|
 |
f036d1 |
From 907da069c450ca20442839d9e95e3661a5c06b61 Mon Sep 17 00:00:00 2001
|
|
 |
f036d1 |
From: Daniel Stenberg <daniel@haxx.se>
|
|
 |
f036d1 |
Date: Sun, 5 Aug 2018 11:51:07 +0200
|
|
 |
f036d1 |
Subject: [PATCH 01/14] URL-API
|
|
 |
f036d1 |
|
|
 |
f036d1 |
See header file and man pages for API. All documented API details work
|
|
 |
f036d1 |
and are tested in the 1560 test case.
|
|
 |
f036d1 |
|
|
 |
f036d1 |
Closes #2842
|
|
 |
f036d1 |
|
|
 |
f036d1 |
Upstream-commit: fb30ac5a2d63773c529c19259754e2b306ac2e2e
|
|
 |
f036d1 |
Signed-off-by: Kamil Dudka <kdudka@redhat.com>
|
|
 |
f036d1 |
---
|
|
 |
f036d1 |
docs/libcurl/Makefile.inc | 1 +
|
|
 |
f036d1 |
docs/libcurl/curl_url.3 | 61 ++
|
|
 |
f036d1 |
docs/libcurl/curl_url_cleanup.3 | 44 +
|
|
 |
f036d1 |
docs/libcurl/curl_url_dup.3 | 52 ++
|
|
 |
f036d1 |
docs/libcurl/curl_url_get.3 | 110 +++
|
|
 |
f036d1 |
docs/libcurl/curl_url_set.3 | 120 +++
|
|
 |
f036d1 |
docs/libcurl/symbols-in-versions | 30 +
|
|
 |
f036d1 |
include/curl/Makefile.am | 4 +-
|
|
 |
f036d1 |
include/curl/curl.h | 1 +
|
|
 |
f036d1 |
include/curl/urlapi.h | 121 +++
|
|
 |
f036d1 |
lib/Makefile.inc | 5 +-
|
|
 |
f036d1 |
lib/escape.c | 20 +-
|
|
 |
f036d1 |
lib/escape.h | 3 +-
|
|
 |
f036d1 |
lib/imap.c | 3 +-
|
|
 |
f036d1 |
lib/transfer.c | 314 +------
|
|
 |
f036d1 |
lib/url.c | 44 +-
|
|
 |
f036d1 |
lib/url.h | 2 +
|
|
 |
f036d1 |
lib/{escape.h => urlapi-int.h} | 22 +-
|
|
 |
f036d1 |
lib/urlapi.c | 1315 ++++++++++++++++++++++++++++++
|
|
 |
f036d1 |
tests/data/Makefile.inc | 2 +
|
|
 |
f036d1 |
tests/data/test1560 | 28 +
|
|
 |
f036d1 |
tests/libtest/Makefile.am | 5 +
|
|
 |
f036d1 |
tests/libtest/Makefile.inc | 4 +
|
|
 |
f036d1 |
tests/libtest/lib1560.c | 760 +++++++++++++++++
|
|
 |
f036d1 |
24 files changed, 2716 insertions(+), 355 deletions(-)
|
|
 |
f036d1 |
create mode 100644 docs/libcurl/curl_url.3
|
|
 |
f036d1 |
create mode 100644 docs/libcurl/curl_url_cleanup.3
|
|
 |
f036d1 |
create mode 100644 docs/libcurl/curl_url_dup.3
|
|
 |
f036d1 |
create mode 100644 docs/libcurl/curl_url_get.3
|
|
 |
f036d1 |
create mode 100644 docs/libcurl/curl_url_set.3
|
|
 |
f036d1 |
create mode 100644 include/curl/urlapi.h
|
|
 |
f036d1 |
copy lib/{escape.h => urlapi-int.h} (66%)
|
|
 |
f036d1 |
create mode 100644 lib/urlapi.c
|
|
 |
f036d1 |
create mode 100644 tests/data/test1560
|
|
 |
f036d1 |
create mode 100644 tests/libtest/lib1560.c
|
|
 |
f036d1 |
|
|
 |
f036d1 |
diff --git a/docs/libcurl/Makefile.inc b/docs/libcurl/Makefile.inc
|
|
 |
f036d1 |
index eea48c4..955492c 100644
|
|
 |
f036d1 |
--- a/docs/libcurl/Makefile.inc
|
|
 |
f036d1 |
+++ b/docs/libcurl/Makefile.inc
|
|
 |
f036d1 |
@@ -22,4 +22,5 @@ man_MANS = curl_easy_cleanup.3 curl_easy_getinfo.3 curl_easy_init.3 \
|
|
 |
f036d1 |
curl_mime_data.3 curl_mime_data_cb.3 curl_mime_filedata.3 \
|
|
 |
f036d1 |
curl_mime_filename.3 curl_mime_subparts.3 \
|
|
 |
f036d1 |
curl_mime_type.3 curl_mime_headers.3 curl_mime_encoder.3 libcurl-env.3 \
|
|
 |
f036d1 |
+ curl_url.3 curl_url_cleanup.3 curl_url_dup.3 curl_url_get.3 curl_url_set.3 \
|
|
 |
f036d1 |
libcurl-security.3
|
|
 |
f036d1 |
diff --git a/docs/libcurl/curl_url.3 b/docs/libcurl/curl_url.3
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..0a56264
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/docs/libcurl/curl_url.3
|
|
 |
f036d1 |
@@ -0,0 +1,61 @@
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.\" * _ _ ____ _
|
|
 |
f036d1 |
+.\" * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+.\" * / __| | | | |_) | |
|
|
 |
f036d1 |
+.\" * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+.\" * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+.\" * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+.\" * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+.\" * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+.\" * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+.\" * KIND, either express or implied.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.TH curl_url 3 "6 Aug 2018" "libcurl" "libcurl Manual"
|
|
 |
f036d1 |
+.SH NAME
|
|
 |
f036d1 |
+curl_url - returns a new CURLU handle
|
|
 |
f036d1 |
+.SH SYNOPSIS
|
|
 |
f036d1 |
+.B #include <curl/curl.h>
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+CURLU *curl_url();
|
|
 |
f036d1 |
+.SH EXPERIMENTAL
|
|
 |
f036d1 |
+The URL API is considered \fBEXPERIMENTAL\fP until further notice. Please test
|
|
 |
f036d1 |
+it, report bugs and help us perfect it. Once proven to be reliable, the
|
|
 |
f036d1 |
+experimental label will be removed.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+While this API is marked experimental, we reserve the right to modify the API
|
|
 |
f036d1 |
+slightly if we deem it necessary and it makes it notably better or easier to
|
|
 |
f036d1 |
+use.
|
|
 |
f036d1 |
+.SH DESCRIPTION
|
|
 |
f036d1 |
+This function will allocates and returns a pointer to a fresh CURLU handle, to
|
|
 |
f036d1 |
+be used for further use of the URL API.
|
|
 |
f036d1 |
+.SH RETURN VALUE
|
|
 |
f036d1 |
+Returns a \fBCURLU *\fP if successful, or NULL if out of memory.
|
|
 |
f036d1 |
+.SH EXAMPLE
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+ CURLUcode rc;
|
|
 |
f036d1 |
+ CURLU *url = curl_url();
|
|
 |
f036d1 |
+ rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ char *scheme;
|
|
 |
f036d1 |
+ rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ printf("the scheme is %s\n", scheme);
|
|
 |
f036d1 |
+ curl_free(scheme);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ curl_url_cleanup(url);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH AVAILABILITY
|
|
 |
f036d1 |
+Added in curl 7.63.0
|
|
 |
f036d1 |
+.SH "SEE ALSO"
|
|
 |
f036d1 |
+.BR curl_url_cleanup "(3), " curl_url_get "(3), " curl_url_set "(3), "
|
|
 |
f036d1 |
+.BR curl_url_dup "(3), "
|
|
 |
f036d1 |
diff --git a/docs/libcurl/curl_url_cleanup.3 b/docs/libcurl/curl_url_cleanup.3
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..a8158b7
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/docs/libcurl/curl_url_cleanup.3
|
|
 |
f036d1 |
@@ -0,0 +1,44 @@
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.\" * _ _ ____ _
|
|
 |
f036d1 |
+.\" * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+.\" * / __| | | | |_) | |
|
|
 |
f036d1 |
+.\" * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+.\" * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+.\" * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+.\" * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+.\" * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+.\" * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+.\" * KIND, either express or implied.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.TH curl_url_cleanup 3 "6 Aug 2018" "libcurl" "libcurl Manual"
|
|
 |
f036d1 |
+.SH NAME
|
|
 |
f036d1 |
+curl_url_cleanup - free a CURLU handle
|
|
 |
f036d1 |
+.SH SYNOPSIS
|
|
 |
f036d1 |
+.B #include <curl/curl.h>
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+void curl_url_cleanup(CURLU *handle);
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH DESCRIPTION
|
|
 |
f036d1 |
+Frees all the resources associated with the given CURLU handle!
|
|
 |
f036d1 |
+.SH RETURN VALUE
|
|
 |
f036d1 |
+none
|
|
 |
f036d1 |
+.SH EXAMPLE
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+ CURLU *url = curl_url();
|
|
 |
f036d1 |
+ curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
|
|
 |
f036d1 |
+ curl_url_cleanup(url);
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH AVAILABILITY
|
|
 |
f036d1 |
+Added in curl 7.63.0
|
|
 |
f036d1 |
+.SH "SEE ALSO"
|
|
 |
f036d1 |
+.BR curl_url_dup "(3), " curl_url "(3), " curl_url_set "(3), "
|
|
 |
f036d1 |
+.BR curl_url_get "(3), "
|
|
 |
f036d1 |
diff --git a/docs/libcurl/curl_url_dup.3 b/docs/libcurl/curl_url_dup.3
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..4815dbd
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/docs/libcurl/curl_url_dup.3
|
|
 |
f036d1 |
@@ -0,0 +1,52 @@
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.\" * _ _ ____ _
|
|
 |
f036d1 |
+.\" * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+.\" * / __| | | | |_) | |
|
|
 |
f036d1 |
+.\" * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+.\" * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+.\" * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+.\" * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+.\" * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+.\" * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+.\" * KIND, either express or implied.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.TH curl_url_dup 3 "6 Aug 2018" "libcurl" "libcurl Manual"
|
|
 |
f036d1 |
+.SH NAME
|
|
 |
f036d1 |
+curl_url_dup - duplicate a CURLU handle
|
|
 |
f036d1 |
+.SH SYNOPSIS
|
|
 |
f036d1 |
+.B #include <curl/curl.h>
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+CURLU *curl_url_dup(CURLU *inhandle);
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH DESCRIPTION
|
|
 |
f036d1 |
+Duplicates a given CURLU \fIinhandle\fP and all its contents and returns a
|
|
 |
f036d1 |
+pointer to a new CURLU handle. The new handle also needs to be freed with
|
|
 |
f036d1 |
+\fIcurl_url_cleanup(3)\fP.
|
|
 |
f036d1 |
+.SH RETURN VALUE
|
|
 |
f036d1 |
+Returns a new handle or NULL if out of memory.
|
|
 |
f036d1 |
+.SH EXAMPLE
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+ CURLUcode rc;
|
|
 |
f036d1 |
+ CURLU *url = curl_url();
|
|
 |
f036d1 |
+ CURLU *url2;
|
|
 |
f036d1 |
+ rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ url2 = curl_url_dup(url); /* clone it! */
|
|
 |
f036d1 |
+ curl_url_cleanup(url2);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ curl_url_cleanup(url);
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH AVAILABILITY
|
|
 |
f036d1 |
+Added in curl 7.63.0
|
|
 |
f036d1 |
+.SH "SEE ALSO"
|
|
 |
f036d1 |
+.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), "
|
|
 |
f036d1 |
+.BR curl_url_get "(3), "
|
|
 |
f036d1 |
diff --git a/docs/libcurl/curl_url_get.3 b/docs/libcurl/curl_url_get.3
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..824d496
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/docs/libcurl/curl_url_get.3
|
|
 |
f036d1 |
@@ -0,0 +1,110 @@
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.\" * _ _ ____ _
|
|
 |
f036d1 |
+.\" * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+.\" * / __| | | | |_) | |
|
|
 |
f036d1 |
+.\" * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+.\" * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+.\" * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+.\" * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+.\" * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+.\" * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+.\" * KIND, either express or implied.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.TH curl_url_get 3 "6 Aug 2018" "libcurl" "libcurl Manual"
|
|
 |
f036d1 |
+.SH NAME
|
|
 |
f036d1 |
+curl_url_get - extract a part from a URL
|
|
 |
f036d1 |
+.SH SYNOPSIS
|
|
 |
f036d1 |
+.B #include <curl/curl.h>
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+CURLUcode curl_url_get(CURLU *url,
|
|
 |
f036d1 |
+ CURLUPart what,
|
|
 |
f036d1 |
+ char **part,
|
|
 |
f036d1 |
+ unsigned int flags)
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH DESCRIPTION
|
|
 |
f036d1 |
+Given the \fIurl\fP handle of an already parsed URL, this function lets the
|
|
 |
f036d1 |
+user extract individual pieces from it.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The \fIwhat\fP argument should be the particular part to extract (see list
|
|
 |
f036d1 |
+below) and \fIpart\fP points to a 'char *' to get updated to point to a newly
|
|
 |
f036d1 |
+allocated string with the contents.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The \fIflags\fP argument is a bitmask with individual features.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The returned part pointer must be freed with \fIcurl_free(3)\fP after use.
