--- name: "curl_url_get" section: "3" platform: "linux" package: "curl" version: "8.23.0-1" release: "Arch Linux 2026-11-01" platforms: ["linux", "macos"] url: "https://new.carta.tech/curl_url_get" --- # curl\_url\_get(3) extract a part from a URL ## NAME curl\_url\_get - extract a part from a URL ## SYNOPSIS ``` #include CURLUcode curl_url_get(const CURLU *url, CURLUPart part, char **content, unsigned int flags); ``` ## DESCRIPTION Given a *url* handle of a URL object, this function extracts an individual piece or the full URL from it. The *part* argument specifies which part to extract (see list below) and *content* points to a 'char \*' to get updated to point to a newly allocated string with the contents. The *flags* argument is a bitmask with individual features. The returned content pointer must be freed with *[curl\_free(3)](/curl_free/3.md)* after use. Although not constrained by its type, the pointed string may not be altered. ## FLAGS The flags argument is zero, one or more bits set in a bitmask. - CURLU\_DEFAULT\_PORT If the handle has no port stored, this option makes *[curl\_url\_get(3)](/curl_url_get/3.md)* return the default port for the used scheme. - CURLU\_DEFAULT\_SCHEME If the handle has no scheme stored, this option makes *[curl\_url\_get(3)](/curl_url_get/3.md)* return the default scheme instead of error. - CURLU\_NO\_DEFAULT\_PORT Instructs *[curl\_url\_get(3)](/curl_url_get/3.md)* to not return a port number if it matches the default port for the scheme. - CURLU\_URLDECODE Asks *[curl\_url\_get(3)](/curl_url_get/3.md)* to URL decode the contents before returning it. It does not decode the scheme, the port number or the full URL. The query component also gets plus-to-space conversion as a bonus when this bit is set. Note that this URL decoding is charset unaware and you get a null-terminated string back with data that could be intended for a particular encoding. If there are byte values lower than 32 in the decoded string, the get operation returns an error instead. - CURLU\_URLENCODE If set, *[curl\_url\_get(3)](/curl_url_get/3.md)* URL encodes the hostname part when a full URL is retrieved. If not set (default), libcurl returns the URL with the hostname raw to support IDN names to appear as-is. IDN hostnames are typically using non-ASCII bytes that otherwise gets percent-encoded. Note that even when not asking for URL encoding, the '%' (byte 37) is URL encoded to make sure the hostname remains valid. - CURLU\_PUNYCODE If set and *CURLU\_URLENCODE* is not set, and asked to retrieve the **CURLUPART\_HOST** or **CURLUPART\_URL** parts, libcurl returns the host name in its punycode version if it contains any non-ASCII octets (and is an IDN name). If libcurl is built without IDN capabilities, using this bit makes *[curl\_url\_get(3)](/curl_url_get/3.md)* return *CURLUE\_LACKS\_IDN* if the hostname contains anything outside the ASCII range. (Added in curl 7.88.0) - CURLU\_PUNY2IDN If set and asked to retrieve the **CURLUPART\_HOST** or **CURLUPART\_URL** parts, libcurl returns the hostname in its IDN (International Domain Name) UTF-8 version if it otherwise is a punycode version. If the punycode name cannot be converted to IDN correctly, libcurl returns *CURLUE\_BAD\_HOSTNAME*. If libcurl is built without IDN capabilities, using this bit makes *[curl\_url\_get(3)](/curl_url_get/3.md)* return *CURLUE\_LACKS\_IDN* if the hostname is using punycode. (Added in curl 8.3.0) - CURLU\_GET\_EMPTY When this flag is used in curl\_url\_get(), it makes the function return empty query and fragments parts or when used in the full URL. By default, libcurl otherwise considers empty parts non-existing. An empty query part is one where this is nothing following the question mark (before the possible fragment). An empty fragments part is one where there is nothing following the hash sign. (Added in curl 