aboutsummaryrefslogtreecommitdiff
path: root/lib
diff options
context:
space:
mode:
authorDaniel Stenberg <daniel@haxx.se>2018-08-05 11:51:07 +0200
committerDaniel Stenberg <daniel@haxx.se>2018-09-08 15:36:11 +0200
commitfb30ac5a2d63773c529c19259754e2b306ac2e2e (patch)
tree0e26d4e3f085a2f50b19f3eba53bafce509e3826 /lib
parent17ca0ccff4aeacc63bf7fa90314ea58d23464617 (diff)
URL-API
See header file and man pages for API. All documented API details work and are tested in the 1560 test case. Closes #2842
Diffstat (limited to 'lib')
-rw-r--r--lib/Makefile.inc4
-rw-r--r--lib/escape.c20
-rw-r--r--lib/escape.h3
-rw-r--r--lib/imap.c3
-rw-r--r--lib/transfer.c314
-rw-r--r--lib/url.c44
-rw-r--r--lib/url.h2
-rw-r--r--lib/urlapi-int.h29
-rw-r--r--lib/urlapi.c1315
9 files changed, 1394 insertions, 340 deletions
diff --git a/lib/Makefile.inc b/lib/Makefile.inc
index ac0affeae..83a27dd0b 100644
--- a/lib/Makefile.inc
+++ b/lib/Makefile.inc
@@ -55,7 +55,7 @@ LIB_CFILES = file.c timeval.c base64.c hostip.c progress.c formdata.c \
curl_multibyte.c hostcheck.c conncache.c pipeline.c dotdot.c \
x509asn1.c http2.c smb.c curl_endian.c curl_des.c system_win32.c \
mime.c sha256.c setopt.c curl_path.c curl_ctype.c curl_range.c psl.c \
- doh.c
+ doh.c urlapi.c
LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \
formdata.h cookie.h http.h sendf.h ftp.h url.h dict.h if2ip.h \
@@ -75,7 +75,7 @@ LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \
curl_setup_once.h multihandle.h setup-vms.h pipeline.h dotdot.h \
x509asn1.h http2.h sigpipe.h smb.h curl_endian.h curl_des.h \
curl_printf.h system_win32.h rand.h mime.h curl_sha256.h setopt.h \
- curl_path.h curl_ctype.h curl_range.h psl.h doh.h
+ curl_path.h curl_ctype.h curl_range.h psl.h doh.h urlapi-int.h
LIB_RCFILES = libcurl.rc
diff --git a/lib/escape.c b/lib/escape.c
index 10774f058..afd3899f9 100644
--- a/lib/escape.c
+++ b/lib/escape.c
@@ -5,7 +5,7 @@
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
- * Copyright (C) 1998 - 2017, Daniel Stenberg, <daniel@haxx.se>, et al.
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
*
* This software is licensed as described in the file COPYING, which
* you should have received as part of this distribution. The terms
@@ -41,7 +41,7 @@
its behavior is altered by the current locale.
See https://tools.ietf.org/html/rfc3986#section-2.3
*/
-static bool Curl_isunreserved(unsigned char in)
+bool Curl_isunreserved(unsigned char in)
{
switch(in) {
case '0': case '1': case '2': case '3': case '4':
@@ -141,6 +141,8 @@ char *curl_easy_escape(struct Curl_easy *data, const char *string,
* Returns a pointer to a malloced string in *ostring with length given in
* *olen. If length == 0, the length is assumed to be strlen(string).
*
+ * 'data' can be set to NULL but then this function can't convert network
+ * data to host for non-ascii.
*/
CURLcode Curl_urldecode(struct Curl_easy *data,
const char *string, size_t length,
@@ -151,7 +153,7 @@ CURLcode Curl_urldecode(struct Curl_easy *data,
char *ns = malloc(alloc);
size_t strindex = 0;
unsigned long hex;
- CURLcode result;
+ CURLcode result = CURLE_OK;
if(!ns)
return CURLE_OUT_OF_MEMORY;
@@ -171,11 +173,13 @@ CURLcode Curl_urldecode(struct Curl_easy *data,
in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */
- result = Curl_convert_from_network(data, (char *)&in, 1);
- if(result) {
- /* Curl_convert_from_network calls failf if unsuccessful */
- free(ns);
- return result;
+ if(data) {
+ result = Curl_convert_from_network(data, (char *)&in, 1);
+ if(result) {
+ /* Curl_convert_from_network calls failf if unsuccessful */
+ free(ns);
+ return result;
+ }
}
string += 2;
diff --git a/lib/escape.h b/lib/escape.h
index 638666f03..666f1ad71 100644
--- a/lib/escape.h
+++ b/lib/escape.h
@@ -7,7 +7,7 @@
* | (__| |_| | _ <| |___
* \___|\___/|_| \_\_____|
*
- * Copyright (C) 1998 - 2011, Daniel Stenberg, <daniel@haxx.se>, et al.
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
*
* This software is licensed as described in the file COPYING, which
* you should have received as part of this distribution. The terms
@@ -24,6 +24,7 @@
/* Escape and unescape URL encoding in strings. The functions return a new
* allocated string or NULL if an error occurred. */
+bool Curl_isunreserved(unsigned char in);
CURLcode Curl_urldecode(struct Curl_easy *data,
const char *string, size_t length,
char **ostring, size_t *olen,
diff --git a/lib/imap.c b/lib/imap.c
index 4a9c01ae7..63fcb4d41 100644
--- a/lib/imap.c
+++ b/lib/imap.c
@@ -159,7 +159,8 @@ const struct Curl_handler Curl_handler_imaps = {
ZERO_NULL, /* connection_check */
PORT_IMAPS, /* defport */
CURLPROTO_IMAPS, /* protocol */
- PROTOPT_CLOSEACTION | PROTOPT_SSL /* flags */
+ PROTOPT_CLOSEACTION | PROTOPT_SSL | /* flags */
+ PROTOPT_URLOPTIONS
};
#endif
diff --git a/lib/transfer.c b/lib/transfer.c
index 63d01efd9..3d8089ee6 100644
--- a/lib/transfer.c
+++ b/lib/transfer.c
@@ -75,6 +75,7 @@
#include "http2.h"
#include "mime.h"
#include "strcase.h"
+#include "urlapi-int.h"
/* The last 3 #include files should be in this order */
#include "curl_printf.h"
@@ -1454,311 +1455,6 @@ CURLcode Curl_posttransfer(struct Curl_easy *data)
return CURLE_OK;
}
-#ifndef CURL_DISABLE_HTTP
-/*
- * Find the separator at the end of the host name, or the '?' in cases like
- * http://www.url.com?id=2380
- */
-static const char *find_host_sep(const char *url)
-{
- const char *sep;
- const char *query;
-
- /* Find the start of the hostname */
- sep = strstr(url, "//");
- if(!sep)
- sep = url;
- else
- sep += 2;
-
- query = strchr(sep, '?');
- sep = strchr(sep, '/');
-
- if(!sep)
- sep = url + strlen(url);
-
- if(!query)
- query = url + strlen(url);
-
- return sep < query ? sep : query;
-}
-
-/*
- * Decide in an encoding-independent manner whether a character in an
- * URL must be escaped. The same criterion must be used in strlen_url()
- * and strcpy_url().
- */
-static bool urlchar_needs_escaping(int c)
-{
- return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c));
-}
-
-/*
- * strlen_url() returns the length of the given URL if the spaces within the
- * URL were properly URL encoded.
- * URL encoding should be skipped for host names, otherwise IDN resolution
- * will fail.
