+Implemented bug #806: Deal with HTTP clients sending LF in HTTP headers.
+
+Added ZOOM_connection_peek_event.
+
Implemented Generic select hook for ZOOM (bug #803). This is achieved with
the following new functions:
ZOOM_process_event,
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
-/* $Id: comstack.h,v 1.25 2007-01-03 08:42:14 adam Exp $ */
+/* $Id: comstack.h,v 1.26 2007-01-11 10:30:40 adam Exp $ */
/**
* \file comstack.h
YAZ_EXPORT int cs_set_ssl_certificate_file(COMSTACK cs, const char *fname);
YAZ_EXPORT int cs_get_peer_certificate_x509(COMSTACK cs, char **buf, int *len);
YAZ_EXPORT void cs_set_max_recv_bytes(COMSTACK cs, int max_recv_bytes);
+YAZ_EXPORT int cs_complete_http(const char *buf, int len);
/*
* error management.
* Copyright (C) 1995-2007, Index Data ApS
* See the file LICENSE for details.
*
- * $Id: comstack.c,v 1.17 2007-01-03 08:42:15 adam Exp $
+ * $Id: comstack.c,v 1.18 2007-01-11 10:30:41 adam Exp $
*/
/**
return cs->event;
}
-#define CHUNK_DEBUG 0
-int cs_complete_auto(const unsigned char *buf, int len)
+static int skip_crlf(const char *buf, int len, int *i)
{
- if (len > 5 && buf[0] >= 0x20 && buf[0] < 0x7f
- && buf[1] >= 0x20 && buf[1] < 0x7f
- && buf[2] >= 0x20 && buf[2] < 0x7f)
+ if (*i < len)
{
- /* deal with HTTP request/response */
- int i = 2, content_len = 0, chunked = 0;
+ if (buf[*i] == '\r' && *i < len-1 && buf[*i + 1] == '\n')
+ {
+ (*i) += 2;
+ return 1;
+ }
+ else if (buf[*i] == '\n')
+ {
+ (*i)++;
+ return 1;
+ }
+ }
+ return 0;
+}
- /* if dealing with HTTP responses - then default
- content length is unlimited (socket close) */
- if (!memcmp(buf, "HTTP/", 5))
- content_len = -1;
+#define CHUNK_DEBUG 0
- while (i <= len-4)
+int cs_complete_http(const char *buf, int len)
+{
+ /* deal with HTTP request/response */
+ int i = 2, content_len = 0, chunked = 0;
+
+ if (len < 6)
+ return 0;
+
+ /* if dealing with HTTP responses - then default
+ content length is unlimited (socket close) */
+ if (!memcmp(buf, "HTTP/", 5))
+ content_len = -1;
+
+#if 0
+ printf("len = %d\n", len);
+ fwrite (buf, 1, len, stdout);
+ printf("----------\n");
+#endif
+ while (i <= len-2)
+ {
+ if (i > 8192)
{
- if (i > 8192)
- {
- return i; /* do not allow more than 8K HTTP header */
- }
- if (buf[i] == '\r' && buf[i+1] == '\n')
+ return i; /* do not allow more than 8K HTTP header */
+ }
+ if (skip_crlf(buf, len, &i))
+ {
+ if (skip_crlf(buf, len, &i))
{
- i += 2;
- if (buf[i] == '\r' && buf[i+1] == '\n')
- {
- if (chunked)
- {
- /* inside chunked body .. */
- while(1)
- {
- int j, chunk_len = 0;
- i += 2;
+ /* inside content */
+ if (chunked)
+ {
+ /* inside chunked body .. */
+ while(1)
+ {
+ int chunk_len = 0;
#if CHUNK_DEBUG
-/* debugging */
- if (i <len-2)
- {
- printf ("\n<<<");
- int j;
- for (j = i; j <= i+4; j++)
- printf ("%c", buf[j]);
- printf (">>>\n");
- }
+ if (i < len-2)
+ {
+ printf ("\n<<<");
+ int j;
+ for (j = i; j <= i+3; j++)
+ printf ("%c", buf[j]);
+ printf (">>>\n");
+ }
#endif
- /* read chunk length */
- while (1)
- if (i >= len-2) {
+ /* read chunk length */
+ while (1)
+ if (i >= len-2) {
#if CHUNK_DEBUG
-/* debugging */
- printf ("XXXXXXXX not there yet 1\n");
- printf ("i=%d len=%d\n", i, len);
+ printf ("returning incomplete read at 1\n");
+ printf ("i=%d len=%d\n", i, len);
#endif
- return 0;
- } else if (isdigit(buf[i]))
- chunk_len = chunk_len * 16 +
- (buf[i++] - '0');
- else if (isupper(buf[i]))
- chunk_len = chunk_len * 16 +
- (buf[i++] - ('A'-10));
- else if (islower(buf[i]))
- chunk_len = chunk_len * 16 +
- (buf[i++] - ('a'-10));
- else
