From 907da069c450ca20442839d9e95e3661a5c06b61 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Sun, 5 Aug 2018 11:51:07 +0200 Subject: [PATCH 01/14] URL-API See header file and man pages for API. All documented API details work and are tested in the 1560 test case. Closes #2842 Upstream-commit: fb30ac5a2d63773c529c19259754e2b306ac2e2e Signed-off-by: Kamil Dudka --- docs/libcurl/Makefile.inc | 1 + docs/libcurl/curl_url.3 | 61 ++ docs/libcurl/curl_url_cleanup.3 | 44 + docs/libcurl/curl_url_dup.3 | 52 ++ docs/libcurl/curl_url_get.3 | 110 +++ docs/libcurl/curl_url_set.3 | 120 +++ docs/libcurl/symbols-in-versions | 30 + include/curl/Makefile.am | 4 +- include/curl/curl.h | 1 + include/curl/urlapi.h | 121 +++ lib/Makefile.inc | 5 +- lib/escape.c | 20 +- lib/escape.h | 3 +- lib/imap.c | 3 +- lib/transfer.c | 314 +------ lib/url.c | 44 +- lib/url.h | 2 + lib/{escape.h => urlapi-int.h} | 22 +- lib/urlapi.c | 1315 ++++++++++++++++++++++++++++++ tests/data/Makefile.inc | 2 + tests/data/test1560 | 28 + tests/libtest/Makefile.am | 5 + tests/libtest/Makefile.inc | 4 + tests/libtest/lib1560.c | 760 +++++++++++++++++ 24 files changed, 2716 insertions(+), 355 deletions(-) create mode 100644 docs/libcurl/curl_url.3 create mode 100644 docs/libcurl/curl_url_cleanup.3 create mode 100644 docs/libcurl/curl_url_dup.3 create mode 100644 docs/libcurl/curl_url_get.3 create mode 100644 docs/libcurl/curl_url_set.3 create mode 100644 include/curl/urlapi.h copy lib/{escape.h => urlapi-int.h} (66%) create mode 100644 lib/urlapi.c create mode 100644 tests/data/test1560 create mode 100644 tests/libtest/lib1560.c diff --git a/docs/libcurl/Makefile.inc b/docs/libcurl/Makefile.inc index eea48c4..955492c 100644 --- a/docs/libcurl/Makefile.inc +++ b/docs/libcurl/Makefile.inc @@ -22,4 +22,5 @@ man_MANS = curl_easy_cleanup.3 curl_easy_getinfo.3 curl_easy_init.3 \ curl_mime_data.3 curl_mime_data_cb.3 curl_mime_filedata.3 \ curl_mime_filename.3 curl_mime_subparts.3 \ curl_mime_type.3 curl_mime_headers.3 curl_mime_encoder.3 libcurl-env.3 \ + curl_url.3 curl_url_cleanup.3 curl_url_dup.3 curl_url_get.3 curl_url_set.3 \ libcurl-security.3 diff --git a/docs/libcurl/curl_url.3 b/docs/libcurl/curl_url.3 new file mode 100644 index 0000000..0a56264 --- /dev/null +++ b/docs/libcurl/curl_url.3 @@ -0,0 +1,61 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH curl_url 3 "6 Aug 2018" "libcurl" "libcurl Manual" +.SH NAME +curl_url - returns a new CURLU handle +.SH SYNOPSIS +.B #include + +CURLU *curl_url(); +.SH EXPERIMENTAL +The URL API is considered \fBEXPERIMENTAL\fP until further notice. Please test +it, report bugs and help us perfect it. Once proven to be reliable, the +experimental label will be removed. + +While this API is marked experimental, we reserve the right to modify the API +slightly if we deem it necessary and it makes it notably better or easier to +use. +.SH DESCRIPTION +This function will allocates and returns a pointer to a fresh CURLU handle, to +be used for further use of the URL API. +.SH RETURN VALUE +Returns a \fBCURLU *\fP if successful, or NULL if out of memory. +.SH EXAMPLE +.nf + CURLUcode rc; + CURLU *url = curl_url(); + rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0); + if(!rc) { + char *scheme; + rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0); + if(!rc) { + printf("the scheme is %s\n", scheme); + curl_free(scheme); + } + curl_url_cleanup(url); + } +.fi +.SH AVAILABILITY +Added in curl 7.63.0 +.SH "SEE ALSO" +.BR curl_url_cleanup "(3), " curl_url_get "(3), " curl_url_set "(3), " +.BR curl_url_dup "(3), " diff --git a/docs/libcurl/curl_url_cleanup.3 b/docs/libcurl/curl_url_cleanup.3 new file mode 100644 index 0000000..a8158b7 --- /dev/null +++ b/docs/libcurl/curl_url_cleanup.3 @@ -0,0 +1,44 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH curl_url_cleanup 3 "6 Aug 2018" "libcurl" "libcurl Manual" +.SH NAME +curl_url_cleanup - free a CURLU handle +.SH SYNOPSIS +.B #include + +void curl_url_cleanup(CURLU *handle); +.fi +.SH DESCRIPTION +Frees all the resources associated with the given CURLU handle! +.SH RETURN VALUE +none +.SH EXAMPLE +.nf + CURLU *url = curl_url(); + curl_url_set(url, CURLUPART_URL, "https://example.com", 0); + curl_url_cleanup(url); +.fi +.SH AVAILABILITY +Added in curl 7.63.0 +.SH "SEE ALSO" +.BR curl_url_dup "(3), " curl_url "(3), " curl_url_set "(3), " +.BR curl_url_get "(3), " diff --git a/docs/libcurl/curl_url_dup.3 b/docs/libcurl/curl_url_dup.3 new file mode 100644 index 0000000..4815dbd --- /dev/null +++ b/docs/libcurl/curl_url_dup.3 @@ -0,0 +1,52 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH curl_url_dup 3 "6 Aug 2018" "libcurl" "libcurl Manual" +.SH NAME +curl_url_dup - duplicate a CURLU handle +.SH SYNOPSIS +.B #include + +CURLU *curl_url_dup(CURLU *inhandle); +.fi +.SH DESCRIPTION +Duplicates a given CURLU \fIinhandle\fP and all its contents and returns a +pointer to a new CURLU handle. The new handle also needs to be freed with +\fIcurl_url_cleanup(3)\fP. +.SH RETURN VALUE +Returns a new handle or NULL if out of memory. +.SH EXAMPLE +.nf + CURLUcode rc; + CURLU *url = curl_url(); + CURLU *url2; + rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0); + if(!rc) { + url2 = curl_url_dup(url); /* clone it! */ + curl_url_cleanup(url2); + } + curl_url_cleanup(url); +.fi +.SH AVAILABILITY +Added in curl 7.63.0 +.SH "SEE ALSO" +.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), " +.BR curl_url_get "(3), " diff --git a/docs/libcurl/curl_url_get.3 b/docs/libcurl/curl_url_get.3 new file mode 100644 index 0000000..824d496 --- /dev/null +++ b/docs/libcurl/curl_url_get.3 @@ -0,0 +1,110 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH curl_url_get 3 "6 Aug 2018" "libcurl" "libcurl Manual" +.SH NAME +curl_url_get - extract a part from a URL +.SH SYNOPSIS +.B #include + +.nf +CURLUcode curl_url_get(CURLU *url, + CURLUPart what, + char **part, + unsigned int flags) +.fi +.SH DESCRIPTION +Given the \fIurl\fP handle of an already parsed URL, this function lets the +user extract individual pieces from it. + +The \fIwhat\fP argument should be the particular part to extract (see list +below) and \fIpart\fP points to a 'char *' to get updated to point to a newly +allocated string with the contents. + +The \fIflags\fP argument is a bitmask with individual features. + +The returned part pointer must be freed with \fIcurl_free(3)\fP after use. +.SH FLAGS +The flags argument is zero, one or more bits set in a bitmask. +.IP CURLU_DEFAULT_PORT +If the handle has no port stored, this option will make \fIcurl_url_get(3)\fP +return the default port for the used scheme. +.IP CURLU_DEFAULT_SCHEME +If the handle has no scheme stored, this option will make +\fIcurl_url_get(3)\fP return the default scheme instead of error. +.IP CURLU_NO_DEFAULT_PORT +Instructs \fIcurl_url_get(3)\fP to not return a port number if it matches the +default port for the scheme. +.IP CURLU_URLDECODE +Asks \fIcurl_url_get(3)\fP to URL decode the contents before returning it. It +will not attempt to decode the scheme, the port number or the full URL. + +The query component will also get plus-to-space convertion as a bonus when +this bit is set. + +Note that this URL decoding is charset unaware and you will get a zero +terminated string back with data that could be intended for a particular +encoding. + +If there's any byte values lower than 32 in the decoded string, the get +operation will return an error instead. +.SH PARTS +.IP CURLUPART_URL +When asked to return the full URL, \fIcurl_url_get(3)\fP will return a +normalized and possibly cleaned up version of what was previously parsed. +.IP CURLUPART_SCHEME +Scheme cannot be URL decoded on get. +.IP CURLUPART_USER +.IP CURLUPART_PASSWORD +.IP CURLUPART_OPTIONS +.IP CURLUPART_HOST +.IP CURLUPART_PORT +Port cannot be URL decoded on get. +.IP CURLUPART_PATH +.IP CURLUPART_QUERY +The query part will also get pluses converted to space when asked to URL +decode on get with the CURLU_URLDECODE bit. +.IP CURLUPART_FRAGMENT +.SH RETURN VALUE +Returns a CURLUcode error value, which is CURLUE_OK (0) if everything went +fine. + +If this function returns an error, no URL part is returned. +.SH EXAMPLE +.nf + CURLUcode rc; + CURLU *url = curl_url(); + rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0); + if(!rc) { + char *scheme; + rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0); + if(!rc) { + printf("the scheme is %s\n", scheme); + curl_free(scheme); + } + curl_url_cleanup(url); + } +.fi +.SH AVAILABILITY +Added in curl 7.63.0 +.SH "SEE ALSO" +.