Parse and normalize all URL's inside Dillo. More...

#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include "url.h"
#include "hsts.h"
#include "misc.h"
#include "msg.h"

Include dependency graph for url.c:

Go to the source code of this file.

Macros
#define	URL_STR_FIELD_CMP(s1, s2) (s1) && (s2) ? strcmp(s1,s2) : !(s1) && !(s2) ? 0 : (s1) ? 1 : -1

#define	URL_STR_FIELD_I_CMP(s1, s2) (s1) && (s2) ? dStrAsciiCasecmp(s1,s2) : !(s1) && !(s2) ? 0 : (s1) ? 1 : -1

Functions
char *	a_Url_str (const DilloUrl *u)
	Return the url as a string.

const char *	a_Url_hostname (const DilloUrl *u)
	Return the hostname as a string.

static DilloUrl *	Url_object_new (const char *uri_str)
	Create a DilloUrl object and initialize it.

void	a_Url_free (DilloUrl *url)
	Free a DilloUrl.

static Dstr *	Url_resolve_relative (const char RelStr, const char BaseStr)
	Resolve the URL as RFC3986 suggests.

DilloUrl *	a_Url_new (const char url_str, const char base_url)
	Transform (and resolve) an URL string into the respective DilloURL.

DilloUrl *	a_Url_dup (const DilloUrl *ori)
	Duplicate a Url structure.

int	a_Url_cmp (const DilloUrl A, const DilloUrl B)
	Compare two Url's to check if they're the same, or which one is bigger.

void	a_Url_set_flags (DilloUrl *u, int flags)
	Set DilloUrl flags.

void	a_Url_set_data (DilloUrl u, Dstr *data)
	Set DilloUrl data (like POST info, etc.)

void	a_Url_set_ismap_coords (DilloUrl u, char coord_str)
	Set DilloUrl ismap coordinates.

static int	Url_decode_hex_octet (const char *s)
	Given an hex octet (e.g., e3, 2F, 20), return the corresponding character if the octet is valid, and -1 otherwise.

char *	a_Url_decode_hex_str (const char *str)
	Parse possible hexadecimal octets in the URI path.

char *	a_Url_encode_hex_str (const char *str)
	Urlencode 'str'.

char *	a_Url_string_strip_delimiters (const char *str)
	RFC-3986 suggests this stripping when "importing" URLs from other media.

int	a_Url_host_type (const char *host)
	What type of host is this?

static uint_t	Url_host_public_internal_dots (const char *host)
	How many internal dots are in the public portion of this hostname?.

static const char *	Url_host_find_public_suffix (const char *host)
	Given a URL host string, return the portion that is public.

bool_t	a_Url_same_organization (const DilloUrl u1, const DilloUrl u2)

Variables
static const char *	HEX = "0123456789ABCDEF"

Detailed Description

Parse and normalize all URL's inside Dillo.

<scheme> <authority> <path> <query> and <fragment> point to 'buffer'.
'url_string' is built upon demand (transparent to the caller).
'hostname' and 'port' are also being handled on demand.

Definition in file url.c.

Macro Definition Documentation

◆ URL_STR_FIELD_CMP

#define URL_STR_FIELD_CMP	(	s1,
		s2
	)	(s1) && (s2) ? strcmp(s1,s2) : !(s1) && !(s2) ? 0 : (s1) ? 1 : -1

Definition at line 57 of file url.c.

◆ URL_STR_FIELD_I_CMP

#define URL_STR_FIELD_I_CMP	(	s1,
		s2
	)	(s1) && (s2) ? dStrAsciiCasecmp(s1,s2) : !(s1) && !(s2) ? 0 : (s1) ? 1 : -1

Definition at line 59 of file url.c.

Function Documentation

◆ a_Url_cmp()

int a_Url_cmp	(	const DilloUrl *	A,
		const DilloUrl *	B
	)

Compare two Url's to check if they're the same, or which one is bigger.

The fields which are compared here are: <scheme>, <authority>, <path>, <query> and <data> Other fields are left for the caller to check

Return value: 0 if equal, > 0 if A > B, < 0 if A < B.

Note: this function defines a sorting order different from strcmp!

Definition at line 506 of file url.c.

References DilloUrl::authority, DilloUrl::data, dReturn_val_if_fail, dStr_cmp(), DilloUrl::path, DilloUrl::query, DilloUrl::scheme, URL_STR_FIELD_CMP, and URL_STR_FIELD_I_CMP.

Referenced by a_Bw_add_url(), a_Bw_get_url_doc(), a_Capi_conn_abort_by_url(), a_History_add_url(), a_History_get_title_by_url(), a_History_set_title_by_url(), a_Nav_cancel_expect_if_eq(), a_Nav_push(), Cache_entry_by_url_cmp(), Cache_entry_cmp(), Dicache_entry_cmp(), Html_tag_open_meta(), DilloHtml::loadImages(), and Nav_open_url().

◆ a_Url_decode_hex_str()

char * a_Url_decode_hex_str ( const char * str )

Parse possible hexadecimal octets in the URI path.

Returns a new allocated string.

Definition at line 586 of file url.c.

References dNew, dRealloc(), dStrdup(), and Url_decode_hex_octet().

◆ a_Url_dup()

DilloUrl * a_Url_dup ( const DilloUrl * ori )

Duplicate a Url structure.

Definition at line 477 of file url.c.