- break;
- /* move forward until CRLF - skip chunk ext */
- j = 0;
- while (buf[i] != '\r' && buf[i+1] != '\n')
- {
- if (i >= len-2)
- return 0; /* need more buffer .. */
- if (++j > 1000)
- return i; /* enough.. stop */
- i++;
- }
- /* got CRLF */
-#if CHUNK_DEBUG
- printf ("XXXXXX chunk_len=%d\n", chunk_len);
-#endif
- if (chunk_len < 0)
- return i+2; /* bad chunk_len */
- if (chunk_len == 0)
+ return 0;
+ } else if (isdigit(buf[i]))
+ chunk_len = chunk_len * 16 +
+ (buf[i++] - '0');
+ else if (isupper(buf[i]))
+ chunk_len = chunk_len * 16 +
+ (buf[i++] - ('A'-10));
+ else if (islower(buf[i]))
+ chunk_len = chunk_len * 16 +
+ (buf[i++] - ('a'-10));
+ else
break;
- i += chunk_len+2;
- }
- /* consider trailing headers .. */
- while(i <= len-4)
+ if (chunk_len == 0)
+ break;
+ if (chunk_len < 0)
+ return i;
+
+ while (1)
{
- if (buf[i] == '\r' && buf[i+1] == '\n' &&
- buf[i+2] == '\r' && buf[i+3] == '\n')
- if (len >= i+4)
- return i+4;
+ if (i >= len -1)
+ return 0;
+ if (skip_crlf(buf, len, &i))
+ break;
i++;
}
+ /* got CRLF */
#if CHUNK_DEBUG
-/* debugging */
- printf ("XXXXXXXXX not there yet 2\n");
- printf ("i=%d len=%d\n", i, len);
-#endif
- return 0;
+ printf ("chunk_len=%d\n", chunk_len);
+#endif
+ i += chunk_len;
+ if (i >= len-2)
+ return 0;
+ if (!skip_crlf(buf, len, &i))
+ return 0;
}
- else
- { /* not chunked ; inside body */
- /* i += 2 seems not to work with GCC -O2 ..
- so i+2 is used instead .. */
- if (content_len == -1)
- return 0; /* no content length */
- else if (len >= (i+2)+ content_len)
+ /* consider trailing headers .. */
+ while (i < len)
+ {
+ if (skip_crlf(buf, len, &i))
{
- return (i+2)+ content_len;
+ if (skip_crlf(buf, len, &i))
+ return i;
}
+ else
+ i++;
}
- break;
- }
- else if (i < len - 20 &&
- !strncasecmp((const char *) buf+i, "Transfer-Encoding:", 18))
- {
- i+=18;
- while (buf[i] == ' ')
- i++;
- if (i < len - 8)
- if (!strncasecmp((const char *) buf+i, "chunked", 7))
- chunked = 1;
- }
- else if (i < len - 17 &&
- !strncasecmp((const char *)buf+i, "Content-Length:", 15))
- {
- i+= 15;
- while (buf[i] == ' ')
- i++;
- content_len = 0;
- while (i <= len-4 && isdigit(buf[i]))
- content_len = content_len*10 + (buf[i++] - '0');
- if (content_len < 0) /* prevent negative offsets */
- content_len = 0;
+#if CHUNK_DEBUG
+ printf ("returning incomplete read at 2\n");
+ printf ("i=%d len=%d\n", i, len);
+#endif
+ return 0;
}
else
+ { /* not chunked ; inside body */
+ if (content_len == -1)
+ return 0; /* no content length */
+ else if (len >= i + content_len)
+ {
+ return i + content_len;
+ }
+ }
+ break;
+ }
+ else if (i < len - 20 &&
+ !strncasecmp((const char *) buf+i, "Transfer-Encoding:", 18))
+ {
+ i+=18;
+ while (buf[i] == ' ')
+ i++;
+ if (i < len - 8)
+ if (!strncasecmp((const char *) buf+i, "chunked", 7))
+ chunked = 1;
+ }
+ else if (i < len - 17 &&
+ !strncasecmp((const char *)buf+i, "Content-Length:", 15))
+ {
+ i+= 15;
+ while (buf[i] == ' ')
i++;
+ content_len = 0;
+ while (i <= len-4 && isdigit(buf[i]))
+ content_len = content_len*10 + (buf[i++] - '0');
+ if (content_len < 0) /* prevent negative offsets */
+ content_len = 0;
}
else
i++;
}
- return 0;
+ else
+ i++;
+ }
+ return 0;
+}
+
+int cs_complete_auto(const unsigned char *buf, int len)
+{
+ if (len > 5 && buf[0] >= 0x20 && buf[0] < 0x7f
+ && buf[1] >= 0x20 && buf[1] < 0x7f
+ && buf[2] >= 0x20 && buf[2] < 0x7f)
+ {
+ int r = cs_complete_http((const char *) buf, len);
+ return r;
}
return completeBER(buf, len);
}
* Copyright (C) 1995-2007, Index Data ApS
* See the file LICENSE for details.