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), " +.BR curl_url_dup "(3), " diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3 new file mode 100644 index 0000000..75fc0d9 --- /dev/null +++ b/docs/libcurl/curl_url_set.3 @@ -0,0 +1,120 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH curl_url_set 3 "6 Aug 2018" "libcurl" "libcurl Manual" +.SH NAME +curl_url_set - set a part from a URL +.SH SYNOPSIS +.B #include + +CURLUcode curl_url_set(CURLU *url, + CURLUPart part, + const char *content, + unsigned int flags) +.fi +.SH DESCRIPTION +Given the \fIurl\fP handle of an already parsed URL, this function lets the +user set/update individual pieces of it. + +The \fIpart\fP argument should identify the particular URL part (see list +below) to set or change, with \fIcontent\fP pointing to a zero terminated +string with the new contents for that URL part. The contents should be in the +form and encoding they'd use in a URL: URL encoded. + +Setting a part to a NULL pointer will effectively remove that part's contents +from the CURLU handle. + +The \fIflags\fP argument is a bitmask with independent features. +.SH PARTS +.IP CURLUPART_URL +Allows the full URL of the handle to be replaced. If the handle already is +populated with a URL, the new URL can be relative to the previous. + +When successfully setting a new URL, relative or absolute, the handle contents +will be replaced with the information of the newly set URL. + +Pass a pointer to a zero terminated string to the \fIurl\fP parameter. The +string must point to a correctly formatted "RFC 3986+" URL or be a NULL +pointer. +.IP CURLUPART_SCHEME +Scheme cannot be URL decoded on set. +.IP CURLUPART_USER +.IP CURLUPART_PASSWORD +.IP CURLUPART_OPTIONS +.IP CURLUPART_HOST +The host name can use IDNA. The string must then be encoded as your locale +says or UTF-8 (when winidn is used). +.IP CURLUPART_PORT +Port cannot be URL encoded on set. +.IP CURLUPART_PATH +If a path is set in the URL without a leading slash, a slash will be inserted +automatically when this URL is read from the handle. +.IP CURLUPART_QUERY +The query part will also get spaces converted to pluses when asked to URL +encode on set with the CURLU_URLENCODE bit. + +If used in with \fICURLU_APPENDQUERY\fP, the provided part will be appended on +the end of the existing query - and if the previous part didn't end with an +ampersand (&), an ampersand will be inserted before the new appended part. + +When \fCURLU_APPENDQUERY\fP is used together with \fICURLU_URLENCODE\fP, +the '=' symbols will not be URL encoded. + +The question mark in the URL is not part of the actual query contents. +.IP CURLUPART_FRAGMENT +The hash sign in the URL is not part of the actual fragment contents. +.SH FLAGS +The flags argument is zero, one or more bits set in a bitmask. +.IP CURLU_NON_SUPPORT_SCHEME +If set, allows \fIcurl_url_set(3)\fP to set a non-supported scheme. +.IP CURLU_URLENCODE +When set, \fIcurl_url_set(3)\fP URL encodes the part on entry, except for +scheme, port and URL. + +When setting the path component with URL encoding enabled, the slash character +will be skipped. + +The query part gets space-to-plus conversion before the URL conversion. + +This URL encoding is charset unaware and will convert the input on a +byte-by-byte manner. +.SH RETURN VALUE +Returns a CURLUcode error value, which is CURLUE_OK (0) if everything went +fine. + +If this function returns an error, no URL part is returned. +.SH EXAMPLE +.nf + CURLUcode rc; + CURLU *url = curl_url(); + rc = curl_url_set(url, CURLUPART_URL, "https://example.com", 0); + if(!rc) { + char *scheme; + /* change it to an FTP URL */ + rc = curl_url_set(url, CURLUPART_SCHEME, "ftp", 0); + } + curl_url_cleanup(url); +.fi +.SH AVAILABILITY +Added in curl 7.63.0 +.SH "SEE ALSO" +.BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_get "(3), " +.BR curl_url_dup "(3), " diff --git a/docs/libcurl/symbols-in-versions b/docs/libcurl/symbols-in-versions index 7448b4f..c797cb7 100644 --- a/docs/libcurl/symbols-in-versions +++ b/docs/libcurl/symbols-in-versions @@ -718,6 +718,36 @@ CURLSSLSET_NO_BACKENDS 7.56.0 CURLSSLSET_OK 7.56.0 CURLSSLSET_TOO_LATE 7.56.0 CURLSSLSET_UNKNOWN_BACKEND 7.56.0 +CURLUPART_FRAGMENT 7.62.0 +CURLUPART_HOST 7.62.0 +CURLUPART_OPTIONS 7.62.0 +CURLUPART_PASSWORD 7.62.0 +CURLUPART_PATH 7.62.0 +CURLUPART_PORT 7.62.0 +CURLUPART_QUERY 7.62.0 +CURLUPART_SCHEME 7.62.0 +CURLUPART_URL 7.62.0 +CURLUPART_USER 7.62.0 +CURLUE_BAD_HANDLE 7.62.0 +CURLUE_BAD_PARTPOINTER 7.62.0 +CURLUE_BAD_PORT_NUMBER 7.62.0 +CURLUE_MALFORMED_INPUT 7.62.0 +CURLUE_NO_FRAGMENT 7.62.0 +CURLUE_NO_HOST 7.62.0 +CURLUE_NO_OPTIONS 7.62.0 +CURLUE_NO_PASSWORD 7.62.0 +CURLUE_NO_PATH 7.62.0 +CURLUE_NO_PORT 7.62.0 +CURLUE_NO_QUERY 7.62.0 +CURLUE_NO_SCHEME 7.62.0 +CURLUE_NO_USER 7.62.0 +CURLUE_OK 7.62.0 +CURLUE_OUT_OF_MEMORY 7.62.0 +CURLUE_RELATIVE 7.62.0 +CURLUE_UNKNOWN_PART 7.62.0 +CURLUE_UNSUPPORTED_SCHEME 7.62.0 +CURLUE_URLDECODE 7.62.0 +CURLUE_USER_NOT_ALLOWED 7.62.0 CURLUSESSL_ALL 7.17.0 CURLUSESSL_CONTROL 7.17.0 CURLUSESSL_NONE 7.17.0 diff --git a/include/curl/Makefile.am b/include/curl/Makefile.am index 989d4a2..bf5f061 100644 --- a/include/curl/Makefile.am +++ b/include/curl/Makefile.am @@ -5,7 +5,7 @@ # | (__| |_| | _ <| |___ # \___|\___/|_| \_\_____| # -# Copyright (C) 1998 - 2017, Daniel Stenberg, , et al. +# Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. # # This software is licensed as described in the file COPYING, which # you should have received as part of this distribution. The terms @@ -21,7 +21,7 @@ ########################################################################### pkginclude_HEADERS = \ curl.h curlver.h easy.h mprintf.h stdcheaders.h multi.h \ - typecheck-gcc.h system.h + typecheck-gcc.h system.h urlapi.h pkgincludedir= $(includedir)/curl diff --git a/include/curl/curl.h b/include/curl/curl.h index 067b34d..8f473e2 100644 --- a/include/curl/curl.h +++ b/include/curl/curl.h @@ -2779,6 +2779,7 @@ CURL_EXTERN CURLcode curl_easy_pause(CURL *handle, int bitmask); stuff before they can be included! */ #include "easy.h" /* nothing in curl is fun without the easy stuff */ #include "multi.h" +#include "urlapi.h" /* the typechecker doesn't work in C++ (yet) */ #if defined(__GNUC__) && defined(__GNUC_MINOR__) && \ diff --git a/include/curl/urlapi.h b/include/curl/urlapi.h new file mode 100644 index 0000000..b16cfce --- /dev/null +++ b/include/curl/urlapi.h @@ -0,0 +1,121 @@ +#ifndef __CURL_URLAPI_H +#define __CURL_URLAPI_H +/*************************************************************************** + * _ _ ____ _ + * Project ___| | | | _ \| | + * / __| | | | |_) | | + * | (__| |_| | _ <| |___ + * \___|\___/|_| \_\_____| + * + * Copyright (C) 2018, Daniel Stenberg, , et al. + * + * This software is licensed as described in the file COPYING, which + * you should have received as part of this distribution. The terms + * are also available at https://curl.haxx.se/docs/copyright.html. + * + * You may opt to