References DilloUrl::data, dReturn_val_if_fail, dStr_append_l(), dStr_new(), dStr_sized_new(), DilloUrl::flags, DilloUrl::illegal_chars, DilloUrl::illegal_chars_spc, DilloUrl::ismap_url_len, DilloUrl::port, URL_DATA, Url_object_new(), URL_STR, URL_STR_, and DilloUrl::url_string.

Referenced by a_Bw_add_url(), a_Bw_expect(), a_Capi_set_vsource_url(), a_History_add_url(), a_Http_init(), a_Menu_bugmeter_popup(), a_Menu_form_popup(), a_Menu_image_popup(), a_Menu_link_popup(), a_Menu_page_popup(), a_Nav_expect_done(), a_Nav_redirection0(), a_Web_new(), DilloHtml::addCssUrl(), Auth_do_auth_dialog(), Cache_auth_entry(), Cache_entry_init(), Cache_process_queue(), Capi_conn_new(), Dicache_add_entry(), DilloHtml::DilloHtml(), Html_tag_open_form(), Html_tag_open_isindex(), Http_get(), Nav_reload_callback(), Nav_repush(), StyleEngine::StyleEngine(), Tls_conn_new(), and Tls_conn_new().

◆ a_Url_encode_hex_str()

char * a_Url_encode_hex_str ( const char * str )

Urlencode 'str'.

-RL :: According to the RFC 1738, only alphanumerics, the special characters "$-_.+!*'(),", and reserved characters ";/?:@=&" used for their reserved purposes may be used unencoded within a URL. We'll escape everything but alphanumeric and "-_.*" (as lynx). –Jcid

Note: the content type "application/x-www-form-urlencoded" is used: i.e., ' ' -> '+' and '
' -> CR LF (see HTML 4.01, Sec. 17.13.4)

Definition at line 620 of file url.c.

References d_isascii, dIsalnum, dNew, and HEX.

Referenced by Menu_bugmeter_validate(), and UIcmd_make_search_str().

◆ a_Url_free()

void a_Url_free ( DilloUrl * url )

Free a DilloUrl.

Do nothing if the argument is NULL

Definition at line 208 of file url.c.

References DilloUrl::authority, DilloUrl::buffer, DilloUrl::data, dFree(), dStr_free(), DilloUrl::hostname, TRUE, and DilloUrl::url_string.

Referenced by a_Bw_cancel_expect(), a_Bw_cleanup(), a_Bw_expect(), a_Bw_free(), a_Cache_init(), a_Capi_set_vsource_url(), a_Dicache_freeall(), a_History_freeall(), a_Http_freeall(), a_Menu_bugmeter_popup(), a_Menu_file_popup(), a_Menu_form_popup(), a_Menu_image_popup(), a_Menu_link_popup(), a_Menu_page_popup(), a_Nav_expect_done(), a_Nav_redirection0(), a_Prefs_freeall(), a_UIcmd_book(), a_UIcmd_open_file(), a_UIcmd_open_urlstr(), a_UIcmd_view_page_source(), a_Web_free(), StyleEngine::apply(), Auth_do_auth_dialog(), Cache_auth_callback(), Cache_entry_free(), Cache_process_queue(), Cache_redirect(), Cache_savelink_cb(), Capi_conn_unref(), Dicache_remove(), Html_tag_content_img(), Html_tag_content_map(), Html_tag_open_base(), Html_tag_open_form(), Html_tag_open_isindex(), Html_tag_open_link(), Html_tag_open_meta(), Html_tag_open_object(), Http_socket_free(), main(), Nav_redirection0_callback(), Nav_reload_callback(), Nav_repush(), CssParser::parseImport(), parseOption(), CssParser::parseUrl(), Tls_close_by_key(), Tls_close_by_key(), Url_resolve_relative(), DilloHtml::~DilloHtml(), and StyleEngine::~StyleEngine().

◆ a_Url_host_type()

int a_Url_host_type ( const char * host )

What type of host is this?

Definition at line 683 of file url.c.

References URL_HOST_ERROR, URL_HOST_IPV4, URL_HOST_IPV6, and URL_HOST_NAME.

Referenced by a_Tls_openssl_connect(), Cache_parse_header(), Tls_cert_authorities_print_summary(), and Url_host_find_public_suffix().

◆ a_Url_hostname()

const char * a_Url_hostname ( const DilloUrl * u )

Return the hostname as a string.

(initializing 'hostname' and 'port' fields if necessary) Note: a similar approach can be taken for user:password auth.

Definition at line 98 of file url.c.

References DilloUrl::authority, dStrAsciiCasecmp(), dStrndup(), DilloUrl::hostname, DilloUrl::port, DilloUrl::scheme, URL_HTTP_PORT, and URL_HTTPS_PORT.

Referenced by a_Url_new().

◆ a_Url_new()

DilloUrl * a_Url_new	(	const char *	url_str,
		const char *	base_url
	)

Transform (and resolve) an URL string into the respective DilloURL.

If URL = "http://dillo.sf.net:8080/index.html?long#part2" then the resulting DilloURL should be:

DilloURL = {
   url_string         = "http://dillo.sf.net:8080/index.html?long#part2"
   scheme             = "http"
   authority          = "dillo.sf.net:8080:
   path               = "/index.html"
   query              = "long"
   fragment           = "part2"
   hostname           = "dillo.sf.net"
   port               = 8080
   flags              = URL_Get
   data               = Dstr * ("")
   ismap_url_len      = 0
}