*
- * $Id: zgdu.c,v 1.17 2007-01-03 08:42:15 adam Exp $
+ * $Id: zgdu.c,v 1.18 2007-01-11 10:30:41 adam Exp $
*/
/**
int chunked = 0;
*headers = 0;
- while (i < o->size-1 && o->buf[i] == '\r')
+ while (i < o->size-1 && o->buf[i] == '\n')
{
int po;
i++;
- if (o->buf[i] != '\n')
+ if (o->buf[i] == '\r' && i < o->size-1 && o->buf[i+1] == '\n')
{
- o->error = OHTTP;
- return 0;
+ i++;
+ break;
}
- i++;
- if (o->buf[i] == '\r')
+ if (o->buf[i] == '\n')
break;
for (po = i; ; i++)
{
i++;
while (i < o->size-1 && o->buf[i] == ' ')
i++;
- for (po = i; i < o->size-1 && o->buf[i] != '\r' ; i++)
+ for (po = i; i < o->size-1 && !strchr("\r\n", o->buf[i]); i++)
;
(*headers)->value = (char*) odr_malloc(o, i - po + 1);
!strcasecmp((*headers)->value, "chunked"))
chunked = 1;
headers = &(*headers)->next;
+ if (i < o->size-1 && o->buf[i] == '\r')
+ i++;
}
*headers = 0;
- i++;
if (o->buf[i] != '\n')
{
o->error = OHTTP;
hr->content_len = 0;
po = i = 5;
- while (i < o->size-2 && o->buf[i] != ' ' && o->buf[i] != '\r')
+ while (i < o->size-2 && !strchr(" \r\n", o->buf[i]))
i++;
hr->version = (char *) odr_malloc(o, i - po + 1);
if (i - po)
hr->code = hr->code*10 + (o->buf[i] - '0');
i++;
}
- while (i < o->size-1 && o->buf[i] != '\r')
+ while (i < o->size-1 && o->buf[i] != '\n')
i++;
return decode_headers_content(o, i, &hr->headers,
&hr->content_buf, &hr->content_len);
}
i+= 5;
po = i;
- while (o->buf[i] != '\r')
- {
- if (i >= o->size-1)
- {
- o->error = OHTTP;
- return 0;
- }
+ while (i < o->size && !strchr("\r\n", o->buf[i]))
i++;
- }
hr->version = (char *) odr_malloc(o, i - po + 1);
memcpy(hr->version, o->buf + po, i - po);
hr->version[i - po] = '\0';
/* headers */
+ if (i < o->size-1 && o->buf[i] == '\r')
+ i++;
+ if (o->buf[i] != '\n')
+ {
+ o->error = OHTTP;
+ return 0;
+ }
return decode_headers_content(o, i, &hr->headers,
&hr->content_buf, &hr->content_len);