use, copy, modify, merge, publish, distribute and/or sell + * copies of the Software, and permit persons to whom the Software is + * furnished to do so, under the terms of the COPYING file. + * + * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY + * KIND, either express or implied. + * + ***************************************************************************/ + +#ifdef __cplusplus +extern "C" { +#endif + +/* the error codes for the URL API */ +typedef enum { + CURLUE_OK, + CURLUE_BAD_HANDLE, /* 1 */ + CURLUE_BAD_PARTPOINTER, /* 2 */ + CURLUE_MALFORMED_INPUT, /* 3 */ + CURLUE_BAD_PORT_NUMBER, /* 4 */ + CURLUE_UNSUPPORTED_SCHEME, /* 5 */ + CURLUE_URLDECODE, /* 6 */ + CURLUE_RELATIVE, /* 7 */ + CURLUE_USER_NOT_ALLOWED, /* 8 */ + CURLUE_UNKNOWN_PART, /* 9 */ + CURLUE_NO_SCHEME, /* 10 */ + CURLUE_NO_USER, /* 11 */ + CURLUE_NO_PASSWORD, /* 12 */ + CURLUE_NO_OPTIONS, /* 13 */ + CURLUE_NO_HOST, /* 14 */ + CURLUE_NO_PORT, /* 15 */ + CURLUE_NO_PATH, /* 16 */ + CURLUE_NO_QUERY, /* 17 */ + CURLUE_NO_FRAGMENT, /* 18 */ + CURLUE_OUT_OF_MEMORY /* 19 */ +} CURLUcode; + +typedef enum { + CURLUPART_URL, + CURLUPART_SCHEME, + CURLUPART_USER, + CURLUPART_PASSWORD, + CURLUPART_OPTIONS, + CURLUPART_HOST, + CURLUPART_PORT, + CURLUPART_PATH, + CURLUPART_QUERY, + CURLUPART_FRAGMENT +} CURLUPart; + +#define CURLU_DEFAULT_PORT (1<<0) /* return default port number */ +#define CURLU_NO_DEFAULT_PORT (1<<1) /* act as if no port number was set, + if the port number matches the + default for the scheme */ +#define CURLU_DEFAULT_SCHEME (1<<2) /* return default scheme if + missing */ +#define CURLU_NON_SUPPORT_SCHEME (1<<3) /* allow non-supported scheme */ +#define CURLU_PATH_AS_IS (1<<4) /* leave dot sequences */ +#define CURLU_DISALLOW_USER (1<<5) /* no user+password allowed */ +#define CURLU_URLDECODE (1<<6) /* URL decode on get */ +#define CURLU_URLENCODE (1<<7) /* URL encode on set */ +#define CURLU_APPENDQUERY (1<<8) /* append a form style part */ + +typedef struct Curl_URL CURLU; + +/* + * curl_url() creates a new CURLU handle and returns a pointer to it. + * Must be freed with curl_url_cleanup(). + */ +CURL_EXTERN CURLU *curl_url(void); + +/* + * curl_url_cleanup() frees the CURLU handle and related resources used for + * the URL parsing. It will not free strings previously returned with the URL + * API. + */ +CURL_EXTERN void curl_url_cleanup(CURLU *handle); + +/* + * curl_url_dup() duplicates a CURLU handle and returns a new copy. The new + * handle must also be freed with curl_url_cleanup(). + */ +CURL_EXTERN CURLU *curl_url_dup(CURLU *in); + +/* + * curl_url_get() extracts a specific part of the URL from a CURLU + * handle. Returns error code. The returned pointer MUST be freed with + * curl_free() afterwards. + */ +CURL_EXTERN CURLUcode curl_url_get(CURLU *handle, CURLUPart what, + char **part, unsigned int flags); + +/* + * curl_url_set() sets a specific part of the URL in a CURLU handle. Returns + * error code. The passed in string will be copied. Passing a NULL instead of + * a part string, clears that part. + */ +CURL_EXTERN CURLUcode curl_url_set(CURLU *handle, CURLUPart what, + const char *part, unsigned int flags); + + +#ifdef __cplusplus +} /* end of extern "C" */ +#endif + +#endif diff --git a/lib/Makefile.inc b/lib/Makefile.inc index 76ca6d0..1ff82e1 100644 --- a/lib/Makefile.inc +++ b/lib/Makefile.inc @@ -54,7 +54,8 @@ LIB_CFILES = file.c timeval.c base64.c hostip.c progress.c formdata.c \ http_ntlm.c curl_ntlm_wb.c curl_ntlm_core.c curl_sasl.c rand.c \ curl_multibyte.c hostcheck.c conncache.c pipeline.c dotdot.c \ x509asn1.c http2.c smb.c curl_endian.c curl_des.c system_win32.c \ - mime.c sha256.c setopt.c curl_path.c curl_ctype.c curl_range.c psl.c + mime.c sha256.c setopt.c curl_path.c curl_ctype.c curl_range.c psl.c \ + urlapi.c LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \ formdata.h cookie.h http.h sendf.h ftp.h url.h dict.h if2ip.h \ @@ -74,7 +75,7 @@ LIB_HFILES = arpa_telnet.h netrc.h file.h timeval.h hostip.h progress.h \ curl_setup_once.h multihandle.h setup-vms.h pipeline.h dotdot.h \ x509asn1.h http2.h sigpipe.h smb.h curl_endian.h curl_des.h \ curl_printf.h system_win32.h rand.h mime.h curl_sha256.h setopt.h \ - curl_path.h curl_ctype.h curl_range.h psl.h + curl_path.h curl_ctype.h curl_range.h psl.h urlapi-int.h LIB_RCFILES = libcurl.rc diff --git a/lib/escape.c b/lib/escape.c index 10774f0..afd3899 100644 --- a/lib/escape.c +++ b/lib/escape.c @@ -5,7 +5,7 @@ * | (__| |_| | _ <| |___ * \___|\___/|_| \_\_____| * - * Copyright (C) 1998 - 2017, Daniel Stenberg, , et al. + * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. * * This software is licensed as described in the file COPYING, which * you should have received as part of this distribution. The terms @@ -41,7 +41,7 @@ its behavior is altered by the current locale. See https://tools.ietf.org/html/rfc3986#section-2.3 */ -static bool Curl_isunreserved(unsigned char in) +bool Curl_isunreserved(unsigned char in) { switch(in) { case '0': case '1': case '2': case '3': case '4': @@ -141,6 +141,8 @@ char *curl_easy_escape(struct Curl_easy *data, const char *string, * Returns a pointer to a malloced string in *ostring with length given in * *olen. If length == 0, the length is assumed to be strlen(string). * + * 'data' can be set to NULL but then this function can't convert network + * data to host for non-ascii. */ CURLcode Curl_urldecode(struct Curl_easy *data, const char *string, size_t length, @@ -151,7 +153,7 @@ CURLcode Curl_urldecode(struct Curl_easy *data, char *ns = malloc(alloc); size_t strindex = 0; unsigned long hex; - CURLcode result; + CURLcode result = CURLE_OK; if(!ns) return CURLE_OUT_OF_MEMORY; @@ -171,11 +173,13 @@ CURLcode Curl_urldecode(struct Curl_easy *data, in = curlx_ultouc(hex); /* this long is never bigger than 255 anyway */ - result = Curl_convert_from_network(data, (char *)&in, 1); - if(result) { - /* Curl_convert_from_network calls failf if unsuccessful */ - free(ns); - return result; + if(data) { + result = Curl_convert_from_network(data, (char *)&in, 1); + if(result) { + /* Curl_convert_from_network calls failf if unsuccessful */ + free(ns); + return result; + } } string += 2; diff --git a/lib/escape.h b/lib/escape.h index 638666f..666f1ad 100644 --- a/lib/escape.h +++ b/lib/escape.h @@ -7,7 +7,7 @@ * | (__| |_| | _ <| |___ * \___|\___/|_| \_\_____| * - * Copyright (C) 1998 - 2011, Daniel Stenberg, , et al. + * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. * * This software is licensed as described in the file COPYING, which * you should have received as part of this distribution. The terms @@ -24,6 +24,7 @@ /* Escape and unescape URL encoding in strings. The functions return a new * allocated string or NULL if an error occurred. */ +bool Curl_isunreserved(unsigned char in); CURLcode Curl_urldecode(struct Curl_easy *data, const char *string, size_t length, char **ostring, size_t *olen, diff --git a/lib/imap.c b/lib/imap.c index 942fe7d..28962c1 100644 --- a/lib/imap.c +++ b/lib/imap.c @@ -159,7 +159,8 @@ const struct Curl_handler Curl_handler_imaps = { ZERO_NULL, /* connection_check */ PORT_IMAPS, /* defport */ CURLPROTO_IMAPS, /* protocol */ - PROTOPT_CLOSEACTION | PROTOPT_SSL /* flags */ + PROTOPT_CLOSEACTION | PROTOPT_SSL | /* flags */ + PROTOPT_URLOPTIONS }; #endif diff --git a/lib/transfer.c b/lib/transfer.c index 7159d5c..ecd1063 100644 --- a/lib/transfer.c +++ b/lib/transfer.c @@ -75,6 +75,7 @@ #include "http2.h" #include "mime.h" #include "strcase.h" +#include "urlapi-int.h" /* The last 3 #include files should be in this order */ #include "curl_printf.h" @@ -1453,311 +1454,6 @@ CURLcode Curl_posttransfer(struct Curl_easy *data) return CURLE_OK; } -#ifndef CURL_DISABLE_HTTP -/* - * Find the separator at the