int oid_dotstring_to_oid(const char *name, Odr_oid *oid)
{
int i = 0;
int val = 0;
while (yaz_isdigit (*name))
{
val = val*10 + (*name - '0');
name++;
if (*name == '.')
{
if (i < OID_SIZE-1)
oid[i++] = val;
val = 0;
name++;
}
}
if (i == 0)
return -1;
oid[i] = val;
oid[i+1] = -1;
return 0;
}
static int cs_complete_http(const char *buf, int len, int head_only)
{
/* deal with HTTP request/response */
int i, content_len = 0, chunked = 0;
/* need at least one line followed by \n or \r .. */
for (i = 0; ; i++)
if (i == len)
return 0; /* incomplete */
else if (buf[i] == '\n' || buf[i] == '\r')
break;
/* check to see if it's a response with content */
if (!head_only && !memcmp(buf, "HTTP/", 5))
{
int j;
for (j = 5; j < i; j++)
if (buf[j] == ' ')
{
++j;
if (buf[j] == '1') /* 1XX */
;
else if (!memcmp(buf + j, "204", 3))
;
else if (!memcmp(buf + j, "304", 3))
;
else
content_len = -1;
break;
}
}
#if 0
printf("len = %d\n", len);
fwrite (buf, 1, len, stdout);
printf("----------\n");
#endif
for (i = 2; i <= len-2; )
{
if (i > 8192)
{
return i; /* do not allow more than 8K HTTP header */
}
if (skip_crlf(buf, len, &i))
{
if (skip_crlf(buf, len, &i))
{
/* inside content */
if (chunked)
return cs_read_chunk(buf, i, len);
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 &&
!yaz_strncasecmp((const char *) buf+i,
"Transfer-Encoding:", 18))
{
i+=18;
while (buf[i] == ' ')
i++;
if (i < len - 8)
if (!yaz_strncasecmp((const char *) buf+i, "chunked", 7))
chunked = 1;
}
else if (i < len - 17 &&
!yaz_strncasecmp((const char *)buf+i,
"Content-Length:", 15))
{
i+= 15;
while (buf[i] == ' ')
i++;
content_len = 0;
while (i <= len-4 && yaz_isdigit(buf[i]))
content_len = content_len*10 + (buf[i++] - '0');
if (content_len < 0) /* prevent negative offsets */
content_len = 0;
}
else
i++;
}
else
i++;
}
return 0;
}
static int cs_read_chunk(const char *buf, int i, int len)
{
/* inside chunked body .. */
while (1)
{
int chunk_len = 0;
#if CHUNK_DEBUG
if (i < len-2)
{
int j;
printf ("\n<<<");
for (j = i; j <= i+3; j++)
printf ("%c", buf[j]);
printf (">>>\n");
}
#endif
/* read chunk length */
while (1)
if (i >= len-2) {
#if CHUNK_DEBUG
printf ("returning incomplete read at 1\n");
printf ("i=%d len=%d\n", i, len);
#endif
return 0;
} else if (yaz_isdigit(buf[i]))
chunk_len = chunk_len * 16 +
(buf[i++] - '0');
else if (yaz_isupper(buf[i]))
chunk_len = chunk_len * 16 +
(buf[i++] - ('A'-10));
else if (yaz_islower(buf[i]))
chunk_len = chunk_len * 16 +
(buf[i++] - ('a'-10));
else
break;
if (chunk_len == 0)
break;
if (chunk_len < 0)
return i;
while (1)
{
if (i >= len -1)
return 0;
if (skip_crlf(buf, len, &i))
break;
i++;
}
/* got CRLF */
#if CHUNK_DEBUG
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;
}
/* consider trailing headers .. */
while (i < len)
{
if (skip_crlf(buf, len, &i))
{
if (skip_crlf(buf, len, &i))
return i;
}
else
i++;
}
#if CHUNK_DEBUG
printf ("returning incomplete read at 2\n");
printf ("i=%d len=%d\n", i, len);
#endif
return 0;
}
static int cql_transform_parse_tok_line(cql_transform_t ct,
