X-Git-Url: http://lists.indexdata.dk/cgi-bin?a=blobdiff_plain;f=src%2Fhttp.c;h=71cddf3e749332682293cd9d921f2e57a6b35a4c;hb=eeeac3dc5704fa9e9af4e51dee8b3a256b080e0e;hp=98b9489ac6be3a607301d70711d2e5e3ed8375ba;hpb=c99908a4b07dafc96ee813375971fd9b589bc551;p=pazpar2-moved-to-github.git

diff --git a/src/http.c b/src/http.c
index 98b9489..71cddf3 100644
--- a/src/http.c
+++ b/src/http.c
@@ -1,5 +1,22 @@
-/*
- * $Id: http.c,v 1.20 2007-03-31 20:27:15 marc Exp $
+/* $Id: http.c,v 1.38 2007-09-23 15:39:24 adam Exp $
+   Copyright (c) 2006-2007, Index Data.
+
+This file is part of Pazpar2.
+
+Pazpar2 is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Pazpar2 is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Pazpar2; see the file LICENSE.  If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
  */
 
 #include <stdio.h>
@@ -21,10 +38,12 @@
 #endif
 
 #include <netinet/in.h>
+#include <arpa/inet.h>
+#include <netdb.h>
 
 #include <yaz/yaz-util.h>
 #include <yaz/comstack.h>
-#include <netdb.h>
+#include <yaz/nmem.h>
 
 #include "cconfig.h"
 #include "util.h"
@@ -33,13 +52,12 @@
 #include "http.h"
 #include "http_command.h"
 
+#define MAX_HTTP_HEADER 4096
+
 static void proxy_io(IOCHAN i, int event);
-static struct http_channel *http_create(void);
+static struct http_channel *http_create(const char *addr);
 static void http_destroy(IOCHAN i);
 
-extern IOCHAN channel_list;
-extern struct parameters global_parameters;
-
 // If this is set, we proxy normal HTTP requests
 static struct sockaddr_in *proxy_addr = 0; 
 static char proxy_url[256] = "";
@@ -47,6 +65,23 @@ static char myurl[256] = "";
 static struct http_buf *http_buf_freelist = 0;
 static struct http_channel *http_channel_freelist = 0;
 
+struct http_channel_observer_s {
+    void *data;
+    void (*destroy)(void *data, struct http_channel *chan);
+    struct http_channel_observer_s *next;
+    struct http_channel *chan;
+};
+
+
+static const char *http_lookup_header(struct http_header *header,
+                                      const char *name)
+{
+    for (; header; header = header->next)
+        if (!strcasecmp(name, header->name))
+            return header->value;
+    return 0;
+}
+
 static struct http_buf *http_buf_create()
 {
     struct http_buf *r;
@@ -128,7 +163,7 @@ static struct http_buf *http_buf_bywrbuf(WRBUF wrbuf)
 
 // Non-destructively collapse chain of buffers into a string (max *len)
 // Return
-static int http_buf_peek(struct http_buf *b, char *buf, int len)
+static void http_buf_peek(struct http_buf *b, char *buf, int len)
 {
     int rd = 0;
     while (b && rd < len)
@@ -141,7 +176,14 @@ static int http_buf_peek(struct http_buf *b, char *buf, int len)
         b = b->next;
     }
     buf[rd] = '\0';
-    return rd;
+}
+
+static int http_buf_size(struct http_buf *b)
+{
+    int sz = 0;
+    for (; b; b = b->next)
+        sz += b->len;
+    return sz;
 }
 
 // Ddestructively munch up to len  from head of queue.
@@ -253,20 +295,38 @@ struct http_response *http_create_response(struct http_channel *c)
 }
 