|
|
 |
f036d1 |
+.SH FLAGS
|
|
 |
f036d1 |
+The flags argument is zero, one or more bits set in a bitmask.
|
|
 |
f036d1 |
+.IP CURLU_DEFAULT_PORT
|
|
 |
f036d1 |
+If the handle has no port stored, this option will make \fIcurl_url_get(3)\fP
|
|
 |
f036d1 |
+return the default port for the used scheme.
|
|
 |
f036d1 |
+.IP CURLU_DEFAULT_SCHEME
|
|
 |
f036d1 |
+If the handle has no scheme stored, this option will make
|
|
 |
f036d1 |
+\fIcurl_url_get(3)\fP return the default scheme instead of error.
|
|
 |
f036d1 |
+.IP CURLU_NO_DEFAULT_PORT
|
|
 |
f036d1 |
+Instructs \fIcurl_url_get(3)\fP to not return a port number if it matches the
|
|
 |
f036d1 |
+default port for the scheme.
|
|
 |
f036d1 |
+.IP CURLU_URLDECODE
|
|
 |
f036d1 |
+Asks \fIcurl_url_get(3)\fP to URL decode the contents before returning it. It
|
|
 |
f036d1 |
+will not attempt to decode the scheme, the port number or the full URL.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The query component will also get plus-to-space convertion as a bonus when
|
|
 |
f036d1 |
+this bit is set.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+Note that this URL decoding is charset unaware and you will get a zero
|
|
 |
f036d1 |
+terminated string back with data that could be intended for a particular
|
|
 |
f036d1 |
+encoding.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+If there's any byte values lower than 32 in the decoded string, the get
|
|
 |
f036d1 |
+operation will return an error instead.
|
|
 |
f036d1 |
+.SH PARTS
|
|
 |
f036d1 |
+.IP CURLUPART_URL
|
|
 |
f036d1 |
+When asked to return the full URL, \fIcurl_url_get(3)\fP will return a
|
|
 |
f036d1 |
+normalized and possibly cleaned up version of what was previously parsed.
|
|
 |
f036d1 |
+.IP CURLUPART_SCHEME
|
|
 |
f036d1 |
+Scheme cannot be URL decoded on get.
|
|
 |
f036d1 |
+.IP CURLUPART_USER
|
|
 |
f036d1 |
+.IP CURLUPART_PASSWORD
|
|
 |
f036d1 |
+.IP CURLUPART_OPTIONS
|
|
 |
f036d1 |
+.IP CURLUPART_HOST
|
|
 |
f036d1 |
+.IP CURLUPART_PORT
|
|
 |
f036d1 |
+Port cannot be URL decoded on get.
|
|
 |
f036d1 |
+.IP CURLUPART_PATH
|
|
 |
f036d1 |
+.IP CURLUPART_QUERY
|
|
 |
f036d1 |
+The query part will also get pluses converted to space when asked to URL
|
|
 |
f036d1 |
+decode on get with the CURLU_URLDECODE bit.
|
|
 |
f036d1 |
+.IP CURLUPART_FRAGMENT
|
|
 |
f036d1 |
+.SH RETURN VALUE
|
|
 |
f036d1 |
+Returns a CURLUcode error value, which is CURLUE_OK (0) if everything went
|
|
 |
f036d1 |
+fine.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+If this function returns an error, no URL part is returned.
|
|
 |
f036d1 |
+.SH EXAMPLE
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+ CURLUcode rc;
|
|
 |
f036d1 |
+ CURLU *url = curl_url();
|
|
 |
f036d1 |
+ rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ char *scheme;
|
|
 |
f036d1 |
+ rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ printf("the scheme is %s\n", scheme);
|
|
 |
f036d1 |
+ curl_free(scheme);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ curl_url_cleanup(url);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH AVAILABILITY
|
|
 |
f036d1 |
+Added in curl 7.63.0
|
|
 |
f036d1 |
+.SH "SEE ALSO"
|
|
 |
f036d1 |
+.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), "
|
|
 |
f036d1 |
+.BR curl_url_dup "(3), "
|
|
 |
f036d1 |
diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..75fc0d9
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/docs/libcurl/curl_url_set.3
|
|
 |
f036d1 |
@@ -0,0 +1,120 @@
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.\" * _ _ ____ _
|
|
 |
f036d1 |
+.\" * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+.\" * / __| | | | |_) | |
|
|
 |
f036d1 |
+.\" * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+.\" * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+.\" * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+.\" * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+.\" * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+.\" * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+.\" * KIND, either express or implied.
|
|
 |
f036d1 |
+.\" *
|
|
 |
f036d1 |
+.\" **************************************************************************
|
|
 |
f036d1 |
+.TH curl_url_set 3 "6 Aug 2018" "libcurl" "libcurl Manual"
|
|
 |
f036d1 |
+.SH NAME
|
|
 |
f036d1 |
+curl_url_set - set a part from a URL
|
|
 |
f036d1 |
+.SH SYNOPSIS
|
|
 |
f036d1 |
+.B #include <curl/curl.h>
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+CURLUcode curl_url_set(CURLU *url,
|
|
 |
f036d1 |
+ CURLUPart part,
|
|
 |
f036d1 |
+ const char *content,
|
|
 |
f036d1 |
+ unsigned int flags)
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH DESCRIPTION
|
|
 |
f036d1 |
+Given the \fIurl\fP handle of an already parsed URL, this function lets the
|
|
 |
f036d1 |
+user set/update individual pieces of it.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The \fIpart\fP argument should identify the particular URL part (see list
|
|
 |
f036d1 |
+below) to set or change, with \fIcontent\fP pointing to a zero terminated
|
|
 |
f036d1 |
+string with the new contents for that URL part. The contents should be in the
|
|
 |
f036d1 |
+form and encoding they'd use in a URL: URL encoded.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+Setting a part to a NULL pointer will effectively remove that part's contents
|
|
 |
f036d1 |
+from the CURLU handle.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The \fIflags\fP argument is a bitmask with independent features.
|
|
 |
f036d1 |
+.SH PARTS
|
|
 |
f036d1 |
+.IP CURLUPART_URL
|
|
 |
f036d1 |
+Allows the full URL of the handle to be replaced. If the handle already is
|
|
 |
f036d1 |
+populated with a URL, the new URL can be relative to the previous.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+When successfully setting a new URL, relative or absolute, the handle contents
|
|
 |
f036d1 |
+will be replaced with the information of the newly set URL.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+Pass a pointer to a zero terminated string to the \fIurl\fP parameter. The
|
|
 |
f036d1 |
+string must point to a correctly formatted "RFC 3986+" URL or be a NULL
|
|
 |
f036d1 |
+pointer.
|
|
 |
f036d1 |
+.IP CURLUPART_SCHEME
|
|
 |
f036d1 |
+Scheme cannot be URL decoded on set.
|
|
 |
f036d1 |
+.IP CURLUPART_USER
|
|
 |
f036d1 |
+.IP CURLUPART_PASSWORD
|
|
 |
f036d1 |
+.IP CURLUPART_OPTIONS
|
|
 |
f036d1 |
+.IP CURLUPART_HOST
|
|
 |
f036d1 |
+The host name can use IDNA. The string must then be encoded as your locale
|
|
 |
f036d1 |
+says or UTF-8 (when winidn is used).
|
|
 |
f036d1 |
+.IP CURLUPART_PORT
|
|
 |
f036d1 |
+Port cannot be URL encoded on set.
|
|
 |
f036d1 |
+.IP CURLUPART_PATH
|
|
 |
f036d1 |
+If a path is set in the URL without a leading slash, a slash will be inserted
|
|
 |
f036d1 |
+automatically when this URL is read from the handle.
|
|
 |
f036d1 |
+.IP CURLUPART_QUERY
|
|
 |
f036d1 |
+The query part will also get spaces converted to pluses when asked to URL
|
|
 |
f036d1 |
+encode on set with the CURLU_URLENCODE bit.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+If used in with \fICURLU_APPENDQUERY\fP, the provided part will be appended on
|
|
 |
f036d1 |
+the end of the existing query - and if the previous part didn't end with an
|
|
 |
f036d1 |
+ampersand (&), an ampersand will be inserted before the new appended part.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+When \fCURLU_APPENDQUERY\fP is used together with \fICURLU_URLENCODE\fP,
|
|
 |
f036d1 |
+the '=' symbols will not be URL encoded.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The question mark in the URL is not part of the actual query contents.
|
|
 |
f036d1 |
+.IP CURLUPART_FRAGMENT
|
|
 |
f036d1 |
+The hash sign in the URL is not part of the actual fragment contents.
|
|
 |
f036d1 |
+.SH FLAGS
|
|
 |
f036d1 |
+The flags argument is zero, one or more bits set in a bitmask.
|
|
 |
f036d1 |
+.IP CURLU_NON_SUPPORT_SCHEME
|
|
 |
f036d1 |
+If set, allows \fIcurl_url_set(3)\fP to set a non-supported scheme.
|
|
 |
f036d1 |
+.IP CURLU_URLENCODE
|
|
 |
f036d1 |
+When set, \fIcurl_url_set(3)\fP URL encodes the part on entry, except for
|
|
 |
f036d1 |
+scheme, port and URL.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+When setting the path component with URL encoding enabled, the slash character
|
|
 |
f036d1 |
+will be skipped.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+The query part gets space-to-plus conversion before the URL conversion.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+This URL encoding is charset unaware and will convert the input on a
|
|
 |
f036d1 |
+byte-by-byte manner.
|
|
 |
f036d1 |
+.SH RETURN VALUE
|
|
 |
f036d1 |
+Returns a CURLUcode error value, which is CURLUE_OK (0) if everything went
|
|
 |
f036d1 |
+fine.
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+If this function returns an error, no URL part is returned.