const char *pattern,
yaz_tok_parse_t tp)
{
int ae_num = 0;
Z_AttributeElement *ae[20];
int ret = 0; /* 0=OK, != 0 FAIL */
int t;
t = yaz_tok_move(tp);
while (t == YAZ_TOK_STRING && ae_num < 20)
{
WRBUF type_str = wrbuf_alloc();
WRBUF set_str = 0;
Z_AttributeElement *elem = 0;
const char *value_str = 0;
/* attset type=value OR type=value */
elem = (Z_AttributeElement *) nmem_malloc(ct->nmem, sizeof(*elem));
elem->attributeSet = 0;
ae[ae_num] = elem;
wrbuf_puts(ct->w, yaz_tok_parse_string(tp));
wrbuf_puts(type_str, yaz_tok_parse_string(tp));
t = yaz_tok_move(tp);
if (t == YAZ_TOK_EOF)
{
wrbuf_destroy(type_str);
if (set_str)
wrbuf_destroy(set_str);
break;
}
if (t == YAZ_TOK_STRING)
{
wrbuf_puts(ct->w, " ");
wrbuf_puts(ct->w, yaz_tok_parse_string(tp));
set_str = type_str;
elem->attributeSet =
yaz_string_to_oid_nmem(yaz_oid_std(), CLASS_ATTSET,
wrbuf_cstr(set_str), ct->nmem);
type_str = wrbuf_alloc();
wrbuf_puts(type_str, yaz_tok_parse_string(tp));
t = yaz_tok_move(tp);
}
elem->attributeType = nmem_intdup(ct->nmem, 0);
if (sscanf(wrbuf_cstr(type_str), ODR_INT_PRINTF, elem->attributeType)
!= 1)
{
wrbuf_destroy(type_str);
if (set_str)
wrbuf_destroy(set_str);
yaz_log(YLOG_WARN, "Expected numeric attribute type");
ret = -1;
break;
}
wrbuf_destroy(type_str);
if (set_str)
wrbuf_destroy(set_str);
if (t != '=')
{
yaz_log(YLOG_WARN, "Expected = after after attribute type");
ret = -1;
break;
}
t = yaz_tok_move(tp);
if (t != YAZ_TOK_STRING) /* value */
{
yaz_log(YLOG_WARN, "Missing attribute value");
ret = -1;
break;
}
value_str = yaz_tok_parse_string(tp);
if (yaz_isdigit(*value_str))
{
elem->which = Z_AttributeValue_numeric;
elem->value.numeric =
nmem_intdup(ct->nmem, atoi(value_str));
}
else
{
Z_ComplexAttribute *ca = (Z_ComplexAttribute *)
nmem_malloc(ct->nmem, sizeof(*ca));
elem->which = Z_AttributeValue_complex;
elem->value.complex = ca;
ca->num_list = 1;
ca->list = (Z_StringOrNumeric **)
nmem_malloc(ct->nmem, sizeof(Z_StringOrNumeric *));
ca->list[0] = (Z_StringOrNumeric *)
nmem_malloc(ct->nmem, sizeof(Z_StringOrNumeric));
ca->list[0]->which = Z_StringOrNumeric_string;
ca->list[0]->u.string = nmem_strdup(ct->nmem, value_str);
ca->num_semanticAction = 0;
ca->semanticAction = 0;
}
wrbuf_puts(ct->w, "=");
wrbuf_puts(ct->w, yaz_tok_parse_string(tp));
t = yaz_tok_move(tp);
wrbuf_puts(ct->w, " ");
ae_num++;
}
if (ret == 0) /* OK? */
{
struct cql_prop_entry **pp = &ct->entry;
while (*pp)
pp = &(*pp)->next;
*pp = (struct cql_prop_entry *) xmalloc(sizeof(**pp));
(*pp)->pattern = xstrdup(pattern);
(*pp)->value = xstrdup(wrbuf_cstr(ct->w));
(*pp)->attr_list.num_attributes = ae_num;
if (ae_num == 0)
(*pp)->attr_list.attributes = 0;
else
{
(*pp)->attr_list.attributes = (Z_AttributeElement **)
nmem_malloc(ct->nmem,
ae_num * sizeof(Z_AttributeElement *));
memcpy((*pp)->attr_list.attributes, ae,
ae_num * sizeof(Z_AttributeElement *));
}
(*pp)->next = 0;
if (0)
{
ODR pr = odr_createmem(ODR_PRINT);
Z_AttributeList *alp = &(*pp)->attr_list;