 // Check if buf contains a package (minus payload)
-static int package_check(const char *buf)
+static int package_check(const char *buf, int sz)
 {
+    int content_len = 0;
     int len = 0;
+
     while (*buf) // Check if we have a sequence of lines terminated by an empty line
     {
-        char *b = strstr(buf, "\r\n");
+        const char *b = strstr(buf, "\r\n");
 
         if (!b)
             return 0;
 
         len += (b - buf) + 2;
         if (b == buf)
-            return len;
+        {
+            if (len + content_len <= sz)
+                return len + content_len;
+            return 0;
+        }
         buf = b + 2;
+        // following first skip of \r\n so that we don't consider Method
+        if (!strncasecmp(buf, "Content-Length:", 15))
+        {
+            const char *cp = buf+15;
+            while (*cp == ' ')
+                cp++;
+            content_len = 0;
+            while (*cp && isdigit(*cp))
+                content_len = content_len*10 + (*cp++ - '0');
+            if (content_len < 0) /* prevent negative offsets */
+                content_len = 0;
+        }
     }
     return 0;
 }
@@ -277,20 +337,20 @@ static int package_check(const char *buf)
 // other than an empty GET
 static int request_check(struct http_buf *queue)
 {
-    char tmp[4096];
+    char tmp[MAX_HTTP_HEADER];
 
-    http_buf_peek(queue, tmp, 4096);
-    return package_check(tmp);
+    http_buf_peek(queue, tmp, MAX_HTTP_HEADER-1);
+    return package_check(tmp, http_buf_size(queue));
 }
 
 struct http_response *http_parse_response_buf(struct http_channel *c, const char *buf, int len)
 {
-    char tmp[4096];
+    char tmp[MAX_HTTP_HEADER];
     struct http_response *r = http_create_response(c);
     char *p, *p2;
     struct http_header **hp = &r->headers;
 
-    if (len >= 4096)
+    if (len >= MAX_HTTP_HEADER)
         return 0;
     memcpy(tmp, buf, len);
     for (p = tmp; *p && *p != ' '; p++) // Skip HTTP version
@@ -334,23 +394,62 @@ struct http_response *http_parse_response_buf(struct http_channel *c, const char
     return r;
 }
 
-struct http_request *http_parse_request(struct http_channel *c, struct http_buf **queue,
-        int len)
+static int http_parse_arguments(struct http_request *r, NMEM nmem,
+                                const char *args)
+{
+    const char *p2 = args;
+
+    while (*p2)
+    {
+        struct http_argument *a;
+        const char *equal = strchr(p2, '=');
+        const char *eoa = strchr(p2, '&');
+        if (!equal)
+        {
+            yaz_log(YLOG_WARN, "Expected '=' in argument");
+            return -1;
+        }
+        if (!eoa)
+            eoa = equal + strlen(equal); // last argument
+        else if (equal > eoa)
+        {
+            yaz_log(YLOG_WARN, "Missing '&' in argument");
+            return -1;
+        }
+        a = nmem_malloc(nmem, sizeof(struct http_argument));
+        a->name = nmem_strdupn(nmem, p2, equal - p2);
+        a->value = nmem_strdupn(nmem, equal+1, eoa - equal - 1);
+        urldecode(a->name, a->name);
+        urldecode(a->value, a->value);
+        a->next = r->arguments;
+        r->arguments = a;
+        p2 = eoa;
+        while (*p2 == '&')
+            p2++;
+    }
+    return 0;
+}
+
+struct http_request *http_parse_request(struct http_channel *c,
+                                        struct http_buf **queue,
+                                        int len)
 {
     struct http_request *r = nmem_malloc(c->nmem, sizeof(*r));
     char *p, *p2;
-    char tmp[4096];
-    char *buf = tmp;
+    char *start = nmem_malloc(c->nmem, len+1);
+    char *buf = start;
 