end of the host name, or the '?' in cases like - * http://www.url.com?id=2380 - */ -static const char *find_host_sep(const char *url) -{ - const char *sep; - const char *query; - - /* Find the start of the hostname */ - sep = strstr(url, "//"); - if(!sep) - sep = url; - else - sep += 2; - - query = strchr(sep, '?'); - sep = strchr(sep, '/'); - - if(!sep) - sep = url + strlen(url); - - if(!query) - query = url + strlen(url); - - return sep < query ? sep : query; -} - -/* - * Decide in an encoding-independent manner whether a character in an - * URL must be escaped. The same criterion must be used in strlen_url() - * and strcpy_url(). - */ -static bool urlchar_needs_escaping(int c) -{ - return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c)); -} - -/* - * strlen_url() returns the length of the given URL if the spaces within the - * URL were properly URL encoded. - * URL encoding should be skipped for host names, otherwise IDN resolution - * will fail. - */ -static size_t strlen_url(const char *url, bool relative) -{ - const unsigned char *ptr; - size_t newlen = 0; - bool left = TRUE; /* left side of the ? */ - const unsigned char *host_sep = (const unsigned char *) url; - - if(!relative) - host_sep = (const unsigned char *) find_host_sep(url); - - for(ptr = (unsigned char *)url; *ptr; ptr++) { - - if(ptr < host_sep) { - ++newlen; - continue; - } - - switch(*ptr) { - case '?': - left = FALSE; - /* FALLTHROUGH */ - default: - if(urlchar_needs_escaping(*ptr)) - newlen += 2; - newlen++; - break; - case ' ': - if(left) - newlen += 3; - else - newlen++; - break; - } - } - return newlen; -} - -/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in - * the source URL accordingly. - * URL encoding should be skipped for host names, otherwise IDN resolution - * will fail. - */ -static void strcpy_url(char *output, const char *url, bool relative) -{ - /* we must add this with whitespace-replacing */ - bool left = TRUE; - const unsigned char *iptr; - char *optr = output; - const unsigned char *host_sep = (const unsigned char *) url; - - if(!relative) - host_sep = (const unsigned char *) find_host_sep(url); - - for(iptr = (unsigned char *)url; /* read from here */ - *iptr; /* until zero byte */ - iptr++) { - - if(iptr < host_sep) { - *optr++ = *iptr; - continue; - } - - switch(*iptr) { - case '?': - left = FALSE; - /* FALLTHROUGH */ - default: - if(urlchar_needs_escaping(*iptr)) { - snprintf(optr, 4, "%%%02x", *iptr); - optr += 3; - } - else - *optr++=*iptr; - break; - case ' ': - if(left) { - *optr++='%'; /* add a '%' */ - *optr++='2'; /* add a '2' */ - *optr++='0'; /* add a '0' */ - } - else - *optr++='+'; /* add a '+' here */ - break; - } - } - *optr = 0; /* zero terminate output buffer */ - -} - -/* - * Returns true if the given URL is absolute (as opposed to relative) - */ -static bool is_absolute_url(const char *url) -{ - char prot[16]; /* URL protocol string storage */ - char letter; /* used for a silly sscanf */ - - return (2 == sscanf(url, "%15[^?&/:]://%c", prot, &letter)) ? TRUE : FALSE; -} - -/* - * Concatenate a relative URL to a base URL making it absolute. - * URL-encodes any spaces. - * The returned pointer must be freed by the caller unless NULL - * (returns NULL on out of memory). - */ -static char *concat_url(const char *base, const char *relurl) -{ - /*** - TRY to append this new path to the old URL - to the right of the host part. Oh crap, this is doomed to cause - problems in the future... - */ - char *newest; - char *protsep; - char *pathsep; - size_t newlen; - bool host_changed = FALSE; - - const char *useurl = relurl; - size_t urllen; - - /* we must make our own copy of the URL to play with, as it may - point to read-only data */ - char *url_clone = strdup(base); - - if(!url_clone) - return NULL; /* skip out of this NOW */ - - /* protsep points to the start of the host name */ - protsep = strstr(url_clone, "//"); - if(!protsep) - protsep = url_clone; - else - protsep += 2; /* pass the slashes */ - - if('/' != relurl[0]) { - int level = 0; - - /* First we need to find out if there's a ?-letter in the URL, - and cut it and the right-side of that off */ - pathsep = strchr(protsep, '?'); - if(pathsep) - *pathsep = 0; - - /* we have a relative path to append to the last slash if there's one - available, or if the new URL is just a query string (starts with a - '?') we append the new one at the end of the entire currently worked - out URL */ - if(useurl[0] != '?') { - pathsep = strrchr(protsep, '/'); - if(pathsep) - *pathsep = 0; - } - - /* Check if there's any slash after the host name, and if so, remember - that position instead */ - pathsep = strchr(protsep, '/'); - if(pathsep) - protsep = pathsep + 1; - else - protsep = NULL; - - /* now deal with one "./" or any amount of "../" in the newurl - and act accordingly */ - - if((useurl[0] == '.') && (useurl[1] == '/')) - useurl += 2; /* just skip the "./" */ - - while((useurl[0] == '.') && - (useurl[1] == '.') && - (useurl[2] == '/')) { - level++; - useurl += 3; /* pass the "../" */ - } - - if(protsep) { - while(level--) { - /* cut off one more level from the right of the original URL */ - pathsep = strrchr(protsep, '/'); - if(pathsep) - *pathsep = 0; - else { - *protsep = 0; - break; - } - } - } - } - else { - /* We got a new absolute path for this server */ - - if((relurl[0] == '/') && (relurl[1] == '/')) { - /* the new URL starts with //, just keep the protocol part from the - original one */ - *protsep = 0; - useurl = &relurl[2]; /* we keep the slashes from the original, so we - skip the new ones */ - host_changed = TRUE; - } - else { - /* cut off the original URL from the first slash, or deal with URLs - without slash */ - pathsep = strchr(protsep, '/'); - if(pathsep) { - /* When people use badly formatted URLs, such as - "http://www.url.com?dir=/home/daniel" we must not use the first - slash, if there's a ?-letter before it! */ - char *sep = strchr(protsep, '?'); - if(sep && (sep < pathsep)) - pathsep = sep; - *pathsep = 0; - } - else { - /* There was no slash. Now, since we might be operating on a badly - formatted URL, such as "http://www.url.com?id=2380" which doesn't - use a slash separator as it is supposed to, we need to check for a - ?-letter as well! */ - pathsep = strchr(protsep, '?'); - if(pathsep) - *pathsep = 0; - } - } - } - - /* If the new part contains a space, this is a mighty stupid redirect - but we still make an effort to do "right". To the left of a '?' - letter we replace each space with %20 while it is replaced with '+' - on the right side of the '?' letter. - */ - newlen = strlen_url(useurl, !host_changed); - - urllen = strlen(url_clone); - - newest = malloc(urllen + 1 + /* possible slash */ - newlen + 1 /* zero byte */); - - if(!newest) { - free(url_clone); /* don't leak this */ - return NULL; - } - - /* copy over the root url part */ - memcpy(newest, url_clone, urllen); - - /* check if we need to append a slash */ - if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0])) - ; - else - newest[urllen++]='/'; - - /* then append the new piece on the right side */ - strcpy_url(&newest[urllen], useurl, !host_changed); - - free(url_clone); - - return newest; -} -#endif /* CURL_DISABLE_HTTP */ - /* * Curl_follow() handles the URL redirect magic. Pass in the 'newurl' string * as given by the remote server and set up the new URL to request. @@ -1809,12 +1505,12 @@ CURLcode Curl_follow(struct Curl_easy *data, } } - if(!is_absolute_url(newurl)) { + if(!Curl_is_absolute_url(newurl, NULL, 8)) { /*** *DANG* this is an RFC 2068 violation. The URL is supposed to be absolute and this doesn't seem to be that! */ - char *absolute = concat_url(data->change.url, newurl); + char *absolute = Curl_concat_url(data->change.url, newurl); if(!absolute) return