|
|
 |
f036d1 |
+.SH EXAMPLE
|
|
 |
f036d1 |
+.nf
|
|
 |
f036d1 |
+ CURLUcode rc;
|
|
 |
f036d1 |
+ CURLU *url = curl_url();
|
|
 |
f036d1 |
+ rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0);
|
|
 |
f036d1 |
+ if(!rc) {
|
|
 |
f036d1 |
+ char *scheme;
|
|
 |
f036d1 |
+ /* change it to an FTP URL */
|
|
 |
f036d1 |
+ rc = curl_url_set(url, CURLUPART_SCHEME, "ftp", 0);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ curl_url_cleanup(url);
|
|
 |
f036d1 |
+.fi
|
|
 |
f036d1 |
+.SH AVAILABILITY
|
|
 |
f036d1 |
+Added in curl 7.63.0
|
|
 |
f036d1 |
+.SH "SEE ALSO"
|
|
 |
f036d1 |
+.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_get "(3), "
|
|
 |
f036d1 |
+.BR curl_url_dup "(3), "
|
|
 |
f036d1 |
diff --git a/docs/libcurl/symbols-in-versions b/docs/libcurl/symbols-in-versions
|
|
 |
f036d1 |
index 7448b4f..c797cb7 100644
|
|
 |
f036d1 |
--- a/docs/libcurl/symbols-in-versions
|
|
 |
f036d1 |
+++ b/docs/libcurl/symbols-in-versions
|
|
 |
f036d1 |
@@ -718,6 +718,36 @@ CURLSSLSET_NO_BACKENDS 7.56.0
|
|
 |
f036d1 |
CURLSSLSET_OK 7.56.0
|
|
 |
f036d1 |
CURLSSLSET_TOO_LATE 7.56.0
|
|
 |
f036d1 |
CURLSSLSET_UNKNOWN_BACKEND 7.56.0
|
|
 |
f036d1 |
+CURLUPART_FRAGMENT 7.62.0
|
|
 |
f036d1 |
+CURLUPART_HOST 7.62.0
|
|
 |
f036d1 |
+CURLUPART_OPTIONS 7.62.0
|
|
 |
f036d1 |
+CURLUPART_PASSWORD 7.62.0
|
|
 |
f036d1 |
+CURLUPART_PATH 7.62.0
|
|
 |
f036d1 |
+CURLUPART_PORT 7.62.0
|
|
 |
f036d1 |
+CURLUPART_QUERY 7.62.0
|
|
 |
f036d1 |
+CURLUPART_SCHEME 7.62.0
|
|
 |
f036d1 |
+CURLUPART_URL 7.62.0
|
|
 |
f036d1 |
+CURLUPART_USER 7.62.0
|
|
 |
f036d1 |
+CURLUE_BAD_HANDLE 7.62.0
|
|
 |
f036d1 |
+CURLUE_BAD_PARTPOINTER 7.62.0
|
|
 |
f036d1 |
+CURLUE_BAD_PORT_NUMBER 7.62.0
|
|
 |
f036d1 |
+CURLUE_MALFORMED_INPUT 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_FRAGMENT 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_HOST 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_OPTIONS 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_PASSWORD 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_PATH 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_PORT 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_QUERY 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_SCHEME 7.62.0
|
|
 |
f036d1 |
+CURLUE_NO_USER 7.62.0
|
|
 |
f036d1 |
+CURLUE_OK 7.62.0
|
|
 |
f036d1 |
+CURLUE_OUT_OF_MEMORY 7.62.0
|
|
 |
f036d1 |
+CURLUE_RELATIVE 7.62.0
|
|
 |
f036d1 |
+CURLUE_UNKNOWN_PART 7.62.0
|
|
 |
f036d1 |
+CURLUE_UNSUPPORTED_SCHEME 7.62.0
|
|
 |
f036d1 |
+CURLUE_URLDECODE 7.62.0
|
|
 |
f036d1 |
+CURLUE_USER_NOT_ALLOWED 7.62.0
|
|
 |
f036d1 |
CURLUSESSL_ALL 7.17.0
|
|
 |
f036d1 |
CURLUSESSL_CONTROL 7.17.0
|
|
 |
f036d1 |
CURLUSESSL_NONE 7.17.0
|
|
 |
f036d1 |
diff --git a/include/curl/Makefile.am b/include/curl/Makefile.am
|
|
 |
f036d1 |
index 989d4a2..bf5f061 100644
|
|
 |
f036d1 |
--- a/include/curl/Makefile.am
|
|
 |
f036d1 |
+++ b/include/curl/Makefile.am
|
|
 |
f036d1 |
@@ -5,7 +5,7 @@
|
|
 |
f036d1 |
# | (__| |_| | _ <| |___
|
|
 |
f036d1 |
# \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
#
|
|
 |
f036d1 |
-# Copyright (C) 1998 - 2017, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+# Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
#
|
|
 |
f036d1 |
# This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
# you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
@@ -21,7 +21,7 @@
|
|
 |
f036d1 |
###########################################################################
|
|
 |
f036d1 |
pkginclude_HEADERS = \
|
|
 |
f036d1 |
curl.h curlver.h easy.h mprintf.h stdcheaders.h multi.h \
|
|
 |
f036d1 |
- typecheck-gcc.h system.h
|
|
 |
f036d1 |
+ typecheck-gcc.h system.h urlapi.h
|
|
 |
f036d1 |
|
|
 |
f036d1 |
pkgincludedir= $(includedir)/curl
|
|
 |
f036d1 |
|
|
 |
f036d1 |
diff --git a/include/curl/curl.h b/include/curl/curl.h
|
|
 |
f036d1 |
index 067b34d..8f473e2 100644
|
|
 |
f036d1 |
--- a/include/curl/curl.h
|
|
 |
f036d1 |
+++ b/include/curl/curl.h
|
|
 |
f036d1 |
@@ -2779,6 +2779,7 @@ CURL_EXTERN CURLcode curl_easy_pause(CURL *handle, int bitmask);
|
|
 |
f036d1 |
stuff before they can be included! */
|
|
 |
f036d1 |
#include "easy.h" /* nothing in curl is fun without the easy stuff */
|
|
 |
f036d1 |
#include "multi.h"
|
|
 |
f036d1 |
+#include "urlapi.h"
|
|
 |
f036d1 |
|
|
 |
f036d1 |
/* the typechecker doesn't work in C++ (yet) */
|
|
 |
f036d1 |
#if defined(__GNUC__) && defined(__GNUC_MINOR__) && \
|
|
 |
f036d1 |
diff --git a/include/curl/urlapi.h b/include/curl/urlapi.h
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..b16cfce
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/include/curl/urlapi.h
|
|
 |
f036d1 |
@@ -0,0 +1,121 @@
|
|
 |
f036d1 |
+#ifndef __CURL_URLAPI_H
|
|
 |
f036d1 |
+#define __CURL_URLAPI_H
|
|
 |
f036d1 |
+/***************************************************************************
|
|
 |
f036d1 |
+ * _ _ ____ _
|
|
 |
f036d1 |
+ * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+ * / __| | | | |_) | |
|
|
 |
f036d1 |
+ * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+ * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * Copyright (C) 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+ * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+ * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+ * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+ * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+ * KIND, either express or implied.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ ***************************************************************************/
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#ifdef __cplusplus
|
|
 |
f036d1 |
+extern "C" {
|
|
 |
f036d1 |
+#endif
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* the error codes for the URL API */
|
|
 |
f036d1 |
+typedef enum {
|
|
 |
f036d1 |
+ CURLUE_OK,
|
|
 |
f036d1 |
+ CURLUE_BAD_HANDLE, /* 1 */
|
|
 |
f036d1 |
+ CURLUE_BAD_PARTPOINTER, /* 2 */
|
|
 |
f036d1 |
+ CURLUE_MALFORMED_INPUT, /* 3 */
|
|
 |
f036d1 |
+ CURLUE_BAD_PORT_NUMBER, /* 4 */
|
|
 |
f036d1 |
+ CURLUE_UNSUPPORTED_SCHEME, /* 5 */
|
|
 |
f036d1 |
+ CURLUE_URLDECODE, /* 6 */
|
|
 |
f036d1 |
+ CURLUE_RELATIVE, /* 7 */
|
|
 |
f036d1 |
+ CURLUE_USER_NOT_ALLOWED, /* 8 */
|
|
 |
f036d1 |
+ CURLUE_UNKNOWN_PART, /* 9 */
|
|
 |
f036d1 |
+ CURLUE_NO_SCHEME, /* 10 */
|
|
 |
f036d1 |
+ CURLUE_NO_USER, /* 11 */
|
|
 |
f036d1 |
+ CURLUE_NO_PASSWORD, /* 12 */
|
|
 |
f036d1 |
+ CURLUE_NO_OPTIONS, /* 13 */
|
|
 |
f036d1 |
+ CURLUE_NO_HOST, /* 14 */
|
|
 |
f036d1 |
+ CURLUE_NO_PORT, /* 15 */
|
|
 |
f036d1 |
+ CURLUE_NO_PATH, /* 16 */
|
|
 |
f036d1 |
+ CURLUE_NO_QUERY, /* 17 */
|
|
 |
f036d1 |
+ CURLUE_NO_FRAGMENT, /* 18 */
|
|
 |
f036d1 |
+ CURLUE_OUT_OF_MEMORY /* 19 */
|
|
 |
f036d1 |
+} CURLUcode;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+typedef enum {
|
|
 |
f036d1 |
+ CURLUPART_URL,
|
|
 |
f036d1 |
+ CURLUPART_SCHEME,
|
|
 |
f036d1 |
+ CURLUPART_USER,
|
|
 |
f036d1 |
+ CURLUPART_PASSWORD,
|
|
 |
f036d1 |
+ CURLUPART_OPTIONS,
|
|
 |
f036d1 |
+ CURLUPART_HOST,
|
|
 |
f036d1 |
+ CURLUPART_PORT,
|
|
 |
f036d1 |
+ CURLUPART_PATH,
|
|
 |
f036d1 |
+ CURLUPART_QUERY,
|
|
 |
f036d1 |
+ CURLUPART_FRAGMENT
|
|
 |
f036d1 |
+} CURLUPart;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#define CURLU_DEFAULT_PORT (1<<0) /* return default port number */
|
|
 |
f036d1 |
+#define CURLU_NO_DEFAULT_PORT (1<<1) /* act as if no port number was set,
|
|
 |
f036d1 |
+ if the port number matches the
|
|
 |
f036d1 |
+ default for the scheme */
|
|
 |
f036d1 |
+#define CURLU_DEFAULT_SCHEME (1<<2) /* return default scheme if
|
|
 |
f036d1 |
+ missing */
|
|
 |
f036d1 |
+#define CURLU_NON_SUPPORT_SCHEME (1<<3) /* allow non-supported scheme */
|
|
 |
f036d1 |
+#define CURLU_PATH_AS_IS (1<<4) /* leave dot sequences */
|
|
 |
f036d1 |
+#define CURLU_DISALLOW_USER (1<<5) /* no user+password allowed */
|
|
 |
f036d1 |
+#define CURLU_URLDECODE (1<<6) /* URL decode on get */
|
|
 |
f036d1 |
+#define CURLU_URLENCODE (1<<7) /* URL encode on set */
|
|
 |
f036d1 |
+#define CURLU_APPENDQUERY (1<<8) /* append a form style part */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+typedef struct Curl_URL CURLU;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * curl_url() creates a new CURLU handle and returns a pointer to it.
|
|
 |
f036d1 |
+ * Must be freed with curl_url_cleanup().
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+CURL_EXTERN CURLU *curl_url(void);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * curl_url_cleanup() frees the CURLU handle and related resources used for
|
|
 |
f036d1 |
+ * the URL parsing. It will not free strings previously returned with the URL
|
|
 |
f036d1 |
+ * API.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+CURL_EXTERN void curl_url_cleanup(CURLU *handle);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * curl_url_dup() duplicates a CURLU handle and returns a new copy. The new
|
|
 |
f036d1 |
+ * handle must also be freed with curl_url_cleanup().
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+CURL_EXTERN CURLU *curl_url_dup(CURLU *in);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * curl_url_get() extracts a specific part of the URL from a CURLU
|
|
 |
f036d1 |
+ * handle. Returns error code. The returned pointer MUST be freed with
|
|
 |
f036d1 |
+ * curl_free() afterwards.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+CURL_EXTERN CURLUcode curl_url_get(CURLU *handle, CURLUPart what,
|
|
 |
f036d1 |
+ char **part, unsigned int flags);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * curl_url_set() sets a specific part of the URL in a CURLU handle. Returns
|
|
 |
f036d1 |
+ * error code. The passed in string will be copied. Passing a NULL instead of
|
|
 |
f036d1 |
+ * a part string, clears that part.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+CURL_EXTERN CURLUcode curl_url_set(CURLU *handle, CURLUPart what,
|
|
 |
f036d1 |
+ const char *part, unsigned int flags);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#ifdef __cplusplus
|
|
 |
f036d1 |
+} /* end of extern "C" */
|
|
 |
f036d1 |
+#endif
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#endif
|
|
 |
f036d1 |
diff --git a/lib/Makefile.inc b/lib/Makefile.inc
|
|
 |
f036d1 |
index 76ca6d0..1ff82e1 100644
|
|
 |
f036d1 |
--- a/lib/Makefile.inc
|
|
 |
f036d1 |
+++ b/lib/Makefile.inc
|
|
 |
f036d1 |
@@ -54,7 +54,8 @@ LIB_CFILES = file.c timeval.c base64.c hostip.c progress.c formdata.c \
|
|
 |
f036d1 |
http_ntlm.c curl_ntlm_wb.c curl_ntlm_core.c curl_sasl.c rand.c \
|
|
 |
f036d1 |
curl_multibyte.c hostcheck.c conncache.c pipeline.c dotdot.c \
|
|
 |
f036d1 |
x509asn1.c http2.c smb.c curl_endian.c curl_des.c system_win32.c \
|
|
 |
f036d1 |
- mime.c sha256.c setopt.c curl_path.c curl_ctype.c curl_range.c psl.c
|
|
 |
f036d1 |
+ mime.c sha256.c setopt.c curl_path.c curl_ctype.c curl_range.c psl.c \
|
|
 |
f036d1 |
+ urlapi.c
|
|
 |
f036d1 |
|
|
 |
f036d1 |
LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \
|
|
 |
f036d1 |
formdata.h cookie.h http.h sendf.h ftp.h url.h dict.h if2ip.h \
|
|
 |
f036d1 |
@@ -74,7 +75,7 @@ LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \
|
|
 |
f036d1 |
curl_setup_once.h multihandle.h setup-vms.h pipeline.h dotdot.h \
|
|
 |
f036d1 |
x509asn1.h http2.h sigpipe.h smb.h curl_endian.h curl_des.h \
|
|
 |
f036d1 |
curl_printf.h system_win32.h rand.h mime.h curl_sha256.h setopt.h \
|
|
 |
f036d1 |
- curl_path.h curl_ctype.h curl_range.h psl.h
|
|
 |
f036d1 |
+ curl_path.h curl_ctype.h curl_range.h psl.h urlapi-int.h
|
|
 |
f036d1 |
|
|
 |
f036d1 |
LIB_RCFILES = libcurl.rc
|
|
 |
f036d1 |
|
|
 |
f036d1 |
diff --git a/lib/escape.c b/lib/escape.c
|
|
 |
f036d1 |
index 10774f0..afd3899 100644
|
|
 |
f036d1 |
--- a/lib/escape.c
|
|
 |
f036d1 |
+++ b/lib/escape.c
|
|
 |
f036d1 |
@@ -5,7 +5,7 @@
|
|
 |
f036d1 |
* | (__| |_| | _ <| |___
|
|
 |
f036d1 |
* \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
- * Copyright (C) 1998 - 2017, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
* This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
* you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
@@ -41,7 +41,7 @@
|
|
 |
f036d1 |
its behavior is altered by the current locale.