- */
-static size_t strlen_url(const char *url, bool relative)
-{
- const unsigned char *ptr;
- size_t newlen = 0;
- bool left = TRUE; /* left side of the ? */
- const unsigned char *host_sep = (const unsigned char *) url;
-
- if(!relative)
- host_sep = (const unsigned char *) find_host_sep(url);
-
- for(ptr = (unsigned char *)url; *ptr; ptr++) {
-
- if(ptr < host_sep) {
- ++newlen;
- continue;
- }
-
- switch(*ptr) {
- case '?':
- left = FALSE;
- /* FALLTHROUGH */
- default:
- if(urlchar_needs_escaping(*ptr))
- newlen += 2;
- newlen++;
- break;
- case ' ':
- if(left)
- newlen += 3;
- else
- newlen++;
- break;
- }
- }
- return newlen;
-}
-
-/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in
- * the source URL accordingly.
- * URL encoding should be skipped for host names, otherwise IDN resolution
- * will fail.
- */
-static void strcpy_url(char *output, const char *url, bool relative)
-{
- /* we must add this with whitespace-replacing */
- bool left = TRUE;
- const unsigned char *iptr;
- char *optr = output;
- const unsigned char *host_sep = (const unsigned char *) url;
-
- if(!relative)
- host_sep = (const unsigned char *) find_host_sep(url);
-
- for(iptr = (unsigned char *)url; /* read from here */
- *iptr; /* until zero byte */
- iptr++) {
-
- if(iptr < host_sep) {
- *optr++ = *iptr;
- continue;
- }
-
- switch(*iptr) {
- case '?':
- left = FALSE;
- /* FALLTHROUGH */
- default:
- if(urlchar_needs_escaping(*iptr)) {
- snprintf(optr, 4, "%%%02x", *iptr);
- optr += 3;
- }
- else
- *optr++=*iptr;
- break;
- case ' ':
- if(left) {
- *optr++='%'; /* add a '%' */
- *optr++='2'; /* add a '2' */
- *optr++='0'; /* add a '0' */
- }
- else
- *optr++='+'; /* add a '+' here */
- break;
- }
- }
- *optr = 0; /* zero terminate output buffer */
-
-}
-
-/*
- * Returns true if the given URL is absolute (as opposed to relative)
- */
-static bool is_absolute_url(const char *url)
-{
- char prot[16]; /* URL protocol string storage */
- char letter; /* used for a silly sscanf */
-
- return (2 == sscanf(url, "%15[^?&/:]://%c", prot, &letter)) ? TRUE : FALSE;
-}
-
-/*
- * Concatenate a relative URL to a base URL making it absolute.
- * URL-encodes any spaces.
- * The returned pointer must be freed by the caller unless NULL
- * (returns NULL on out of memory).
- */
-static char *concat_url(const char *base, const char *relurl)
-{
- /***
- TRY to append this new path to the old URL
- to the right of the host part. Oh crap, this is doomed to cause
- problems in the future...
- */
- char *newest;
- char *protsep;
- char *pathsep;
- size_t newlen;
- bool host_changed = FALSE;
-
- const char *useurl = relurl;
- size_t urllen;
-
- /* we must make our own copy of the URL to play with, as it may
- point to read-only data */
- char *url_clone = strdup(base);
-
- if(!url_clone)
- return NULL; /* skip out of this NOW */
-
- /* protsep points to the start of the host name */
- protsep = strstr(url_clone, "//");
- if(!protsep)
- protsep = url_clone;
- else
- protsep += 2; /* pass the slashes */
-
- if('/' != relurl[0]) {
- int level = 0;
-
- /* First we need to find out if there's a ?-letter in the URL,
- and cut it and the right-side of that off */
- pathsep = strchr(protsep, '?');
- if(pathsep)
- *pathsep = 0;
-
- /* we have a relative path to append to the last slash if there's one
- available, or if the new URL is just a query string (starts with a
- '?') we append the new one at the end of the entire currently worked
- out URL */
- if(useurl[0] != '?') {
- pathsep = strrchr(protsep, '/');
- if(pathsep)
- *pathsep = 0;
- }
-
- /* Check if there's any slash after the host name, and if so, remember
- that position instead */
- pathsep = strchr(protsep, '/');
- if(pathsep)
- protsep = pathsep + 1;
- else
- protsep = NULL;
-
- /* now deal with one "./" or any amount of "../" in the newurl
- and act accordingly */
-
- if((useurl[0] == '.') && (useurl[1] == '/'))
- useurl += 2; /* just skip the "./" */
-
- while((useurl[0] == '.') &&
- (useurl[1] == '.') &&
- (useurl[2] == '/')) {
- level++;
- useurl += 3; /* pass the "../" */
- }
-
- if(protsep) {
- while(level--) {
- /* cut off one more level from the right of the original URL */
- pathsep = strrchr(protsep, '/');
- if(pathsep)
- *pathsep = 0;
- else {
- *protsep = 0;
- break;
- }
- }
- }
- }
- else {
- /* We got a new absolute path for this server */
-
- if((relurl[0] == '/') && (relurl[1] == '/')) {
- /* the new URL starts with //, just keep the protocol part from the
- original one */
- *protsep = 0;
- useurl = &relurl[2]; /* we keep the slashes from the original, so we
- skip the new ones */
- host_changed = TRUE;
- }
- else {
- /* cut off the original URL from the first slash, or deal with URLs
- without slash */
- pathsep = strchr(protsep, '/');
- if(pathsep) {
- /* When people use badly formatted URLs, such as
- "http://www.url.com?dir=/home/daniel" we must not use the first
- slash, if there's a ?-letter before it! */
- char *sep = strchr(protsep, '?');
- if(sep && (sep < pathsep))
- pathsep = sep;
- *pathsep = 0;
- }
- else {
- /* There was no slash. Now, since we might be operating on a badly
- formatted URL, such as "http://www.url.com?id=2380" which doesn't
- use a slash separator as it is supposed to, we need to check for a
- ?-letter as well! */
- pathsep = strchr(protsep, '?');
- if(pathsep)
- *pathsep = 0;
- }
- }
- }
-
- /* If the new part contains a space, this is a mighty stupid redirect
- but we still make an effort to do "right". To the left of a '?'
- letter we replace each space with %20 while it is replaced with '+'
- on the right side of the '?' letter.
- */
- newlen = strlen_url(useurl, !host_changed);
-
- urllen = strlen(url_clone);
-
- newest = malloc(urllen + 1 + /* possible slash */
- newlen + 1 /* zero byte */);
-
- if(!newest) {
- free(url_clone); /* don't leak this */
- return NULL;
- }
-
- /* copy over the root url part */
- memcpy(newest, url_clone, urllen);
-
- /* check if we need to append a slash */
- if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0]))
- ;
- else
- newest[urllen++]='/';
-
- /* then append the new piece on the right side */
- strcpy_url(&newest[urllen], useurl, !host_changed);
-
- free(url_clone);
-
- return newest;
-}
-#endif /* CURL_DISABLE_HTTP */
-
/*
* Curl_follow() handles the URL redirect magic. Pass in the 'newurl' string
* as given by the remote server and set up the new URL to request.
@@ -1810,12 +1506,12 @@ CURLcode Curl_follow(struct Curl_easy *data,
}
}
- if(!is_absolute_url(newurl)) {
+ if(!Curl_is_absolute_url(newurl, NULL, 8)) {
/***
*DANG* this is an RFC 2068 violation. The URL is supposed
to be absolute and this doesn't seem to be that!
