Merge pull request #38 from gnustep/nsstring_and_nscharacterset_changes

NSString and NSCharacterSet changes

ChangeLog

@@ -1,4 +1,16 @@
+2019-04-12 Gregory John Casamento <greg.casamento@gmail.com>
+
+	* Source/NSCharacterSet.m: Added declarations/implementations for
+	URL*AllowedCharacterSet.
+	* Headers/NSCharacterSet.h: Added declarations here for
+	URL*AllowedCharacterSet
+	* Source/CharSets/*: Added headers to define new charactersets.
+	* Source/NSString.m: Added methods stringByRemovingPercentEncoding/
+	stringByAddingPercentEncodingWithAllowedCharacterSet:
+	* Header/NSString.h: Added declrations for methods mentioned above.
+
 2019-03-29 Igor Zhukov <fsb4000@yandex.ru>
+
 	* Source/NSDateFormatter.m:
 	* configure:
 	* configure.ac:

Headers/Foundation/NSCharacterSet.h

@@ -98,6 +98,36 @@ extern "C" {
  * 0x000A and 0x000D and nextline 0x0085 character.
  */
 + (id) newlineCharacterSet;
+
+/**
+ * Returns allowed characers for URL fragment component.
+ */
++ (id) URLFragmentAllowedCharacterSet;
+
+/**
+ * Returns allowed characers for URL host component.
+ */
++ (id) URLHostAllowedCharacterSet;
+
+/**
+ * Returns allowed characers for URL password component.
+ */
++ (id) URLPasswordAllowedCharacterSet;
+
+/**
+ * Returns allowed characers for URL path component.
+ */
++ (id) URLPathAllowedCharacterSet;
+
+/**
+ * Returns allowed characers for URL query component.
+ */
++ (id) URLQueryAllowedCharacterSet;
+
+/**
+ * Returns allowed characers for URL USER component.
+ */
++ (id) URLUserAllowedCharacterSet;
 #endif
 
 /**

Headers/Foundation/NSString.h

@@ -802,6 +802,11 @@ typedef NSUInteger NSStringEncodingConversionOptions;
 - (const char *)UTF8String;
 #endif
 
+#if OS_API_VERSION(MAC_OS_X_VERSION_10_9,GS_API_LATEST)
+- (NSString *) stringByAddingPercentEncodingWithAllowedCharacters: (NSCharacterSet *)aSet;
+- (NSString *) stringByRemovingPercentEncoding;
+#endif
+
 #if OS_API_VERSION(MAC_OS_X_VERSION_10_3,GS_API_LATEST) 
 /** Not implemented */
 - (void) getParagraphStart: (NSUInteger *)startIndex

Source/NSCharacterSet.m

@@ -49,6 +49,12 @@
 #undef	GNUSTEP_INDEX_CHARSET
 
 #import "NSCharacterSetData.h"
+#import "CharSets/URLFragmentAllowedCharSet.h"
+#import "CharSets/URLHostAllowedCharSet.h"
+#import "CharSets/URLPasswordAllowedCharSet.h"
+#import "CharSets/URLPathAllowedCharSet.h"
+#import "CharSets/URLQueryAllowedCharSet.h"
+#import "CharSets/URLUserAllowedCharSet.h"
 
 #define GSUNICODE_MAX	1114112
 #define GSBITMAP_SIZE	8192
@@ -537,7 +543,7 @@
 
 
 /* A simple array for caching standard bitmap sets */
-#define MAX_STANDARD_SETS 15
+#define MAX_STANDARD_SETS 21
 static NSCharacterSet *cache_set[MAX_STANDARD_SETS];
 static Class abstractClass = nil;
 static Class abstractMutableClass = nil;
@@ -833,6 +839,48 @@ static Class concreteMutableClass = nil;
     return nil;
 }
 