|
|
 |
f036d1 |
See https://tools.ietf.org/html/rfc3986#section-2.3
|
|
 |
f036d1 |
*/
|
|
 |
f036d1 |
-static bool Curl_isunreserved(unsigned char in)
|
|
 |
f036d1 |
+bool Curl_isunreserved(unsigned char in)
|
|
 |
f036d1 |
{
|
|
 |
f036d1 |
switch(in) {
|
|
 |
f036d1 |
case '0': case '1': case '2': case '3': case '4':
|
|
 |
f036d1 |
@@ -141,6 +141,8 @@ char *curl_easy_escape(struct Curl_easy *data, const char *string,
|
|
 |
f036d1 |
* Returns a pointer to a malloced string in *ostring with length given in
|
|
 |
f036d1 |
* *olen. If length == 0, the length is assumed to be strlen(string).
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
+ * 'data' can be set to NULL but then this function can't convert network
|
|
 |
f036d1 |
+ * data to host for non-ascii.
|
|
 |
f036d1 |
*/
|
|
 |
f036d1 |
CURLcode Curl_urldecode(struct Curl_easy *data,
|
|
 |
f036d1 |
const char *string, size_t length,
|
|
 |
f036d1 |
@@ -151,7 +153,7 @@ CURLcode Curl_urldecode(struct Curl_easy *data,
|
|
 |
f036d1 |
char *ns = malloc(alloc);
|
|
 |
f036d1 |
size_t strindex = 0;
|
|
 |
f036d1 |
unsigned long hex;
|
|
 |
f036d1 |
- CURLcode result;
|
|
 |
f036d1 |
+ CURLcode result = CURLE_OK;
|
|
 |
f036d1 |
|
|
 |
f036d1 |
if(!ns)
|
|
 |
f036d1 |
return CURLE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
@@ -171,11 +173,13 @@ CURLcode Curl_urldecode(struct Curl_easy *data,
|
|
 |
f036d1 |
|
|
 |
f036d1 |
in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */
|
|
 |
f036d1 |
|
|
 |
f036d1 |
- result = Curl_convert_from_network(data, (char *)&in, 1);
|
|
 |
f036d1 |
- if(result) {
|
|
 |
f036d1 |
- /* Curl_convert_from_network calls failf if unsuccessful */
|
|
 |
f036d1 |
- free(ns);
|
|
 |
f036d1 |
- return result;
|
|
 |
f036d1 |
+ if(data) {
|
|
 |
f036d1 |
+ result = Curl_convert_from_network(data, (char *)&in, 1);
|
|
 |
f036d1 |
+ if(result) {
|
|
 |
f036d1 |
+ /* Curl_convert_from_network calls failf if unsuccessful */
|
|
 |
f036d1 |
+ free(ns);
|
|
 |
f036d1 |
+ return result;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
string += 2;
|
|
 |
f036d1 |
diff --git a/lib/escape.h b/lib/escape.h
|
|
 |
f036d1 |
index 638666f..666f1ad 100644
|
|
 |
f036d1 |
--- a/lib/escape.h
|
|
 |
f036d1 |
+++ b/lib/escape.h
|
|
 |
f036d1 |
@@ -7,7 +7,7 @@
|
|
 |
f036d1 |
* | (__| |_| | _ <| |___
|
|
 |
f036d1 |
* \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
- * Copyright (C) 1998 - 2011, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
* This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
* you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
@@ -24,6 +24,7 @@
|
|
 |
f036d1 |
/* Escape and unescape URL encoding in strings. The functions return a new
|
|
 |
f036d1 |
* allocated string or NULL if an error occurred. */
|
|
 |
f036d1 |
|
|
 |
f036d1 |
+bool Curl_isunreserved(unsigned char in);
|
|
 |
f036d1 |
CURLcode Curl_urldecode(struct Curl_easy *data,
|
|
 |
f036d1 |
const char *string, size_t length,
|
|
 |
f036d1 |
char **ostring, size_t *olen,
|
|
 |
f036d1 |
diff --git a/lib/imap.c b/lib/imap.c
|
|
 |
f036d1 |
index 942fe7d..28962c1 100644
|
|
 |
f036d1 |
--- a/lib/imap.c
|
|
 |
f036d1 |
+++ b/lib/imap.c
|
|
 |
f036d1 |
@@ -159,7 +159,8 @@ const struct Curl_handler Curl_handler_imaps = {
|
|
 |
f036d1 |
ZERO_NULL, /* connection_check */
|
|
 |
f036d1 |
PORT_IMAPS, /* defport */
|
|
 |
f036d1 |
CURLPROTO_IMAPS, /* protocol */
|
|
 |
f036d1 |
- PROTOPT_CLOSEACTION | PROTOPT_SSL /* flags */
|
|
 |
f036d1 |
+ PROTOPT_CLOSEACTION | PROTOPT_SSL | /* flags */
|
|
 |
f036d1 |
+ PROTOPT_URLOPTIONS
|
|
 |
f036d1 |
};
|
|
 |
f036d1 |
#endif
|
|
 |
f036d1 |
|
|
 |
f036d1 |
diff --git a/lib/transfer.c b/lib/transfer.c
|
|
 |
f036d1 |
index 7159d5c..ecd1063 100644
|
|
 |
f036d1 |
--- a/lib/transfer.c
|
|
 |
f036d1 |
+++ b/lib/transfer.c
|
|
 |
f036d1 |
@@ -75,6 +75,7 @@
|
|
 |
f036d1 |
#include "http2.h"
|
|
 |
f036d1 |
#include "mime.h"
|
|
 |
f036d1 |
#include "strcase.h"
|
|
 |
f036d1 |
+#include "urlapi-int.h"
|
|
 |
f036d1 |
|
|
 |
f036d1 |
/* The last 3 #include files should be in this order */
|
|
 |
f036d1 |
#include "curl_printf.h"
|
|
 |
f036d1 |
@@ -1453,311 +1454,6 @@ CURLcode Curl_posttransfer(struct Curl_easy *data)
|
|
 |
f036d1 |
return CURLE_OK;
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
-#ifndef CURL_DISABLE_HTTP
|
|
 |
f036d1 |
-/*
|
|
 |
f036d1 |
- * Find the separator at the end of the host name, or the '?' in cases like
|
|
 |
f036d1 |
- * http://www.url.com?id=2380
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static const char *find_host_sep(const char *url)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- const char *sep;
|
|
 |
f036d1 |
- const char *query;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* Find the start of the hostname */
|
|
 |
f036d1 |
- sep = strstr(url, "//");
|
|
 |
f036d1 |
- if(!sep)
|
|
 |
f036d1 |
- sep = url;
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- sep += 2;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- query = strchr(sep, '?');
|
|
 |
f036d1 |
- sep = strchr(sep, '/');
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!sep)
|
|
 |
f036d1 |
- sep = url + strlen(url);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!query)
|
|
 |
f036d1 |
- query = url + strlen(url);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- return sep < query ? sep : query;
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-/*
|
|
 |
f036d1 |
- * Decide in an encoding-independent manner whether a character in an
|
|
 |
f036d1 |
- * URL must be escaped. The same criterion must be used in strlen_url()
|
|
 |
f036d1 |
- * and strcpy_url().
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static bool urlchar_needs_escaping(int c)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c));
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-/*
|
|
 |
f036d1 |
- * strlen_url() returns the length of the given URL if the spaces within the
|
|
 |
f036d1 |
- * URL were properly URL encoded.
|
|
 |
f036d1 |
- * URL encoding should be skipped for host names, otherwise IDN resolution
|
|
 |
f036d1 |
- * will fail.
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static size_t strlen_url(const char *url, bool relative)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- const unsigned char *ptr;
|
|
 |
f036d1 |
- size_t newlen = 0;
|
|
 |
f036d1 |
- bool left = TRUE; /* left side of the ? */
|
|
 |
f036d1 |
- const unsigned char *host_sep = (const unsigned char *) url;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!relative)
|
|
 |
f036d1 |
- host_sep = (const unsigned char *) find_host_sep(url);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- for(ptr = (unsigned char *)url; *ptr; ptr++) {
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(ptr < host_sep) {
|
|
 |
f036d1 |
- ++newlen;
|
|
 |
f036d1 |
- continue;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- switch(*ptr) {
|
|
 |
f036d1 |
- case '?':
|
|
 |
f036d1 |
- left = FALSE;
|
|
 |
f036d1 |
- /* FALLTHROUGH */
|
|
 |
f036d1 |
- default:
|
|
 |
f036d1 |
- if(urlchar_needs_escaping(*ptr))
|
|
 |
f036d1 |
- newlen += 2;
|
|
 |
f036d1 |
- newlen++;
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
- case ' ':
|
|
 |
f036d1 |
- if(left)
|
|
 |
f036d1 |
- newlen += 3;
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- newlen++;
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- return newlen;
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in
|
|
 |
f036d1 |
- * the source URL accordingly.
|
|
 |
f036d1 |
- * URL encoding should be skipped for host names, otherwise IDN resolution
|
|
 |
f036d1 |
- * will fail.
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static void strcpy_url(char *output, const char *url, bool relative)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- /* we must add this with whitespace-replacing */
|
|
 |
f036d1 |
- bool left = TRUE;
|
|
 |
f036d1 |
- const unsigned char *iptr;
|
|
 |
f036d1 |
- char *optr = output;
|
|
 |
f036d1 |
- const unsigned char *host_sep = (const unsigned char *) url;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!relative)
|
|
 |
f036d1 |
- host_sep = (const unsigned char *) find_host_sep(url);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- for(iptr = (unsigned char *)url; /* read from here */
|
|
 |
f036d1 |
- *iptr; /* until zero byte */
|
|
 |
f036d1 |
- iptr++) {
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(iptr < host_sep) {
|
|
 |
f036d1 |
- *optr++ = *iptr;
|
|
 |
f036d1 |
- continue;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- switch(*iptr) {
|
|
 |
f036d1 |
- case '?':
|
|
 |
f036d1 |
- left = FALSE;
|
|
 |
f036d1 |
- /* FALLTHROUGH */
|
|
 |
f036d1 |
- default:
|
|
 |
f036d1 |
- if(urlchar_needs_escaping(*iptr)) {
|
|
 |
f036d1 |
- snprintf(optr, 4, "%%%02x", *iptr);
|
|
 |
f036d1 |
- optr += 3;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- *optr++=*iptr;
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
- case ' ':
|
|
 |
f036d1 |
- if(left) {
|
|
 |
f036d1 |
- *optr++='%'; /* add a '%' */
|
|
 |
f036d1 |
- *optr++='2'; /* add a '2' */
|
|
 |
f036d1 |
- *optr++='0'; /* add a '0' */
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- *optr++='+'; /* add a '+' here */
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- *optr = 0; /* zero terminate output buffer */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-/*
|
|
 |
f036d1 |
- * Returns true if the given URL is absolute (as opposed to relative)
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static bool is_absolute_url(const char *url)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- char prot[16]; /* URL protocol string storage */
|
|
 |
f036d1 |
- char letter; /* used for a silly sscanf */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- return (2 == sscanf(url, "%15[^?&/:]://%c", prot, &letter)) ? TRUE : FALSE;
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-/*
|
|
 |
f036d1 |
- * Concatenate a relative URL to a base URL making it absolute.
|
|
 |
f036d1 |
- * URL-encodes any spaces.
|
|
 |
f036d1 |
- * The returned pointer must be freed by the caller unless NULL
|
|
 |
f036d1 |
- * (returns NULL on out of memory).
