Changeset ca8037e

Timestamp:

2013-06-09T21:17:45Z (11 years ago)

Author:

Wilmer van der Gaast <wilmer@…>

Branches:

master

Children:

777461b

Parents:

41a94dd

Message:

Add better handling of HTTP/1.1 and/or keepalive connections. This should
let me close #641, and more importantly, prepares the Twitter module for
an upcoming API change. ​https://dev.twitter.com/blog/deprecating-http-1.0-streaming-api

Location:

lib

Files:

• http_client.c (modified) (15 diffs)
• http_client.h (modified) (2 diffs)
• oauth.c (modified) (1 diff)
• oauth2.c (modified) (1 diff)
• proxy.c (modified) (1 diff)

lib/http_client.c

@@ -2 +2 @@
   * BitlBee -- An IRC to other IM-networks gateway                     *
   *                                                                    *
-  * Copyright 2002-2012 Wilmer van der Gaast and others                *
+  * Copyright 2002-2013 Wilmer van der Gaast and others                *
   \********************************************************************/
 
@@ -69 +69 @@
         req->request_length = strlen( request );
         req->redir_ttl = 3;
+        req->content_length = -1;
         
         if( getenv( "BITLBEE_DEBUG" ) )
@@ -96 +97 @@
         request = g_strdup_printf( "GET %s HTTP/1.0\r\n"
                                    "Host: %s\r\n"
-                                   "Connection: close\r\n"
                                    "User-Agent: BitlBee " BITLBEE_VERSION " " ARCH "/" CPU "\r\n"
                                    "\r\n", url->file, url->host );
@@ -193 +193 @@
 }
 
+typedef enum {
+        CR_OK,
+        CR_EOF,
+        CR_ERROR,
+        CR_ABORT,
+} http_ret_t;
+
 static gboolean http_handle_headers( struct http_request *req );
+static http_ret_t http_process_chunked_data( struct http_request *req, const char *buffer, int len );
+static http_ret_t http_process_data( struct http_request *req, const char *buffer, int len );
 
 static gboolean http_incoming_data( gpointer data, int source, b_input_condition cond )
@@ -199 +208 @@
         struct http_request *req = data;
         char buffer[4096];
-        char *s;
-        size_t content_length;
         int st;
         
@@ -244 +251 @@
         }
         
-        if( st > 0 && !req->sbuf )
-        {
-                req->reply_headers = g_realloc( req->reply_headers, req->bytes_read + st + 1 );
-                memcpy( req->reply_headers + req->bytes_read, buffer, st );
-                req->bytes_read += st;
-                
-                st = 0;
-        }
-        
-        if( st >= 0 && ( req->flags & HTTPC_STREAMING ) )
-        {
-                if( !req->reply_body &&
-                    ( strstr( req->reply_headers, "\r\n\r\n" ) ||
-                      strstr( req->reply_headers, "\n\n" ) ) )
-                {
-                        size_t hlen;
-                        
-                        /* We've now received all headers, so process them once
-                           before we start feeding back data. */
-                        if( !http_handle_headers( req ) )
-                                return FALSE;
-                        
-                        hlen = req->reply_body - req->reply_headers;
-                        
-                        req->sblen = req->bytes_read - hlen;
-                        req->sbuf = g_memdup( req->reply_body, req->sblen + 1 );
-                        req->reply_headers = g_realloc( req->reply_headers, hlen + 1 );
-                        
-                        req->reply_body = req->sbuf;
-                }
-                
-                if( st > 0 )
-                {
-                        int pos = req->reply_body - req->sbuf;
-                        req->sbuf = g_realloc( req->sbuf, req->sblen + st + 1 );
-                        memcpy( req->sbuf + req->sblen, buffer, st );
-                        req->bytes_read += st;
-                        req->sblen += st;
-                        req->sbuf[req->sblen] = '\0';
-                        req->reply_body = req->sbuf + pos;
-                        req->body_size = req->sblen - pos;
-                }
-                
-                if( req->reply_body )
-                        req->func( req );
-        }
+        if( st > 0 )
+        {
+                http_ret_t c;
+                
+                if( req->flags & HTTPC_CHUNKED )
+                        c = http_process_chunked_data( req, buffer, st );
+                else
+                        c = http_process_data( req, buffer, st );
+                
+                if( c == CR_EOF )
+                        goto eof;
+                else if( c == CR_ERROR || c == CR_ABORT )
+                        return FALSE;
+        }
+        
+        if( req->content_length != -1 &&
+            req->body_size >= req->content_length )
+                goto eof;
         
         if( ssl_pending( req->ssl ) )
@@ -311 +290 @@
                 goto cleanup;
         }
-        
-        if( !( req->flags & HTTPC_STREAMING ) )
-        {
-                /* Returns FALSE if we were redirected, in which case we should abort
-                   and not run any callback yet. */
-                if( !http_handle_headers( req ) )
-                        return FALSE;
-        }
 
 cleanup:
@@ -326 +297 @@
                 closesocket( req->fd );
         