-    if (len > 4096)
-        return 0;
     if (http_buf_read(queue, buf, len) < len)
+    {
+        yaz_log(YLOG_WARN, "http_buf_read < len (%d)", len);
         return 0;
-
+    }
     r->search = "";
     r->channel = c;
     r->arguments = 0;
     r->headers = 0;
+    r->content_buf = 0;
+    r->content_len = 0;
     // Parse first line
     for (p = buf, p2 = r->method; *p && *p != ' ' && p - buf < 19; p++)
         *(p2++) = *p;
@@ -380,29 +479,7 @@ struct http_request *http_parse_request(struct http_channel *c, struct http_buf
     {
         r->search = nmem_strdup(c->nmem, p2);
         // Parse Arguments
-        while (*p2)
-        {
-            struct http_argument *a;
-            char *equal = strchr(p2, '=');
-            char *eoa = strchr(p2, '&');
-            if (!equal)
-            {
-                yaz_log(YLOG_WARN, "Expected '=' in argument");
-                return 0;
-            }
-            if (!eoa)
-                eoa = equal + strlen(equal); // last argument
-            else
-                *(eoa++) = '\0';
-            a = nmem_malloc(c->nmem, sizeof(struct http_argument));
-            *(equal++) = '\0';
-            a->name = nmem_strdup(c->nmem, p2);
-            urldecode(equal, equal);
-            a->value = nmem_strdup(c->nmem, equal);
-            a->next = r->arguments;
-            r->arguments = a;
-            p2 = eoa;
-        }
+        http_parse_arguments(r, c->nmem, p2);
     }
     buf = p;
 
@@ -412,7 +489,10 @@ struct http_request *http_parse_request(struct http_channel *c, struct http_buf
     {
         buf += 5;
         if (!(p = strstr(buf, "\r\n")))
+        {
+            yaz_log(YLOG_WARN, "Did not see \\r\\n (1)");
             return 0;
+        }
         *(p++) = '\0';
         p++;
         strcpy(r->http_version, buf);
@@ -424,9 +504,15 @@ struct http_request *http_parse_request(struct http_channel *c, struct http_buf
     while (*buf)
     {
         if (!(p = strstr(buf, "\r\n")))
+        {
+            yaz_log(YLOG_WARN, "Did not see \\r\\n (2)");
             return 0;
+        }
         if (p == buf)
+        {
+            buf += 2;
             break;
+        }
         else
         {
             struct http_header *h = nmem_malloc(c->nmem, sizeof(*h));
@@ -449,6 +535,18 @@ struct http_request *http_parse_request(struct http_channel *c, struct http_buf
         }
     }
 
+    if (buf < start + len)
+    {
+        const char *content_type = http_lookup_header(r->headers,
+                                                      "Content-Type");
+        r->content_len = start + len - buf;
+        r->content_buf = buf;
+
+        if (!strcmp(content_type, "application/x-www-form-urlencoded"))
+        {
+            http_parse_arguments(r, c->nmem, r->content_buf);
+        }
+    }
     return r;
 }
 
@@ -466,6 +564,20 @@ static struct http_buf *http_serialize_response(struct http_channel *c,
         wrbuf_printf(c->wrbuf, "Content-length: %d\r\n", r->payload ?
                 (int) strlen(r->payload) : 0);
         wrbuf_printf(c->wrbuf, "Content-type: text/xml\r\n");
+        if (1)
+        {
+            xmlDoc *doc = xmlParseMemory(r->payload, strlen(r->payload));
+            if (doc)
+            {
+                xmlFreeDoc(doc);
+            }
+            else
+            {
+                yaz_log(YLOG_WARN, "Sending non-wellformed "
+                        "response (bug #1162");
+                yaz_log(YLOG_WARN, "payload: %s", r->payload);
+            }
+        }
     }
     wrbuf_puts(c->wrbuf, "\r\n");
 
@@ -480,20 +592,10 @@ static struct http_buf *http_serialize_request(struct http_request *r)
 {
     struct http_channel *c = r->channel;
     struct http_header *h;
-    struct http_argument *a;
 
     wrbuf_rewind(c->wrbuf);
-    wrbuf_printf(c->wrbuf, "%s %s", r->method, r->path);
-
-    if (r->arguments)
-    {
-        wrbuf_putc(c->wrbuf, '?');
-        for (a = r->arguments; a; a = a->next) {
-            if (a != r->arguments)
-                wrbuf_putc(c->wrbuf, '&');
-            wrbuf_printf(c->wrbuf, "%s=%s", a->name, a->value);
-        }
-    }
+    wrbuf_printf(c->wrbuf, "%s %s%s%s", r->method, r->path,
+                 *r->search ? "?" : "", r->search);
 
     wrbuf_printf(c->wrbuf, " HTTP/%s\r\n", r->http_version);
 