CURLE_OUT_OF_MEMORY; newurl = absolute; @@ -1823,7 +1519,7 @@ CURLcode Curl_follow(struct Curl_easy *data, /* The new URL MAY contain space or high byte values, that means a mighty stupid redirect URL but we still make an effort to do "right". */ char *newest; - size_t newlen = strlen_url(newurl, FALSE); + size_t newlen = Curl_strlen_url(newurl, FALSE); /* This is an absolute URL, don't allow the custom port number */ disallowport = TRUE; @@ -1832,7 +1528,7 @@ CURLcode Curl_follow(struct Curl_easy *data, if(!newest) return CURLE_OUT_OF_MEMORY; - strcpy_url(newest, newurl, FALSE); /* create a space-free URL */ + Curl_strcpy_url(newest, newurl, FALSE); /* create a space-free URL */ newurl = newest; /* use this instead now */ } diff --git a/lib/url.c b/lib/url.c index dcc1ecc..4f75f11 100644 --- a/lib/url.c +++ b/lib/url.c @@ -1939,30 +1939,37 @@ static struct connectdata *allocate_conn(struct Curl_easy *data) return NULL; } -static CURLcode findprotocol(struct Curl_easy *data, - struct connectdata *conn, - const char *protostr) +/* returns the handdler if the given scheme is built-in */ +const struct Curl_handler *Curl_builtin_scheme(const char *scheme) { const struct Curl_handler * const *pp; const struct Curl_handler *p; - - /* Scan protocol handler table and match against 'protostr' to set a few - variables based on the URL. Now that the handler may be changed later - when the protocol specific setup function is called. */ - for(pp = protocols; (p = *pp) != NULL; pp++) { - if(strcasecompare(p->scheme, protostr)) { + /* Scan protocol handler table and match against 'scheme'. The handler may + be changed later when the protocol specific setup function is called. */ + for(pp = protocols; (p = *pp) != NULL; pp++) + if(strcasecompare(p->scheme, scheme)) /* Protocol found in table. Check if allowed */ - if(!(data->set.allowed_protocols & p->protocol)) - /* nope, get out */ - break; + return p; + return NULL; /* not found */ +} - /* it is allowed for "normal" request, now do an extra check if this is - the result of a redirect */ - if(data->state.this_is_a_follow && - !(data->set.redir_protocols & p->protocol)) - /* nope, get out */ - break; +static CURLcode findprotocol(struct Curl_easy *data, + struct connectdata *conn, + const char *protostr) +{ + const struct Curl_handler *p = Curl_builtin_scheme(protostr); + + if(p && /* Protocol found in table. Check if allowed */ + (data->set.allowed_protocols & p->protocol)) { + + /* it is allowed for "normal" request, now do an extra check if this is + the result of a redirect */ + if(data->state.this_is_a_follow && + !(data->set.redir_protocols & p->protocol)) + /* nope, get out */ + ; + else { /* Perform setup complement if some. */ conn->handler = conn->given = p; @@ -1971,7 +1978,6 @@ static CURLcode findprotocol(struct Curl_easy *data, } } - /* The protocol was not found in the table, but we don't have to assign it to anything since it is already assigned to a dummy-struct in the create_conn() function when the connectdata struct is allocated. */ diff --git a/lib/url.h b/lib/url.h index ef3ebf0..0034f82 100644 --- a/lib/url.h +++ b/lib/url.h @@ -69,6 +69,8 @@ void Curl_getoff_all_pipelines(struct Curl_easy *data, void Curl_close_connections(struct Curl_easy *data); +const struct Curl_handler *Curl_builtin_scheme(const char *scheme); + #define CURL_DEFAULT_PROXY_PORT 1080 /* default proxy port unless specified */ #define CURL_DEFAULT_HTTPS_PROXY_PORT 443 /* default https proxy port unless specified */ diff --git a/lib/escape.h b/lib/urlapi-int.h similarity index 66% copy from lib/escape.h copy to lib/urlapi-int.h index 638666f..7ac09fd 100644 --- a/lib/escape.h +++ b/lib/urlapi-int.h @@ -1,5 +1,5 @@ -#ifndef HEADER_CURL_ESCAPE_H -#define HEADER_CURL_ESCAPE_H +#ifndef HEADER_CURL_URLAPI_INT_H +#define HEADER_CURL_URLAPI_INT_H /*************************************************************************** * _ _ ____ _ * Project ___| | | | _ \| | @@ -7,7 +7,7 @@ * | (__| |_| | _ <| |___ * \___|\___/|_| \_\_____| * - * Copyright (C) 1998 - 2011, Daniel Stenberg, , et al. + * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. * * This software is licensed as described in the file COPYING, which * you should have received as part of this distribution. The terms @@ -21,13 +21,9 @@ * KIND, either express or implied. * ***************************************************************************/ -/* Escape and unescape URL encoding in strings. The functions return a new - * allocated string or NULL if an error occurred. */ - -CURLcode Curl_urldecode(struct Curl_easy *data, - const char *string, size_t length, - char **ostring, size_t *olen, - bool reject_crlf); - -#endif /* HEADER_CURL_ESCAPE_H */ - +#include "curl_setup.h" +bool Curl_is_absolute_url(const char *url, char *scheme, size_t buflen); +char *Curl_concat_url(const char *base, const char *relurl); +size_t Curl_strlen_url(const char *url, bool relative); +void Curl_strcpy_url(char *output, const char *url, bool relative); +#endif diff --git a/lib/urlapi.c b/lib/urlapi.c new file mode 100644 index 0000000..8287861 --- /dev/null +++ b/lib/urlapi.c @@ -0,0 +1,1315 @@ +/*************************************************************************** + * _ _ ____ _ + * Project ___| | | | _ \| | + * / __| | | | |_) | | + * | (__| |_| | _ <| |___ + * \___|\___/|_| \_\_____| + * + * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. + * + * This software is licensed as described in the file COPYING, which + * you should have received as part of this distribution. The terms + * are also available at https://curl.haxx.se/docs/copyright.html. + * + * You may opt to use, copy, modify, merge, publish, distribute and/or sell + * copies of the Software, and permit persons to whom the Software is + * furnished to do so, under the terms of the COPYING file. + * + * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY + * KIND, either express or implied. + * + ***************************************************************************/ + +#include "curl_setup.h" + +#include "urldata.h" +#include "urlapi-int.h" +#include "strcase.h" +#include "dotdot.h" +#include "url.h" +#include "escape.h" +#include "curl_ctype.h" + +/* The last 3 #include files should be in this order */ +#include "curl_printf.h" +#include "curl_memory.h" +#include "memdebug.h" + +/* Internal representation of CURLU. Point to URL-encoded strings. */ +struct Curl_URL { + char *scheme; + char *user; + char *password; + char *options; /* IMAP only? */ + char *host; + char *port; + char *path; + char *query; + char *fragment; + + char *scratch; /* temporary scratch area */ + long portnum; /* the numerical version */ +}; + +#define DEFAULT_SCHEME "https" + +/* scheme is not URL encoded, the longest libcurl supported ones are 6 + letters */ +#define MAX_SCHEME_LEN 8 + +static void free_urlhandle(struct Curl_URL *u) +{ + free(u->scheme); + free(u->user); + free(u->password); + free(u->options); + free(u->host); + free(u->port); + free(u->path); + free(u->query); + free(u->fragment); + free(u->scratch); +} + +/* move the full contents of one handle onto another and + free the original */ +static void mv_urlhandle(struct Curl_URL *from, + struct Curl_URL *to) +{ + free_urlhandle(to); + *to = *from; + free(from); +} + +/* + * Find the separator at the end of the host name, or the '?' in cases like + * http://www.url.com?id=2380 + */ +static const char *find_host_sep(const char *url) +{ + const char *sep; + const char *query; + + /* Find the start of the hostname */ + sep = strstr(url, "//"); + if(!sep) + sep = url; + else + sep += 2; + + query = strchr(sep, '?'); + sep = strchr(sep, '/'); + + if(!sep) + sep = url + strlen(url); + + if(!query) + query = url + strlen(url); + + return sep < query ? sep : query; +} + +/* + * Decide in an encoding-independent manner whether a character in an + * URL must be escaped. The same criterion must be used in strlen_url() + * and strcpy_url(). + */ +static bool urlchar_needs_escaping(int c) +{ + return !