-        if( ( s = get_rfc822_header( req->reply_headers, "Content-Length", 0 ) ) &&
-            sscanf( s, "%zd", &content_length ) == 1 )
-        {
-                if( content_length < req->body_size )
-                {
-                        req->status_code = -1;
-                        g_free( req->status_string );
-                        req->status_string = g_strdup( "Response truncated" );
-                }
-        }
-        g_free( s );
+        if( req->body_size < req->content_length )
+        {
+                req->status_code = -1;
+                g_free( req->status_string );
+                req->status_string = g_strdup( "Response truncated" );
+        }
         
         if( getenv( "BITLBEE_DEBUG" ) && req )
@@ -347 +313 @@
 }
 
+static http_ret_t http_process_chunked_data( struct http_request *req, const char *buffer, int len )
+{
+        char *chunk, *eos, *s;
+        
+        if( len < 0 )
+                return TRUE;
+        
+        if( len > 0 )
+        {
+                req->cbuf = g_realloc( req->cbuf, req->cblen + len + 1 );
+                memcpy( req->cbuf + req->cblen, buffer, len );
+                req->cblen += len;
+                req->cbuf[req->cblen] = '\0';
+        }
+        
+        /* Turns out writing a proper chunked-encoding state machine is not
+           that simple. :-( */
+        chunk = req->cbuf;
+        eos = req->cbuf + req->cblen;
+        while( TRUE )
+        {
+                int clen = 0;
+                
+                /* Might be a \r\n from the last chunk. */
+                s = chunk;
+                while( isspace( *s ) )
+                        s ++;
+                /* Chunk length. Might be incomplete. */
+                if( s < eos && sscanf( s, "%x", &clen ) != 1 )
+                        return CR_ERROR;
+                while( isxdigit( *s ) )
+                        s ++;
+                
+                /* If we read anything here, it *must* be \r\n. */
+                if( strncmp( s, "\r\n", MIN( 2, eos - s ) ) != 0 )
+                        return CR_ERROR;
+                s += 2;
+                
+                if( s >= eos )
+                        break;
+                
+                /* 0-length chunk means end of response. */     
+                if( clen == 0 )
+                        return CR_EOF;
+                
+                if( s + clen > eos )
+                        break;
+                if( http_process_data( req, s, clen ) != CR_OK )
+                        return CR_ABORT;
+                
+                chunk = s + clen;
+        }
+        
+        if( chunk != req->cbuf )
+        {
+                req->cblen = eos - chunk;
+                s = g_memdup( chunk, req->cblen + 1 );
+                g_free( req->cbuf );
+                req->cbuf = s;
+        }
+        
+        return CR_OK;
+}
+
+static http_ret_t http_process_data( struct http_request *req, const char *buffer, int len )
+{
+        if( len <= 0 )
+                return CR_OK;
+        
+        if( !req->reply_body )
+        {
+                req->reply_headers = g_realloc( req->reply_headers, req->bytes_read + len + 1 );
+                memcpy( req->reply_headers + req->bytes_read, buffer, len );
+                req->bytes_read += len;
+                req->reply_headers[req->bytes_read] = '\0';
+                
+                if( strstr( req->reply_headers, "\r\n\r\n" ) ||
+                    strstr( req->reply_headers, "\n\n" ) )
+                {
+                        /* We've now received all headers. Look for something
+                           interesting. */
+                        if( !http_handle_headers( req ) )
+                                return CR_ABORT;
+                        
+                        /* Start parsing the body as chunked if required. */
+                        if( req->flags & HTTPC_CHUNKED )
+                                return http_process_chunked_data( req, NULL, 0 );
+                }
+        }
+        else
+        {
+                int pos = req->reply_body - req->sbuf;
+                req->sbuf = g_realloc( req->sbuf, req->sblen + len + 1 );
+                memcpy( req->sbuf + req->sblen, buffer, len );
+                req->bytes_read += len;
+                req->sblen += len;
+                req->sbuf[req->sblen] = '\0';
+                req->reply_body = req->sbuf + pos;
+                req->body_size = req->sblen - pos;
+        }
+        
+        if( ( req->flags & HTTPC_STREAMING ) && req->reply_body )
+                req->func( req );
+        
+        return CR_OK;
+}
+
 /* Splits headers and body. Checks result code, in case of 300s it'll handle
    redirects. If this returns FALSE, don't call any callbacks! */
 static gboolean http_handle_headers( struct http_request *req )
 {
-        char *end1, *end2;
+        char *end1, *end2, *s;
         int evil_server = 0;
         