++ (id) URLFragmentAllowedCharacterSet
+{
+  return [self _staticSet: urlFragmentAllowedCharSet
+		   length: sizeof(urlFragmentAllowedCharSet)
+		   number: 15]; 
+}
+
++ (id) URLPasswordAllowedCharacterSet
+{
+  return [self _staticSet: urlPasswordAllowedCharSet
+		   length: sizeof(urlPasswordAllowedCharSet)
+		   number: 16];
+}
+
++ (id) URLPathAllowedCharacterSet
+{
+  return [self _staticSet: urlPathAllowedCharSet
+		   length: sizeof(urlPathAllowedCharSet)
+		   number: 17];
+}
+
++ (id) URLQueryAllowedCharacterSet
+{
+  return [self _staticSet: urlQueryAllowedCharSet
+		   length: sizeof(urlQueryAllowedCharSet)
+		   number: 18];
+}
+
++ (id) URLUserAllowedCharacterSet
+{
+  return [self _staticSet: urlUserAllowedCharSet
+		   length: sizeof(urlUserAllowedCharSet)
+		   number: 19];
+}
+
++ (id) URLHostAllowedCharacterSet
+{
+  return [self _staticSet: urlHostAllowedCharSet
+		   length: sizeof(urlHostAllowedCharSet)
+		   number: 20];
+}
+
 - (NSData*) bitmapRepresentation
 {
   BOOL		(*imp)(id, SEL, unichar);
@@ -1118,6 +1166,36 @@ static Class concreteMutableClass = nil;
   return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
 }
 
++ (id) URLFragmentAllowedCharacterSet;
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
++ (id) URLHostAllowedCharacterSet;
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
++ (id) URLPasswordAllowedCharacterSet;
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
++ (id) URLPathAllowedCharacterSet;
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
++ (id) URLQueryAllowedCharacterSet;
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
++ (id) URLUserAllowedCharacterSet
+{
+  return AUTORELEASE([[abstractClass performSelector: _cmd] mutableCopy]);
+}
+
 + (id) characterSetWithCharactersInString: (NSString*)aString
 {
   NSMutableCharacterSet	*ms;

Source/NSString.m

@@ -276,6 +276,42 @@ GSPathHandling(const char *mode)
     }
 }
 
+// Test if the digit is hex..
+inline int ishex(int x)
+{
+  return
+    (x >= '0' && x <= '9')	||
+    (x >= 'a' && x <= 'f')	||
+    (x >= 'A' && x <= 'F');
+}
+
+// URL decoding...
+int urldecode(const char *s, char *dec)
+{
+	char *o;
+	const char *end = s + strlen(s);
+	int c;
+	int cx; 
+	for (o = dec; s <= end; o++)
+	  {
+	    c = *s++;
+	    cx = c;  // preserve original value if we are not to decode
+	    if (c == '%' && (!ishex(*s++)	||
+			     !ishex(*s++)	||
+			     !sscanf(s - 2, "%2x", &c)))
+	      {
+		c = cx; // reset c to original value.
+		s--;
+	      }
+	    if (dec)
+	      {
+		*o = c;
+	      }
+	  }
+	
+	return o - dec;
+}
+
 #define	GSPathHandlingRight()	\
   ((pathHandling == PH_DO_THE_RIGHT_THING) ? YES : NO)
 #define	GSPathHandlingUnix()	\
@@ -1886,6 +1922,70 @@ GSICUCollatorOpen(NSStringCompareOptions mask, NSLocale *locale)
     }
 }
 