8.8.0) - CURLU\_NO\_GUESS\_SCHEME When this flag is used in curl\_url\_get(), it treats the scheme as non-existing if it was set as a result of a previous guess; when CURLU\_GUESS\_SCHEME was used parsing a URL. Using this flag when getting CURLUPART\_SCHEME if the scheme was set as the result of a guess makes curl\_url\_get() return CURLUE\_NO\_SCHEME. Using this flag when getting CURLUPART\_URL if the scheme was set as the result of a guess makes curl\_url\_get() return the full URL without the scheme component. Such a URL can then only be parsed with curl\_url\_set() if CURLU\_GUESS\_SCHEME is used. (Added in curl 8.9.0) ## PARTS - CURLUPART\_URL When asked to return the full URL, *[curl\_url\_get(3)](/curl_url_get/3.md)* returns a slightly cleaned up version of the complete URL using all available parts. We advise using the *CURLU\_PUNYCODE* option to get the URL as "normalized" as possible since IDN allows hostnames to be written in many different ways that still end up the same punycode version. Zero-length queries and fragments are excluded from the URL unless CURLU\_GET\_EMPTY is set. - CURLUPART\_SCHEME Scheme cannot be URL decoded on get. - CURLUPART\_USER - CURLUPART\_PASSWORD - CURLUPART\_OPTIONS The options field is an optional field that might follow the password in the userinfo part. It is only recognized/used when parsing URLs for the following schemes: pop3, smtp and imap. The URL API still allows users to set and get this field independently of scheme when not parsing full URLs. - CURLUPART\_HOST The hostname. If it is an IPv6 numeric address, the zone id is not part of it but is provided separately in *CURLUPART\_ZONEID*. IPv6 numerical addresses are returned within brackets (\[\]). IPv6 names are normalized when set, which should make them as short as possible while maintaining correct syntax. - CURLUPART\_ZONEID If the hostname is a numeric IPv6 address, this field might also be set. - CURLUPART\_PORT A port cannot be URL decoded on get. This number is returned in a string like all other parts. That string is guaranteed to hold a valid port number in ASCII using base 10. - CURLUPART\_PATH The *part* is always at least a slash ('/') even if no path was supplied in the URL. A URL path always starts with a slash. - CURLUPART\_QUERY The initial question mark that denotes the beginning of the query part is a delimiter only. It is not part of the query contents. A not-present query returns *part* set to NULL. A zero-length query returns *part* as NULL unless CURLU\_GET\_EMPTY is set. The query part gets pluses converted to space when asked to URL decode on get with the CURLU\_URLDECODE bit. - CURLUPART\_FRAGMENT The initial hash sign that denotes the beginning of the fragment is a delimiter only. It is not part of the fragment contents. A not-present fragment returns *part* set to NULL. A zero-length fragment returns *part* as NULL unless CURLU\_GET\_EMPTY is set. ## PROTOCOLS This functionality affects all supported protocols ## EXAMPLE ``` int main(void) { CURLUcode rc; CURLU *url = curl_url(); rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0); if(!rc) { char *scheme; rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0); if(!rc) { printf("the scheme is %s\n", scheme); curl_free(scheme); } curl_url_cleanup(url); } } ``` ## AVAILABILITY Added in curl 7.62.0 ## RETURN VALUE Returns a CURLUcode error value, which is CURLUE\_OK (0) if everything went fine. See the *[libcurl-errors(3)](/libcurl-errors/3.md)* man page for the full list with descriptions. If this function returns an error, no URL part is returned. ## SEE ALSO [**CURLOPT\_CURLU**(3)](/CURLOPT_CURLU/3.md), [**curl\_url**(3)](/curl_url/3.md), [**curl\_url\_cleanup**(3)](/curl_url_cleanup/3.md), [**curl\_url\_dup**(3)](/curl_url_dup/3.md), [**curl\_url\_set**(3)](/curl_url_set/3.md), [**curl\_url\_strerror**(3)](/curl_url_strerror/3.md)