*/
- char *absolute = concat_url(data->change.url, newurl);
+ char *absolute = Curl_concat_url(data->change.url, newurl);
if(!absolute)
return CURLE_OUT_OF_MEMORY;
newurl = absolute;
@@ -1824,7 +1520,7 @@ CURLcode Curl_follow(struct Curl_easy *data,
/* The new URL MAY contain space or high byte values, that means a mighty
stupid redirect URL but we still make an effort to do "right". */
char *newest;
- size_t newlen = strlen_url(newurl, FALSE);
+ size_t newlen = Curl_strlen_url(newurl, FALSE);
/* This is an absolute URL, don't allow the custom port number */
disallowport = TRUE;
@@ -1833,7 +1529,7 @@ CURLcode Curl_follow(struct Curl_easy *data,
if(!newest)
return CURLE_OUT_OF_MEMORY;
- strcpy_url(newest, newurl, FALSE); /* create a space-free URL */
+ Curl_strcpy_url(newest, newurl, FALSE); /* create a space-free URL */
newurl = newest; /* use this instead now */
}
diff --git a/lib/url.c b/lib/url.c
index 977f1d6ad..66faa400d 100644
--- a/lib/url.c
+++ b/lib/url.c
@@ -1944,30 +1944,37 @@ static struct connectdata *allocate_conn(struct Curl_easy *data)
return NULL;
}
-static CURLcode findprotocol(struct Curl_easy *data,
- struct connectdata *conn,
- const char *protostr)
+/* returns the handdler if the given scheme is built-in */
+const struct Curl_handler *Curl_builtin_scheme(const char *scheme)
{
const struct Curl_handler * const *pp;
const struct Curl_handler *p;
-
- /* Scan protocol handler table and match against 'protostr' to set a few
- variables based on the URL. Now that the handler may be changed later
- when the protocol specific setup function is called. */
- for(pp = protocols; (p = *pp) != NULL; pp++) {
- if(strcasecompare(p->scheme, protostr)) {
+ /* Scan protocol handler table and match against 'scheme'. The handler may
+ be changed later when the protocol specific setup function is called. */
+ for(pp = protocols; (p = *pp) != NULL; pp++)
+ if(strcasecompare(p->scheme, scheme))
/* Protocol found in table. Check if allowed */
- if(!(data->set.allowed_protocols & p->protocol))
- /* nope, get out */
- break;
+ return p;
+ return NULL; /* not found */
+}
- /* it is allowed for "normal" request, now do an extra check if this is
- the result of a redirect */
- if(data->state.this_is_a_follow &&
- !(data->set.redir_protocols & p->protocol))
- /* nope, get out */
- break;
+static CURLcode findprotocol(struct Curl_easy *data,
+ struct connectdata *conn,
+ const char *protostr)
+{
+ const struct Curl_handler *p = Curl_builtin_scheme(protostr);
+
+ if(p && /* Protocol found in table. Check if allowed */
+ (data->set.allowed_protocols & p->protocol)) {
+
+ /* it is allowed for "normal" request, now do an extra check if this is
+ the result of a redirect */
+ if(data->state.this_is_a_follow &&
+ !(data->set.redir_protocols & p->protocol))
+ /* nope, get out */
+ ;
+ else {
/* Perform setup complement if some. */
conn->handler = conn->given = p;
@@ -1976,7 +1983,6 @@ static CURLcode findprotocol(struct Curl_easy *data,
}
}
-
/* The protocol was not found in the table, but we don't have to assign it
to anything since it is already assigned to a dummy-struct in the
create_conn() function when the connectdata struct is allocated. */
diff --git a/lib/url.h b/lib/url.h
index 3eb733a0a..7b9aff5c4 100644
--- a/lib/url.h
+++ b/lib/url.h
@@ -79,6 +79,8 @@ void Curl_getoff_all_pipelines(struct Curl_easy *data,
CURLcode Curl_upkeep(struct conncache *conn_cache, void *data);
+const struct Curl_handler *Curl_builtin_scheme(const char *scheme);
+
#define CURL_DEFAULT_PROXY_PORT 1080 /* default proxy port unless specified */
#define CURL_DEFAULT_HTTPS_PROXY_PORT 443 /* default https proxy port unless
specified */
diff --git a/lib/urlapi-int.h b/lib/urlapi-int.h
new file mode 100644
index 000000000..7ac09fdb5
--- /dev/null
+++ b/lib/urlapi-int.h
@@ -0,0 +1,29 @@
+#ifndef HEADER_CURL_URLAPI_INT_H
+#define HEADER_CURL_URLAPI_INT_H
+/***************************************************************************
+ * _ _ ____ _
+ * Project ___| | | | _ \| |
+ * / __| | | | |_) | |
+ * | (__| |_| | _ <| |___
+ * \___|\___/|_| \_\_____|
+ *
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
+ *
+ * This software is licensed as described in the file COPYING, which
+ * you should have received as part of this distribution. The terms
+ * are also available at https://curl.haxx.se/docs/copyright.html.
+ *
+ * You may opt to use, copy, modify, merge, publish, distribute and/or sell
+ * copies of the Software, and permit persons to whom the Software is
+ * furnished to do so, under the terms of the COPYING file.
+ *
+ * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
+ * KIND, either express or implied.
+ *
+ ***************************************************************************/
+#include "curl_setup.h"
+bool Curl_is_absolute_url(const char *url, char *scheme, size_t buflen);
+char *Curl_concat_url(const char *base, const char *relurl);
+size_t Curl_strlen_url(const char *url, bool relative);
+void Curl_strcpy_url(char *output, const char *url, bool relative);
+#endif
diff --git a/lib/urlapi.c b/lib/urlapi.c
new file mode 100644
index 000000000..8287861f7
--- /dev/null
+++ b/lib/urlapi.c
@@ -0,0 +1,1315 @@
+/***************************************************************************
+ * _ _ ____ _
+ * Project ___| | | | _ \| |
+ * / __| | | | |_) | |
+ * | (__| |_| | _ <| |___
+ * \___|\___/|_| \_\_____|
+ *
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
+ *
+ * This software is licensed as described in the file COPYING, which
+ * you should have received as part of this distribution. The terms
+ * are also available at https://curl.haxx.se/docs/copyright.html.
+ *
+ * You may opt to use, copy, modify, merge, publish, distribute and/or sell
+ * copies of the Software, and permit persons to whom the Software is
+ * furnished to do so, under the terms of the COPYING file.
+ *
+ * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
+ * KIND, either express or implied.
+ *
+ ***************************************************************************/
+
+#include "curl_setup.h"
+
+#include "urldata.h"
+#include "urlapi-int.h"
+#include "strcase.h"
+#include "dotdot.h"
+#include "url.h"
+#include "escape.h"
+#include "curl_ctype.h"
+
+/* The last 3 #include files should be in this order */
+#include "curl_printf.h"
+#include "curl_memory.h"
+#include "memdebug.h"
+
+/* Internal representation of CURLU. Point to URL-encoded strings. */
+struct Curl_URL {
+ char *scheme;
+ char *user;
+ char *password;
+ char *options; /* IMAP only? */
+ char *host;
+ char *port;
+ char *path;
+ char *query;
+ char *fragment;
+
+ char *scratch; /* temporary scratch area */
+ long portnum; /* the numerical version */
+};
+
+#define DEFAULT_SCHEME "https"
+
+/* scheme is not URL encoded, the longest libcurl supported ones are 6
+ letters */
+#define MAX_SCHEME_LEN 8
+
+static void free_urlhandle(struct Curl_URL *u)
+{
+ free(u->scheme);
+ free(u->user);
+ free(u->password);
+ free(u->options);
+ free(u->host);
+ free(u->port);
+ free(u->path);
+ free(u->query);
+ free(u->fragment);
+ free(u->scratch);
+}
+
+/* move the full contents of one handle onto another and
+ free the original */
+static void mv_urlhandle(struct Curl_URL *from,
+ struct Curl_URL *to)
+{
+ free_urlhandle(to);
+ *to = *from;
+ free(from);
+}
+
+/*
+ * Find the separator at the end of the host name, or the '?' in cases like
+ * http://www.url.com?id=2380
+ */
+static const char *find_host_sep(const char *url)
+{
+ const char *sep;
+ const char *query;
+
+ /* Find the start of the hostname */
+ sep = strstr(url, "//");
+ if(!sep)
+ sep = url;
+ else
+ sep += 2;
+
+ query = strchr(sep, '?');
+ sep = strchr(sep, '/');
+
+ if(!sep)
+ sep = url + strlen(url);
+
+ if(!query)
+ query = url + strlen(url);
+
+ return sep < query ? sep : query;
+}
+
+/*
+ * Decide in an encoding-independent manner whether a character in an
+ * URL must be escaped. The same criterion must be used in strlen_url()
+ * and strcpy_url().
