bstring uri_normalize(const char *uri)
{
int st, origSize, normSize;
char *normal_str, *cleaned_uri;
bstring normalized;
UriParserStateA state;
UriUriA u;
state.uri = &u;
cleaned_uri = malloc((sizeof(char) * strlen(uri)) + 1);
clean_uri(uri, cleaned_uri);
/* Parse URI */
if (uriParseUriA(&state, cleaned_uri) != URI_SUCCESS) {
free(cleaned_uri);
uriFreeUriMembersA(&u);
return bfromcstr("");
}
free(cleaned_uri);
/* Normalize URI */
st = uriNormalizeSyntaxA(&u);
if (st != URI_SUCCESS) {
uriFreeUriMembersA(&u);
return bfromcstr("");
}
origSize = strlen(uri);
normal_str = malloc(sizeof(char) * origSize + 1);
if (normal_str == NULL) {
uriFreeUriMembersA(&u);
return bfromcstr("");
}
st = uriToStringA(normal_str, &u, origSize + 1, &normSize);
if (st != URI_SUCCESS) {
uriFreeUriMembersA(&u);
return bfromcstr("");
}
normalized = bfromcstr(normal_str);
free(normal_str);
uriFreeUriMembersA(&u);
return normalized;
}
Datum parse_uri(PG_FUNCTION_ARGS)
{
TupleDesc tupledesc;
AttInMetadata *attinmeta;
text* input=PG_GETARG_TEXT_P(0);
/* Function is defined STRICT in SQL, so no NULL check is needed. */
bool normalize=PG_GETARG_BOOL(1);
bool parse_query=PG_GETARG_BOOL(2);
char* inp=palloc((1+VARSIZE(input)-VARHDRSZ)*sizeof(char));
if (!inp) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpy(inp,VARDATA(input),VARSIZE(input)-VARHDRSZ);
inp[VARSIZE(input)-VARHDRSZ]='\0';
/* Function internals start here */
int i;
int memctr;
UriPathSegmentA* pathseg;
int quit;
char* writehere;
UriParserStateA state;
UriUriA uri;
state.uri = &uri;
if (uriParseUriA(&state, inp) != URI_SUCCESS) {
uriFreeUriMembersA(&uri);
/*
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("Unable to parse URI.")));
*/
PG_RETURN_NULL();
}
if (normalize) {
if (uriNormalizeSyntaxA(&uri) != URI_SUCCESS) {
uriFreeUriMembersA(&uri);
PG_RETURN_NULL();
}
}
UriQueryListA* queryList=NULL;
int itemCount;
if (parse_query&&(uri.query.afterLast!=uri.query.first)) {
if (uriDissectQueryMallocA(&queryList, &itemCount, uri.query.first,
uri.query.afterLast) != URI_SUCCESS) {
uriFreeUriMembersA(&uri);
uriFreeQueryListA(queryList);
PG_RETURN_NULL();
}
}
/* Function internals finish here */
if (get_call_result_type(fcinfo, NULL, &tupledesc) != TYPEFUNC_COMPOSITE) {
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("function returning record called in context "
"that cannot accept type record")));
PG_RETURN_NULL();
}
/* This error should never happen, because the SQL function is defined
as returning a uri_type. */
attinmeta = TupleDescGetAttInMetadata(tupledesc);
char** retval=(char**) palloc(13*sizeof(char*));
if (!retval) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
if (uri.scheme.afterLast==uri.scheme.first) {
retval[0]=NULL;
} else {
retval[0]=(char*) palloc((1+(uri.scheme.afterLast-uri.scheme.first))*sizeof(char)); /* scheme, e.g. "http" */
if (!retval[0]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[0],uri.scheme.first,uri.scheme.afterLast-uri.scheme.first);
}
if (uri.userInfo.afterLast==uri.userInfo.first) {
retval[1]=NULL;
} else {
retval[1]=(char*) palloc((1+(uri.userInfo.afterLast-uri.userInfo.first))*sizeof(char)); /* userInfo, e.g. "gpadmin" */
if (!retval[1]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[1],uri.userInfo.first,uri.userInfo.afterLast-uri.userInfo.first);
}
if (uri.hostText.afterLast==uri.hostText.first) {
retval[2]=NULL;
} else {
