From deba4c0c91696b431bd95e0536d32c621b36a57b Mon Sep 17 00:00:00 2001
From: oliskoli <oliskoli>
Date: Thu, 4 Sep 2008 20:49:39 +0000
Subject: [PATCH] gbfile: Add support for reading UNICODE text files (using
 gbfgetstr).

---
 gbfile.c | 70 +++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 gbfile.h |  8 ++++++-
 2 files changed, 76 insertions(+), 2 deletions(-)

diff --git a/gbfile.c b/gbfile.c
index 678093d84..73f76e7cb 100644
--- a/gbfile.c
+++ b/gbfile.c
@@ -257,7 +257,7 @@ gbfread(void *buf, const gbsize_t size, const gbsize_t members, gbfile *file)
 		
 		/* Check for an incomplete READ */
 		if ((members == 1) && (size > 1) && (result > 0) && (result < (int)size))
-			fatal("%s: Unexpected end of files (EOF)!\n", file->module);
+			fatal("%s: Unexpected end of file (EOF)!\n", file->module);
 
 		result /= size;
 
@@ -750,6 +750,46 @@ gbfgetpstr(gbfile *file)
 	return result;
 }
 
+static char *
+gbfgetucs2str(gbfile *file)
+{
+	int len = 0;
+	char *result = file->line;
+	
+	for (;;) {
+		char buff[8];
+		int clen;
+		int c = gbfgetc(file);
+		
+		if ((c == EOF) && (len == 0)) return NULL;
+		
+		c = c | (gbfgetc(file) << 8);
+		if (file->big_endian) c = be_read16(&c);
+
+		if (c == '\r') {
+			c = gbfgetc(file) | (gbfgetc(file) << 8);
+			if (file->big_endian) c = be_read16(&c);
+			if (c != '\n')
+				fatal("%s: Invalid unicode (UCS-2/%s endian) line break!\n", 
+					file->module,
+					file->big_endian ? "Big" : "Little");
+			break;
+		}
+		
+		clen = cet_ucs4_to_utf8(buff, sizeof(buff), c);
+
+		if (len+clen >= file->linesz) {
+			file->linesz += 64;
+			result = file->line = xrealloc(file->line, file->linesz + 1);
+		}
+		memcpy(&result[len], buff, clen);
+		len += clen;
+	}
+	result[len] = '\0';	// terminate resulting string
+	
+	return result;
+}
+
 /*
  * gbfgetstr: Reads a string from file (util any type of line-breaks or eof or error)
  *            except xfree and free you can do all possible things with the result
@@ -761,6 +801,8 @@ gbfgetstr(gbfile *file)
 	int len = 0;
 	char *result = file->line;
 	
+	if (file->unicode) return gbfgetucs2str(file);
+	
 	for (;;) {
 		char c = gbfgetc(file);
 		
@@ -887,4 +929,30 @@ gbfputpstr(const char *s, gbfile *file)
 	return (len + 1);
 }
 
+int
+gbfunicode(gbfile *file)
+{
+	if (! file->unicode_checked) {
+		int c;
+		size_t pos;
+		
+		file->unicode_checked = 1;
+		
+		pos = gbftell(file);
+		gbfrewind(file);
+		c = gbfgetc(file) | (gbfgetc(file) << 8);
+		
+		if (c == 0xFEFF) file->big_endian = 0;
+		else if (c == 0xFFFE) file->big_endian = 1;
+		else {
+			gbfseek(file, pos, SEEK_SET);
+			return 0;
+		}
+		file->unicode = 1;
+		if (pos != 0) gbfseek(file, pos, SEEK_SET);
+	}
+	return file->unicode;
+}
+
+
 /* Thats all, sorry. */
diff --git a/gbfile.h b/gbfile.h
index 545ebfc68..662282ad9 100644
--- a/gbfile.h
+++ b/gbfile.h
@@ -23,9 +23,12 @@
 #ifndef GBFILE_H
 #define GBFILE_H
 
+#include <ctype.h>
+#include <stdarg.h>
+#include <string.h>
 #include "config.h"
 #include "defs.h"
-#include <stdarg.h>
+#include "cet.h"
 
 typedef struct gbfile_s {
 #ifdef DEBUG_MEM
@@ -48,6 +51,8 @@ typedef struct gbfile_s {
 	unsigned char big_endian:1;
 	unsigned char binary:1;
 	unsigned char gzapi:1;
+	unsigned char unicode:1;
+	unsigned char unicode_checked:1;
 } gbfile;
 
 
@@ -95,4 +100,5 @@ int gbfputflt(const float f, gbfile *file);	// write a float value
 int gbfputcstr(const char *s, gbfile *file);	// write string including '\0'
 int gbfputpstr(const char *s, gbfile *file);	// write as pascal string
 
+int gbfunicode(gbfile *file);
 #endif
-- 
2.30.2