+ */
+static bool urlchar_needs_escaping(int c)
+{
+ return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c));
+}
+
+/*
+ * strlen_url() returns the length of the given URL if the spaces within the
+ * URL were properly URL encoded.
+ * URL encoding should be skipped for host names, otherwise IDN resolution
+ * will fail.
+ */
+size_t Curl_strlen_url(const char *url, bool relative)
+{
+ const unsigned char *ptr;
+ size_t newlen = 0;
+ bool left = TRUE; /* left side of the ? */
+ const unsigned char *host_sep = (const unsigned char *) url;
+
+ if(!relative)
+ host_sep = (const unsigned char *) find_host_sep(url);
+
+ for(ptr = (unsigned char *)url; *ptr; ptr++) {
+
+ if(ptr < host_sep) {
+ ++newlen;
+ continue;
+ }
+
+ switch(*ptr) {
+ case '?':
+ left = FALSE;
+ /* FALLTHROUGH */
+ default:
+ if(urlchar_needs_escaping(*ptr))
+ newlen += 2;
+ newlen++;
+ break;
+ case ' ':
+ if(left)
+ newlen += 3;
+ else
+ newlen++;
+ break;
+ }
+ }
+ return newlen;
+}
+
+/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in
+ * the source URL accordingly.
+ * URL encoding should be skipped for host names, otherwise IDN resolution
+ * will fail.
+ */
+void Curl_strcpy_url(char *output, const char *url, bool relative)
+{
+ /* we must add this with whitespace-replacing */
+ bool left = TRUE;
+ const unsigned char *iptr;
+ char *optr = output;
+ const unsigned char *host_sep = (const unsigned char *) url;
+
+ if(!relative)
+ host_sep = (const unsigned char *) find_host_sep(url);
+
+ for(iptr = (unsigned char *)url; /* read from here */
+ *iptr; /* until zero byte */
+ iptr++) {
+
+ if(iptr < host_sep) {
+ *optr++ = *iptr;
+ continue;
+ }
+
+ switch(*iptr) {
+ case '?':
+ left = FALSE;
+ /* FALLTHROUGH */
+ default:
+ if(urlchar_needs_escaping(*iptr)) {
+ snprintf(optr, 4, "%%%02x", *iptr);
+ optr += 3;
+ }
+ else
+ *optr++=*iptr;
+ break;
+ case ' ':
+ if(left) {
+ *optr++='%'; /* add a '%' */
+ *optr++='2'; /* add a '2' */
+ *optr++='0'; /* add a '0' */
+ }
+ else
+ *optr++='+'; /* add a '+' here */
+ break;
+ }
+ }
+ *optr = 0; /* zero terminate output buffer */
+
+}
+
+/*
+ * Returns true if the given URL is absolute (as opposed to relative) within
+ * the buffer size. Returns the scheme in the buffer if TRUE and 'buf' is
+ * non-NULL.
+ */
+bool Curl_is_absolute_url(const char *url, char *buf, size_t buflen)
+{
+ size_t i;
+ for(i = 0; i < buflen && url[i]; ++i) {
+ char s = url[i];
+ if(s == ':') {
+ if(buf)
+ buf[i] = 0;
+ return TRUE;
+ }
+ /* RFC 3986 3.1 explains:
+ scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
+ */
+ else if(ISALNUM(s) || (s == '+') || (s == '-') || (s == '.') ) {
+ if(buf)
+ buf[i] = (char)TOLOWER(s);
+ }
+ else
+ break;
+ }
+ return FALSE;
+}
+
+/*
+ * Concatenate a relative URL to a base URL making it absolute.
+ * URL-encodes any spaces.
+ * The returned pointer must be freed by the caller unless NULL
+ * (returns NULL on out of memory).
+ */
+char *Curl_concat_url(const char *base, const char *relurl)
+{
+ /***
+ TRY to append this new path to the old URL
+ to the right of the host part. Oh crap, this is doomed to cause
+ problems in the future...
+ */
+ char *newest;
+ char *protsep;
+ char *pathsep;
+ size_t newlen;
+ bool host_changed = FALSE;
+
+ const char *useurl = relurl;
+ size_t urllen;
+
+ /* we must make our own copy of the URL to play with, as it may
+ point to read-only data */
+ char *url_clone = strdup(base);
+
+ if(!url_clone)
+ return NULL; /* skip out of this NOW */
+
+ /* protsep points to the start of the host name */
+ protsep = strstr(url_clone, "//");
+ if(!protsep)
+ protsep = url_clone;
+ else
+ protsep += 2; /* pass the slashes */
+
+ if('/' != relurl[0]) {
+ int level = 0;
+
+ /* First we need to find out if there's a ?-letter in the URL,
+ and cut it and the right-side of that off */
+ pathsep = strchr(protsep, '?');
+ if(pathsep)
+ *pathsep = 0;
+
+ /* we have a relative path to append to the last slash if there's one
+ available, or if the new URL is just a query string (starts with a
+ '?') we append the new one at the end of the entire currently worked
+ out URL */
+ if(useurl[0] != '?') {
+ pathsep = strrchr(protsep, '/');
+ if(pathsep)
+ *pathsep = 0;
+ }
+
+ /* Check if there's any slash after the host name, and if so, remember
+ that position instead */
+ pathsep = strchr(protsep, '/');
+ if(pathsep)
+ protsep = pathsep + 1;
+ else
+ protsep = NULL;
+
+ /* now deal with one "./" or any amount of "../" in the newurl
+ and act accordingly */
+
+ if((useurl[0] == '.') && (useurl[1] == '/'))
+ useurl += 2; /* just skip the "./" */
+
+ while((useurl[0] == '.') &&
+ (useurl[1] == '.') &&
+ (useurl[2] == '/')) {
+ level++;
+ useurl += 3; /* pass the "../" */
+ }
+
+ if(protsep) {
+ while(level--) {
+ /* cut off one more level from the right of the original URL */
+ pathsep = strrchr(protsep, '/');
+ if(pathsep)
+ *pathsep = 0;
+ else {
+ *protsep = 0;
+ break;
+ }
+ }
+ }
+ }
+ else {
+ /* We got a new absolute path for this server */
+
+ if((relurl[0] == '/') && (relurl[1] == '/')) {
+ /* the new URL starts with //, just keep the protocol part from the
+ original one */
+ *protsep = 0;
+ useurl = &relurl[2]; /* we keep the slashes from the original, so we
+ skip the new ones */
+ host_changed = TRUE;
+ }
+ else {
+ /* cut off the original URL from the first slash, or deal with URLs
+ without slash */
+ pathsep = strchr(protsep, '/');
+ if(pathsep) {
+ /* When people use badly formatted URLs, such as
+ "http://www.url.com?dir=/home/daniel" we must not use the first
+ slash, if there's a ?-letter before it! */
+ char *sep = strchr(protsep, '?');
+ if(sep && (sep < pathsep))
+ pathsep = sep;
+ *pathsep = 0;
+ }
+ else {
+ /* There was no slash. Now, since we might be operating on a badly
+ formatted URL, such as "http://www.url.com?id=2380" which doesn't
+ use a slash separator as it is supposed to, we need to check for a
+ ?-letter as well! */
+ pathsep = strchr(protsep, '?');
+ if(pathsep)
+ *pathsep = 0;
+ }
+ }
+ }
+
+ /* If the new part contains a space, this is a mighty stupid redirect
+ but we still make an effort to do "right". To the left of a '?'