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
-static char *concat_url(const char *base, const char *relurl)
|
|
 |
f036d1 |
-{
|
|
 |
f036d1 |
- /***
|
|
 |
f036d1 |
- TRY to append this new path to the old URL
|
|
 |
f036d1 |
- to the right of the host part. Oh crap, this is doomed to cause
|
|
 |
f036d1 |
- problems in the future...
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
- char *newest;
|
|
 |
f036d1 |
- char *protsep;
|
|
 |
f036d1 |
- char *pathsep;
|
|
 |
f036d1 |
- size_t newlen;
|
|
 |
f036d1 |
- bool host_changed = FALSE;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- const char *useurl = relurl;
|
|
 |
f036d1 |
- size_t urllen;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* we must make our own copy of the URL to play with, as it may
|
|
 |
f036d1 |
- point to read-only data */
|
|
 |
f036d1 |
- char *url_clone = strdup(base);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!url_clone)
|
|
 |
f036d1 |
- return NULL; /* skip out of this NOW */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* protsep points to the start of the host name */
|
|
 |
f036d1 |
- protsep = strstr(url_clone, "//");
|
|
 |
f036d1 |
- if(!protsep)
|
|
 |
f036d1 |
- protsep = url_clone;
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- protsep += 2; /* pass the slashes */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if('/' != relurl[0]) {
|
|
 |
f036d1 |
- int level = 0;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* First we need to find out if there's a ?-letter in the URL,
|
|
 |
f036d1 |
- and cut it and the right-side of that off */
|
|
 |
f036d1 |
- pathsep = strchr(protsep, '?');
|
|
 |
f036d1 |
- if(pathsep)
|
|
 |
f036d1 |
- *pathsep = 0;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* we have a relative path to append to the last slash if there's one
|
|
 |
f036d1 |
- available, or if the new URL is just a query string (starts with a
|
|
 |
f036d1 |
- '?') we append the new one at the end of the entire currently worked
|
|
 |
f036d1 |
- out URL */
|
|
 |
f036d1 |
- if(useurl[0] != '?') {
|
|
 |
f036d1 |
- pathsep = strrchr(protsep, '/');
|
|
 |
f036d1 |
- if(pathsep)
|
|
 |
f036d1 |
- *pathsep = 0;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* Check if there's any slash after the host name, and if so, remember
|
|
 |
f036d1 |
- that position instead */
|
|
 |
f036d1 |
- pathsep = strchr(protsep, '/');
|
|
 |
f036d1 |
- if(pathsep)
|
|
 |
f036d1 |
- protsep = pathsep + 1;
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- protsep = NULL;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* now deal with one "./" or any amount of "../" in the newurl
|
|
 |
f036d1 |
- and act accordingly */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if((useurl[0] == '.') && (useurl[1] == '/'))
|
|
 |
f036d1 |
- useurl += 2; /* just skip the "./" */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- while((useurl[0] == '.') &&
|
|
 |
f036d1 |
- (useurl[1] == '.') &&
|
|
 |
f036d1 |
- (useurl[2] == '/')) {
|
|
 |
f036d1 |
- level++;
|
|
 |
f036d1 |
- useurl += 3; /* pass the "../" */
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(protsep) {
|
|
 |
f036d1 |
- while(level--) {
|
|
 |
f036d1 |
- /* cut off one more level from the right of the original URL */
|
|
 |
f036d1 |
- pathsep = strrchr(protsep, '/');
|
|
 |
f036d1 |
- if(pathsep)
|
|
 |
f036d1 |
- *pathsep = 0;
|
|
 |
f036d1 |
- else {
|
|
 |
f036d1 |
- *protsep = 0;
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- else {
|
|
 |
f036d1 |
- /* We got a new absolute path for this server */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if((relurl[0] == '/') && (relurl[1] == '/')) {
|
|
 |
f036d1 |
- /* the new URL starts with //, just keep the protocol part from the
|
|
 |
f036d1 |
- original one */
|
|
 |
f036d1 |
- *protsep = 0;
|
|
 |
f036d1 |
- useurl = &relurl[2]; /* we keep the slashes from the original, so we
|
|
 |
f036d1 |
- skip the new ones */
|
|
 |
f036d1 |
- host_changed = TRUE;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- else {
|
|
 |
f036d1 |
- /* cut off the original URL from the first slash, or deal with URLs
|
|
 |
f036d1 |
- without slash */
|
|
 |
f036d1 |
- pathsep = strchr(protsep, '/');
|
|
 |
f036d1 |
- if(pathsep) {
|
|
 |
f036d1 |
- /* When people use badly formatted URLs, such as
|
|
 |
f036d1 |
- "http://www.url.com?dir=/home/daniel" we must not use the first
|
|
 |
f036d1 |
- slash, if there's a ?-letter before it! */
|
|
 |
f036d1 |
- char *sep = strchr(protsep, '?');
|
|
 |
f036d1 |
- if(sep && (sep < pathsep))
|
|
 |
f036d1 |
- pathsep = sep;
|
|
 |
f036d1 |
- *pathsep = 0;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- else {
|
|
 |
f036d1 |
- /* There was no slash. Now, since we might be operating on a badly
|
|
 |
f036d1 |
- formatted URL, such as "http://www.url.com?id=2380" which doesn't
|
|
 |
f036d1 |
- use a slash separator as it is supposed to, we need to check for a
|
|
 |
f036d1 |
- ?-letter as well! */
|
|
 |
f036d1 |
- pathsep = strchr(protsep, '?');
|
|
 |
f036d1 |
- if(pathsep)
|
|
 |
f036d1 |
- *pathsep = 0;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* If the new part contains a space, this is a mighty stupid redirect
|
|
 |
f036d1 |
- but we still make an effort to do "right". To the left of a '?'
|
|
 |
f036d1 |
- letter we replace each space with %20 while it is replaced with '+'
|
|
 |
f036d1 |
- on the right side of the '?' letter.
|
|
 |
f036d1 |
- */
|
|
 |
f036d1 |
- newlen = strlen_url(useurl, !host_changed);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- urllen = strlen(url_clone);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- newest = malloc(urllen + 1 + /* possible slash */
|
|
 |
f036d1 |
- newlen + 1 /* zero byte */);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- if(!newest) {
|
|
 |
f036d1 |
- free(url_clone); /* don't leak this */
|
|
 |
f036d1 |
- return NULL;
|
|
 |
f036d1 |
- }
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* copy over the root url part */
|
|
 |
f036d1 |
- memcpy(newest, url_clone, urllen);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* check if we need to append a slash */
|
|
 |
f036d1 |
- if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0]))
|
|
 |
f036d1 |
- ;
|
|
 |
f036d1 |
- else
|
|
 |
f036d1 |
- newest[urllen++]='/';
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* then append the new piece on the right side */
|
|
 |
f036d1 |
- strcpy_url(&newest[urllen], useurl, !host_changed);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- free(url_clone);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- return newest;
|
|
 |
f036d1 |
-}
|
|
 |
f036d1 |
-#endif /* CURL_DISABLE_HTTP */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
/*
|
|
 |
f036d1 |
* Curl_follow() handles the URL redirect magic. Pass in the 'newurl' string
|
|
 |
f036d1 |
* as given by the remote server and set up the new URL to request.
|
|
 |
f036d1 |
@@ -1809,12 +1505,12 @@ CURLcode Curl_follow(struct Curl_easy *data,
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
- if(!is_absolute_url(newurl)) {
|
|
 |
f036d1 |
+ if(!Curl_is_absolute_url(newurl, NULL, 8)) {
|
|
 |
f036d1 |
/***
|
|
 |
f036d1 |
*DANG* this is an RFC 2068 violation. The URL is supposed
|
|
 |
f036d1 |
to be absolute and this doesn't seem to be that!
|
|
 |
f036d1 |
*/
|
|
 |
f036d1 |
- char *absolute = concat_url(data->change.url, newurl);
|
|
 |
f036d1 |
+ char *absolute = Curl_concat_url(data->change.url, newurl);
|
|
 |
f036d1 |
if(!absolute)
|
|
 |
f036d1 |
return CURLE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
newurl = absolute;
|
|
 |
f036d1 |
@@ -1823,7 +1519,7 @@ CURLcode Curl_follow(struct Curl_easy *data,
|
|
 |
f036d1 |
/* The new URL MAY contain space or high byte values, that means a mighty
|
|
 |
f036d1 |
stupid redirect URL but we still make an effort to do "right". */
|
|
 |
f036d1 |
char *newest;
|
|
 |
f036d1 |
- size_t newlen = strlen_url(newurl, FALSE);
|
|
 |
f036d1 |
+ size_t newlen = Curl_strlen_url(newurl, FALSE);
|
|
 |
f036d1 |
|
|
 |
f036d1 |
/* This is an absolute URL, don't allow the custom port number */
|
|
 |
f036d1 |
disallowport = TRUE;
|
|
 |
f036d1 |
@@ -1832,7 +1528,7 @@ CURLcode Curl_follow(struct Curl_easy *data,
|
|
 |
f036d1 |
if(!newest)
|
|
 |
f036d1 |
return CURLE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
|
|
 |
f036d1 |
- strcpy_url(newest, newurl, FALSE); /* create a space-free URL */
|
|
 |
f036d1 |
+ Curl_strcpy_url(newest, newurl, FALSE); /* create a space-free URL */
|
|
 |
f036d1 |
newurl = newest; /* use this instead now */
|
|
 |
f036d1 |
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
diff --git a/lib/url.c b/lib/url.c
|
|
 |
f036d1 |
index dcc1ecc..4f75f11 100644
|
|
 |
f036d1 |
--- a/lib/url.c
|
|
 |
f036d1 |
+++ b/lib/url.c
|
|
 |
f036d1 |
@@ -1939,30 +1939,37 @@ static struct connectdata *allocate_conn(struct Curl_easy *data)
|
|
 |
f036d1 |
return NULL;
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
-static CURLcode findprotocol(struct Curl_easy *data,
|
|
 |
f036d1 |
- struct connectdata *conn,
|
|
 |
f036d1 |
- const char *protostr)
|
|
 |
f036d1 |
+/* returns the handdler if the given scheme is built-in */
|
|
 |
f036d1 |
+const struct Curl_handler *Curl_builtin_scheme(const char *scheme)
|
|
 |
f036d1 |
{
|
|
 |
f036d1 |
const struct Curl_handler * const *pp;
|
|
 |
f036d1 |
const struct Curl_handler *p;
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
- /* Scan protocol handler table and match against 'protostr' to set a few
|
|
 |
f036d1 |
- variables based on the URL. Now that the handler may be changed later
|
|
 |
f036d1 |
- when the protocol specific setup function is called. */
|
|
 |
f036d1 |
- for(pp = protocols; (p = *pp) != NULL; pp++) {
|
|
 |
f036d1 |
- if(strcasecompare(p->scheme, protostr)) {
|
|
 |
f036d1 |
+ /* Scan protocol handler table and match against 'scheme'. The handler may
|
|
 |
f036d1 |
+ be changed later when the protocol specific setup function is called. */
|
|
 |
f036d1 |
+ for(pp = protocols; (p = *pp) != NULL; pp++)
|
|
 |
f036d1 |
+ if(strcasecompare(p->scheme, scheme))
|
|
 |
f036d1 |
/* Protocol found in table. Check if allowed */
|
|
 |
f036d1 |
- if(!(data->set.allowed_protocols & p->protocol))
|
|
 |
f036d1 |
- /* nope, get out */
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
+ return p;
|
|
 |
f036d1 |
+ return NULL; /* not found */
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
- /* it is allowed for "normal" request, now do an extra check if this is
|
|
 |
f036d1 |
- the result of a redirect */
|
|
 |
f036d1 |
- if(data->state.this_is_a_follow &&
|
|
 |
f036d1 |
- !(data->set.redir_protocols & p->protocol))
|
|
 |
f036d1 |
- /* nope, get out */
|
|
 |
f036d1 |
- break;
|
|
 |
f036d1 |
|
|
 |
f036d1 |
+static CURLcode findprotocol(struct Curl_easy *data,
|
|
 |
f036d1 |
+ struct connectdata *conn,
|
|
 |
f036d1 |
+ const char *protostr)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ const struct Curl_handler *p = Curl_builtin_scheme(protostr);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(p && /* Protocol found in table. Check if allowed */
|
|
 |
f036d1 |
+ (data->set.allowed_protocols & p->protocol)) {
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* it is allowed for "normal" request, now do an extra check if this is
|
|
 |
f036d1 |
+ the result of a redirect */
|
|
 |
f036d1 |
+ if(data->state.this_is_a_follow &&
|
|
 |
f036d1 |
+ !(data->set.redir_protocols & p->protocol))
|
|
 |
f036d1 |
+ /* nope, get out */
|
|
 |
f036d1 |
+ ;
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
/* Perform setup complement if some. */
|
|
 |
f036d1 |
conn->handler = conn->given = p;
|
|
 |
f036d1 |
|
|
 |
f036d1 |
@@ -1971,7 +1978,6 @@ static CURLcode findprotocol(struct Curl_easy *data,
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
}
|
|
 |
f036d1 |
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
/* The protocol was not found in the table, but we don't have to assign it
|
|
 |
f036d1 |
to anything since it is already assigned to a dummy-struct in the
|
|
 |
f036d1 |
create_conn() function when the connectdata struct is allocated. */
|
|
 |
f036d1 |
diff --git a/lib/url.h b/lib/url.h
|
|
 |
f036d1 |
index ef3ebf0..0034f82 100644
|
|
 |
f036d1 |
--- a/lib/url.h
|
|
 |
f036d1 |
+++ b/lib/url.h
|
|
 |
f036d1 |
@@ -69,6 +69,8 @@ void Curl_getoff_all_pipelines(struct Curl_easy *data,
|
|
 |
f036d1 |
|
|
 |
f036d1 |
void Curl_close_connections(struct Curl_easy *data);
|
|
 |
f036d1 |
|
|
 |
f036d1 |
+const struct Curl_handler *Curl_builtin_scheme(const char *scheme);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
#define CURL_DEFAULT_PROXY_PORT 1080 /* default proxy port unless specified */
|
|
 |
f036d1 |
#define CURL_DEFAULT_HTTPS_PROXY_PORT 443 /* default https proxy port unless
|
|
 |
f036d1 |
specified */
|
|
 |
f036d1 |
diff --git a/lib/escape.h b/lib/urlapi-int.h
|
|
 |
f036d1 |
similarity index 66%
|
|
 |
f036d1 |
copy from lib/escape.h
|
|
 |
f036d1 |
copy to lib/urlapi-int.h
|
|
 |
f036d1 |
index 638666f..7ac09fd 100644
|
|
 |
f036d1 |
--- a/lib/escape.h
|
|
 |
f036d1 |
+++ b/lib/urlapi-int.h
|
|
 |
f036d1 |
@@ -1,5 +1,5 @@
|
|
 |
f036d1 |
-#ifndef HEADER_CURL_ESCAPE_H
|
|
 |
f036d1 |
-#define HEADER_CURL_ESCAPE_H
|
|
 |
f036d1 |
+#ifndef HEADER_CURL_URLAPI_INT_H
|
|
 |
f036d1 |
+#define HEADER_CURL_URLAPI_INT_H
|
|
 |
f036d1 |
/***************************************************************************
|
|
 |
f036d1 |
* _ _ ____ _
|
|
 |
f036d1 |
* Project ___| | | | _ \| |
|
|
 |
f036d1 |
@@ -7,7 +7,7 @@
|
|
 |
f036d1 |
* | (__| |_| | _ <| |___
|
|
 |
f036d1 |
* \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
- * Copyright (C) 1998 - 2011, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
* This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
* you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
@@ -21,13 +21,9 @@
|
|
 |
f036d1 |
* KIND, either express or implied.