@@ -501,7 +603,14 @@ static struct http_buf *http_serialize_request(struct http_request *r)
         wrbuf_printf(c->wrbuf, "%s: %s\r\n", h->name, h->value);
 
     wrbuf_puts(c->wrbuf, "\r\n");
-    
+
+    if (r->content_buf)
+        wrbuf_write(c->wrbuf, r->content_buf, r->content_len);
+
+#if 0
+    yaz_log(YLOG_LOG, "WRITING TO PROXY:\n%s\n----",
+            wrbuf_cstr(c->wrbuf));
+#endif
     return http_buf_bywrbuf(c->wrbuf);
 }
 
@@ -590,17 +699,14 @@ static int http_proxy(struct http_request *rq)
         p->first_response = 1;
         c->proxy = p;
         // We will add EVENT_OUTPUT below
-        p->iochan = iochan_create(sock, 0, proxy_io, EVENT_INPUT);
+        p->iochan = iochan_create(sock, proxy_io, EVENT_INPUT);
         iochan_setdata(p->iochan, p);
-        p->iochan->next = channel_list;
-        channel_list = p->iochan;
+        pazpar2_add_channel(p->iochan);
     }
 
     // Do _not_ modify Host: header, just checking it's existence
-    for (hp = rq->headers; hp; hp = hp->next)
-        if (!strcmp(hp->name, "Host"))
-            break;
-    if (!hp)
+
+    if (!http_lookup_header(rq->headers, "Host"))
     {
         yaz_log(YLOG_WARN, "Failed to find Host header in proxy");
         return -1;
@@ -619,10 +725,11 @@ static int http_proxy(struct http_request *rq)
         sprintf(server_via,  "1.1 %s:%s (%s/%s)",  
                 ser->host, server_port, PACKAGE_NAME, PACKAGE_VERSION);
         hp = http_header_append(c, hp, "Via" , server_via);
-        hp = http_header_append(c, hp,"X-Forwarded-For", c->iochan->addr_str);
-      }
-
+        hp = http_header_append(c, hp, "X-Forwarded-For", c->addr);
+    }
+    
     requestbuf = http_serialize_request(rq);
+
     http_buf_enqueue(&p->oqueue, requestbuf);
     iochan_setflag(p->iochan, EVENT_OUTPUT);
     return 0;
@@ -671,46 +778,39 @@ static void http_io(IOCHAN i, int event)
                 http_destroy(i);
                 return;
             }
-            if (res > 0)
-            {
-                htbuf->buf[res] = '\0';
-                htbuf->len = res;
-                http_buf_enqueue(&hc->iqueue, htbuf);
-            }
+            htbuf->buf[res] = '\0';
+            htbuf->len = res;
+            http_buf_enqueue(&hc->iqueue, htbuf);
 