+ letter we replace each space with %20 while it is replaced with '+'
+ on the right side of the '?' letter.
+ */
+ newlen = Curl_strlen_url(useurl, !host_changed);
+
+ urllen = strlen(url_clone);
+
+ newest = malloc(urllen + 1 + /* possible slash */
+ newlen + 1 /* zero byte */);
+
+ if(!newest) {
+ free(url_clone); /* don't leak this */
+ return NULL;
+ }
+
+ /* copy over the root url part */
+ memcpy(newest, url_clone, urllen);
+
+ /* check if we need to append a slash */
+ if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0]))
+ ;
+ else
+ newest[urllen++]='/';
+
+ /* then append the new piece on the right side */
+ Curl_strcpy_url(&newest[urllen], useurl, !host_changed);
+
+ free(url_clone);
+
+ return newest;
+}
+
+/*
+ * parse_hostname_login()
+ *
+ * Parse the login details (user name, password and options) from the URL and
+ * strip them out of the host name
+ *
+ */
+static CURLUcode parse_hostname_login(struct Curl_URL *u,
+ const struct Curl_handler *h,
+ char **hostname,
+ unsigned int flags)
+{
+ CURLUcode result = CURLUE_OK;
+ CURLcode ccode;
+ char *userp = NULL;
+ char *passwdp = NULL;
+ char *optionsp = NULL;
+
+ /* At this point, we're hoping all the other special cases have
+ * been taken care of, so conn->host.name is at most
+ * [user[:password][;options]]@]hostname
+ *
+ * We need somewhere to put the embedded details, so do that first.
+ */
+
+ char *ptr = strchr(*hostname, '@');
+ char *login = *hostname;
+
+ if(!ptr)
+ goto out;
+
+ /* We will now try to extract the
+ * possible login information in a string like:
+ * ftp://user:password@ftp.my.site:8021/README */
+ *hostname = ++ptr;
+
+ /* We could use the login information in the URL so extract it. Only parse
+ options if the handler says we should. */
+ ccode = Curl_parse_login_details(login, ptr - login - 1,
+ &userp, &passwdp,
+ h->flags & PROTOPT_URLOPTIONS ?
+ &optionsp:NULL);
+ if(ccode) {
+ result = CURLUE_MALFORMED_INPUT;
+ goto out;
+ }
+
+ if(userp) {
+ if(flags & CURLU_DISALLOW_USER) {
+ /* Option DISALLOW_USER is set and url contains username. */
+ result = CURLUE_USER_NOT_ALLOWED;
+ goto out;
+ }
+
+ u->user = userp;
+ }
+
+ if(passwdp)
+ u->password = passwdp;
+
+ if(optionsp)
+ u->options = optionsp;
+
+ return CURLUE_OK;
+ out:
+
+ free(userp);
+ free(passwdp);
+ free(optionsp);
+
+ return result;
+}
+
+static CURLUcode parse_port(struct Curl_URL *u, char *hostname)
+{
+ char *portptr;
+ char endbracket;
+ int len;
+
+ if((1 == sscanf(hostname, "[%*45[0123456789abcdefABCDEF:.]%c%n",
+ &endbracket, &len)) &&
+ (']' == endbracket)) {
+ /* this is a RFC2732-style specified IP-address */
+ portptr = &hostname[len];
+ if (*portptr != ':')
+ return CURLUE_MALFORMED_INPUT;
+ }
+ else
+ portptr = strchr(hostname, ':');
+
+ if(portptr) {
+ char *rest;
+ long port;
+ char portbuf[7];
+
+ if(!ISDIGIT(portptr[1]))
+ return CURLUE_BAD_PORT_NUMBER;
+
+ port = strtol(portptr + 1, &rest, 10); /* Port number must be decimal */
+
+ if((port <= 0) || (port > 0xffff))
+ /* Single unix standard says port numbers are 16 bits long, but we don't
+ treat port zero as OK. */
+ return CURLUE_BAD_PORT_NUMBER;
+
+ if(rest[0])
+ return CURLUE_BAD_PORT_NUMBER;
+
+ if(rest != &portptr[1]) {
+ *portptr++ = '\0'; /* cut off the name there */
+ *rest = 0;
+ /* generate a new to get rid of leading zeroes etc */
+ snprintf(portbuf, sizeof(portbuf), "%ld", port);
+ u->portnum = port;
+ u->port = strdup(portbuf);
+ if(!u->port)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ else {
+ /* Browser behavior adaptation. If there's a colon with no digits after,
+ just cut off the name there which makes us ignore the colon and just
+ use the default port. Firefox and Chrome both do that. */
+ *portptr = '\0';
+ }
+ }
+
+ return CURLUE_OK;
+}
+
+/* scan for byte values < 31 or 127 */
+static CURLUcode junkscan(char *part)
+{
+ char badbytes[]={
+ /* */ 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+ 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
+ 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
+ 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
+ 0x7f,
+ 0x00 /* zero terminate */
+ };
+ if(part) {
+ size_t n = strlen(part);
+ size_t nfine = strcspn(part, badbytes);
+ if(nfine != n)
+ /* since we don't know which part is scanned, return a generic error
+ code */
+ return CURLUE_MALFORMED_INPUT;
+ }
+ return CURLUE_OK;
+}
+
+static CURLUcode hostname_check(char *hostname, unsigned int flags)
+{
+ const char *l; /* accepted characters */
+ size_t len;
+ size_t hlen = strlen(hostname);
+ (void)flags;
+
+ if(hostname[0] == '[') {
+ hostname++;
+ l = "0123456789abcdefABCDEF::.";
+ hlen -= 2;
+ }
+ else /* % for URL escaped letters */
+ l = "0123456789abcdefghijklimnopqrstuvwxyz-_.ABCDEFGHIJKLIMNOPQRSTUVWXYZ%";
+
+ len = strspn(hostname, l);
+ if(hlen != len)
+ /* hostname with bad content */
+ return CURLUE_MALFORMED_INPUT;
+
+ return CURLUE_OK;
+}
+
+#define HOSTNAME_END(x) (((x) == '/') || ((x) == '?') || ((x) == '#'))
+
+static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags)
+{
+ char *path;
+ bool path_alloced = FALSE;
+ char *hostname;
+ char *query = NULL;
+ char *fragment = NULL;
+ CURLUcode result;
+ bool url_has_scheme = FALSE;
+ char schemebuf[MAX_SCHEME_LEN];
+ char *schemep;
+ size_t schemelen = 0;
+ size_t urllen;
+ const struct Curl_handler *h = NULL;
+
+ if(!url)
+ return CURLUE_MALFORMED_INPUT;
+
+ /*************************************************************
+ * Parse the URL.
