- support for gzip files

[packages/elinks.git] / elinks-gzip-http.patch

--- elinks/src/document/session.h	Wed May 15 06:40:09 2002
+++ elinks_with_gzip/src/document/session.h	Mon May 13 13:55:41 2002
@@ -2,6 +2,15 @@
 
 #ifndef EL__DOCUMENT_SESSION_H
 #define EL__DOCUMENT_SESSION_H
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+#ifdef HAVE_SSL
+#include <openssl/ssl.h>
+#endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
 
 /* We need to declare these first :/. Damn cross-dependencies. */
 struct session;
--- elinks/src/lowlevel/sched.c	Wed May 15 06:40:15 2002
+++ elinks_with_gzip/src/lowlevel/sched.c	Mon May 13 21:42:33 2002
@@ -5,10 +5,15 @@
 #include "config.h"
 #endif
 
+
 #include <string.h>
 #ifdef HAVE_SSL
 #include <openssl/ssl.h>
 #endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
+
 #ifdef HAVE_UNISTD_H
 #include <unistd.h>
 #endif
@@ -391,6 +396,12 @@
 {
 	del_from_list(c);
 	send_connection_info(c);
+#ifdef HAVE_ZLIB_H
+	if (c->z) {
+		inflateEnd(c->z);
+		mem_free(c->z);
+	}
+#endif
 	mem_free(c->url);
 	mem_free(c);
 }
--- elinks/src/lowlevel/sched.h	Wed May 15 06:40:15 2002
+++ elinks_with_gzip/src/lowlevel/sched.h	Mon May 13 13:41:12 2002
@@ -2,7 +2,15 @@
 
 #ifndef EL__LOWLEVEL_SCHED_H
 #define EL__LOWLEVEL_SCHED_H
-
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+#ifdef HAVE_SSL
+#include <openssl/ssl.h>
+#endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
 #include "links.h" /* tcount, list_head */
 #include "document/cache.h"
 #include "lowlevel/ttime.h"
@@ -74,6 +82,10 @@
 	SSL *ssl;
 	int no_tsl;
 #endif
+#ifdef HAVE_ZLIB_H
+	int gzip;
+	z_streamp z;
+#endif
 };
 
 #define S_WAIT		0
--- elinks/src/protocol/http/http.c	Wed May 15 06:40:22 2002
+++ elinks_with_gzip/src/protocol/http/http.c	Wed May 15 06:32:28 2002
@@ -8,6 +8,9 @@
 #ifdef HAVE_SSL
 #include <openssl/ssl.h>
 #endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
 #include <stdlib.h>
 #include <string.h>
 
@@ -29,6 +32,7 @@
 #include "protocol/url.h"
 #include "util/base64.h"
 #include "util/blacklist.h"
+#include "util/compress.h"
 
 struct http_connection_info {
 	enum blacklist_flags bl_flags;
@@ -121,6 +125,7 @@
 #endif
 		}
 	}
+
 	if (c->info && !((struct http_connection_info *)c->info)->close
 #ifdef HAVE_SSL
 	&& (!c->ssl) /* We won't keep alive ssl connections */
@@ -401,7 +406,9 @@
 	}
 
 	add_to_str(&hdr, &l, "Accept: */*\r\n");
-
+#ifdef HAVE_ZLIB_H
+	add_to_str(&hdr, &l, "Accept-Encoding: gzip\r\n");
+#endif
 	if (!accept_charset) {
 		unsigned char *cs, *ac;
 		int aclen = 0;
@@ -558,9 +565,27 @@
 		int l = rb->len;
 		if (info->length >= 0 && info->length < l) l = info->length;
 		c->received += l;
-		if (add_fragment(c->cache, c->from, rb->data, l) == 1) c->tries = 0;
+		if (l) {
+#ifdef HAVE_ZLIB_H
+		    if (c->gzip) {
+			    int dl;
+			    unsigned char *data = decompress_gzip(&c->z, rb->data, l, &dl);
+			    if (!data) {
+				    setcstate(c, S_OUT_OF_MEM);
+				    abort_connection(c);
+				    return;
+			    }
+			    if (add_fragment(c->cache, c->from, data, dl) == 1) c->tries = 0;
+			    mem_free(data);
+			    c->from += dl;
+		    } else
+#endif
+		    {		
+			if (add_fragment(c->cache, c->from, rb->data, l) == 1) c->tries = 0;
+			c->from += l;
+		    }
+		}
 		if (info->length >= 0) info->length -= l;
-		c->from += l;
 		kill_buffer_data(rb, l);
 		if (!info->length && !rb->close) {
 			setcstate(c, S_OK);
@@ -604,9 +629,25 @@
 			int l = info->chunk_remaining;
 			if (l > rb->len) l = rb->len;
 			c->received += l;
-			if (add_fragment(c->cache, c->from, rb->data, l) == 1) c->tries = 0;
 			info->chunk_remaining -= l;
-			c->from += l;
+#ifdef HAVE_ZLIB_H
+			if (c->gzip) {
+				int dl;
+				unsigned char *data = decompress_gzip(&c->z, rb->data, l, &dl);
+				if (!data) {
+					setcstate(c, S_OUT_OF_MEM);
+					abort_connection(c);
+					return;
+				}
+				if (add_fragment(c->cache, c->from, data, dl) == 1) c->tries = 0;
+				mem_free(data);
+				c->from += dl;
+			} else
+#endif
+			{
+				if (add_fragment(c->cache, c->from, rb->data, l) == 1) c->tries = 0;
+				c->from += l;
+			}
 			kill_buffer_data(rb, l);
 			if (!info->chunk_remaining && rb->len >= 1) {
 				if (rb->data[0] == 10) kill_buffer_data(rb, 1);
@@ -846,6 +887,14 @@
 	if (!e->last_modified && (d = parse_http_header(e->head, "Date", NULL)))
 		e->last_modified = d;
 	if (info->length == -1 || (version < 11 && info->close)) rb->close = 1;
+#ifdef HAVE_ZLIB_H
+	d = parse_http_header(e->head, "Content-Encoding", NULL);
+	c->gzip = 0;
+	if (d) {
+		if (!strcasecmp(d, "gzip") || !strcasecmp(d, "x-gzip")) c->gzip = 1;
+		mem_free(d);
+	}
+#endif
 	read_http_data(c, rb);
 }
 