|
|
 |
f036d1 |
*
|
|
 |
f036d1 |
***************************************************************************/
|
|
 |
f036d1 |
-/* Escape and unescape URL encoding in strings. The functions return a new
|
|
 |
f036d1 |
- * allocated string or NULL if an error occurred. */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-CURLcode Curl_urldecode(struct Curl_easy *data,
|
|
 |
f036d1 |
- const char *string, size_t length,
|
|
 |
f036d1 |
- char **ostring, size_t *olen,
|
|
 |
f036d1 |
- bool reject_crlf);
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
-#endif /* HEADER_CURL_ESCAPE_H */
|
|
 |
f036d1 |
-
|
|
 |
f036d1 |
+#include "curl_setup.h"
|
|
 |
f036d1 |
+bool Curl_is_absolute_url(const char *url, char *scheme, size_t buflen);
|
|
 |
f036d1 |
+char *Curl_concat_url(const char *base, const char *relurl);
|
|
 |
f036d1 |
+size_t Curl_strlen_url(const char *url, bool relative);
|
|
 |
f036d1 |
+void Curl_strcpy_url(char *output, const char *url, bool relative);
|
|
 |
f036d1 |
+#endif
|
|
 |
f036d1 |
diff --git a/lib/urlapi.c b/lib/urlapi.c
|
|
 |
f036d1 |
new file mode 100644
|
|
 |
f036d1 |
index 0000000..8287861
|
|
 |
f036d1 |
--- /dev/null
|
|
 |
f036d1 |
+++ b/lib/urlapi.c
|
|
 |
f036d1 |
@@ -0,0 +1,1315 @@
|
|
 |
f036d1 |
+/***************************************************************************
|
|
 |
f036d1 |
+ * _ _ ____ _
|
|
 |
f036d1 |
+ * Project ___| | | | _ \| |
|
|
 |
f036d1 |
+ * / __| | | | |_) | |
|
|
 |
f036d1 |
+ * | (__| |_| | _ <| |___
|
|
 |
f036d1 |
+ * \___|\___/|_| \_\_____|
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * Copyright (C) 1998 - 2018, Daniel Stenberg, <daniel@haxx.se>, et al.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * This software is licensed as described in the file COPYING, which
|
|
 |
f036d1 |
+ * you should have received as part of this distribution. The terms
|
|
 |
f036d1 |
+ * are also available at https://curl.haxx.se/docs/copyright.html.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * You may opt to use, copy, modify, merge, publish, distribute and/or sell
|
|
 |
f036d1 |
+ * copies of the Software, and permit persons to whom the Software is
|
|
 |
f036d1 |
+ * furnished to do so, under the terms of the COPYING file.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
|
|
 |
f036d1 |
+ * KIND, either express or implied.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ ***************************************************************************/
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#include "curl_setup.h"
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#include "urldata.h"
|
|
 |
f036d1 |
+#include "urlapi-int.h"
|
|
 |
f036d1 |
+#include "strcase.h"
|
|
 |
f036d1 |
+#include "dotdot.h"
|
|
 |
f036d1 |
+#include "url.h"
|
|
 |
f036d1 |
+#include "escape.h"
|
|
 |
f036d1 |
+#include "curl_ctype.h"
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* The last 3 #include files should be in this order */
|
|
 |
f036d1 |
+#include "curl_printf.h"
|
|
 |
f036d1 |
+#include "curl_memory.h"
|
|
 |
f036d1 |
+#include "memdebug.h"
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* Internal representation of CURLU. Point to URL-encoded strings. */
|
|
 |
f036d1 |
+struct Curl_URL {
|
|
 |
f036d1 |
+ char *scheme;
|
|
 |
f036d1 |
+ char *user;
|
|
 |
f036d1 |
+ char *password;
|
|
 |
f036d1 |
+ char *options; /* IMAP only? */
|
|
 |
f036d1 |
+ char *host;
|
|
 |
f036d1 |
+ char *port;
|
|
 |
f036d1 |
+ char *path;
|
|
 |
f036d1 |
+ char *query;
|
|
 |
f036d1 |
+ char *fragment;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ char *scratch; /* temporary scratch area */
|
|
 |
f036d1 |
+ long portnum; /* the numerical version */
|
|
 |
f036d1 |
+};
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#define DEFAULT_SCHEME "https"
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* scheme is not URL encoded, the longest libcurl supported ones are 6
|
|
 |
f036d1 |
+ letters */
|
|
 |
f036d1 |
+#define MAX_SCHEME_LEN 8
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+static void free_urlhandle(struct Curl_URL *u)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ free(u->scheme);
|
|
 |
f036d1 |
+ free(u->user);
|
|
 |
f036d1 |
+ free(u->password);
|
|
 |
f036d1 |
+ free(u->options);
|
|
 |
f036d1 |
+ free(u->host);
|
|
 |
f036d1 |
+ free(u->port);
|
|
 |
f036d1 |
+ free(u->path);
|
|
 |
f036d1 |
+ free(u->query);
|
|
 |
f036d1 |
+ free(u->fragment);
|
|
 |
f036d1 |
+ free(u->scratch);
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* move the full contents of one handle onto another and
|
|
 |
f036d1 |
+ free the original */
|
|
 |
f036d1 |
+static void mv_urlhandle(struct Curl_URL *from,
|
|
 |
f036d1 |
+ struct Curl_URL *to)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ free_urlhandle(to);
|
|
 |
f036d1 |
+ *to = *from;
|
|
 |
f036d1 |
+ free(from);
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * Find the separator at the end of the host name, or the '?' in cases like
|
|
 |
f036d1 |
+ * http://www.url.com?id=2380
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+static const char *find_host_sep(const char *url)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ const char *sep;
|
|
 |
f036d1 |
+ const char *query;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* Find the start of the hostname */
|
|
 |
f036d1 |
+ sep = strstr(url, "//");
|
|
 |
f036d1 |
+ if(!sep)
|
|
 |
f036d1 |
+ sep = url;
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ sep += 2;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ query = strchr(sep, '?');
|
|
 |
f036d1 |
+ sep = strchr(sep, '/');
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!sep)
|
|
 |
f036d1 |
+ sep = url + strlen(url);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!query)
|
|
 |
f036d1 |
+ query = url + strlen(url);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return sep < query ? sep : query;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * Decide in an encoding-independent manner whether a character in an
|
|
 |
f036d1 |
+ * URL must be escaped. The same criterion must be used in strlen_url()
|
|
 |
f036d1 |
+ * and strcpy_url().
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+static bool urlchar_needs_escaping(int c)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c));
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * strlen_url() returns the length of the given URL if the spaces within the
|
|
 |
f036d1 |
+ * URL were properly URL encoded.
|
|
 |
f036d1 |
+ * URL encoding should be skipped for host names, otherwise IDN resolution
|
|
 |
f036d1 |
+ * will fail.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+size_t Curl_strlen_url(const char *url, bool relative)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ const unsigned char *ptr;
|
|
 |
f036d1 |
+ size_t newlen = 0;
|
|
 |
f036d1 |
+ bool left = TRUE; /* left side of the ? */
|
|
 |
f036d1 |
+ const unsigned char *host_sep = (const unsigned char *) url;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!relative)
|
|
 |
f036d1 |
+ host_sep = (const unsigned char *) find_host_sep(url);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ for(ptr = (unsigned char *)url; *ptr; ptr++) {
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(ptr < host_sep) {
|
|
 |
f036d1 |
+ ++newlen;
|
|
 |
f036d1 |
+ continue;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ switch(*ptr) {
|
|
 |
f036d1 |
+ case '?':
|
|
 |
f036d1 |
+ left = FALSE;
|
|
 |
f036d1 |
+ /* FALLTHROUGH */
|
|
 |
f036d1 |
+ default:
|
|
 |
f036d1 |
+ if(urlchar_needs_escaping(*ptr))
|
|
 |
f036d1 |
+ newlen += 2;
|
|
 |
f036d1 |
+ newlen++;
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ case ' ':
|
|
 |
f036d1 |
+ if(left)
|
|
 |
f036d1 |
+ newlen += 3;
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ newlen++;
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ return newlen;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in
|
|
 |
f036d1 |
+ * the source URL accordingly.