odr_setprint(pr, yaz_log_file());
z_AttributeList(pr, &alp, 0, 0);
odr_setprint(pr, 0);
odr_destroy(pr);
}
}
return ret;
}
static void emit_term(cql_transform_t ct,
struct cql_node *cn,
const char *term, int length,
void (*pr)(const char *buf, void *client_data),
void *client_data)
{
int i;
const char *ns = cn->u.st.index_uri;
int z3958_mode = 0;
int process_term = 1;
if (has_modifier(cn, "regexp"))
process_term = 0;
else if (cql_lookup_property(ct, "truncation", 0, "cql"))
{
process_term = 0;
cql_pr_attr(ct, "truncation", "cql", 0,
pr, client_data, YAZ_SRW_MASKING_CHAR_UNSUPP);
}
assert(cn->which == CQL_NODE_ST);
if (process_term)
{
unsigned anchor = 0;
unsigned trunc = 0;
for (i = 0; i < length; i++)
{
if (term[i] == '\\' && i < length - 1)
i++;
else
{
switch (term[i])
{
case '^':
if (i == 0)
anchor |= 1;
else if (i == length - 1)
anchor |= 2;
break;
case '*':
if (i == 0)
trunc |= 1;
else if (i == length - 1)
trunc |= 2;
else
z3958_mode = 1;
break;
case '?':
z3958_mode = 1;
break;
}
}
}
if (anchor == 3)
{
cql_pr_attr(ct, "position", "firstAndLast", 0,
pr, client_data, YAZ_SRW_ANCHORING_CHAR_IN_UNSUPP_POSITION);
term++;
length -= 2;
}
else if (anchor == 1)
{
cql_pr_attr(ct, "position", "first", 0,
pr, client_data, YAZ_SRW_ANCHORING_CHAR_IN_UNSUPP_POSITION);
term++;
length--;
}
else if (anchor == 2)
{
cql_pr_attr(ct, "position", "last", 0,
pr, client_data, YAZ_SRW_ANCHORING_CHAR_IN_UNSUPP_POSITION);
length--;
}
else
{
cql_pr_attr(ct, "position", "any", 0,
pr, client_data, YAZ_SRW_ANCHORING_CHAR_IN_UNSUPP_POSITION);
}
if (z3958_mode == 0)
{
if (trunc == 3 && cql_pr_attr(ct, "truncation",
"both", 0, pr, client_data, 0))
{
term++;
length -= 2;
}
else if (trunc == 1 && cql_pr_attr(ct, "truncation",
"left", 0, pr, client_data, 0))
{
term++;
length--;
}
else if (trunc == 2 && cql_pr_attr(ct, "truncation", "right", 0,
pr, client_data, 0))
{
length--;
}
else if (trunc)
z3958_mode = 1;
else
cql_pr_attr(ct, "truncation", "none", 0,
pr, client_data, 0);
}
if (z3958_mode)
cql_pr_attr(ct, "truncation", "z3958", 0,
pr, client_data, YAZ_SRW_MASKING_CHAR_UNSUPP);
}
if (ns) {
cql_pr_attr_uri(ct, "index", ns,
cn->u.st.index, "serverChoice",
pr, client_data, YAZ_SRW_UNSUPP_INDEX);
}
if (cn->u.st.modifiers)
{
struct cql_node *mod = cn->u.st.modifiers;
for (; mod; mod = mod->u.st.modifiers)
{
cql_pr_attr(ct, "relationModifier", mod->u.st.index, 0,
pr, client_data, YAZ_SRW_UNSUPP_RELATION_MODIFIER);
}
}
(*pr)("\"", client_data);
if (process_term)
for (i = 0; i < length; i++)
{
char x[2]; /* temp buffer */
if (term[i] == '\\' && i < length - 1)
{
i++;
if (strchr("\"\\", term[i]))
pr("\\", client_data);
if (z3958_mode && strchr("#?", term[i]))
pr("\\\\", client_data); /* double \\ to survive PQF parse */
x[0] = term[i];
x[1] = '\0';
pr(x, client_data);
}
else if (z3958_mode && term[i] == '*')
{
pr("?", client_data);
if (i < length - 1 && yaz_isdigit(term[i+1]))
pr("\\\\", client_data); /* dbl \\ to survive PQF parse */
}
else if (z3958_mode && term[i] == '?')