-            if (hc->state == Http_Busy)
-                return;
-            if ((reqlen = request_check(hc->iqueue)) <= 2)
-                return;
-
-            nmem_reset(hc->nmem);
-            if (!(hc->request = http_parse_request(hc, &hc->iqueue, reqlen)))
+            while (1)
             {
-                yaz_log(YLOG_WARN, "Failed to parse request");
-                http_destroy(i);
-                return;
-            }
-            hc->response = 0;
-            yaz_log(YLOG_LOG, "Request: %s %s%s%s", hc->request->method,
-                    hc->request->path,
-                    *hc->request->search ? "?" : "",
-                    hc->request->search);
-            if (http_weshouldproxy(hc->request))
-                http_proxy(hc->request);
-            else
-            {
-                // Execute our business logic!
-                hc->state = Http_Busy;
-                http_command(hc);
-            }
-            if (hc->iqueue)
-            {
-                yaz_log(YLOG_DEBUG, "We think we have more input to read. Forcing event");
-                iochan_setevent(i, EVENT_INPUT);
+                if (hc->state == Http_Busy)
+                    return;
+                if ((reqlen = request_check(hc->iqueue)) <= 2)
+                    return;
+                // we have a complete HTTP request
+                nmem_reset(hc->nmem);
+                if (!(hc->request = http_parse_request(hc, &hc->iqueue, reqlen)))
+                {
+                    yaz_log(YLOG_WARN, "Failed to parse request");
+                    http_destroy(i);
+                    return;
+                }
+                hc->response = 0;
+                yaz_log(YLOG_LOG, "Request: %s %s%s%s", hc->request->method,
+                        hc->request->path,
+                        *hc->request->search ? "?" : "",
+                        hc->request->search);
+                if (http_weshouldproxy(hc->request))
+                    http_proxy(hc->request);
+                else
+                {
+                    // Execute our business logic!
+                    hc->state = Http_Busy;
+                    http_command(hc);
+                }
             }
-
             break;
-
         case EVENT_OUTPUT:
             if (hc->oqueue)
             {
@@ -820,19 +920,19 @@ static void proxy_io(IOCHAN pi, int event)
                         struct http_response *res = http_parse_response_buf(hc, htbuf->buf, len);
                         if (res)
                         {
-                            struct http_header *h;
-                            for (h = res->headers; h; h = h->next)
-                                if (!strcmp(h->name, "Location"))
-                                {
-                                    // We found a location header. Rewrite it.
-                                    struct http_buf *buf;
-                                    h->value = sub_hostname(hc, h->value);
-                                    buf = http_serialize_response(hc, res);
-                                    yaz_log(YLOG_LOG, "Proxy rewrite");
-                                    http_buf_enqueue(&hc->oqueue, buf);
-                                    htbuf->offset = len;
-                                    break;
-                                }
+                            const char *location = http_lookup_header(
+                                res->header, "Location");
+                            if (location)
+                            {
+                                // We found a location header. Rewrite it.
+                                struct http_buf *buf;
+                                h->value = sub_hostname(hc, location);
+                                buf = http_serialize_response(hc, res);
+                                yaz_log(YLOG_LOG, "Proxy rewrite");
+                                http_buf_enqueue(&hc->oqueue, buf);
+                                htbuf->offset = len;
+                                break;
+                            }
                         }
                     }
                     pc->first_response = 0;
@@ -879,6 +979,9 @@ static void proxy_io(IOCHAN pi, int event)
     }
 }
 
+static void http_fire_observers(struct http_channel *c);
+static void http_destroy_observers(struct http_channel *c);
+
 // Cleanup channel
 static void http_destroy(IOCHAN i)
 {
@@ -894,13 +997,17 @@ static void http_destroy(IOCHAN i)
         http_buf_destroy_queue(s->proxy->oqueue);
         xfree(s->proxy);
     }
+    http_buf_destroy_queue(s->iqueue);
+    http_buf_destroy_queue(s->oqueue);
+    http_fire_observers(s);
+    http_destroy_observers(s);
     s->next = http_channel_freelist;
     http_channel_freelist = s;
     close(iochan_getfd(i));
     iochan_destroy(i);
 }
 
-static struct http_channel *http_create(void)
+static struct http_channel *http_create(const char *addr)
 {
     struct http_channel *r = http_channel_freelist;
 
@@ -922,6 +1029,13 @@ static struct http_channel *http_create(void)
     r->state = Http_Idle;
     r->request = 0;
     r->response = 0;
+    if (!addr)
+    {
+        yaz_log(YLOG_WARN, "Invalid HTTP forward address");
+        exit(1);
+    }
+    strcpy(r->addr, addr);
+    r->observers = 0;
     return r;
 }
 