--- elinks/src/util/compress.c	Wed May 15 06:40:24 2002
+++ elinks_with_gzip/src/util/compress.c	Wed May 15 06:37:29 2002
@@ -23,7 +23,6 @@
 #include "util/compress.h"
 
 
-#if 0
 static void *
 z_mem_alloc(void *opaque, int items, int size)
 {
@@ -35,7 +34,6 @@
 {
 	mem_free(address);
 }
-#endif
 
 
 struct decoding_handlers {
@@ -111,74 +109,162 @@
 	gzclose((gzFile *) stream->data);
 }
 
-#if 0
-static unsigned char *
-decompress_gzip(unsigned char *stream, int cur_size, int *new_size)
-{
-	z_stream z;
-	char *stream_pos = stream;
-	char method, flags;
-	char *output;
-	int size;
-	int ret;
+struct decoding_handlers gzip_handlers = {
+	gzip_open,
+	gzip_read,
+	gzip_close,
+};
 
-	output = mem_alloc(cur_size * 4);
-	if (!output) return stream;
+#define WMAXBITS 15
+#define ASCII_FLAG 0x01
+#define HEAD_CRC 0x02
+#define EXTRA_FIELD 0x04
+#define ORIG_NAME 0x08
+#define COMMENT 0x10
+#define RESERVED 0xE0
 
-	z.opaque = NULL;
-	z.zalloc = (alloc_func) z_mem_alloc;
-	z.zfree = z_mem_free;
-	z.next_in = stream_pos;
-	z.next_out = output;
-	z.avail_out = size = cur_size * 4;
-	z.avail_in = cur_size + 1;
+static z_streamp gzip_init(unsigned char *buf_old, int l) {
+	
+/* check header */
+/* gzip magic */
+	unsigned char method;
+	unsigned char flags;
+	unsigned char *buf = buf_old;
+	int len;
+	int ret;
+	z_streamp z;
+	
+	if (buf[0] != 0x1f || buf[1] != 0x8b) return NULL;
+	
+	method = buf[2];
+	flags = buf[3];
+	
+	if (method != Z_DEFLATED || (flags & RESERVED) != 0) return NULL;
 
-	/* XXX: Why -15? --pasky */
-	ret = inflateInit2(&z, -15);
+/* Comments are borrowed from gzio.c - zlib */
+/* Discard time, xflags and OS code: */
+	buf += 10;
+	l -= 10;
+                               
+	if ((flags & EXTRA_FIELD) != 0) { /* skip the extra field */
+		len  =  2 + buf[0] + (buf[1] << 8);
+		buf += len;
+		l -= len;
+	}
+	if (l <= 0) return NULL; 
+	
+	if ((flags & ORIG_NAME) != 0) {/* skip the original file name */
+		len = strlen(buf) + 1;
+		buf += len;
+		l -= len;
+	} 
+	if (l <= 0) return NULL; 
+	
+	if ((flags & COMMENT) != 0) {/* skip the .gz file comment */
+		len = strlen(buf) + 1;
+		buf += len;
+		l -= len;
+	}
+	if (l <= 0) return NULL;
 
