ObjFW  Check-in [8f4d7a5b74]

/*
 * Copyright (c) 2008 - 2009
 *   Jonathan Schleifer <js@webkeks.org>
 *
 * All rights reserved.
 *
 * This file is part of libobjfw. It may be distributed under the terms of the
 * Q Public License 1.0, which can be found in the file LICENSE included in
 * the packaging of this file.
 */

#include "config.h"


#include <stdlib.h>
#include <string.h>
#include <ctype.h>

#import "OFURLEncoding.h"
#import "OFExceptions.h"

						    andSize: (length * 3) + 1];

	for (i = 0; *s != '\0'; s++) {
		if (isalnum(*s) || *s == '-' || *s == '_' || *s == '.' ||
		    *s == '~')
			ret_c[i++] = *s;
		else {
			uint8_t high, low;

			high = *s >> 4;
			low = *s & 0x0F;

			ret_c[i++] = '%';
			ret_c[i++] = (high > 9 ? high - 10 + 'A' : high + '0');
			ret_c[i++] = (low  > 9 ? low  - 10 + 'A' : low  + '0');
		}
	}
	ret_c[i] = '\0';

	@try {
		ret = [OFString stringWithCString: ret_c];
	} @finally {

				ret_c[i++] = ' ';
			else
				ret_c[i++] = *s;
			break;
		case 1:
		case 2:
			if (*s >= '0' && *s <= '9')
				c += (*s - '0') << (st == 1 ? 4 : 0);
			else if (*s >= 'A' && *s <= 'F')
				c += (*s - 'A' + 10) << (st == 1 ? 4 : 0);
			else if (*s >= 'a' && *s <= 'f')
				c += (*s - 'a' + 10) << (st == 1 ? 4 : 0);
			else {
				free(ret_c);
				@throw [OFInvalidEncodingException
				    newWithClass: isa];
			}

			if (++st == 3) {

- (BOOL)xmlParser: (OFXMLParser*)parser
      foundString: (OFString*)string;
@end

@interface OFXMLParser: OFObject
{
	OFObject <OFXMLParserDelegate> *delegate;
	enum {
		OF_XMLPARSER_OUTSIDE_TAG,
		OF_XMLPARSER_TAG_OPENED,
		OF_XMLPARSER_IN_TAG_NAME,
		OF_XMLPARSER_IN_CLOSE_TAG_NAME,
		OF_XMLPARSER_IN_TAG,
		OF_XMLPARSER_IN_ATTR_NAME,
		OF_XMLPARSER_EXPECT_DELIM,
		OF_XMLPARSER_IN_ATTR_VALUE,
		OF_XMLPARSER_EXPECT_CLOSE,
		OF_XMLPARSER_EXPECT_SPACE_OR_CLOSE
	} state;
	OFString *cache;
	OFString *name;
	OFString *prefix;
	OFString *ns;
	OFDictionary *attrs;
	OFString *attr_name;
	char delim;
}

+ xmlParser;
- (id)delegate;
- setDelegate: (OFObject <OFXMLParserDelegate>*)delegate;
- parseBuffer: (const char*)buf
     withSize: (size_t)size;
@end

@protocol OFXMLUnescapingDelegate
- (OFString*)foundUnknownEntityNamed: (OFString*)entitiy;
@end

@interface OFString (OFXMLUnescaping)
- stringByXMLUnescaping;
- stringByXMLUnescapingWithHandler: (OFObject <OFXMLUnescapingDelegate>*)h;
@end

			return nil;

		entity++;
		length--;

		for (i = 0; i < length; i++) {
			if (entity[i] >= '0' && entity[i] <= '9')
				c = (c << 4) + (entity[i] - '0');
			else if (entity[i] >= 'A' && entity[i] <= 'F')
				c = (c << 4) + (entity[i] - 'A' + 10);
			else if (entity[i] >= 'a' && entity[i] <= 'f')
				c = (c << 4) + (entity[i] - 'A' + 10);
			else
				return nil;
		}
	} else {
		for (i = 0; i < length; i++) {
			if (entity[i] >= '0' && entity[i] <= '9')
				c = (c * 10) + (entity[i] - '0');

}

@implementation OFXMLParser
+ xmlParser
{
	return [[[self alloc] init] autorelease];
}

- init
{
	self = [super init];