(ISCNTRL(c) || ISSPACE(c) || ISGRAPH(c)); +} + +/* + * strlen_url() returns the length of the given URL if the spaces within the + * URL were properly URL encoded. + * URL encoding should be skipped for host names, otherwise IDN resolution + * will fail. + */ +size_t Curl_strlen_url(const char *url, bool relative) +{ + const unsigned char *ptr; + size_t newlen = 0; + bool left = TRUE; /* left side of the ? */ + const unsigned char *host_sep = (const unsigned char *) url; + + if(!relative) + host_sep = (const unsigned char *) find_host_sep(url); + + for(ptr = (unsigned char *)url; *ptr; ptr++) { + + if(ptr < host_sep) { + ++newlen; + continue; + } + + switch(*ptr) { + case '?': + left = FALSE; + /* FALLTHROUGH */ + default: + if(urlchar_needs_escaping(*ptr)) + newlen += 2; + newlen++; + break; + case ' ': + if(left) + newlen += 3; + else + newlen++; + break; + } + } + return newlen; +} + +/* strcpy_url() copies a url to a output buffer and URL-encodes the spaces in + * the source URL accordingly. + * URL encoding should be skipped for host names, otherwise IDN resolution + * will fail. + */ +void Curl_strcpy_url(char *output, const char *url, bool relative) +{ + /* we must add this with whitespace-replacing */ + bool left = TRUE; + const unsigned char *iptr; + char *optr = output; + const unsigned char *host_sep = (const unsigned char *) url; + + if(!relative) + host_sep = (const unsigned char *) find_host_sep(url); + + for(iptr = (unsigned char *)url; /* read from here */ + *iptr; /* until zero byte */ + iptr++) { + + if(iptr < host_sep) { + *optr++ = *iptr; + continue; + } + + switch(*iptr) { + case '?': + left = FALSE; + /* FALLTHROUGH */ + default: + if(urlchar_needs_escaping(*iptr)) { + snprintf(optr, 4, "%%%02x", *iptr); + optr += 3; + } + else + *optr++=*iptr; + break; + case ' ': + if(left) { + *optr++='%'; /* add a '%' */ + *optr++='2'; /* add a '2' */ + *optr++='0'; /* add a '0' */ + } + else + *optr++='+'; /* add a '+' here */ + break; + } + } + *optr = 0; /* zero terminate output buffer */ + +} + +/* + * Returns true if the given URL is absolute (as opposed to relative) within + * the buffer size. Returns the scheme in the buffer if TRUE and 'buf' is + * non-NULL. + */ +bool Curl_is_absolute_url(const char *url, char *buf, size_t buflen) +{ + size_t i; + for(i = 0; i < buflen && url[i]; ++i) { + char s = url[i]; + if(s == ':') { + if(buf) + buf[i] = 0; + return TRUE; + } + /* RFC 3986 3.1 explains: + scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." ) + */ + else if(ISALNUM(s) || (s == '+') || (s == '-') || (s == '.') ) { + if(buf) + buf[i] = (char)TOLOWER(s); + } + else + break; + } + return FALSE; +} + +/* + * Concatenate a relative URL to a base URL making it absolute. + * URL-encodes any spaces. + * The returned pointer must be freed by the caller unless NULL + * (returns NULL on out of memory). + */ +char *Curl_concat_url(const char *base, const char *relurl) +{ + /*** + TRY to append this new path to the old URL + to the right of the host part. Oh crap, this is doomed to cause + problems in the future... + */ + char *newest; + char *protsep; + char *pathsep; + size_t newlen; + bool host_changed = FALSE; + + const char *useurl = relurl; + size_t urllen; + + /* we must make our own copy of the URL to play with, as it may + point to read-only data */ + char *url_clone = strdup(base); + + if(!url_clone) + return NULL; /* skip out of this NOW */ + + /* protsep points to the start of the host name */ + protsep = strstr(url_clone, "//"); + if(!protsep) + protsep = url_clone; + else + protsep += 2; /* pass the slashes */ + + if('/' != relurl[0]) { + int level = 0; + + /* First we need to find out if there's a ?-letter in the URL, + and cut it and the right-side of that off */ + pathsep = strchr(protsep, '?'); + if(pathsep) + *pathsep = 0; + + /* we have a relative path to append to the last slash if there's one + available, or if the new URL is just a query string (starts with a + '?') we append the new one at the end of the entire currently worked + out URL */ + if(useurl[0] != '?') { + pathsep = strrchr(protsep, '/'); + if(pathsep) + *pathsep = 0; + } + + /* Check if there's any slash after the host name, and if so, remember + that position instead */ + pathsep = strchr(protsep, '/'); + if(pathsep) + protsep = pathsep + 1; + else + protsep = NULL; + + /* now deal with one "./" or any amount of "../" in the newurl + and act accordingly */ + + if((useurl[0] == '.') && (useurl[1] == '/')) + useurl += 2; /* just skip the "./" */ + + while((useurl[0] == '.') && + (useurl[1] == '.') && + (useurl[2] == '/')) { + level++; + useurl += 3; /* pass the "../" */ + } + + if(protsep) { + while(level--) { + /* cut off one more level from the right of the original URL */ + pathsep = strrchr(protsep, '/'); + if(pathsep) + *pathsep = 0; + else { + *protsep = 0; + break; + } + } + } + } + else { + /* We got a new absolute path for this server */ + + if((relurl[0] == '/') && (relurl[1] == '/')) { + /* the new URL starts with //, just keep the protocol part from the + original one */ + *protsep = 0; + useurl = &relurl[2]; /* we keep the slashes from the original, so we + skip the new ones */ + host_changed = TRUE; + } + else { + /* cut off the original URL from the first slash, or deal with URLs + without slash */ + pathsep = strchr(protsep, '/'); + if(pathsep) { + /* When people use badly formatted URLs, such as + "http://www.url.com?dir=/home/daniel" we must not use the first + slash, if there's a ?-letter before it! */ + char *sep = strchr(protsep, '?'); + if(sep && (sep < pathsep)) + pathsep = sep; + *pathsep = 0; + } + else { + /* There was no slash. Now, since we might be operating on a badly + formatted URL, such as "http://www.url.com?id=2380" which doesn't + use a slash separator as it is supposed to, we need to check for a + ?-letter as well! */ + pathsep = strchr(protsep, '?'); + if(pathsep) + *pathsep = 0; + } + } + } + + /* If the new part contains a space, this is a mighty stupid redirect + but we still make an effort to do "right". To the left of a '?' + letter we replace each space with %20 while it is replaced with '+' + on the right side of the '?' letter. + */ + newlen = Curl_strlen_url(useurl, !host_changed); + + urllen = strlen(url_clone); + + newest = malloc(urllen + 1 + /* possible slash */ + newlen + 1 /* zero byte */); + + if(!newest) { + free(url_clone); /* don't leak this */ + return NULL; + } + + /* copy over the root url part */ + memcpy(newest, url_clone, urllen); + + /* check if we need to append a slash */ + if(('/' == useurl[0]) || (protsep && !