{
pr("#", client_data);
}
else
{
if (term[i] == '\"')
pr("\\", client_data);
if (z3958_mode && strchr("#?", term[i]))
pr("\\\\", client_data); /* dbl \\ to survive PQF parse */
x[0] = term[i];
x[1] = '\0';
pr(x, client_data);
}
}
else
{
for (i = 0; i < length; i++)
{
char x[2];
x[0] = term[i];
x[1] = '\0';
pr(x, client_data);
}
}
(*pr)("\" ", client_data);
}
static int decode_headers_content(ODR o, int off, Z_HTTP_Header **headers,
char **content_buf, int *content_len)
{
int i = off;
int chunked = 0;
*headers = 0;
while (i < o->size-1 && o->buf[i] == '\n')
{
int po;
i++;
if (o->buf[i] == '\r' && i < o->size-1 && o->buf[i+1] == '\n')
{
i++;
break;
}
if (o->buf[i] == '\n')
break;
for (po = i; ; i++)
{
if (i == o->size)
{
o->error = OHTTP;
return 0;
}
else if (o->buf[i] == ':')
break;
}
*headers = (Z_HTTP_Header *) odr_malloc(o, sizeof(**headers));
(*headers)->name = (char*) odr_malloc(o, i - po + 1);
memcpy ((*headers)->name, o->buf + po, i - po);
(*headers)->name[i - po] = '\0';
i++;
while (i < o->size-1 && o->buf[i] == ' ')
i++;
for (po = i; i < o->size-1 && !strchr("\r\n", o->buf[i]); i++)
;
(*headers)->value = (char*) odr_malloc(o, i - po + 1);
memcpy ((*headers)->value, o->buf + po, i - po);
(*headers)->value[i - po] = '\0';
if (!strcasecmp((*headers)->name, "Transfer-Encoding")
&&
!strcasecmp((*headers)->value, "chunked"))
chunked = 1;
headers = &(*headers)->next;
if (i < o->size-1 && o->buf[i] == '\r')
i++;
}
*headers = 0;
if (o->buf[i] != '\n')
{
o->error = OHTTP;
return 0;
}
i++;
if (chunked)
{
int off = 0;
/* we know buffer will be smaller than o->size - i*/
*content_buf = (char*) odr_malloc(o, o->size - i);
while (1)
{
/* chunk length .. */
int chunk_len = 0;
for (; i < o->size-2; i++)
if (yaz_isdigit(o->buf[i]))
chunk_len = chunk_len * 16 +
(o->buf[i] - '0');
else if (yaz_isupper(o->buf[i]))
chunk_len = chunk_len * 16 +
(o->buf[i] - ('A'-10));
else if (yaz_islower(o->buf[i]))
chunk_len = chunk_len * 16 +
(o->buf[i] - ('a'-10));
else
break;
/* chunk extension ... */
while (o->buf[i] != '\r' && o->buf[i+1] != '\n')
{
if (i >= o->size-2)
{
o->error = OHTTP;
return 0;
}
i++;
}
i += 2; /* skip CRLF */
if (chunk_len == 0)
break;
if (chunk_len < 0 || off + chunk_len > o->size)
{
o->error = OHTTP;
return 0;
}
/* copy chunk .. */
memcpy (*content_buf + off, o->buf + i, chunk_len);
i += chunk_len + 2; /* skip chunk+CRLF */
off += chunk_len;
}
if (!off)
*content_buf = 0;
*content_len = off;
}
else
{
if (i > o->size)
{
o->error = OHTTP;
return 0;
}
else if (i == o->size)
{
*content_buf = 0;
*content_len = 0;
}
else
{
*content_len = o->size - i;
*content_buf = (char*) odr_malloc(o, *content_len + 1);
memcpy(*content_buf, o->buf + i, *content_len);
(*content_buf)[*content_len] = '\0';
}
}
return 1;
}
int yaz_marc_read_iso2709(yaz_marc_t mt, const char *buf, int bsize)
{
int entry_p;
int record_length;
int indicator_length;
int identifier_length;
int end_of_directory;
int base_address;
int length_data_entry;
int length_starting;
int length_implementation;
yaz_marc_reset(mt);
if (!atoi_n_check(buf, 5, &record_length))
{
yaz_marc_cprintf(mt, "Bad leader");
return -1;
}
if (record_length < 25)
{
yaz_marc_cprintf(mt, "Record length %d < 24", record_length);
return -1;
}
/* ballout if bsize is known and record_length is less than that */
if (bsize != -1 && record_length > bsize)
{
yaz_marc_cprintf(mt, "Record appears to be larger than buffer %d < %d",
record_length, bsize);
return -1;
}
if (yaz_marc_get_debug(mt))
yaz_marc_cprintf(mt, "Record length %5d", record_length);
yaz_marc_set_leader(mt, buf,
&indicator_length,
&identifier_length,
&base_address,
&length_data_entry,
&length_starting,
&length_implementation);
/* First pass. determine length of directory & base of data */
for (entry_p = 24; buf[entry_p] != ISO2709_FS; )
{
/* length of directory entry */
int l = 3 + length_data_entry + length_starting;
if (entry_p + l >= record_length)
{
yaz_marc_cprintf(mt, "Directory offset %d: end of record."