	@try {
		cache = [[OFMutableString alloc] init];
	} @catch (OFException *e) {
		/* We can't use [super dealloc] on OS X here. Compiler bug? */
		[self dealloc];
		@throw e;
	}

	return self;
}

- (void)dealloc
{
	[cache release];
	[name release];
	[prefix release];
	[ns release];
	[attrs release];
	[attr_name release];

	[super dealloc];
}

- (id)delegate
{
	return [[delegate retain] autorelease];
}

- setDelegate: (OFObject <OFXMLParserDelegate>*)delegate_
{
	[delegate release];
	delegate = [delegate_ retain];

	return self;
}

- parseBuffer: (const char*)buf
     withSize: (size_t)size
{
	OFAutoreleasePool *pool;
	size_t i, last, len;

	last = 0;

	for (i = 0; i < size; i++) {
		switch (state) {
		/* Not in a tag */
		case OF_XMLPARSER_OUTSIDE_TAG:
			if (buf[i] == '<') {
				len = i - last;

				if (len > 0)
					[cache appendCString: buf + last
						  withLength: len];

				if ([cache length] > 0) {
					OFString *str;

					pool = [[OFAutoreleasePool alloc] init];
					str = [cache stringByXMLUnescaping];

					[delegate xmlParser: self
						foundString: str];

					[pool release];
				}

				[cache setToCString: ""];

				last = i + 1;
				state = OF_XMLPARSER_TAG_OPENED;
			}
			break;

		/* Tag was just opened */
		case OF_XMLPARSER_TAG_OPENED:
			if (buf[i] == '/') {
				last = i + 1;
				state = OF_XMLPARSER_IN_CLOSE_TAG_NAME;
			} else {
				state = OF_XMLPARSER_IN_TAG_NAME;
				i--;
			}
			break;

		/* Inside a tag, no name yet */
		case OF_XMLPARSER_IN_TAG_NAME:
			if (buf[i] == ' ' || buf[i] == '>' || buf[i] == '/') {
				const char *cache_c, *tmp;
				size_t cache_len;

				len = i - last;
				if (len > 0)
					[cache appendCString: buf + last
						  withLength: len];
				cache_c = [cache cString];
				cache_len = [cache length];

				if ((tmp = memchr(cache_c, ':',
				    cache_len)) != NULL) {
					prefix = [[OFString alloc]
					    initWithCString: cache_c
						  andLength: tmp - cache_c];
					name = [[OFString alloc]
					    initWithCString: tmp + 1
						  andLength: cache_len - (tmp -
							     cache_c) - 1];
				} else {
					prefix = nil;
					name = [[OFString alloc]
					    initWithCString: cache_c
						  andLength: cache_len];
				}

				[cache setToCString: ""];

				if (buf[i] == '>' || buf[i] == '/') {
					pool = [[OFAutoreleasePool alloc] init];

					[delegate xmlParser: self
					didStartTagWithName: name
						  andPrefix: prefix
					       andNamespace: ns
					      andAttributes: nil];

					if (buf[i] == '/')
						[delegate xmlParser: self
						  didEndTagWithName: name
							  andPrefix: prefix
						       andNamespace: ns];

					[pool release];

					[name release];
					[prefix release];
					[ns release];

					state = (buf[i] == '/'
					    ? OF_XMLPARSER_EXPECT_CLOSE
					    : OF_XMLPARSER_OUTSIDE_TAG);
				} else
					state = OF_XMLPARSER_IN_TAG;

				last = i + 1;
			}
			break;

		/* Inside a close tag, no name yet */
		case OF_XMLPARSER_IN_CLOSE_TAG_NAME:
			if (buf[i] == ' ' || buf[i] == '>') {
				const char *cache_c, *tmp;
				size_t cache_len;

				len = i - last;
				if (len > 0)
					[cache appendCString: buf + last
						  withLength: len];
				cache_c = [cache cString];
				cache_len = [cache length];

				if ((tmp = memchr(cache_c, ':',
				    cache_len)) != NULL) {
					prefix = [[OFString alloc]
					    initWithCString: cache_c
						  andLength: tmp - cache_c];
					name = [[OFString alloc]
					    initWithCString: tmp + 1
						  andLength: cache_len - (tmp -
							     cache_c) - 1];
				} else {
					prefix = nil;
					name = [[OFString alloc]
					    initWithCString: cache_c
						  andLength: cache_len];
				}