+- (NSString *) stringByAddingPercentEncodingWithAllowedCharacters: (NSCharacterSet *)aSet
+{
+  NSData	*data = [self dataUsingEncoding: NSUTF8StringEncoding];
+  NSString	*s = nil;
+
+  if (data != nil)
+    {
+      unsigned char	*src = (unsigned char*)[data bytes];
+      unsigned int	slen = [data length];
+      unsigned char	*dst;
+      unsigned int	spos = 0;
+      unsigned int	dpos = 0;
+
+      dst = (unsigned char*)NSZoneMalloc(NSDefaultMallocZone(), slen * 3);
+      while (spos < slen)
+	{
+	  unichar	c = src[spos++];
+	  unsigned int	hi;
+	  unsigned int	lo;
+
+	  if([aSet characterIsMember: c]) // if the character is in the allowed set, put it in
+	    {
+	      dst[dpos++] = c;
+	    }
+	  else // if not, then encode it...
+	    {
+	      dst[dpos++] = '%';
+	      hi = (c & 0xf0) >> 4;
+	      dst[dpos++] = (hi > 9) ? 'A' + hi - 10 : '0' + hi;
+	      lo = (c & 0x0f);
+	      dst[dpos++] = (lo > 9) ? 'A' + lo - 10 : '0' + lo;
+	    }
+	}
+      s = [[NSString alloc] initWithBytes: dst
+				   length: dpos
+				 encoding: NSUTF8StringEncoding];
+      NSZoneFree(NSDefaultMallocZone(), dst);
+      IF_NO_GC([s autorelease];)
+    }
+  return s;
+}
+
+- (NSString *) stringByRemovingPercentEncoding
+{
+  NSData *data = [self dataUsingEncoding: NSUTF8StringEncoding];
+  NSString *result = nil;
+  char *s = (char *)[data bytes];
+  NSUInteger slen = 0; 
+  char *o = NULL;
+
+  // Allocate memory...
+  slen = strlen(s);
+  o = (char *)NSZoneMalloc(NSDefaultMallocZone(), slen);
+
+  // Decode...
+  urldecode(s,o);
+
+  // Free up temporary space...
+  result = [NSString stringWithCString: o encoding: NSUTF8StringEncoding];
+  NSZoneFree(NSDefaultMallocZone(), o);
+  
+  return result; 
+}
+
 /**
  * Constructs a new ASCII string which is a representation of the receiver
  * in which characters are escaped where necessary in order to produce a

Tests/base/NSCharacterSet/general.m

@@ -88,7 +88,52 @@ int main()
        ![theSet characterIsMember: '#'],
        "Check custom set");
 
+  theSet = [NSCharacterSet URLFragmentAllowedCharacterSet];
+  NSString *setString = @"0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ?/:@-._~!$&'()*+,=";
+  NSUInteger i = 0;
+  BOOL p = YES;
+  for(i = 0; i < [setString length]; i++)
+    {
+      char c = [setString characterAtIndex: i];
+      if([theSet characterIsMember: c] == NO) // , [msg cStringUsingEncoding: NSUTF8Encoding]);
+	{
+	  NSLog(@"%c is not in set", c);
+	  PASS(NO,"URLFragmentAllowedCharacterSet char not found");
+	  p = NO;
+	}
+    }
 
+  if(!p)
+    {
+      PASS(YES, "URLFragmentAllowedCharacterSet passwed");
+    }
+  /*
+      PASS(//[theSet characterIsMember: 'A'] &&
+       //[theSet characterIsMember: 'Z'] &&
+       //[theSet characterIsMember: 'a'] &&
+       //[theSet characterIsMember: 'z'] &&
+       //[theSet characterIsMember: '9'] &&
+       // [theSet characterIsMember: '0'] &&
+       [theSet characterIsMember: '?'] &&
+       [theSet characterIsMember: '/ '] &&
+       [theSet characterIsMember: ':'] &&
+       [theSet characterIsMember: '@ '] &&
+       [theSet characterIsMember: '-'] &&
+       [theSet characterIsMember: '.'] &&
+       [theSet characterIsMember: '_'] &&
+       [theSet characterIsMember: '~'] &&
+       [theSet characterIsMember: '!'] &&
+       [theSet characterIsMember: '$ '] &&
+       [theSet characterIsMember: '&'] &&
+       // [theSet characterIsMember: '\''] &&
+       [theSet characterIsMember: '('] &&
+       [theSet characterIsMember: ')'] &&
+       [theSet characterIsMember: '*'] &&
+       [theSet characterIsMember: '+'] &&
+       [theSet characterIsMember: ','] &&
+       [theSet characterIsMember: ';'] &&
+       [theSet characterIsMember: '='],
+       "Check some characters from URLFramgmentAllowedCharacterSet"); */
     
   [arp release]; arp = nil;
   return 0;