" Missing FS char", entry_p);
return -1;
}
if (yaz_marc_get_debug(mt))
{
WRBUF hex = wrbuf_alloc();
wrbuf_puts(hex, "Tag ");
wrbuf_write_escaped(hex, buf + entry_p, 3);
wrbuf_puts(hex, ", length ");
wrbuf_write_escaped(hex, buf + entry_p + 3,
length_data_entry);
wrbuf_puts(hex, ", starting ");
wrbuf_write_escaped(hex, buf + entry_p + 3 + length_data_entry,
length_starting);
yaz_marc_cprintf(mt, "Directory offset %d: %s",
entry_p, wrbuf_cstr(hex));
wrbuf_destroy(hex);
}
/* Check for digits in length+starting info */
while (--l >= 3)
if (!yaz_isdigit(buf[entry_p + l]))
break;
if (l >= 3)
{
WRBUF hex = wrbuf_alloc();
/* Not all digits, so stop directory scan */
wrbuf_write_escaped(hex, buf + entry_p,
length_data_entry + length_starting + 3);
yaz_marc_cprintf(mt, "Directory offset %d: Bad value for data"
" length and/or length starting (%s)", entry_p,
wrbuf_cstr(hex));
wrbuf_destroy(hex);
break;
}
entry_p += 3 + length_data_entry + length_starting;
}
end_of_directory = entry_p;
if (base_address != entry_p+1)
{
yaz_marc_cprintf(mt, "Base address not at end of directory,"
" base %d, end %d", base_address, entry_p+1);
}
/* Second pass. parse control - and datafields */
for (entry_p = 24; entry_p != end_of_directory; )
{
int data_length;
int data_offset;
int end_offset;
int i;
char tag[4];
int identifier_flag = 0;
int entry_p0 = entry_p;
memcpy (tag, buf+entry_p, 3);
entry_p += 3;
tag[3] = '\0';
data_length = atoi_n(buf+entry_p, length_data_entry);
entry_p += length_data_entry;
data_offset = atoi_n(buf+entry_p, length_starting);
entry_p += length_starting;
i = data_offset + base_address;
end_offset = i+data_length-1;
if (data_length <= 0 || data_offset < 0)
break;
if (yaz_marc_get_debug(mt))
{
yaz_marc_cprintf(mt, "Tag: %s. Directory offset %d: data-length %d,"
" data-offset %d",
tag, entry_p0, data_length, data_offset);
}
if (end_offset >= record_length)
{
yaz_marc_cprintf(mt, "Directory offset %d: Data out of bounds %d >= %d",
entry_p0, end_offset, record_length);
break;
}
if (memcmp (tag, "00", 2))
identifier_flag = 1; /* if not 00X assume subfields */
else if (indicator_length < 4 && indicator_length > 0)
{
/* Danmarc 00X have subfields */
if (buf[i + indicator_length] == ISO2709_IDFS)
identifier_flag = 1;
else if (buf[i + indicator_length + 1] == ISO2709_IDFS)
identifier_flag = 2;
}
if (identifier_flag)
{
/* datafield */
i += identifier_flag-1;
if (indicator_length)
{
/* skip RS/FS bytes in indicator. They are not allowed there */
int j;
for (j = indicator_length; --j >= 0; )
if (buf[j+i] < ' ')
{
j++;
i += j;
end_offset += j;
yaz_marc_cprintf(mt, "Bad indicator data. "
"Skipping %d bytes", j);
break;
}
yaz_marc_add_datafield(mt, tag, buf+i, indicator_length);
i += indicator_length;
}
while (i < end_offset &&
buf[i] != ISO2709_RS && buf[i] != ISO2709_FS)
{
int code_offset = i+1;
i ++;
while (i < end_offset &&
buf[i] != ISO2709_RS && buf[i] != ISO2709_IDFS &&
buf[i] != ISO2709_FS)
i++;
if (i > code_offset)
yaz_marc_add_subfield(mt, buf+code_offset, i - code_offset);
}
}
else
{
/* controlfield */
int i0 = i;
while (i < end_offset &&
buf[i] != ISO2709_RS && buf[i] != ISO2709_FS)
i++;
yaz_marc_add_controlfield(mt, tag, buf+i0, i-i0);
}
if (i < end_offset)
{
yaz_marc_cprintf(mt, "Separator but not at end of field length=%d",
data_length);
}
if (buf[i] != ISO2709_RS && buf[i] != ISO2709_FS)
{
yaz_marc_cprintf(mt, "No separator at end of field length=%d",
data_length);
}
}
return record_length;
}
