18:43:29 +00:00
"KW_WHILE",
2006-04-24 17:41:27 +00:00
"LC",
"LP",
"RC",
"RP",
"SEMI",
};
2006-04-26 18:43:29 +00:00
static char *token_equivs2[] =
2006-04-24 17:41:27 +00:00
{
"&",
"@",
"|",
":",
",",
"=",
"=>",
2006-04-26 18:43:29 +00:00
"break",
"case",
"catch",
2006-04-24 17:41:27 +00:00
"context",
2006-04-26 18:43:29 +00:00
"continue",
"default",
2006-04-24 17:41:27 +00:00
"else",
"eswitches",
2006-04-26 18:43:29 +00:00
"for",
2006-04-24 17:41:27 +00:00
"globals",
"goto",
"hint",
"ifTime",
"if",
"ignorepat",
"includes"
"jump",
"macro",
2006-04-26 18:43:29 +00:00
"pattern",
"regexten",
"return",
"switches",
2006-04-24 17:41:27 +00:00
"switch",
2006-04-26 18:43:29 +00:00
"while",
2006-04-24 17:41:27 +00:00
"{",
"(",
"}",
")",
";",
};
2007-09-27 00:06:06 +00:00
static char *ael_token_subst(const char *mess)
2006-04-24 17:41:27 +00:00
{
/* calc a length, malloc, fill, and return; yyerror had better free it! */
int len=0,i;
2007-09-27 00:06:06 +00:00
const char *p;
2006-04-24 17:41:27 +00:00
char *res, *s,*t;
int token_equivs_entries = sizeof(token_equivs1)/sizeof(char*);
for (p=mess; *p; p++) {
for (i=0; i<token_equivs_entries; i++) {
if ( strncmp(p,token_equivs1[i],strlen(token_equivs1[i])) == 0 )
{
len+=strlen(token_equivs2[i])+2;
p += strlen(token_equivs1[i])-1;
break;
}
}
len++;
}
2006-05-02 03:03:27 +00:00
res = calloc(1, len+1);
2006-04-24 17:41:27 +00:00
res[0] = 0;
s = res;
for (p=mess; *p;) {
int found = 0;
for (i=0; i<token_equivs_entries; i++) {
if ( strncmp(p,token_equivs1[i],strlen(token_equivs1[i])) == 0 ) {
*s++ = '\'';
for (t=token_equivs2[i]; *t;) {
*s++ = *t++;
}
*s++ = '\'';
p += strlen(token_equivs1[i]);
found = 1;
break;
}
}
if( !found )
*s++ = *p++;
}
*s++ = 0;
return res;
}
void yyerror(YYLTYPE *locp, struct parse_io *parseio, char const *s)
{
char *s2 = ael_token_subst((char *)s);
if (locp->first_line == locp->last_line) {
ast_log(LOG_ERROR, "==== File: %s, Line %d, Cols: %d-%d: Error: %s\n", my_file, locp->first_line, locp->first_column, locp->last_column, s2);
} else {
ast_log(LOG_ERROR, "==== File: %s, Line %d Col %d to Line %d Col %d: Error: %s\n", my_file, locp->first_line, locp->first_column, locp->last_line, locp->last_column, s2);
}
free(s2);
parseio->syntax_error_count++;
}
2007-08-15 19:21:27 +00:00
struct pval *npval(pvaltype type, int first_line, int last_line,
2006-04-27 17:39:55 +00:00
int first_column, int last_column)
2006-04-24 17:41:27 +00:00
{
2006-05-02 03:03:27 +00:00
pval *z = calloc(1, sizeof(struct pval));
2006-04-24 17:41:27 +00:00
z->type = type;
z->startline = first_line;
z->endline = last_line;
z->startcol = first_column;
z->endcol = last_column;
z->filename = strdup(my_file);
return z;
}
2006-04-27 18:09:20 +00:00
static struct pval *npval2(pvaltype type, YYLTYPE *first, YYLTYPE *last)
{
return npval(type, first->first_line, last->last_line,
first->first_column, last->last_column);
}
2006-04-30 12:12:39 +00:00
static struct pval *update_last(pval *obj, YYLTYPE *last)
{
obj->endline = last->last_line;
obj->endcol = last->last_column;
return obj;
}
2006-04-30 13:57:08 +00:00
/* frontend for npval to create a PV_WORD string from the given token */
static pval *nword(char *string, YYLTYPE *pos)
{
pval *p = npval2(PV_WORD, pos, pos);
if (p)
p->u1.str = string;
return p;
}
2006-08-07 12:59:47 +00:00
/* this routine adds a dad ptr to each element in the list */
static void set_dads(struct pval *dad, struct pval *child_list)
{
struct pval *t;
for(t=child_list;t;t=t->next) /* simple stuff */
t->dad = dad;
}