-	while (ret == Z_OK) {
-		char *output_new;
+	if ((flags & HEAD_CRC) != 0) {  /* skip the header crc */
+		buf += 2;
+		l -= 2;
+	}
+	if (l <= 0) return NULL;
 
-		ret = inflate(&z, Z_SYNC_FLUSH);
+/* initialization of z_stream */
+	z = (z_streamp)mem_alloc(sizeof(z_stream));
+	if (!z) return NULL;
+	
+	z->opaque = NULL;
+	z->zalloc = (alloc_func)z_mem_alloc;
+	z->zfree = (free_func)z_mem_free;
+	z->avail_in = l;
+	z->next_in = buf;
+/* windowBits is passed < 0 to tell that there is no zlib header.
+ * Note that in this case inflate *requires* an extra "dummy" byte
+ * after the compressed stream in order to complete decompression and
+ * return Z_STREAM_END. Here the gzip CRC32 ensures that 4 bytes are
+ * present after the compressed stream.
+ */
+	ret = inflateInit2(z, -WMAXBITS);
+	if (ret == Z_OK) return z;
+	
+	mem_free(z);
+	return NULL;
+}
 
-		if (ret == Z_STREAM_END) {
-			mem_free(stream);
-			*new_size = (int) z.total_out;
-			output = mem_realloc(output, z.total_out);
-			inflateEnd(&z);
-			return output;
-		}
+#define OUTPUT_BUFFER_SIZE 65536
 
-		if (ret != Z_OK) {
-			inflateEnd(&z);
-			break;
+unsigned char *decompress_gzip(z_streamp *z, unsigned char *buf, int l, int *dl)
+{
+	unsigned char *output;
+	int cur_size;
+	int new_size;
+	int ret;
+	
+	if (!*z) {
+		*z = gzip_init(buf, l);
+		if (!*z) {
+			*dl = -1;
+			return NULL;
 		}
+	}
+	else {
+		(*z)->next_in = buf;
+		(*z)->avail_in = l;
+	}
 
-		size += cur_size * 4;
-
-		output_new = mem_realloc(output, size);
-		if (!output_new) {
-			mem_free(output);
-			inflateEnd(&z);
-			return stream;
+	(*z)->total_out = 0L;
+	cur_size = OUTPUT_BUFFER_SIZE;
+	output = mem_alloc(cur_size); /* output will be freed in http_read_data */
+	if (!output) {
+		inflateEnd(*z);
+		mem_free(*z);
+		*z = NULL;
+		*dl = -2;
+		return NULL;
+	}
+	
+	(*z)->next_out = output;
+	(*z)->avail_out = 65536;
+	
+	ret = inflate(*z, Z_SYNC_FLUSH); 
+	while (ret == Z_OK) {
+		if (!(*z)->avail_in) {
+			*dl = (int)(*z)->total_out;
+			return output;
 		}
-
-		output = output_new;
-		z.avail_out += cur_size * 4;
-		z.next_out = output + z.total_out;
+		
+		new_size = cur_size + OUTPUT_BUFFER_SIZE;
+		output = mem_realloc(output, new_size);
+		if (!output) {
+			inflateEnd(*z);
+			mem_free(*z);
+			*z = NULL;
+			*dl = -3;
+			return NULL;
+		}
+		
+		(*z)->next_out = output + cur_size; /* assume that z->avail_out == 0 */
+		(*z)->avail_out = OUTPUT_BUFFER_SIZE;
+		cur_size = new_size;
+		ret = inflate(*z, Z_SYNC_FLUSH);
 	}
 
-	mem_free(output);
+	if (ret == Z_STREAM_END) *dl = (int)(*z)->total_out;
+	else { /* something went wrong */
+		*dl = -4;
+		mem_free(output);
+		output = NULL;
+	}
 	
-	return stream;
+	inflateEnd(*z);
+	mem_free(*z);
+	*z = NULL;
+	return output;
 }
-#endif
-
-struct decoding_handlers gzip_handlers = {
-	gzip_open,
-	gzip_read,
-	gzip_close,
-};
 
 #endif
 
--- elinks/src/util/compress.h	Wed May 15 06:40:24 2002
+++ elinks_with_gzip/src/util/compress.h	Wed May 15 01:35:11 2002
@@ -3,6 +3,13 @@
 #ifndef EL__UTIL_COMPRESS_H
 #define EL__UTIL_COMPRESS_H
 
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
+
 enum stream_encoding {
 	ENCODING_NONE,
 	ENCODING_GZIP,
@@ -17,5 +24,8 @@
 struct stream_encoded *open_encoded(int, enum stream_encoding);
 int read_encoded(struct stream_encoded *, unsigned char *, int);
 void close_encoded(struct stream_encoded *);
+#ifdef HAVE_ZLIB_H
+unsigned char *decompress_gzip(z_streamp *, unsigned char *, int, int *);
+#endif
 
 #endif