@@ -377 +450 @@
         }
         
-        *end1 = 0;
+        *end1 = '\0';
         
         if( getenv( "BITLBEE_DEBUG" ) )
@@ -387 +460 @@
                 req->reply_body = end1 + 2;
         
-        req->body_size = req->reply_headers + req->bytes_read - req->reply_body;
+        /* Separately allocated space for headers and body. */
+        req->sblen = req->body_size = req->reply_headers + req->bytes_read - req->reply_body;
+        req->sbuf = req->reply_body = g_memdup( req->reply_body, req->body_size + 1 );
+        req->reply_headers = g_realloc( req->reply_headers, end1 - req->reply_headers + 1 );
         
         if( ( end1 = strchr( req->reply_headers, ' ' ) ) != NULL )
@@ -452 +528 @@
                            don't need this yet anyway, I won't implement it. */
                         
-                        req->status_string = g_strdup( "Can't handle recursive redirects" );
+                        req->status_string = g_strdup( "Can't handle relative redirects" );
                         
                         return TRUE;
@@ -508 +584 @@
                         
                         /* Okay, this isn't fun! We have to rebuild the request... :-( */
-                        new_request = g_strdup_printf( "%s %s HTTP/1.0\r\nHost: %s%s",
+                        new_request = g_strdup_printf( "%s %s HTTP/1.1\r\nHost: %s%s",
                                                        new_method, url->file, url->host, s );
                         
@@ -564 +640 @@
                 return FALSE;
         }
+
+        if( ( s = get_rfc822_header( req->reply_headers, "Content-Length", 0 ) ) &&
+            sscanf( s, "%d", &req->content_length ) != 1 )
+                req->content_length = -1;
+        g_free( s );
+        
+        if( ( s = get_rfc822_header( req->reply_headers, "Transfer-Encoding", 0 ) ) )
+        {
+                if( strcasestr( s, "chunked" ) )
+                {
+                        req->flags |= HTTPC_CHUNKED;
+                        req->cbuf = req->sbuf;
+                        req->cblen = req->sblen;
+                        
+                        req->reply_body = req->sbuf = g_strdup( "" );
+                        req->body_size = req->sblen = 0;
+                }
+                g_free( s );
+        }
         
         return TRUE;
@@ -607 +702 @@
         g_free( req->status_string );
         g_free( req->sbuf );
+        g_free( req->cbuf );
         g_free( req );
 }

lib/http_client.h

@@ -42 +42 @@
         HTTPC_STREAMING = 1,
         HTTPC_EOF = 2,
+        HTTPC_CHUNKED = 4,
         
         /* Let's reserve 0x1000000+ for lib users. */
@@ -77 +78 @@
         int bytes_written;
         int bytes_read;
+        int content_length;     /* "Content-Length:" header or -1 */
         
         /* Used in streaming mode. Caller should read from reply_body. */
         char *sbuf;
         size_t sblen;
+        
+        /* Chunked encoding only. Raw chunked stream is decoded from here. */
+        char *cbuf;
+        size_t cblen;
 };
 

lib/oauth.c

@@ -262 +262 @@
                              "Content-Type: application/x-www-form-urlencoded\r\n"
                              "Content-Length: %zd\r\n"
-                             "Connection: close\r\n"
                              "\r\n"
                              "%s", url_p.file, url_p.host, strlen( post ), post );

lib/oauth2.c

@@ -96 +96 @@
                              "Content-Type: application/x-www-form-urlencoded\r\n"
                              "Content-Length: %zd\r\n"
-                             "Connection: close\r\n"
                              "\r\n"
                              "%s", url_p.file, url_p.host, strlen( args_s ), args_s );

lib/proxy.c

@@ -158 +158 @@
                 }
 
-                event_debug("proxy_connect_none( \"%s\", %d ) = %d\n", host, port, fd);
+                event_debug("proxy_connect_none( \"%s\", %d ) = %d\n", host, port_, fd);
         
                 if (connect(fd, phb->gai_cur->ai_addr, phb->gai_cur->ai_addrlen) < 0 && !sockerr_again()) {