@@ -949,14 +1063,13 @@ static void http_accept(IOCHAN i, int event)
         yaz_log(YLOG_FATAL|YLOG_ERRNO, "fcntl2");
 
     yaz_log(YLOG_DEBUG, "New command connection");
-    c = iochan_create(s, &addr, http_io, EVENT_INPUT | EVENT_EXCEPT);
-
-    ch = http_create();
+    c = iochan_create(s, http_io, EVENT_INPUT | EVENT_EXCEPT);
+    
+    ch = http_create(inet_ntoa(addr.sin_addr));
     ch->iochan = c;
     iochan_setdata(c, ch);
 
-    c->next = channel_list;
-    channel_list = c;
+    pazpar2_add_channel(c);
 }
 
 /* Create a http-channel listener, syntax [host:]port */
@@ -983,11 +1096,11 @@ void http_init(const char *addr)
 
         strncpy(hostname, addr, len);
         hostname[len] = '\0';
-        if (!(he = gethostbyname(hostname)))
-        {
+        if (!(he = gethostbyname(hostname))){
             yaz_log(YLOG_FATAL, "Unable to resolve '%s'", hostname);
             exit(1);
         }
+        
         memcpy(&myaddr.sin_addr.s_addr, he->h_addr_list[0], he->h_length);
         port = atoi(pp + 1);
     }
@@ -996,6 +1109,7 @@ void http_init(const char *addr)
         port = atoi(addr);
         myaddr.sin_addr.s_addr = INADDR_ANY;
     }
+
     myaddr.sin_port = htons(port);
 
     if (!(p = getprotobyname("tcp"))) {
@@ -1008,13 +1122,18 @@ void http_init(const char *addr)
         abort();
 
     if (bind(l, (struct sockaddr *) &myaddr, sizeof myaddr) < 0) 
+    {
         yaz_log(YLOG_FATAL|YLOG_ERRNO, "bind");
+        exit(1);
+    }
     if (listen(l, SOMAXCONN) < 0) 
+    {
         yaz_log(YLOG_FATAL|YLOG_ERRNO, "listen");
+        exit(1);
+    }
 
-    c = iochan_create(l, &myaddr, http_accept, EVENT_INPUT | EVENT_EXCEPT);
-    c->next = channel_list;
-    channel_list = c;
+    c = iochan_create(l, http_accept, EVENT_INPUT | EVENT_EXCEPT);
+    pazpar2_add_channel(c);
 }
 
 void http_set_proxyaddr(char *host, char *base_url)
@@ -1045,6 +1164,55 @@ void http_set_proxyaddr(char *host, char *base_url)
     proxy_addr->sin_port = htons(port);
 }
 
+static void http_fire_observers(struct http_channel *c)
+{
+    http_channel_observer_t p = c->observers;
+    while (p)
+    {
+        p->destroy(p->data, c);
+        p = p->next;
+    }
+}
+
+static void http_destroy_observers(struct http_channel *c)
+{
+    while (c->observers)
+    {
+        http_channel_observer_t obs = c->observers;
+        c->observers = obs->next;
+        xfree(obs);
+    }
+}
+
+http_channel_observer_t http_add_observer(struct http_channel *c, void *data,
+                                          http_channel_destroy_t des)
+{
+    http_channel_observer_t obs = xmalloc(sizeof(*obs));
+    obs->chan = c;
+    obs->data = data;
+    obs->destroy= des;
+    obs->next = c->observers;
+    c->observers = obs;
+    return obs;
+}
+
+void http_remove_observer(http_channel_observer_t obs)
+{
+    struct http_channel *c = obs->chan;
+    http_channel_observer_t found, *p = &c->observers;
+    while (*p != obs)
+        p = &(*p)->next;
+    found = *p;
+    assert(found);
+    *p = (*p)->next;
+    xfree(found);
+}
+
+struct http_channel *http_channel_observer_chan(http_channel_observer_t obs)
+{
+    return obs->chan;
+}
+
 /*
  * Local variables:
  * c-basic-offset: 4