*protsep) || ('?' == useurl[0])) + ; + else + newest[urllen++]='/'; + + /* then append the new piece on the right side */ + Curl_strcpy_url(&newest[urllen], useurl, !host_changed); + + free(url_clone); + + return newest; +} + +/* + * parse_hostname_login() + * + * Parse the login details (user name, password and options) from the URL and + * strip them out of the host name + * + */ +static CURLUcode parse_hostname_login(struct Curl_URL *u, + const struct Curl_handler *h, + char **hostname, + unsigned int flags) +{ + CURLUcode result = CURLUE_OK; + CURLcode ccode; + char *userp = NULL; + char *passwdp = NULL; + char *optionsp = NULL; + + /* At this point, we're hoping all the other special cases have + * been taken care of, so conn->host.name is at most + * [user[:password][;options]]@]hostname + * + * We need somewhere to put the embedded details, so do that first. + */ + + char *ptr = strchr(*hostname, '@'); + char *login = *hostname; + + if(!ptr) + goto out; + + /* We will now try to extract the + * possible login information in a string like: + * ftp://user:password@ftp.my.site:8021/README */ + *hostname = ++ptr; + + /* We could use the login information in the URL so extract it. Only parse + options if the handler says we should. */ + ccode = Curl_parse_login_details(login, ptr - login - 1, + &userp, &passwdp, + h->flags & PROTOPT_URLOPTIONS ? + &optionsp:NULL); + if(ccode) { + result = CURLUE_MALFORMED_INPUT; + goto out; + } + + if(userp) { + if(flags & CURLU_DISALLOW_USER) { + /* Option DISALLOW_USER is set and url contains username. */ + result = CURLUE_USER_NOT_ALLOWED; + goto out; + } + + u->user = userp; + } + + if(passwdp) + u->password = passwdp; + + if(optionsp) + u->options = optionsp; + + return CURLUE_OK; + out: + + free(userp); + free(passwdp); + free(optionsp); + + return result; +} + +static CURLUcode parse_port(struct Curl_URL *u, char *hostname) +{ + char *portptr; + char endbracket; + int len; + + if((1 == sscanf(hostname, "[%*45[0123456789abcdefABCDEF:.]%c%n", + &endbracket, &len)) && + (']' == endbracket)) { + /* this is a RFC2732-style specified IP-address */ + portptr = &hostname[len]; + if (*portptr != ':') + return CURLUE_MALFORMED_INPUT; + } + else + portptr = strchr(hostname, ':'); + + if(portptr) { + char *rest; + long port; + char portbuf[7]; + + if(!ISDIGIT(portptr[1])) + return CURLUE_BAD_PORT_NUMBER; + + port = strtol(portptr + 1, &rest, 10); /* Port number must be decimal */ + + if((port <= 0) || (port > 0xffff)) + /* Single unix standard says port numbers are 16 bits long, but we don't + treat port zero as OK. */ + return CURLUE_BAD_PORT_NUMBER; + + if(rest[0]) + return CURLUE_BAD_PORT_NUMBER; + + if(rest != &portptr[1]) { + *portptr++ = '\0'; /* cut off the name there */ + *rest = 0; + /* generate a new to get rid of leading zeroes etc */ + snprintf(portbuf, sizeof(portbuf), "%ld", port); + u->portnum = port; + u->port = strdup(portbuf); + if(!u->port) + return CURLUE_OUT_OF_MEMORY; + } + else { + /* Browser behavior adaptation. If there's a colon with no digits after, + just cut off the name there which makes us ignore the colon and just + use the default port. Firefox and Chrome both do that. */ + *portptr = '\0'; + } + } + + return CURLUE_OK; +} + +/* scan for byte values < 31 or 127 */ +static CURLUcode junkscan(char *part) +{ + char badbytes[]={ + /* */ 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, + 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, + 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, + 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, + 0x7f, + 0x00 /* zero terminate */ + }; + if(part) { + size_t n = strlen(part); + size_t nfine = strcspn(part, badbytes); + if(nfine != n) + /* since we don't know which part is scanned, return a generic error + code */ + return CURLUE_MALFORMED_INPUT; + } + return CURLUE_OK; +} + +static CURLUcode hostname_check(char *hostname, unsigned int flags) +{ + const char *l; /* accepted characters */ + size_t len; + size_t hlen = strlen(hostname); + (void)flags; + + if(hostname[0] == '[') { + hostname++; + l = "0123456789abcdefABCDEF::."; + hlen -= 2; + } + else /* % for URL escaped letters */ + l = "0123456789abcdefghijklimnopqrstuvwxyz-_.ABCDEFGHIJKLIMNOPQRSTUVWXYZ%"; + + len = strspn(hostname, l); + if(hlen != len) + /* hostname with bad content */ + return CURLUE_MALFORMED_INPUT; + + return CURLUE_OK; +} + +#define HOSTNAME_END(x) (((x) == '/') || ((x) == '?') || ((x) == '#')) + +static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags) +{ + char *path; + bool path_alloced = FALSE; + char *hostname; + char *query = NULL; + char *fragment = NULL; + CURLUcode result; + bool url_has_scheme = FALSE; + char schemebuf[MAX_SCHEME_LEN]; + char *schemep; + size_t schemelen = 0; + size_t urllen; + const struct Curl_handler *h = NULL; + + if(!url) + return CURLUE_MALFORMED_INPUT; + + /************************************************************* + * Parse the URL. + ************************************************************/ + /* allocate scratch area */ + urllen = strlen(url); + path = u->scratch = malloc(urllen * 2 + 2); + if(!path) + return CURLUE_OUT_OF_MEMORY; + + hostname = &path[urllen + 1]; + hostname[0] = 0; + + /* MSDOS/Windows style drive prefix, eg c: in c:foo */ +#define STARTS_WITH_DRIVE_PREFIX(str) \ + ((('a' <= str[0] && str[0] <= 'z') || \ + ('A' <= str[0] && str[0] <= 'Z')) && \ + (str[1] == ':')) + + /* MSDOS/Windows style drive prefix, optionally with + * a '|' instead of ':', followed by a slash or NUL */ +#define STARTS_WITH_URL_DRIVE_PREFIX(str) \ + ((('a' <= (str)[0] && (str)[0] <= 'z') || \ + ('A' <= (str)[0] && (str)[0] <= 'Z')) && \ + ((str)[1] == ':' || (str)[1] == '|') && \ + ((str)[2] == '/' || (str)[2] == '\\' || (str)[2] == 0)) + + if(Curl_is_absolute_url(url, schemebuf, sizeof(schemebuf))) { + url_has_scheme = TRUE; + schemelen = strlen(schemebuf); + } + + /* handle the file: scheme */ + if(url_has_scheme && strcasecompare(schemebuf, "file")) { + /* path has been allocated large anough to hold this */ + strcpy(path, &url[5]); + + hostname = NULL; /* no host for file: URLs */ + u->scheme = strdup("file"); + if(!u->scheme) + return CURLUE_OUT_OF_MEMORY; + + /* Extra handling URLs with an authority component (i.e. that start with + * "file://") + * + * We allow omitted hostname (e.g. file:/) -- valid according to + * RFC 8089, but not the (current) WHAT-WG URL spec. + */ + if(path[0] == '/' && path[1] == '/') { + /* swallow the two slashes */ + char *ptr = &path[2]; + + /* + * According to RFC 8089, a file: URL can be reliably dereferenced if: + * + * o it has no/blank hostname, or + * + * o the hostname matches "localhost" (case-insensitively), or + * + * o the hostname is a FQDN that resolves to this machine. + * + * For brevity, we only consider URLs with empty, "localhost", or + * "127.0.0.1" hostnames as local. + * + * Additionally, there is an exception for URLs with a Windows drive + * letter in the authority (which was accidentally omitted from RFC 8089 + * Appendix E, but believe me, it was meant to be there. --MK) + */ + if(ptr[0] != '/' && !STARTS_WITH_URL_DRIVE_PREFIX(ptr)) { + /* the URL includes a host name, it must match "localhost" or + "127.0.0.1" to be valid */ + if(!checkprefix("localhost/", ptr) && + !checkprefix("127.0.0.1/", ptr)) { + /* Invalid file://hostname/, expected localhost or 127.0.0.1 or + none */ + return CURLUE_MALFORMED_INPUT; + } + ptr += 9; /* now points to the slash after the host */ + } + + path = ptr; + } + +#if !defined(MSDOS) && !defined(WIN32) && !defined(__CYGWIN__) + /* Don't allow Windows drive letters when not in Windows. + * This catches both "file:/c:" and "file:c:" */ + if(('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) || + STARTS_WITH_URL_DRIVE_PREFIX(path)) { + /* File drive letters are only accepted in MSDOS/Windows */ + return CURLUE_MALFORMED_INPUT; + } +#else + /* If the path starts with a slash and a drive letter, ditch the slash */ + if('/' == path[0] && STARTS_WITH_URL_DRIVE_PREFIX(&path[1])) { + /* This cannot be done with strcpy, as the memory chunks overlap! */ + memmove(path, &path[1], strlen(&path[1]) + 1); + } +#endif + + } + else { + /* clear path */ + const char *p; + const char *hostp; + size_t len; + path[0] = 0; + + if(url_has_scheme) { + int i = 0; + p = &url[schemelen + 1]; + while(p && (*p == '/') && (i < 4)) { + p++; + i++; + } + if((i < 1) || (i>3)) + /* less than one or more than three slashes */ + return CURLUE_MALFORMED_INPUT; + + schemep = schemebuf; + if(!Curl_builtin_scheme(schemep) && + !(flags & CURLU_NON_SUPPORT_SCHEME)) + return CURLUE_UNSUPPORTED_SCHEME; + + if(junkscan(schemep)) + return CURLUE_MALFORMED_INPUT; + } + else { + /* no scheme! */ + + if(!