## Copyright (C) 1995-2007, Index Data ApS
## All rights reserved.
-## $Id: Makefile.am,v 1.30 2007-01-05 11:45:11 adam Exp $
+## $Id: Makefile.am,v 1.31 2007-01-11 10:30:41 adam Exp $
check_PROGRAMS = tsticonv tstnmem tstmatchstr tstwrbuf tstodr tstccl tstlog \
tstsoap1 tstsoap2 tstodrstack tstlogthread tstxmlquery tstpquery \
- tst_filepath tst_record_conv tst_retrieval tst_tpath tst_timing
+ tst_comstack tst_filepath tst_record_conv tst_retrieval tst_tpath \
+ tst_timing
check_SCRIPTS = tstcql.sh tstmarciso.sh tstmarcxml.sh tstmarccol.sh
TESTS = $(check_PROGRAMS) $(check_SCRIPTS)
tstlogthread_SOURCES = tstlogthread.c
tstxmlquery_SOURCES = tstxmlquery.c
tstpquery_SOURCES = tstpquery.c
+tst_comstack_SOURCES = tst_comstack.c
tst_filepath_SOURCES = tst_filepath.c
tst_record_conv_SOURCES = tst_record_conv.c
tst_retrieval_SOURCES = tst_retrieval.c
--- /dev/null
+/*
+ * Copyright (C) 1995-2007, Index Data ApS
+ * See the file LICENSE for details.
+ *
+ * $Id: tst_comstack.c,v 1.1 2007-01-11 10:30:41 adam Exp $
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+
+#include <yaz/test.h>
+#include <yaz/comstack.h>
+
+static void tst_http_request(void)
+{
+ {
+ /* no content, no headers */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\r\n"
+ "\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 16), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 17), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 18), 18);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 19), 18);
+ }
+ {
+ /* one header, no content */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\r\n"
+ "Content-Type: x\r\n"
+ "\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 34), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 35), 35);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 36), 35);
+ }
+ {
+ /* one content-length header, length 0 */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\r\n"
+ "Content-Length: 0\r\n"
+ "\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 35), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 37), 37);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 38), 37);
+ }
+ {
+ /* one content-length header, length 5 */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\r\n"
+ "Content-Length: 5\r\n"
+ "\r\n"
+ "ABCDE"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 41), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 42), 42);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 43), 42);
+ }
+ {
+ /* LF only in GET, one content-length header, length 5 */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\n"
+ "Content-Length: 5\r\n"
+ "\r\n"
+ "ABCDE"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 40), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 41), 41);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 42), 41);
+ }
+ {
+ /* LF only in all places, one content-length header, length 5 */
+ const char *http_buf =
+ /*123456789012345678 */
+ "GET / HTTP/1.1\n"
+ "Content-Length: 5\n"
+ "\n"
+ "ABCDE"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 38), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 39), 39);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 40), 39);
+ }
+
+ {
+ /* one header, unknown transfer-encoding (no content) */
+ const char *http_buf =
+ /*12345678901234567890123456789 */
+ "GET / HTTP/1.1\r\n"
+ "Transfer-Encoding: chunke_\r\n"
+ "\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 45), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 46), 46);
+ }
+
+ {
+ /* one header, one chunk */
+ const char *http_buf =
+ /*12345678901234567890123456789 */
+ "GET / HTTP/1.1\r\n"
+ "Transfer-Encoding: chunked\r\n"
+ "\r\n"
+ "3\r\n"
+ "123\r\n"
+ "0\r\n\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 58), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 59), 59);
+ }
+
+ {
+ /* one header, two chunks */
+ const char *http_buf =
+ /*12345678901234567890123456789 */
+ "GET / HTTP/1.1\r\n"
+ "Transfer-Encoding: chunked\r\n"
+ "\r\n"
+ "3\r\n"
+ "123\r\n"
+ "2\r\n"
+ "12\n"
+ "0\r\n\r\n"
+ "GET / HTTP/1.0\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 64), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 65), 65);
+ }
+}
+
+static void tst_http_response(void)
+{
+ {
+ /* unlimited content, no headers */
+ const char *http_buf =
+ /*123456789012345678 */
+ "HTTP/1.1 200 OK\r\n"
+ "\r\n"
+ "HTTP/1.1 200 OK\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 24), 0);
+ }
+ {
+ /* response, content */
+ const char *http_buf =
+ /*123456789012345678 */
+ "HTTP/1.1 200 OK\r\n"
+ "Content-Length: 2\r\n"
+ "\r\n"
+ "12"
+ "HTTP/1.1 200 OK\r\n";
+
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 1), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 2), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 39), 0);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 40), 40);
+ YAZ_CHECK_EQ(cs_complete_http(http_buf, 41), 40);
+ }
+}
+
+
+int main (int argc, char **argv)
+{
+ YAZ_CHECK_INIT(argc, argv);
+ YAZ_CHECK_LOG();
+ tst_http_request();
+ tst_http_response();
+ YAZ_CHECK_TERM;
+}
+
+/*
+ * Local variables:
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ * vim: shiftwidth=4 tabstop=8 expandtab
+ */
+
projects / yaz-moved-to-github.git / commitdiff