+ ************************************************************/
+ /* allocate scratch area */
+ urllen = strlen(url);
+ path = u->scratch = malloc(urllen * 2 + 2);
+ if(!path)
+ return CURLUE_OUT_OF_MEMORY;
+
+ hostname = &path[urllen + 1];
+ hostname[0] = 0;
+
+ /* MSDOS/Windows style drive prefix, eg c: in c:foo */
+#define STARTS_WITH_DRIVE_PREFIX(str) \
+ ((('a' <= str[0] && str[0] <= 'z') || \
+ ('A' <= str[0] && str[0] <= 'Z')) && \
+ (str[1] == ':'))
+
+ /* MSDOS/Windows style drive prefix, optionally with
+ * a '|' instead of ':', followed by a slash or NUL */
+#define STARTS_WITH_URL_DRIVE_PREFIX(str) \
+ ((('a' <= (str)[0] && (str)[0] <= 'z') || \
+ ('A' <= (str)[0] && (str)[0] <= 'Z')) && \
+ ((str)[1] == ':' || (str)[1] == '|') && \
+ ((str)[2] == '/' || (str)[2] == '\\' || (str)[2] == 0))
+
+ if(Curl_is_absolute_url(url, schemebuf, sizeof(schemebuf))) {
+ url_has_scheme = TRUE;
+ schemelen = strlen(schemebuf);
+ }
+
+ /* handle the file: scheme */
+ if(url_has_scheme && strcasecompare(schemebuf, "file")) {
+ /* path has been allocated large anough to hold this */
+ strcpy(path, &url[5]);
+
+ hostname = NULL; /* no host for file: URLs */
+ u->scheme = strdup("file");
+ if(!u->scheme)
+ return CURLUE_OUT_OF_MEMORY;
+
+ /* Extra handling URLs with an authority component (i.e. that start with
+ * "file://")
+ *
+ * We allow omitted hostname (e.g. file:/<path>) -- valid according to
+ * RFC 8089, but not the (current) WHAT-WG URL spec.
+ */
+ if(path[0] == '/' && path[1] == '/') {
+ /* swallow the two slashes */
+ char *ptr = &path[2];
+
+ /*
+ * According to RFC 8089, a file: URL can be reliably dereferenced if:
+ *
+ * o it has no/blank hostname, or
+ *
+ * o the hostname matches "localhost" (case-insensitively), or
+ *
+ * o the hostname is a FQDN that resolves to this machine.
+ *
+ * For brevity, we only consider URLs with empty, "localhost", or
+ * "127.0.0.1" hostnames as local.
+ *
+ * Additionally, there is an exception for URLs with a Windows drive
+ * letter in the authority (which was accidentally omitted from RFC 8089
+ * Appendix E, but believe me, it was meant to be there. --MK)
+ */
+ if(ptr[0] != '/' && !STARTS_WITH_URL_DRIVE_PREFIX(ptr)) {
+ /* the URL includes a host name, it must match "localhost" or
+ "127.0.0.1" to be valid */
+ if(!checkprefix("localhost/", ptr) &&
+ !checkprefix("127.0.0.1/", ptr)) {
+ /* Invalid file://hostname/, expected localhost or 127.0.0.1 or
+ none */
+ return CURLUE_MALFORMED_INPUT;
+ }
+ ptr += 9; /* now points to the slash after the host */
+ }
+
+ path = ptr;
+ }
+
+#if !defined(MSDOS) && !defined(WIN32) && !defined(__CYGWIN__)
+ /* Don't allow Windows drive letters when not in Windows.
+ * This catches both "file:/c:" and "file:c:" */
+ if(('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) ||
+ STARTS_WITH_URL_DRIVE_PREFIX(path)) {
+ /* File drive letters are only accepted in MSDOS/Windows */
+ return CURLUE_MALFORMED_INPUT;
+ }
+#else
+ /* If the path starts with a slash and a drive letter, ditch the slash */
+ if('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) {
+ /* This cannot be done with strcpy, as the memory chunks overlap! */
+ memmove(path, &path[1], strlen(&path[1]) + 1);
+ }
+#endif
+
+ }
+ else {
+ /* clear path */
+ const char *p;
+ const char *hostp;
+ size_t len;
+ path[0] = 0;
+
+ if(url_has_scheme) {
+ int i = 0;
+ p = &url[schemelen + 1];
+ while(p && (*p == '/') && (i < 4)) {
+ p++;
+ i++;
+ }
+ if((i < 1) || (i>3))
+ /* less than one or more than three slashes */
+ return CURLUE_MALFORMED_INPUT;
+
+ schemep = schemebuf;
+ if(!Curl_builtin_scheme(schemep) &&
+ !(flags & CURLU_NON_SUPPORT_SCHEME))
+ return CURLUE_UNSUPPORTED_SCHEME;
+
+ if(junkscan(schemep))
+ return CURLUE_MALFORMED_INPUT;
+ }
+ else {
+ /* no scheme! */
+
+ if(!(flags & CURLU_DEFAULT_SCHEME))
+ return CURLUE_MALFORMED_INPUT;
+ schemep = (char *) DEFAULT_SCHEME;
+
+ /*
+ * The URL was badly formatted, let's try without scheme specified.
+ */
+ p = url;
+ }
+ hostp = p; /* host name starts here */
+
+ while(*p && !HOSTNAME_END(*p)) /* find end of host name */
+ p++;
+
+ len = p - hostp;
+ if(!len)
+ return CURLUE_MALFORMED_INPUT;
+
+ memcpy(hostname, hostp, len);
+ hostname[len] = 0;
+
+ len = strlen(p);
+ memcpy(path, p, len);
+ path[len] = 0;
+
+ u->scheme = strdup(schemep);
+ if(!u->scheme)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+
+ /* if this is a known scheme, get some details */
+ h = Curl_builtin_scheme(u->scheme);
+
+ if(junkscan(path))
+ return CURLUE_MALFORMED_INPUT;
+
+ query = strchr(path, '?');
+ if(query)
+ *query++ = 0;
+
+ fragment = strchr(query?query:path, '#');
+ if(fragment)
+ *fragment++ = 0;
+
+ if(!path[0])
+ /* if there's no path set, unset */
+ path = NULL;
+ else if(!(flags & CURLU_PATH_AS_IS)) {
+ /* sanitise paths and remove ../ and ./ sequences according to RFC3986 */
+ char *newp = Curl_dedotdotify(path);
+ if(!newp)
+ return CURLUE_OUT_OF_MEMORY;
+
+ if(strcmp(newp, path)) {
+ /* if we got a new version */
+ path = newp;
+ path_alloced = TRUE;
+ }
+ else
+ free(newp);
+ }
+ if(path) {
+ u->path = path_alloced?path:strdup(path);
+ if(!u->path)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+
+ if(hostname) {
+ /*
+ * Parse the login details and strip them out of the host name.
+ */
+ if(junkscan(hostname))
+ return CURLUE_MALFORMED_INPUT;
+
+ result = parse_hostname_login(u, h, &hostname, flags);
+ if(result)
+ return result;
+
+ result = parse_port(u, hostname);
+ if(result)
+ return result;
+
+ result = hostname_check(hostname, flags);
+ if(result)
+ return result;
+
+ u->host = strdup(hostname);
+ if(!u->host)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+
+ if(query && query[0]) {
+ u->query = strdup(query);
+ if(!u->query)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ if(fragment && fragment[0]) {
+ u->fragment = strdup(fragment);
+ if(!u->fragment)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+
+ free(u->scratch);
+ u->scratch = NULL;
+
+ return CURLUE_OK;
+}
+
+/*
+ * Parse the URL and set the relevant members of the Curl_URL struct.
