[gentoo-commits] proj/libbash:master commit in: bashast/, test/, bashast/gunit/

["Petteri Räty" <betelgeuse@gentoo.org>]

commit:     842c390d16c128f5271f193b26bd35a0a5c97ef2
Author:     Mu Qiao <qiaomuf <AT> gentoo <DOT> org>
AuthorDate: Fri Apr  8 13:53:15 2011 +0000
Commit:     Petteri Räty <betelgeuse <AT> gentoo <DOT> org>
CommitDate: Sat Apr  9 13:02:52 2011 +0000
URL:        http://git.overlays.gentoo.org/gitweb/?p=proj/libbash.git;a=commit;h=842c390d

Transform ESC_CHAR token into a parser rule

ESC_CHAR results in huge lexer file size. After the transformation,
the size of lexer is reduced by 1/3.

---
 bashast/bashast.g         |   16 +++++++++-------
 bashast/gunit/fname.gunit |   10 +++++-----
 test/post_check.cpp       |    2 +-
 3 files changed, 15 insertions(+), 13 deletions(-)

diff --git a/bashast/bashast.g b/bashast/bashast.g
index 6230b66..356142a 100644
--- a/bashast/bashast.g
+++ b/bashast/bashast.g
@@ -363,10 +363,11 @@ ns_str_part
 ns_str_part_no_res
 	:	num
 	|	name
+	|	esc_char
 	|OTHER|EQUALS|PCT|PCTPCT|MINUS|DOT|DOTDOT|COLON|TEST_EXPR|'_'
 	|TILDE|MUL_ASSIGN|DIVIDE_ASSIGN|MOD_ASSIGN|PLUS_ASSIGN|MINUS_ASSIGN
 	|TIME_POSIX|LSHIFT_ASSIGN|RSHIFT_ASSIGN|AND_ASSIGN|XOR_ASSIGN
-	|OR_ASSIGN|ESC_CHAR|CARET;
+	|OR_ASSIGN|CARET;
 //strings with no slashes, used in certain variable expansions
 ns_str	:	ns_str_part* -> ^(STRING ns_str_part*);
 //Generic strings/filenames.
@@ -493,6 +494,7 @@ parens	:	LPAREN BLANK* RPAREN;
 name	:	NAME
 	|	LETTER
 	|	'_';
+esc_char:	ESC (DIGIT DIGIT? DIGIT?|LETTER ALPHANUM ALPHANUM?|.);
 
 //****************
 // TOKENS/LEXER RULES
@@ -602,17 +604,17 @@ LOGICOR	:	'||';
 EXPORT	:	'export';
 //Tokens for strings
 CONTINUE_LINE
-	:	('\\' EOL)+{$channel=HIDDEN;};
+	:	(ESC EOL)+{$channel=HIDDEN;};
 ESC_RPAREN
-	:	'\\' RPAREN;
+	:	ESC RPAREN;
 ESC_LPAREN
-	:	'\\' LPAREN;
-ESC_LT	:	'\\''<';
-ESC_GT	:	'\\''>';
+	:	ESC LPAREN;
+ESC_LT	:	ESC'<';
+ESC_GT	:	ESC'>';
 //For pipeline
 TIME_POSIX
 	:	'-p';
 //Handle ANSI C escaped characters: escaped octal, escaped hex, escaped ctrl+ chars, then all others
-ESC_CHAR:	'\\' (('0'..'7')('0'..'7')('0'..'7')?|'x'('0'..'9'|'a'..'f'|'A'..'F')('0'..'9'|'a'..'f'|'A'..'F')?|'c'.|.);
+ESC	:	'\\';
 NAME	:	(LETTER|'_')(ALPHANUM|'_')+;
 OTHER	:	.;

diff --git a/bashast/gunit/fname.gunit b/bashast/gunit/fname.gunit
index 0272c37..e783504 100644
--- a/bashast/gunit/fname.gunit
+++ b/bashast/gunit/fname.gunit
@@ -42,11 +42,11 @@ fname:
 "ab@(gh|i)" -> (STRING ab (MATCH_EXACTLY_ONE (STRING gh) (STRING i)))
 "ab!(gh|i)" -> (STRING ab (MATCH_NONE (STRING gh) (STRING i)))
 "\"abc\"\'\"\'\"def\"" -> (STRING (DOUBLE_QUOTED_STRING abc) (SINGLE_QUOTED_STRING ") (DOUBLE_QUOTED_STRING def))
-"my\ name\ is" -> (STRING my \  name \  is)
-"octal\007" -> (STRING octal \007)
-"hex\xaF" -> (STRING hex \xaF)
-"ctrlx\cx" -> (STRING ctrlx \cx)
-"tab\\ttab" -> "(STRING tab \\\t tab)"
+"my\ name\ is" -> (STRING my \   name \   is)
+"octal\007" -> (STRING octal \ 007)
+"hex\xaF" -> (STRING hex \ xaF)
+"ctrlx\cx" -> (STRING ctrlx \ cx)
+"tab\\ttab" -> "(STRING tab \\ \t tab)"
 "abc[def]" -> (STRING abc (MATCH_PATTERN def))
 "a[]" -> (STRING a [ ])
 "ab[d-h]" -> (STRING ab (MATCH_PATTERN d - h))

diff --git a/test/post_check.cpp b/test/post_check.cpp
index 1385290..ebe0409 100644
--- a/test/post_check.cpp
+++ b/test/post_check.cpp
@@ -36,7 +36,7 @@ static void check_file_size(const char *path, uintmax_t size_max)
 
 TEST(post_check, lexer_size)
 {
-  check_file_size("libbashLexer.c", 1024 * 1024);
+  check_file_size("libbashLexer.c", 512 * 1024);
 }
 
 TEST(post_check, parser_size)