From 82f47550fe3327cce6f2e0e1bf62e81d9ebcf90c Mon Sep 17 00:00:00 2001
From: sanine <sanine.not@pm.me>
Date: Wed, 18 Jan 2023 13:03:34 -0600
Subject: begin reentrant refactor

---
 yy/kalmia.l | 92 +++++++++++++++++++++++++++++++++++++++++++------------------
 1 file changed, 65 insertions(+), 27 deletions(-)

(limited to 'yy/kalmia.l')

diff --git a/yy/kalmia.l b/yy/kalmia.l
index 2472bdc..916f576 100644
--- a/yy/kalmia.l
+++ b/yy/kalmia.l
@@ -1,26 +1,39 @@
-%{
-#include <stdio.h>
-#include "y.tab.h"
-
-char *copy_str(char *);
-int line_num;
+%option noinput nounput noyywrap 8bit nodefault
+%option reentrant bison-bridge bison-locations
+%start STRING
 
+%{
+#include <string.h>
+#include "kalmia.tab.h"
+#define YY_USER_ACTION \
+	yylloc->first_line = yylloc->last_line; \
+	yylloc->first_column = yylloc->last_column; \
+	if (*yytext == '\n') { \
+		yylloc->last_line += 1; \
+		yylloc->last_column = 0; \
+	} \
+	else { \
+		yylloc->last_column += yyleng; \
+	}
 %}
 
 S    \x20\x0a\x0d\x09
 DATE  [0-9]{4}\-[0-9]{2}\-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}[A-Z]*
+ATTR ([\x20\x0a\x0d\x09]*)?"="
 
 
 
 
-
-%start STRING
-
 %%
 
-<INITIAL>"<?xml".*"?>"      { return PROLOG; }
 
-<INITIAL>[\x20\x0a\x0d\x09]                { if (*yytext == '\n') { line_num += 1; } }
+<INITIAL>"<?xml".*"?>" {
+	/* xml prologue */
+	return PROLOG;
+}
+<INITIAL>[\x20\x0a\x0d\x09] {
+	/* ignore whitespace */
+}
 
 
 <INITIAL>"<"                { return S_TAG_OPEN; }
@@ -29,24 +42,49 @@ DATE  [0-9]{4}\-[0-9]{2}\-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}[A-Z]*
 <INITIAL>"/>"               { return EMPTY_TAG_CLOSE; }
 
 
-<INITIAL>[a-zA-Z_:][a-zA-Z0-9\.\-_:]* { yylval.sval = copy_str(yytext); return NAME; }
-
-<INITIAL>"=" { return *yytext; }
+<INITIAL>[a-zA-Z_:][a-zA-Z0-9\.\-_:]* { 
+	/* generic tag name */
+	yylval->sval = strdup(yytext); return NAME; 
+}
+<INITIAL>[a-zA-Z_:][a-zA-Z0-9\.\-_:]*/([\x20\x0a\x0d\x09])*"=" {
+	/* generic attribute key */
+	yylval->sval = strdup(yytext);
+	return ATTR;
+}
+<INITIAL>"=" {
+	/* attribute "=" */
+	return *yytext;
+}
 
-<INITIAL>"\"" { BEGIN(STRING); return *yytext; }
-<STRING>"\"" { BEGIN(INITIAL); return *yytext; }
-<STRING>[^"]  { yylval.cval = *yytext; return CHAR; }
 
-<INITIAL>-?[0-9]+ { yylval.lval = strtol(yytext, NULL, 10); return INTEGER; }
-<INITIAL>-?[0-9]+\.?[0-9]* { yylval.dval = strtod(yytext, NULL); return DOUBLE; }
-<INITIAL>{DATE} { return DATE; }
+<INITIAL>"\"" { 
+	/* begin a string */
+	BEGIN(STRING);
+	return *yytext;
+}
+<STRING>"\"" {
+	/* end a string */
+	BEGIN(INITIAL);
+	return *yytext;
+}
+<STRING>[^"]+  {
+	/* within a string */
+	yylval->sval = strdup(yytext);
+	return TEXT;
+}
 
-%%
 
-char * copy_str(char *str)
-{
-	size_t len = strlen(str) + 1;
-	char *copy = malloc(sizeof(char) * len);
-	strncpy(copy, str, len);
-	return copy;
+<INITIAL>-?[0-9]+ {
+	/* integers */
+	yylval->lval = strtol(yytext, NULL, 10);
+	return INTEGER;
+}
+<INITIAL>-?[0-9]+\.?[0-9]* {
+	/* doubles */
+	yylval->dval = strtod(yytext, NULL);
+	return DOUBLE;
+}
+<INITIAL>{DATE} {
+	/* dates */
+	return DATE;
 }
-- 
cgit v1.2.1