				[cache setToCString: ""];

				pool = [[OFAutoreleasePool alloc] init];

				[delegate xmlParser: self
				  didEndTagWithName: name
					  andPrefix: prefix
				       andNamespace: ns];

				[pool release];

				[name release];
				[prefix release];
				[ns release];

				last = i + 1;
				state = (buf[i] == ' '
				    ? OF_XMLPARSER_EXPECT_SPACE_OR_CLOSE
				    : OF_XMLPARSER_OUTSIDE_TAG);
			}
			break;

		/* Inside a tag, name found */
		case OF_XMLPARSER_IN_TAG:
			if (buf[i] == '>' || buf[i] == '/') {
				pool = [[OFAutoreleasePool alloc] init];

				[delegate xmlParser: self
				didStartTagWithName: name
					  andPrefix: prefix
				       andNamespace: ns
				      andAttributes: attrs];

				if (buf[i] == '/')
					[delegate xmlParser: self
					  didEndTagWithName: name
						  andPrefix: prefix
					       andNamespace: ns];

				[pool release];

				[name release];
				[prefix release];
				[ns release];
				[attrs release];

				attrs = nil;

				last = i + 1;
				state = (buf[i] == '/'
				    ? OF_XMLPARSER_EXPECT_CLOSE
				    : OF_XMLPARSER_OUTSIDE_TAG);
			} else if (buf[i] != ' ') {
				last = i;
				state = OF_XMLPARSER_IN_ATTR_NAME;
				i--;
			}
			break;

		/* Looking for attribute name */
		case OF_XMLPARSER_IN_ATTR_NAME:
			if (buf[i] == '=') {
				len = i - last;
				if (len > 0)
					[cache appendCString: buf + last
						  withLength: len];

				attr_name = [cache copy];
				[cache setToCString: ""];

				last = i + 1;
				state = OF_XMLPARSER_EXPECT_DELIM;
			}
			break;

		/* Expecting delimiter */
		case OF_XMLPARSER_EXPECT_DELIM:
			if (buf[i] != '\'' && buf[i] != '"')
				@throw [OFInvalidEncodingException
				    newWithClass: isa];

			delim = buf[i];
			last = i + 1;
			state = OF_XMLPARSER_IN_ATTR_VALUE;
			break;

		/* Looking for attribute value */
		case OF_XMLPARSER_IN_ATTR_VALUE:
			if (buf[i] == delim) {
				len = i - last;
				if (len > 0)
					[cache appendCString: buf + last
						  withLength: len];

				if (attrs == nil)
					attrs =
					    [[OFMutableDictionary alloc] init];

				pool = [[OFAutoreleasePool alloc] init];
				[attrs setObject: [cache stringByXMLUnescaping]
					  forKey: attr_name];
				[pool release];

				[cache setToCString: ""];
				[attr_name release];

				last = i + 1;
				state = OF_XMLPARSER_IN_TAG;
			}
			break;

		/* Expecting closing '>' */
		case OF_XMLPARSER_EXPECT_CLOSE:
			if (buf[i] == '>') {
				last = i + 1;
				state = OF_XMLPARSER_OUTSIDE_TAG;
			} else
				@throw [OFInvalidEncodingException
				    newWithClass: isa];
			break;

		/* Expecting closing '>' or space */
		case OF_XMLPARSER_EXPECT_SPACE_OR_CLOSE:
			if (buf[i] == '>') {
				last = i + 1;
				state = OF_XMLPARSER_OUTSIDE_TAG;
			} else if (buf[i] != ' ')
				@throw [OFInvalidEncodingException
				    newWithClass: isa];
			break;
		}
	}

	len = size - last;
	/* In OF_XMLPARSER_IN_TAG, there can be only spaces */
	if (len > 0 && state != OF_XMLPARSER_IN_TAG)
		[cache appendCString: buf + last
			  withLength: len];

	return self;
}
@end

@implementation OFString (OFXMLUnescaping)
- stringByXMLUnescaping
{

	  OFHashes		\
	  ${OFPLUGIN}		\
	  OFString		\
	  OFTCPSocket		\
	  OFThread		\
	  OFList		\
	  OFXMLElement		\
	  OFXMLParser		\
	  ${OBJC_SYNC}

include ../buildsys.mk