|
|
 |
f036d1 |
+ * URL encoding should be skipped for host names, otherwise IDN resolution
|
|
 |
f036d1 |
+ * will fail.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+void Curl_strcpy_url(char *output, const char *url, bool relative)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ /* we must add this with whitespace-replacing */
|
|
 |
f036d1 |
+ bool left = TRUE;
|
|
 |
f036d1 |
+ const unsigned char *iptr;
|
|
 |
f036d1 |
+ char *optr = output;
|
|
 |
f036d1 |
+ const unsigned char *host_sep = (const unsigned char *) url;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!relative)
|
|
 |
f036d1 |
+ host_sep = (const unsigned char *) find_host_sep(url);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ for(iptr = (unsigned char *)url; /* read from here */
|
|
 |
f036d1 |
+ *iptr; /* until zero byte */
|
|
 |
f036d1 |
+ iptr++) {
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(iptr < host_sep) {
|
|
 |
f036d1 |
+ *optr++ = *iptr;
|
|
 |
f036d1 |
+ continue;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ switch(*iptr) {
|
|
 |
f036d1 |
+ case '?':
|
|
 |
f036d1 |
+ left = FALSE;
|
|
 |
f036d1 |
+ /* FALLTHROUGH */
|
|
 |
f036d1 |
+ default:
|
|
 |
f036d1 |
+ if(urlchar_needs_escaping(*iptr)) {
|
|
 |
f036d1 |
+ snprintf(optr, 4, "%%%02x", *iptr);
|
|
 |
f036d1 |
+ optr += 3;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ *optr++=*iptr;
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ case ' ':
|
|
 |
f036d1 |
+ if(left) {
|
|
 |
f036d1 |
+ *optr++='%'; /* add a '%' */
|
|
 |
f036d1 |
+ *optr++='2'; /* add a '2' */
|
|
 |
f036d1 |
+ *optr++='0'; /* add a '0' */
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ *optr++='+'; /* add a '+' here */
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ *optr = 0; /* zero terminate output buffer */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * Returns true if the given URL is absolute (as opposed to relative) within
|
|
 |
f036d1 |
+ * the buffer size. Returns the scheme in the buffer if TRUE and 'buf' is
|
|
 |
f036d1 |
+ * non-NULL.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+bool Curl_is_absolute_url(const char *url, char *buf, size_t buflen)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ size_t i;
|
|
 |
f036d1 |
+ for(i = 0; i < buflen && url[i]; ++i) {
|
|
 |
f036d1 |
+ char s = url[i];
|
|
 |
f036d1 |
+ if(s == ':') {
|
|
 |
f036d1 |
+ if(buf)
|
|
 |
f036d1 |
+ buf[i] = 0;
|
|
 |
f036d1 |
+ return TRUE;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ /* RFC 3986 3.1 explains:
|
|
 |
f036d1 |
+ scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ else if(ISALNUM(s) || (s == '+') || (s == '-') || (s == '.') ) {
|
|
 |
f036d1 |
+ if(buf)
|
|
 |
f036d1 |
+ buf[i] = (char)TOLOWER(s);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ return FALSE;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * Concatenate a relative URL to a base URL making it absolute.
|
|
 |
f036d1 |
+ * URL-encodes any spaces.
|
|
 |
f036d1 |
+ * The returned pointer must be freed by the caller unless NULL
|
|
 |
f036d1 |
+ * (returns NULL on out of memory).
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+char *Curl_concat_url(const char *base, const char *relurl)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ /***
|
|
 |
f036d1 |
+ TRY to append this new path to the old URL
|
|
 |
f036d1 |
+ to the right of the host part. Oh crap, this is doomed to cause
|
|
 |
f036d1 |
+ problems in the future...
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ char *newest;
|
|
 |
f036d1 |
+ char *protsep;
|
|
 |
f036d1 |
+ char *pathsep;
|
|
 |
f036d1 |
+ size_t newlen;
|
|
 |
f036d1 |
+ bool host_changed = FALSE;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ const char *useurl = relurl;
|
|
 |
f036d1 |
+ size_t urllen;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* we must make our own copy of the URL to play with, as it may
|
|
 |
f036d1 |
+ point to read-only data */
|
|
 |
f036d1 |
+ char *url_clone = strdup(base);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!url_clone)
|
|
 |
f036d1 |
+ return NULL; /* skip out of this NOW */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* protsep points to the start of the host name */
|
|
 |
f036d1 |
+ protsep = strstr(url_clone, "//");
|
|
 |
f036d1 |
+ if(!protsep)
|
|
 |
f036d1 |
+ protsep = url_clone;
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ protsep += 2; /* pass the slashes */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if('/' != relurl[0]) {
|
|
 |
f036d1 |
+ int level = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* First we need to find out if there's a ?-letter in the URL,
|
|
 |
f036d1 |
+ and cut it and the right-side of that off */
|
|
 |
f036d1 |
+ pathsep = strchr(protsep, '?');
|
|
 |
f036d1 |
+ if(pathsep)
|
|
 |
f036d1 |
+ *pathsep = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* we have a relative path to append to the last slash if there's one
|
|
 |
f036d1 |
+ available, or if the new URL is just a query string (starts with a
|
|
 |
f036d1 |
+ '?') we append the new one at the end of the entire currently worked
|
|
 |
f036d1 |
+ out URL */
|
|
 |
f036d1 |
+ if(useurl[0] != '?') {
|
|
 |
f036d1 |
+ pathsep = strrchr(protsep, '/');
|
|
 |
f036d1 |
+ if(pathsep)
|
|
 |
f036d1 |
+ *pathsep = 0;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* Check if there's any slash after the host name, and if so, remember
|
|
 |
f036d1 |
+ that position instead */
|
|
 |
f036d1 |
+ pathsep = strchr(protsep, '/');
|
|
 |
f036d1 |
+ if(pathsep)
|
|
 |
f036d1 |
+ protsep = pathsep + 1;
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ protsep = NULL;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* now deal with one "./" or any amount of "../" in the newurl
|
|
 |
f036d1 |
+ and act accordingly */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if((useurl[0] == '.') && (useurl[1] == '/'))
|
|
 |
f036d1 |
+ useurl += 2; /* just skip the "./" */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ while((useurl[0] == '.') &&
|
|
 |
f036d1 |
+ (useurl[1] == '.') &&
|
|
 |
f036d1 |
+ (useurl[2] == '/')) {
|
|
 |
f036d1 |
+ level++;
|
|
 |
f036d1 |
+ useurl += 3; /* pass the "../" */
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(protsep) {
|
|
 |
f036d1 |
+ while(level--) {
|
|
 |
f036d1 |
+ /* cut off one more level from the right of the original URL */
|
|
 |
f036d1 |
+ pathsep = strrchr(protsep, '/');
|
|
 |
f036d1 |
+ if(pathsep)
|
|
 |
f036d1 |
+ *pathsep = 0;
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ *protsep = 0;
|
|
 |
f036d1 |
+ break;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* We got a new absolute path for this server */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if((relurl[0] == '/') && (relurl[1] == '/')) {
|
|
 |
f036d1 |
+ /* the new URL starts with //, just keep the protocol part from the
|
|
 |
f036d1 |
+ original one */
|
|
 |
f036d1 |
+ *protsep = 0;
|
|
 |
f036d1 |
+ useurl = &relurl[2]; /* we keep the slashes from the original, so we
|
|
 |
f036d1 |
+ skip the new ones */
|
|
 |
f036d1 |
+ host_changed = TRUE;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* cut off the original URL from the first slash, or deal with URLs
|
|
 |
f036d1 |
+ without slash */
|
|
 |
f036d1 |
+ pathsep = strchr(protsep, '/');
|
|
 |
f036d1 |
+ if(pathsep) {
|
|
 |
f036d1 |
+ /* When people use badly formatted URLs, such as
|
|
 |
f036d1 |
+ "http://www.url.com?dir=/home/daniel" we must not use the first
|
|
 |
f036d1 |
+ slash, if there's a ?-letter before it! */
|
|
 |
f036d1 |
+ char *sep = strchr(protsep, '?');
|
|
 |
f036d1 |
+ if(sep && (sep < pathsep))
|
|
 |
f036d1 |
+ pathsep = sep;
|
|
 |
f036d1 |
+ *pathsep = 0;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* There was no slash. Now, since we might be operating on a badly
|
|
 |
f036d1 |
+ formatted URL, such as "http://www.url.com?id=2380" which doesn't
|
|
 |
f036d1 |
+ use a slash separator as it is supposed to, we need to check for a
|
|
 |
f036d1 |
+ ?-letter as well! */
|
|
 |
f036d1 |
+ pathsep = strchr(protsep, '?');
|
|
 |
f036d1 |
+ if(pathsep)
|
|
 |
f036d1 |
+ *pathsep = 0;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* If the new part contains a space, this is a mighty stupid redirect
|
|
 |
f036d1 |
+ but we still make an effort to do "right". To the left of a '?'
|
|
 |
f036d1 |
+ letter we replace each space with %20 while it is replaced with '+'
|
|
 |
f036d1 |
+ on the right side of the '?' letter.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ newlen = Curl_strlen_url(useurl, !host_changed);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ urllen = strlen(url_clone);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ newest = malloc(urllen + 1 + /* possible slash */
|
|
 |
f036d1 |
+ newlen + 1 /* zero byte */);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!newest) {
|
|
 |
f036d1 |
+ free(url_clone); /* don't leak this */
|
|
 |
f036d1 |
+ return NULL;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* copy over the root url part */
|
|
 |
f036d1 |
+ memcpy(newest, url_clone, urllen);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* check if we need to append a slash */
|
|
 |
f036d1 |
+ if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0]))
|
|
 |
f036d1 |
+ ;
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ newest[urllen++]='/';
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* then append the new piece on the right side */
|
|
 |
f036d1 |
+ Curl_strcpy_url(&newest[urllen], useurl, !host_changed);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ free(url_clone);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return newest;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/*
|
|
 |
f036d1 |
+ * parse_hostname_login()
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * Parse the login details (user name, password and options) from the URL and
|
|
 |
f036d1 |
+ * strip them out of the host name
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+static CURLUcode parse_hostname_login(struct Curl_URL *u,
|
|
 |
f036d1 |
+ const struct Curl_handler *h,
|
|
 |
f036d1 |
+ char **hostname,
|
|
 |
f036d1 |
+ unsigned int flags)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ CURLUcode result = CURLUE_OK;
|
|
 |
f036d1 |
+ CURLcode ccode;
|
|
 |
f036d1 |
+ char *userp = NULL;
|
|
 |
f036d1 |
+ char *passwdp = NULL;
|
|
 |
f036d1 |
+ char *optionsp = NULL;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* At this point, we're hoping all the other special cases have
|
|
 |
f036d1 |
+ * been taken care of, so conn->host.name is at most
|
|
 |
f036d1 |
+ * [user[:password][;options]]@]hostname
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * We need somewhere to put the embedded details, so do that first.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ char *ptr = strchr(*hostname, '@');
|
|
 |
f036d1 |
+ char *login = *hostname;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!ptr)
|
|
 |
f036d1 |
+ goto out;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* We will now try to extract the
|
|
 |
f036d1 |
+ * possible login information in a string like:
|
|
 |
f036d1 |
+ * ftp://user:password@ftp.my.site:8021/README */
|
|
 |
f036d1 |
+ *hostname = ++ptr;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* We could use the login information in the URL so extract it. Only parse
|
|
 |
f036d1 |
+ options if the handler says we should. */
|
|
 |
f036d1 |
+ ccode = Curl_parse_login_details(login, ptr - login - 1,
|
|
 |
f036d1 |
+ &userp, &passwdp,
|
|
 |
f036d1 |
+ h->flags & PROTOPT_URLOPTIONS ?