+ */
+static CURLUcode parseurl(const char *url, CURLU *u, unsigned int flags)
+{
+ CURLUcode result = seturl(url, u, flags);
+ if(result) {
+ free_urlhandle(u);
+ memset(u, 0, sizeof(struct Curl_URL));
+ }
+ return result;
+}
+
+/*
+ */
+CURLU *curl_url(void)
+{
+ return calloc(sizeof(struct Curl_URL), 1);
+}
+
+void curl_url_cleanup(CURLU *u)
+{
+ if(u) {
+ free_urlhandle(u);
+ free(u);
+ }
+}
+
+#define DUP(dest, src, name) \
+ if(src->name) { \
+ dest->name = strdup(src->name); \
+ if(!dest->name) \
+ goto fail; \
+ }
+
+CURLU *curl_url_dup(CURLU *in)
+{
+ struct Curl_URL *u = calloc(sizeof(struct Curl_URL), 1);
+ if(u) {
+ DUP(u, in, scheme);
+ DUP(u, in, user);
+ DUP(u, in, password);
+ DUP(u, in, options);
+ DUP(u, in, host);
+ DUP(u, in, port);
+ DUP(u, in, path);
+ DUP(u, in, query);
+ DUP(u, in, fragment);
+ u->portnum = in->portnum;
+ }
+ return u;
+ fail:
+ curl_url_cleanup(u);
+ return NULL;
+}
+
+CURLUcode curl_url_get(CURLU *u, CURLUPart what,
+ char **part, unsigned int flags)
+{
+ char *ptr;
+ CURLUcode ifmissing = CURLUE_UNKNOWN_PART;
+ char portbuf[7];
+ bool urldecode = (flags & CURLU_URLDECODE)?1:0;
+ bool plusdecode = FALSE;
+ (void)flags;
+ if(!u)
+ return CURLUE_BAD_HANDLE;
+ if(!part)
+ return CURLUE_BAD_PARTPOINTER;
+ *part = NULL;
+
+ switch(what) {
+ case CURLUPART_SCHEME:
+ ptr = u->scheme;
+ ifmissing = CURLUE_NO_SCHEME;
+ urldecode = FALSE; /* never for schemes */
+ break;
+ case CURLUPART_USER:
+ ptr = u->user;
+ ifmissing = CURLUE_NO_USER;
+ break;
+ case CURLUPART_PASSWORD:
+ ptr = u->password;
+ ifmissing = CURLUE_NO_PASSWORD;
+ break;
+ case CURLUPART_OPTIONS:
+ ptr = u->options;
+ ifmissing = CURLUE_NO_OPTIONS;
+ break;
+ case CURLUPART_HOST:
+ ptr = u->host;
+ ifmissing = CURLUE_NO_HOST;
+ break;
+ case CURLUPART_PORT:
+ ptr = u->port;
+ ifmissing = CURLUE_NO_PORT;
+ urldecode = FALSE; /* never for port */
+ if(!ptr && (flags & CURLU_DEFAULT_PORT) && u->scheme) {
+ /* there's no stored port number, but asked to deliver
+ a default one for the scheme */
+ const struct Curl_handler *h =
+ Curl_builtin_scheme(u->scheme);
+ if(h) {
+ snprintf(portbuf, sizeof(portbuf), "%ld", h->defport);
+ ptr = portbuf;
+ }
+ }
+ else if(ptr && u->scheme) {
+ /* there is a stored port number, but ask to inhibit if
+ it matches the default one for the scheme */
+ const struct Curl_handler *h =
+ Curl_builtin_scheme(u->scheme);
+ if(h && (h->defport == u->portnum) &&
+ (flags & CURLU_NO_DEFAULT_PORT))
+ ptr = NULL;
+ }
+ break;
+ case CURLUPART_PATH:
+ ptr = u->path;
+ if(!ptr) {
+ ptr = u->path = strdup("/");
+ if(!u->path)
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ break;
+ case CURLUPART_QUERY:
+ ptr = u->query;
+ ifmissing = CURLUE_NO_QUERY;
+ plusdecode = urldecode;
+ break;
+ case CURLUPART_FRAGMENT:
+ ptr = u->fragment;
+ ifmissing = CURLUE_NO_FRAGMENT;
+ break;
+ case CURLUPART_URL: {
+ char *url;
+ char *scheme;
+ char *options = u->options;
+ char *port = u->port;
+ urldecode = FALSE; /* not for the whole thing */
+ if(u->scheme && strcasecompare("file", u->scheme)) {
+ url = aprintf("file://%s%s%s",
+ u->path,
+ u->fragment? "#": "",
+ u->fragment? u->fragment : "");
+ }
+ else if(!u->host)
+ return CURLUE_NO_HOST;
+ else {
+ const struct Curl_handler *h = NULL;
+ if(u->scheme)
+ scheme = u->scheme;
+ else if(flags & CURLU_DEFAULT_SCHEME)
+ scheme = (char *) DEFAULT_SCHEME;
+ else
+ return CURLUE_NO_SCHEME;
+
+ if(scheme) {
+ h = Curl_builtin_scheme(scheme);
+ if(!port && (flags & CURLU_DEFAULT_PORT)) {
+ /* there's no stored port number, but asked to deliver
+ a default one for the scheme */
+ if(h) {
+ snprintf(portbuf, sizeof(portbuf), "%ld", h->defport);
+ port = portbuf;
+ }
+ }
+ else if(port) {
+ /* there is a stored port number, but asked to inhibit if it matches
+ the default one for the scheme */
+ if(h && (h->defport == u->portnum) &&
+ (flags & CURLU_NO_DEFAULT_PORT))
+ port = NULL;
+ }
+ }
+ if(h && !(h->flags & PROTOPT_URLOPTIONS))
+ options = NULL;
+
+ url = aprintf("%s://%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s",
+ scheme,
+ u->user ? u->user : "",
+ u->password ? ":": "",
+ u->password ? u->password : "",
+ options ? ";" : "",
+ options ? options : "",
+ (u->user || u->password || options) ? "@": "",
+ u->host,
+ port ? ":": "",
+ port ? port : "",
+ (u->path && (u->path[0] != '/')) ? "/": "",
+ u->path ? u->path : "/",
+ u->query? "?": "",
+ u->query? u->query : "",
+ u->fragment? "#": "",
+ u->fragment? u->fragment : "");
+ }
+ if(!url)
+ return CURLUE_OUT_OF_MEMORY;
+ *part = url;
+ return CURLUE_OK;
+ break;
+ }
+ default:
+ ptr = NULL;
+ }
+ if(ptr) {
+ *part = strdup(ptr);
+ if(!*part)
+ return CURLUE_OUT_OF_MEMORY;
+ if(plusdecode) {
+ /* convert + to space */
+ char *plus;
+ for(plus = *part; *plus; ++plus) {
+ if(*plus == '+')
+ *plus = ' ';
+ }
+ }
+ if(urldecode) {
+ char *decoded;
+ size_t dlen;
+ CURLcode res = Curl_urldecode(NULL, *part, 0, &decoded, &dlen, TRUE);
+ free(*part);
+ if(res) {
+ *part = NULL;
+ return CURLUE_URLDECODE;
+ }
+ *part = decoded;
+ }
+ return CURLUE_OK;
+ }
+ else
+ return ifmissing;
+}
+
+CURLUcode curl_url_set(CURLU *u, CURLUPart what,
+ const char *part, unsigned int flags)
+{
+ char **storep = NULL;
+ long port = 0;
+ bool urlencode = (flags & CURLU_URLENCODE)? 1 : 0;
+ bool plusencode = FALSE;
+ bool urlskipslash = FALSE;