retval[2]=(char*) palloc((1+(uri.hostText.afterLast-uri.hostText.first))*sizeof(char)); /* hostText, e.g. "192.165.0.0" */
if (!retval[2]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[2],uri.hostText.first,uri.hostText.afterLast-uri.hostText.first);
}
if (uri.hostData.ip4==NULL) {
retval[3]=NULL;
} else {
retval[3]=(char*) palloc(17*sizeof(char)); /* IPv4 */
if (!retval[3]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpy(retval[3],"\\000\\000\\000\\000",17);
for(i=0;i<4;++i) {
retval[3][1+4*i]+=uri.hostData.ip4->data[i]>> 6;
retval[3][2+4*i]+=(uri.hostData.ip4->data[i]>> 3)&7;
retval[3][3+4*i]+=uri.hostData.ip4->data[i]&7;
}
}
if (uri.hostData.ip6==NULL) {
retval[4]=NULL;
} else {
retval[4]=(char*) palloc(65*sizeof(char)); /* IPv6 */
if (!retval[4]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpy(retval[4],"\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000\\000",65);
for(i=0;i<16;++i) {
retval[4][1+4*i]+=uri.hostData.ip6->data[i]>> 6;
retval[4][2+4*i]+=(uri.hostData.ip6->data[i]>> 3)&7;
retval[4][3+4*i]+=uri.hostData.ip6->data[i]&7;
}
}
if (uri.hostData.ipFuture.afterLast==uri.hostData.ipFuture.first) {
retval[5]=NULL;
} else {
retval[5]=(char*) palloc((1+(uri.hostData.ipFuture.afterLast-uri.hostData.ipFuture.first))*sizeof(char)); /* ipFuture, text field */
if (!retval[5]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[5],uri.hostData.ipFuture.first,uri.hostData.ipFuture.afterLast-uri.hostData.ipFuture.first);
}
if (uri.portText.afterLast==uri.portText.first) {
retval[6]=NULL;
} else {
retval[6]=(char*) palloc((1+(uri.portText.afterLast-uri.portText.first))*sizeof(char)); /* portText, e.g. "80" */
if (!retval[6]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[6],uri.portText.first,uri.portText.afterLast-uri.portText.first);
}
if (uri.pathHead==NULL) {
retval[7]=NULL;
} else {
memctr=2;
pathseg=uri.pathHead;
do {
quit=((pathseg==uri.pathTail)||(pathseg->next==NULL));
memctr+=3+2*(pathseg->text.afterLast-pathseg->text.first);
pathseg=pathseg->next;
} while (!quit);
if (memctr==2) {
++memctr;
}
retval[7]=(char*) palloc(memctr*sizeof(char)); /* path */
/* e.g. "{usr,local,lib}" */
if (!retval[7]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
writehere=retval[7];
*writehere='{';
++writehere;
pathseg=uri.pathHead;
do {
quit=((pathseg==uri.pathTail)||(pathseg->next==NULL));
writehere=memenc(writehere,pathseg->text.first,pathseg->text.afterLast-pathseg->text.first);
*writehere=',';
++writehere;
pathseg=pathseg->next;
} while (!quit);
if (memctr!=3) {
--writehere;
}
memcpy(writehere,"}",2);
}
if (uri.query.afterLast==uri.query.first) {
retval[8]=NULL;
} else {
retval[8]=(char*) palloc((1+(uri.query.afterLast-uri.query.first))*sizeof(char)); /* query without leading "?" */
if (!retval[8]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[8],uri.query.first,uri.query.afterLast-uri.query.first);
}
if (uri.fragment.afterLast==uri.fragment.first) {
retval[9]=NULL;
} else {
retval[9]=(char*) palloc((1+(uri.fragment.afterLast-uri.fragment.first))*sizeof(char)); /* fragment without leading "#" */
if (!retval[9]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpz(retval[9],uri.fragment.first,uri.fragment.afterLast-uri.fragment.first);
}
if (uri.absolutePath) {
retval[10]=(char*) palloc(5*sizeof(char)); /* absolutePath */
if (!retval[10]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpy(retval[10],"true",5);
} else {
retval[10]=(char*) palloc(6*sizeof(char)); /* absolutePath */
