discover/grub2: Initial grub2 lexer

Add a simple flex lexer to tokenise our grub2 config files.

Signed-off-by: Jeremy Kerr <jk@ozlabs.org>

3 files changed, 186 insertions, 0 deletions

diff --git a/discover/grub2/grub2.h b/discover/grub2/grub2.h
new file mode 100644
index 0000000..f2a780d
--- /dev/null
+++ b/discover/grub2/grub2.h
@@ -0,0 +1,10 @@
+#ifndef GRUB2_H
+#define GRUB2_H
+
+struct grub2_parser {
+	void	*scanner;
+};
+
+
+#endif /* GRUB2_H */
+
diff --git a/discover/grub2/lexer.l b/discover/grub2/lexer.l
new file mode 100644
index 0000000..18d1667
--- /dev/null
+++ b/discover/grub2/lexer.l
@@ -0,0 +1,127 @@
+
+%{
+#include "parser.h"
+#include <talloc/talloc.h>
+%}
+
+%option nounput noinput
+%option warn
+%option noyywrap
+%option stack noyy_top_state
+%option reentrant
+%option bison-bridge
+%option noyyalloc noyyfree noyyrealloc
+%option extra-type="struct grub2_parser *"
+%option header-file="lexer.h"
+%option outfile="lexer.c"
+
+%x sqstring
+%x dqstring
+
+WORD	[^{}|&$;<> \t\n'"]+
+VARNAME ([[:alpha:]][_[:alnum:]]*|[0-9]|[\?@\*#])
+
+%%
+
+ /* discard leading & trailing whitespace, but keep inter-word delimeters */
+^[ \t]+	;
+[ \t]+$	;
+[ \t]*	return TOKEN_DELIM;
+
+ /* reserved words */
+"[["         return TOKEN_LDSQBRACKET;
+"]]"         return TOKEN_RDSQBRACKET;
+"case"       return TOKEN_CASE;
+"do"         return TOKEN_DO;
+"done"       return TOKEN_DONE;
+"elif"       return TOKEN_ELIF;
+"esac"       return TOKEN_ESAC;
+"fi"         return TOKEN_FI;
+"for"        return TOKEN_FOR;
+"function"   return TOKEN_FUNCTION;
+"if"         return TOKEN_IF;
+"in"         return TOKEN_IN;
+"menuentry"  return TOKEN_MENUENTRY;
+"select"     return TOKEN_SELECT;
+"then"       return TOKEN_THEN;
+"time"       return TOKEN_TIME;
+"until"      return TOKEN_UTIL;
+"while"      return TOKEN_WHILE;
+
+ /* anything that's not a metachar: return as a plain word */
+{WORD}	{
+		yylval->strval = talloc_strdup(yyscanner, yytext);
+		yylval->expand = 0;
+		return TOKEN_WORD;
+	}
+
+\${VARNAME} |
+\$\{{VARNAME}\} {
+		yylval->strval = talloc_strdup(yyscanner, yytext);
+		yylval->expand = 1;
+		yylval->split = 1;
+		return TOKEN_WORD;
+	}
+
+ /* single-quoted strings: return a single, non-expanded word token */
+\'	{
+		yy_push_state(sqstring, yyscanner);
+	}
+<sqstring>\' {
+		yy_pop_state(yyscanner);
+		return TOKEN_WORD;
+	}
+<sqstring>[^']+ {
+		yylval->expand = 0;
+		yylval->split = 0;
+		yylval->strval = talloc_strdup(yyscanner, yytext);
+	}
+
+ /* double-quoted strings: return a single, expanded word token */
+\"	{
+		yy_push_state(dqstring, yyscanner);
+	}
+<dqstring>\" {
+		yy_pop_state(yyscanner);
+		return TOKEN_WORD;
+	}
+<dqstring>([^"]|\\\")+ {
+		yylval->expand = 1;
+		yylval->split = 0;
+		yylval->strval = talloc_strdup(yyscanner, yytext);
+	}
+
+
+ /* blocks */
+"{"	return '{';
+"}"	return '}';
+
+ /* end-of-line */
+[ \t]*(;|\n)[ \t]*	return TOKEN_EOL;
+
+ /* strip comments */
+#.*$	;
+
+
+.	printf("unknown token '%s'\n", yytext); exit(1);
+
+%%
+
+struct grub2_parser;
+
+void *yyalloc(size_t bytes, void *yyscanner)
+{
+	struct grub2_parser *parser = yyget_extra(yyscanner);
+	return talloc_size(parser, bytes);
+}
+
+void *yyrealloc(void *ptr, size_t bytes, void *yyscanner)
+{
+	struct grub2_parser *parser = yyget_extra(yyscanner);
+	return talloc_realloc_size(parser, ptr, bytes);
+}
+
+void yyfree(void *ptr, void *yyscanner __attribute__((unused)))
+{
+	talloc_free(ptr);
+}
diff --git a/discover/grub2/parser.y b/discover/grub2/parser.y
new file mode 100644
index 0000000..2af2132
--- /dev/null
+++ b/discover/grub2/parser.y
@@ -0,0 +1,49 @@
+
+%pure-parser
+%lex-param { yyscan_t scanner }
+%parse-param { struct grub2_parser *parser }
+
+%{
+#include "lexer.h"
+%}
+
+%union {
+	struct {
+		char	*strval;
+		int	expand;
+		int	split;
+	};
+}
+
+/* reserved words */
+%token	TOKEN_LDSQBRACKET	"[["
+%token	TOKEN_RDSQBRACKET	"]]"
+%token	TOKEN_CASE		"case"
+%token	TOKEN_DO		"do"
+%token	TOKEN_DONE		"done"
+%token	TOKEN_ELIF		"elif"
+%token	TOKEN_ESAC		"esac"
+%token	TOKEN_FI		"fi"
+%token	TOKEN_FOR		"for"
+%token	TOKEN_FUNCTION		"function"
+%token	TOKEN_IF		"if"
+%token	TOKEN_IN		"in"
+%token	TOKEN_MENUENTRY		"menuentry"
+%token	TOKEN_SELECT		"select"
+%token	TOKEN_THEN		"then"
+%token	TOKEN_TIME		"time"
+%token	TOKEN_UTIL		"until"
+%token	TOKEN_WHILE		"while"
+
+/* syntax */
+%token	TOKEN_EOL
+%token	TOKEN_DELIM
+%token	TOKEN_WORD
+
+%start	script
+
+%%
+
+script: /* empty */
+
+%%