|
|
 |
f036d1 |
+ &optionsp:NULL);
|
|
 |
f036d1 |
+ if(ccode) {
|
|
 |
f036d1 |
+ result = CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ goto out;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(userp) {
|
|
 |
f036d1 |
+ if(flags & CURLU_DISALLOW_USER) {
|
|
 |
f036d1 |
+ /* Option DISALLOW_USER is set and url contains username. */
|
|
 |
f036d1 |
+ result = CURLUE_USER_NOT_ALLOWED;
|
|
 |
f036d1 |
+ goto out;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ u->user = userp;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(passwdp)
|
|
 |
f036d1 |
+ u->password = passwdp;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(optionsp)
|
|
 |
f036d1 |
+ u->options = optionsp;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return CURLUE_OK;
|
|
 |
f036d1 |
+ out:
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ free(userp);
|
|
 |
f036d1 |
+ free(passwdp);
|
|
 |
f036d1 |
+ free(optionsp);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return result;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+static CURLUcode parse_port(struct Curl_URL *u, char *hostname)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ char *portptr;
|
|
 |
f036d1 |
+ char endbracket;
|
|
 |
f036d1 |
+ int len;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if((1 == sscanf(hostname, "[%*45[0123456789abcdefABCDEF:.]%c%n",
|
|
 |
f036d1 |
+ &endbracket, &len)) &&
|
|
 |
f036d1 |
+ (']' == endbracket)) {
|
|
 |
f036d1 |
+ /* this is a RFC2732-style specified IP-address */
|
|
 |
f036d1 |
+ portptr = &hostname[len];
|
|
 |
f036d1 |
+ if (*portptr != ':')
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else
|
|
 |
f036d1 |
+ portptr = strchr(hostname, ':');
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(portptr) {
|
|
 |
f036d1 |
+ char *rest;
|
|
 |
f036d1 |
+ long port;
|
|
 |
f036d1 |
+ char portbuf[7];
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!ISDIGIT(portptr[1]))
|
|
 |
f036d1 |
+ return CURLUE_BAD_PORT_NUMBER;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ port = strtol(portptr + 1, &rest, 10); /* Port number must be decimal */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if((port <= 0) || (port > 0xffff))
|
|
 |
f036d1 |
+ /* Single unix standard says port numbers are 16 bits long, but we don't
|
|
 |
f036d1 |
+ treat port zero as OK. */
|
|
 |
f036d1 |
+ return CURLUE_BAD_PORT_NUMBER;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(rest[0])
|
|
 |
f036d1 |
+ return CURLUE_BAD_PORT_NUMBER;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(rest != &portptr[1]) {
|
|
 |
f036d1 |
+ *portptr++ = '\0'; /* cut off the name there */
|
|
 |
f036d1 |
+ *rest = 0;
|
|
 |
f036d1 |
+ /* generate a new to get rid of leading zeroes etc */
|
|
 |
f036d1 |
+ snprintf(portbuf, sizeof(portbuf), "%ld", port);
|
|
 |
f036d1 |
+ u->portnum = port;
|
|
 |
f036d1 |
+ u->port = strdup(portbuf);
|
|
 |
f036d1 |
+ if(!u->port)
|
|
 |
f036d1 |
+ return CURLUE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* Browser behavior adaptation. If there's a colon with no digits after,
|
|
 |
f036d1 |
+ just cut off the name there which makes us ignore the colon and just
|
|
 |
f036d1 |
+ use the default port. Firefox and Chrome both do that. */
|
|
 |
f036d1 |
+ *portptr = '\0';
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return CURLUE_OK;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+/* scan for byte values < 31 or 127 */
|
|
 |
f036d1 |
+static CURLUcode junkscan(char *part)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ char badbytes[]={
|
|
 |
f036d1 |
+ /* */ 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
|
|
 |
f036d1 |
+ 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
|
|
 |
f036d1 |
+ 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
|
|
 |
f036d1 |
+ 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f,
|
|
 |
f036d1 |
+ 0x7f,
|
|
 |
f036d1 |
+ 0x00 /* zero terminate */
|
|
 |
f036d1 |
+ };
|
|
 |
f036d1 |
+ if(part) {
|
|
 |
f036d1 |
+ size_t n = strlen(part);
|
|
 |
f036d1 |
+ size_t nfine = strcspn(part, badbytes);
|
|
 |
f036d1 |
+ if(nfine != n)
|
|
 |
f036d1 |
+ /* since we don't know which part is scanned, return a generic error
|
|
 |
f036d1 |
+ code */
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ return CURLUE_OK;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+static CURLUcode hostname_check(char *hostname, unsigned int flags)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ const char *l; /* accepted characters */
|
|
 |
f036d1 |
+ size_t len;
|
|
 |
f036d1 |
+ size_t hlen = strlen(hostname);
|
|
 |
f036d1 |
+ (void)flags;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(hostname[0] == '[') {
|
|
 |
f036d1 |
+ hostname++;
|
|
 |
f036d1 |
+ l = "0123456789abcdefABCDEF::.";
|
|
 |
f036d1 |
+ hlen -= 2;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else /* % for URL escaped letters */
|
|
 |
f036d1 |
+ l = "0123456789abcdefghijklimnopqrstuvwxyz-_.ABCDEFGHIJKLIMNOPQRSTUVWXYZ%";
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ len = strspn(hostname, l);
|
|
 |
f036d1 |
+ if(hlen != len)
|
|
 |
f036d1 |
+ /* hostname with bad content */
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ return CURLUE_OK;
|
|
 |
f036d1 |
+}
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#define HOSTNAME_END(x) (((x) == '/') || ((x) == '?') || ((x) == '#'))
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags)
|
|
 |
f036d1 |
+{
|
|
 |
f036d1 |
+ char *path;
|
|
 |
f036d1 |
+ bool path_alloced = FALSE;
|
|
 |
f036d1 |
+ char *hostname;
|
|
 |
f036d1 |
+ char *query = NULL;
|
|
 |
f036d1 |
+ char *fragment = NULL;
|
|
 |
f036d1 |
+ CURLUcode result;
|
|
 |
f036d1 |
+ bool url_has_scheme = FALSE;
|
|
 |
f036d1 |
+ char schemebuf[MAX_SCHEME_LEN];
|
|
 |
f036d1 |
+ char *schemep;
|
|
 |
f036d1 |
+ size_t schemelen = 0;
|
|
 |
f036d1 |
+ size_t urllen;
|
|
 |
f036d1 |
+ const struct Curl_handler *h = NULL;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!url)
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /*************************************************************
|
|
 |
f036d1 |
+ * Parse the URL.
|
|
 |
f036d1 |
+ ************************************************************/
|
|
 |
f036d1 |
+ /* allocate scratch area */
|
|
 |
f036d1 |
+ urllen = strlen(url);
|
|
 |
f036d1 |
+ path = u->scratch = malloc(urllen * 2 + 2);
|
|
 |
f036d1 |
+ if(!path)
|
|
 |
f036d1 |
+ return CURLUE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ hostname = &path[urllen + 1];
|
|
 |
f036d1 |
+ hostname[0] = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* MSDOS/Windows style drive prefix, eg c: in c:foo */
|
|
 |
f036d1 |
+#define STARTS_WITH_DRIVE_PREFIX(str) \
|
|
 |
f036d1 |
+ ((('a' <= str[0] && str[0] <= 'z') || \
|
|
 |
f036d1 |
+ ('A' <= str[0] && str[0] <= 'Z')) && \
|
|
 |
f036d1 |
+ (str[1] == ':'))
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* MSDOS/Windows style drive prefix, optionally with
|
|
 |
f036d1 |
+ * a '|' instead of ':', followed by a slash or NUL */
|
|
 |
f036d1 |
+#define STARTS_WITH_URL_DRIVE_PREFIX(str) \
|
|
 |
f036d1 |
+ ((('a' <= (str)[0] && (str)[0] <= 'z') || \
|
|
 |
f036d1 |
+ ('A' <= (str)[0] && (str)[0] <= 'Z')) && \
|
|
 |
f036d1 |
+ ((str)[1] == ':' || (str)[1] == '|') && \
|
|
 |
f036d1 |
+ ((str)[2] == '/' || (str)[2] == '\\' || (str)[2] == 0))
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(Curl_is_absolute_url(url, schemebuf, sizeof(schemebuf))) {
|
|
 |
f036d1 |
+ url_has_scheme = TRUE;
|
|
 |
f036d1 |
+ schemelen = strlen(schemebuf);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* handle the file: scheme */
|
|
 |
f036d1 |
+ if(url_has_scheme && strcasecompare(schemebuf, "file")) {
|
|
 |
f036d1 |
+ /* path has been allocated large anough to hold this */
|
|
 |
f036d1 |
+ strcpy(path, &url[5]);
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ hostname = NULL; /* no host for file: URLs */
|
|
 |
f036d1 |
+ u->scheme = strdup("file");
|
|
 |
f036d1 |
+ if(!u->scheme)
|
|
 |
f036d1 |
+ return CURLUE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /* Extra handling URLs with an authority component (i.e. that start with
|
|
 |
f036d1 |
+ * "file://")
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * We allow omitted hostname (e.g. file:/<path>) -- valid according to
|
|
 |
f036d1 |
+ * RFC 8089, but not the (current) WHAT-WG URL spec.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ if(path[0] == '/' && path[1] == '/') {
|
|
 |
f036d1 |
+ /* swallow the two slashes */
|
|
 |
f036d1 |
+ char *ptr = &path[2];
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /*
|
|
 |
f036d1 |
+ * According to RFC 8089, a file: URL can be reliably dereferenced if:
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * o it has no/blank hostname, or
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * o the hostname matches "localhost" (case-insensitively), or
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * o the hostname is a FQDN that resolves to this machine.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * For brevity, we only consider URLs with empty, "localhost", or
|
|
 |
f036d1 |
+ * "127.0.0.1" hostnames as local.
|
|
 |
f036d1 |
+ *
|
|
 |
f036d1 |
+ * Additionally, there is an exception for URLs with a Windows drive
|
|
 |
f036d1 |
+ * letter in the authority (which was accidentally omitted from RFC 8089
|
|
 |
f036d1 |
+ * Appendix E, but believe me, it was meant to be there. --MK)
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ if(ptr[0] != '/' && !STARTS_WITH_URL_DRIVE_PREFIX(ptr)) {
|
|
 |
f036d1 |
+ /* the URL includes a host name, it must match "localhost" or
|
|
 |
f036d1 |
+ "127.0.0.1" to be valid */
|
|
 |
f036d1 |
+ if(!checkprefix("localhost/", ptr) &&
|
|
 |
f036d1 |
+ !checkprefix("127.0.0.1/", ptr)) {
|
|
 |
f036d1 |
+ /* Invalid file://hostname/, expected localhost or 127.0.0.1 or
|
|
 |
f036d1 |
+ none */
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ ptr += 9; /* now points to the slash after the host */
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ path = ptr;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+#if !defined(MSDOS) && !defined(WIN32) && !defined(__CYGWIN__)
|
|
 |
f036d1 |
+ /* Don't allow Windows drive letters when not in Windows.
|
|
 |
f036d1 |
+ * This catches both "file:/c:" and "file:c:" */
|
|
 |
f036d1 |
+ if(('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) ||
|
|
 |
f036d1 |
+ STARTS_WITH_URL_DRIVE_PREFIX(path)) {
|
|
 |
f036d1 |
+ /* File drive letters are only accepted in MSDOS/Windows */
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+#else
|
|
 |
f036d1 |
+ /* If the path starts with a slash and a drive letter, ditch the slash */
|
|
 |
f036d1 |
+ if('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) {
|
|
 |
f036d1 |
+ /* This cannot be done with strcpy, as the memory chunks overlap! */
|
|
 |
f036d1 |
+ memmove(path, &path[1], strlen(&path[1]) + 1);
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+#endif
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* clear path */
|
|
 |
f036d1 |
+ const char *p;
|
|
 |
f036d1 |
+ const char *hostp;
|
|
 |
f036d1 |
+ size_t len;
|
|
 |
f036d1 |
+ path[0] = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(url_has_scheme) {
|
|
 |
f036d1 |
+ int i = 0;
|
|
 |
f036d1 |
+ p = &url[schemelen + 1];
|
|
 |
f036d1 |
+ while(p && (*p == '/') && (i < 4)) {
|
|
 |
f036d1 |
+ p++;
|
|
 |
f036d1 |
+ i++;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ if((i < 1) || (i>3))
|
|
 |
f036d1 |
+ /* less than one or more than three slashes */
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ schemep = schemebuf;
|
|
 |
f036d1 |
+ if(!Curl_builtin_scheme(schemep) &&
|
|
 |
f036d1 |
+ !(flags & CURLU_NON_SUPPORT_SCHEME))
|
|
 |
f036d1 |
+ return CURLUE_UNSUPPORTED_SCHEME;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(junkscan(schemep))
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ else {
|
|
 |
f036d1 |
+ /* no scheme! */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ if(!(flags & CURLU_DEFAULT_SCHEME))
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+ schemep = (char *) DEFAULT_SCHEME;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ /*
|
|
 |
f036d1 |
+ * The URL was badly formatted, let's try without scheme specified.
|
|
 |
f036d1 |
+ */
|
|
 |
f036d1 |
+ p = url;
|
|
 |
f036d1 |
+ }
|
|
 |
f036d1 |
+ hostp = p; /* host name starts here */
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ while(*p && !HOSTNAME_END(*p)) /* find end of host name */
|
|
 |
f036d1 |
+ p++;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ len = p - hostp;
|
|
 |
f036d1 |
+ if(!len)
|
|
 |
f036d1 |
+ return CURLUE_MALFORMED_INPUT;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ memcpy(hostname, hostp, len);
|
|
 |
f036d1 |
+ hostname[len] = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ len = strlen(p);
|
|
 |
f036d1 |
+ memcpy(path, p, len);
|
|
 |
f036d1 |
+ path[len] = 0;
|
|
 |
f036d1 |
+
|
|
 |
f036d1 |
+ u->scheme = strdup(schemep);
|
|
 |
f036d1 |
+ if(!u->scheme)
|
|
 |
f036d1 |
+ return CURLUE_OUT_OF_MEMORY;
|
|
 |
f036d1 |
+ }
|
|
![]() |