if (!retval[10]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
memcpy(retval[10],"false",6);
}
if (parse_query) {
int key_counter=2;
int val_counter=2;
int counter=0;
for(UriQueryListA* it=queryList;(counter!=itemCount)&&(it!=NULL);
it=it->next,++counter) {
if (it->key==NULL) {
key_counter+=3; /* should never reach here. */
} else {
key_counter+=3+2*strlen(it->key);
}
if (it->value==NULL) {
val_counter+=3; /* currently no way to distinguish empty string
value (?a=) from null value (?a). This is a GPDB
limitation. */
} else {
val_counter+=3+2*strlen(it->value);
}
}
if (key_counter==2) {
++key_counter;
}
if (val_counter==2) {
++val_counter;
}
retval[11]=palloc(key_counter*sizeof(char));
if (!retval[11]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
retval[12]=palloc(val_counter*sizeof(char));
if (!retval[12]) {
ereport(ERROR,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("Memory allocation failed.")));
PG_RETURN_NULL();
}
retval[11][0]='{';
retval[12][0]='{';
char* key_ptr=retval[11]+1;
char* val_ptr=retval[12]+1;
counter=0;
for(UriQueryListA* it=queryList;(counter!=itemCount)&&(it!=NULL);
it=it->next,++counter) {
if (it->key==NULL) {
*key_ptr='"';
++key_ptr;
*key_ptr='"';
++key_ptr;
*key_ptr=',';
++key_ptr;
/* should never reach here. */
} else {
key_ptr=strenc(key_ptr,it->key);
*key_ptr=',';
++key_ptr;
}
if (it->value==NULL) {
*val_ptr='"';
++val_ptr;
*val_ptr='"';
++val_ptr;
*val_ptr=',';
++val_ptr;
/* currently no way to distinguish empty string
value (?a=) from null value (?a). This is a GPDB
limitation. */
} else {
val_ptr=strenc(val_ptr,it->value);
*val_ptr=',';
++val_ptr;
}
}
if (key_counter!=3) {
--key_ptr;
}
memcpy(key_ptr,"}",2);
if (val_counter!=3) {
--val_ptr;
}
memcpy(val_ptr,"}",2);
uriFreeQueryListA(queryList);
} else {
retval[11]=NULL;
retval[12]=NULL;
}
/* There is no need to call pfree. It's called automatically. */
HeapTuple tuple;
Datum result;
tuple=BuildTupleFromCStrings(attinmeta,retval);
result=HeapTupleGetDatum(tuple);
/* Free memory start */
uriFreeUriMembersA(&uri);
/* Free memory finish */
PG_RETURN_DATUM(result);
}
UriUriA *create_absolute_uri(const UriUriA *base_uri, const char *url)
{
UriParserStateA state;
UriUriA *abs_dest, rel_source;
char *newurl;
if (!url || !*url)
return NULL;
if ((abs_dest = uri_alloc_1()) == NULL)
return NULL;
if (base_uri) {
state.uri = &rel_source;
if (uriParseUriA(&state, url) != URI_SUCCESS) {
uri_free(abs_dest);
uriFreeUriMembersA(&rel_source);
return NULL;
}
if (uriAddBaseUriA(abs_dest, &rel_source, base_uri) != URI_SUCCESS) {
uri_free(abs_dest);
uriFreeUriMembersA(&rel_source);
return NULL;
}
uriFreeUriMembersA(&rel_source);
}
else {
state.uri = abs_dest;
if (uriParseUriA(&state, url) != URI_SUCCESS) {
uri_free(abs_dest);
return NULL;
}
}
if (uriNormalizeSyntaxA(abs_dest) != URI_SUCCESS) {
uri_free(abs_dest);
return NULL;
}
/* http://www.example.com and http://www.example.com/ have to generate the same object */
if (!base_uri && (!abs_dest->pathHead || !abs_dest->pathHead->text.first)
&& !abs_dest->query.first) {
newurl = string_new(url);
string_cat(&newurl, "/");
uriFreeUriMembersA(abs_dest);
state.uri = abs_dest;
if (uriParseUriA(&state, newurl) != URI_SUCCESS) {
uri_free(abs_dest);
string_free(newurl);
return NULL;
}
if (uriNormalizeSyntaxA(abs_dest) != URI_SUCCESS) {
uri_free(abs_dest);
string_free(newurl);
return NULL;
}
string_free(newurl);
}
return abs_dest;
}