Tests/base/NSString/test09.m

@@ -0,0 +1,59 @@
+#import "ObjectTesting.h"
+#import <Foundation/NSAutoreleasePool.h>
+#import <Foundation/NSString.h>
+#import <Foundation/NSCharacterSet.h>
+
+BOOL testUrlCharacterSetEncoding(NSString* decodedString, NSString* encodedString, NSCharacterSet* allowedCharacterSet)
+{
+  NSLog(@"String by adding percent");
+  NSString* testString = [decodedString stringByAddingPercentEncodingWithAllowedCharacters:allowedCharacterSet];
+  NSLog(@"String by adding percent, done. test=%@ decoded=%@", testString, decodedString);
+  return [encodedString isEqualToString: testString];
+}
+
+int main (int argc, const char * argv[])
+{
+  NSAutoreleasePool * pool = [[NSAutoreleasePool alloc] init];
+
+  NSLog(@"Test1");
+  NSString *urlDecodedString = @"Only alphabetic characters should be allowed and not encoded. !@#$%^&*()_+-=";
+  NSString *urlEncodedString =
+    @"Only%20alphabetic%20characters%20should%20be%20allowed%20and%20not%20encoded%2E%20%21%40%23%24%25%5E%26%2A%28%29%5F%2B%2D%3D";
+  NSCharacterSet *allowedCharacterSet = [NSCharacterSet alphanumericCharacterSet];
+  PASS(testUrlCharacterSetEncoding(urlDecodedString, urlEncodedString, allowedCharacterSet), "alphanumericCharacterSet");  
+       
+  NSLog(@"Test2");
+  urlDecodedString = @"https://www.microsoft.com/en-us/!@#$%^&*()_";
+  urlEncodedString = @"https://www.microsoft.com/en-us/!@%23$%25%5E&*()_";
+  allowedCharacterSet = [NSCharacterSet URLFragmentAllowedCharacterSet];
+  PASS(testUrlCharacterSetEncoding(urlDecodedString, urlEncodedString, allowedCharacterSet), "fragmentCharacterSet");  
+  
+  NSLog(@"Test3");
+  urlDecodedString = @"All alphabetic characters should be encoded. Symbols should not be: !@#$%^&*()_+-=";
+  urlEncodedString = @"%41%6C%6C %61%6C%70%68%61%62%65%74%69%63 %63%68%61%72%61%63%74%65%72%73 %73%68%6F%75%6C%64 %62%65 "
+    @"%65%6E%63%6F%64%65%64. %53%79%6D%62%6F%6C%73 %73%68%6F%75%6C%64 %6E%6F%74 %62%65: !@#$%^&*()_+-=";
+  allowedCharacterSet = [[NSCharacterSet alphanumericCharacterSet] invertedSet];
+  PASS(testUrlCharacterSetEncoding(urlDecodedString, urlEncodedString, allowedCharacterSet), "inverted");  
+
+  NSLog(@"Test4");
+  urlDecodedString = @"Here are some Emojis: \U0001F601 \U0001F602 \U0001F638 Emojis done."; // Multibyte encoded characters
+  urlEncodedString = @"Here%20are%20some%20Emojis:%20%F0%9F%98%81%20%F0%9F%98%82%20%F0%9F%98%B8%20Emojis%20done.";
+  allowedCharacterSet = [NSCharacterSet URLFragmentAllowedCharacterSet];
+  PASS(testUrlCharacterSetEncoding(urlDecodedString, urlEncodedString, allowedCharacterSet), "fragmentCharacterSet emojis");  
+
+  NSLog(@"Test5");
+  urlDecodedString = @"\1";
+  urlEncodedString = @"%01";
+  allowedCharacterSet = [NSCharacterSet alphanumericCharacterSet];
+  PASS(testUrlCharacterSetEncoding(urlDecodedString, urlEncodedString, allowedCharacterSet), "alphanumericCharacterSet");  
+
+  NSLog(@"Test5");
+  urlDecodedString = @"All alphabetic characters should be encoded. Symbols should not be: !@#$%^&*()_+-=";
+  urlEncodedString = @"%41%6C%6C %61%6C%70%68%61%62%65%74%69%63 %63%68%61%72%61%63%74%65%72%73 %73%68%6F%75%6C%64 %62%65 "
+    @"%65%6E%63%6F%64%65%64. %53%79%6D%62%6F%6C%73 %73%68%6F%75%6C%64 %6E%6F%74 %62%65: !@#$%^&*()_+-=";
+  NSString *result = [urlEncodedString stringByRemovingPercentEncoding];
+  PASS([urlDecodedString isEqualToString: result], "stringByRemovingPercentEncoding");
+  NSLog(@"Result = \"%@\",\ndecodedString = \"%@\",\nencodedString = \"%@\"", result, urlDecodedString, urlEncodedString);
+  [pool drain];
+  return 0;
+}