+ bool appendquery = FALSE;
+
+ if(!u)
+ return CURLUE_BAD_HANDLE;
+ if(!part) {
+ /* setting a part to NULL clears it */
+ switch(what) {
+ case CURLUPART_URL:
+ break;
+ case CURLUPART_SCHEME:
+ storep = &u->scheme;
+ break;
+ case CURLUPART_USER:
+ storep = &u->user;
+ break;
+ case CURLUPART_PASSWORD:
+ storep = &u->password;
+ break;
+ case CURLUPART_OPTIONS:
+ storep = &u->options;
+ break;
+ case CURLUPART_HOST:
+ storep = &u->host;
+ break;
+ case CURLUPART_PORT:
+ storep = &u->port;
+ break;
+ case CURLUPART_PATH:
+ storep = &u->path;
+ break;
+ case CURLUPART_QUERY:
+ storep = &u->query;
+ break;
+ case CURLUPART_FRAGMENT:
+ storep = &u->fragment;
+ break;
+ default:
+ return CURLUE_UNKNOWN_PART;
+ }
+ if(storep && *storep) {
+ free(*storep);
+ *storep = NULL;
+ }
+ return CURLUE_OK;
+ }
+
+ switch(what) {
+ case CURLUPART_SCHEME:
+ if(!(flags & CURLU_NON_SUPPORT_SCHEME) &&
+ /* verify that it is a fine scheme */
+ !Curl_builtin_scheme(part))
+ return CURLUE_UNSUPPORTED_SCHEME;
+ storep = &u->scheme;
+ urlencode = FALSE; /* never */
+ break;
+ case CURLUPART_USER:
+ storep = &u->user;
+ break;
+ case CURLUPART_PASSWORD:
+ storep = &u->password;
+ break;
+ case CURLUPART_OPTIONS:
+ storep = &u->options;
+ break;
+ case CURLUPART_HOST:
+ storep = &u->host;
+ break;
+ case CURLUPART_PORT:
+ urlencode = FALSE; /* never */
+ port = strtol(part, NULL, 10); /* Port number must be decimal */
+ if((port <= 0) || (port > 0xffff))
+ return CURLUE_BAD_PORT_NUMBER;
+ storep = &u->port;
+ break;
+ case CURLUPART_PATH:
+ urlskipslash = TRUE;
+ storep = &u->path;
+ break;
+ case CURLUPART_QUERY:
+ plusencode = urlencode;
+ appendquery = (flags & CURLU_APPENDQUERY)?1:0;
+ storep = &u->query;
+ break;
+ case CURLUPART_FRAGMENT:
+ storep = &u->fragment;
+ break;
+ case CURLUPART_URL: {
+ /*
+ * Allow a new URL to replace the existing (if any) contents.
+ *
+ * If the existing contents is enough for a URL, allow a relative URL to
+ * replace it.
+ */
+ CURLUcode result;
+ char *oldurl;
+ char *redired_url;
+ CURLU *handle2;
+
+ if(Curl_is_absolute_url(part, NULL, MAX_SCHEME_LEN)) {
+ handle2 = curl_url();
+ if(!handle2)
+ return CURLUE_OUT_OF_MEMORY;
+ result = parseurl(part, handle2, flags);
+ if(!result)
+ mv_urlhandle(handle2, u);
+ else
+ curl_url_cleanup(handle2);
+ return result;
+ }
+ /* extract the full "old" URL to do the redirect on */
+ result = curl_url_get(u, CURLUPART_URL, &oldurl, flags);
+ if(result) {
+ /* couldn't get the old URL, just use the new! */
+ handle2 = curl_url();
+ if(!handle2)
+ return CURLUE_OUT_OF_MEMORY;
+ result = parseurl(part, handle2, flags);
+ if(!result)
+ mv_urlhandle(handle2, u);
+ else
+ curl_url_cleanup(handle2);
+ return result;
+ }
+
+ /* apply the relative part to create a new URL */
+ redired_url = Curl_concat_url(oldurl, part);
+ free(oldurl);
+ if(!redired_url)
+ return CURLUE_OUT_OF_MEMORY;
+
+ /* now parse the new URL */
+ handle2 = curl_url();
+ if(!handle2) {
+ free(redired_url);
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ result = parseurl(redired_url, handle2, flags);
+ free(redired_url);
+ if(!result)
+ mv_urlhandle(handle2, u);
+ else
+ curl_url_cleanup(handle2);
+ return result;
+ }
+ default:
+ return CURLUE_UNKNOWN_PART;
+ }
+ if(storep) {
+ const char *newp = part;
+ size_t nalloc = strlen(part);
+
+ if(urlencode) {
+ const char *i;
+ char *o;
+ bool free_part = FALSE;
+ char *enc = malloc(nalloc * 3 + 1); /* for worst case! */
+ if(!enc)
+ return CURLUE_OUT_OF_MEMORY;
+ if(plusencode) {
+ /* space to plus */
+ i = part;
+ for(o = enc; *i; ++o, ++i)
+ *o = (*i == ' ') ? '+' : *i;
+ *o = 0; /* zero terminate */
+ part = strdup(enc);
+ if(!part) {
+ free(enc);
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ free_part = TRUE;
+ }
+ for(i = part, o = enc; *i; i++) {
+ if(Curl_isunreserved(*i) ||
+ ((*i == '/') && urlskipslash) ||
+ ((*i == '=') && appendquery) ||
+ ((*i == '+') && plusencode)) {
+ *o = *i;
+ o++;
+ }
+ else {
+ snprintf(o, 4, "%%%02x", *i);
+ o += 3;
+ }
+ }
+ *o = 0; /* zero terminate */
+ newp = enc;
+ if(free_part)
+ free((char *)part);
+ }
+ else {
+ char *p;
+ newp = strdup(part);
+ if(!newp)
+ return CURLUE_OUT_OF_MEMORY;
+ p = (char *)newp;
+ while(*p) {
+ /* make sure percent encoded are lower case */
+ if((*p == '%') && ISXDIGIT(p[1]) && ISXDIGIT(p[2]) &&
+ (ISUPPER(p[1]) || ISUPPER(p[2]))) {
+ p[1] = (char)TOLOWER(p[1]);
+ p[2] = (char)TOLOWER(p[2]);
+ p += 3;
+ }
+ else
+ p++;
+ }
+ }
+
+ if(appendquery) {
+ /* Append the string onto the old query. Add a '&' separator if none is
+ present at the end of the exsting query already */
+ size_t querylen = u->query ? strlen(u->query) : 0;
+ bool addamperand = querylen && (u->query[querylen -1] != '&');
+ if(querylen) {
+ size_t newplen = strlen(newp);
+ char *p = malloc(querylen + addamperand + newplen + 1);
+ if(!p) {
+ free((char *)newp);
+ return CURLUE_OUT_OF_MEMORY;
+ }
+ strcpy(p, u->query); /* original query */
+ if(addamperand)
+ p[querylen] = '&'; /* ampersand */
+ strcpy(&p[querylen + addamperand], newp); /* new suffix */
+ free((char *)newp);
+ free(*storep);
+ *storep = p;
+ return CURLUE_OK;
+ }
+ }
+
+ free(*storep);
+ *storep = (char *)newp;
+ }
+ /* set after the string, to make it not assigned if the allocation above
+ fails */
+ if(port)
+ u->portnum = port;
+ return CURLUE_OK;
+}