(flags & CURLU_DEFAULT_SCHEME)) + return CURLUE_MALFORMED_INPUT; + schemep = (char *) DEFAULT_SCHEME; + + /* + * The URL was badly formatted, let's try without scheme specified. + */ + p = url; + } + hostp = p; /* host name starts here */ + + while(*p && !HOSTNAME_END(*p)) /* find end of host name */ + p++; + + len = p - hostp; + if(!len) + return CURLUE_MALFORMED_INPUT; + + memcpy(hostname, hostp, len); + hostname[len] = 0; + + len = strlen(p); + memcpy(path, p, len); + path[len] = 0; + + u->scheme = strdup(schemep); + if(!u->scheme) + return CURLUE_OUT_OF_MEMORY; + } + + /* if this is a known scheme, get some details */ + h = Curl_builtin_scheme(u->scheme); + + if(junkscan(path)) + return CURLUE_MALFORMED_INPUT; + + query = strchr(path, '?'); + if(query) + *query++ = 0; + + fragment = strchr(query?query:path, '#'); + if(fragment) + *fragment++ = 0; + + if(!path[0]) + /* if there's no path set, unset */ + path = NULL; + else if(!(flags & CURLU_PATH_AS_IS)) { + /* sanitise paths and remove ../ and ./ sequences according to RFC3986 */ + char *newp = Curl_dedotdotify(path); + if(!newp) + return CURLUE_OUT_OF_MEMORY; + + if(strcmp(newp, path)) { + /* if we got a new version */ + path = newp; + path_alloced = TRUE; + } + else + free(newp); + } + if(path) { + u->path = path_alloced?path:strdup(path); + if(!u->path) + return CURLUE_OUT_OF_MEMORY; + } + + if(hostname) { + /* + * Parse the login details and strip them out of the host name. + */ + if(junkscan(hostname)) + return CURLUE_MALFORMED_INPUT; + + result = parse_hostname_login(u, h, &hostname, flags); + if(result) + return result; + + result = parse_port(u, hostname); + if(result) + return result; + + result = hostname_check(hostname, flags); + if(result) + return result; + + u->host = strdup(hostname); + if(!u->host) + return CURLUE_OUT_OF_MEMORY; + } + + if(query && query[0]) { + u->query = strdup(query); + if(!u->query) + return CURLUE_OUT_OF_MEMORY; + } + if(fragment && fragment[0]) { + u->fragment = strdup(fragment); + if(!u->fragment) + return CURLUE_OUT_OF_MEMORY; + } + + free(u->scratch); + u->scratch = NULL; + + return CURLUE_OK; +} + +/* + * Parse the URL and set the relevant members of the Curl_URL struct. + */ +static CURLUcode parseurl(const char *url, CURLU *u, unsigned int flags) +{ + CURLUcode result = seturl(url, u, flags); + if(result) { + free_urlhandle(u); + memset(u, 0, sizeof(struct Curl_URL)); + } + return result; +} + +/* + */ +CURLU *curl_url(void) +{ + return calloc(sizeof(struct Curl_URL), 1); +} + +void curl_url_cleanup(CURLU *u) +{ + if(u) { + free_urlhandle(u); + free(u); + } +} + +#define DUP(dest, src, name) \ + if(src->name) { \ + dest->name = strdup(src->name); \ + if(!dest->name) \ + goto fail; \ + } + +CURLU *curl_url_dup(CURLU *in) +{ + struct Curl_URL *u = calloc(sizeof(struct Curl_URL), 1); + if(u) { + DUP(u, in, scheme); + DUP(u, in, user); + DUP(u, in, password); + DUP(u, in, options); + DUP(u, in, host); + DUP(u, in, port); + DUP(u, in, path); + DUP(u, in, query); + DUP(u, in, fragment); + u->portnum = in->portnum; + } + return u; + fail: + curl_url_cleanup(u); + return NULL; +} + +CURLUcode curl_url_get(CURLU *u, CURLUPart what, + char **part, unsigned int flags) +{ + char *ptr; + CURLUcode ifmissing = CURLUE_UNKNOWN_PART; + char portbuf[7]; + bool urldecode = (flags & CURLU_URLDECODE)?1:0; + bool plusdecode = FALSE; + (void)flags; + if(!u) + return CURLUE_BAD_HANDLE; + if(!part) + return CURLUE_BAD_PARTPOINTER; + *part = NULL; + + switch(what) { + case CURLUPART_SCHEME: + ptr = u->scheme; + ifmissing = CURLUE_NO_SCHEME; + urldecode = FALSE; /* never for schemes */ + break; + case CURLUPART_USER: + ptr = u->user; + ifmissing = CURLUE_NO_USER; + break; + case CURLUPART_PASSWORD: + ptr = u->password; + ifmissing = CURLUE_NO_PASSWORD; + break; + case CURLUPART_OPTIONS: + ptr = u->options; + ifmissing = CURLUE_NO_OPTIONS; + break; + case CURLUPART_HOST: + ptr = u->host; + ifmissing = CURLUE_NO_HOST; + break; + case CURLUPART_PORT: + ptr = u->port; + ifmissing = CURLUE_NO_PORT; + urldecode = FALSE; /* never for port */ + if(!ptr && (flags & CURLU_DEFAULT_PORT) && u->scheme) { + /* there's no stored port number, but asked to deliver + a default one for the scheme */ + const struct Curl_handler *h = + Curl_builtin_scheme(u->scheme); + if(h) { + snprintf(portbuf, sizeof(portbuf), "%ld", h->defport); + ptr = portbuf; + } + } + else if(ptr && u->scheme) { + /* there is a stored port number, but ask to inhibit if + it matches the default one for the scheme */ + const struct Curl_handler *h = + Curl_builtin_scheme(u->scheme); + if(h && (h->defport == u->portnum) && + (flags & CURLU_NO_DEFAULT_PORT)) + ptr = NULL; + } + break; + case CURLUPART_PATH: + ptr = u->path; + if(!ptr) { + ptr = u->path = strdup("/"); + if(!u->path) + return CURLUE_OUT_OF_MEMORY; + } + break; + case CURLUPART_QUERY: + ptr = u->query; + ifmissing = CURLUE_NO_QUERY; + plusdecode = urldecode; + break; + case CURLUPART_FRAGMENT: + ptr = u->fragment; + ifmissing = CURLUE_NO_FRAGMENT; + break; + case CURLUPART_URL: { + char *url; + char *scheme; + char *options = u->options; + char *port = u->port; + urldecode = FALSE; /* not for the whole thing */ + if(u->scheme && strcasecompare("file", u->scheme)) { + url = aprintf("file://%s%s%s", + u->path, + u->fragment? "#": "", + u->fragment? u->fragment : ""); + } + else if(!u->host) + return CURLUE_NO_HOST; + else { + const struct Curl_handler *h = NULL; + if(u->scheme) + scheme = u->scheme; + else if(flags & CURLU_DEFAULT_SCHEME) + scheme = (char *) DEFAULT_SCHEME; + else + return CURLUE_NO_SCHEME; + + if(scheme) { + h = Curl_builtin_scheme(scheme); + if(!port && (flags & CURLU_DEFAULT_PORT)) { + /* there's no stored port number, but asked to deliver + a default one for the scheme */ + if(h) { + snprintf(portbuf, sizeof(portbuf), "%ld", h->defport); + port = portbuf; + } + } + else if(port) { + /* there is a stored port number, but asked to inhibit if it matches + the default one for the scheme */ + if(h && (h->defport == u->portnum) && + (flags & CURLU_NO_DEFAULT_PORT)) + port = NULL; + } + } + if(h && !(h->flags & PROTOPT_URLOPTIONS)) + options = NULL; + + url = aprintf("%s://%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s", + scheme, + u->user ? u->user : "", + u->password ? ":": "", + u->password ? u->password : "", + options ? ";" : "", + options ? options : "", + (u->user || u->password || options) ? "@": "", + u->host, + port ? ":": "", + port ? port : "", + (u->path && (u->path[0] != '/')) ? "/": "", + u->path ? u->path : "/", + u->query? "?": "", + u->query? u->query : "", + u->fragment? "#": "", + u->fragment? u->fragment : ""); + } + if(!url) + return CURLUE_OUT_OF_MEMORY; + *part = url; + return CURLUE_OK; + break; + } + default: + ptr = NULL; + } + if(ptr) { + *part = strdup(ptr); + if(!*part) + return CURLUE_OUT_OF_MEMORY; + if(plusdecode) { + /* convert + to space */ + char *plus; + for(plus = *part; *plus; ++plus) { + if(*plus == '+') + *plus = ' '; + } + } + if(urldecode) { + char *decoded; + size_t dlen; + CURLcode res = Curl_urldecode(NULL, *part, 0, &decoded, &dlen, TRUE); + free(*part); + if(res) { + *part = NULL; + return CURLUE_URLDECODE; + } + *part = decoded; + } + return CURLUE_OK; + } + else + return ifmissing; +} + +CURLUcode curl_url_set(CURLU *u, CURLUPart what, + const char *part, unsigned int flags) +{ + char **storep = NULL; + long port = 0; + bool urlencode = (flags & CURLU_URLENCODE)? 1 : 0; + bool plusencode = FALSE; + bool urlskipslash = FALSE; + bool appendquery = FALSE; + + if(!u) + return CURLUE_BAD_HANDLE; + if(!part) { + /* setting a part to NULL clears it */ + switch(what) { + case CURLUPART_URL: + break; + case CURLUPART_SCHEME: + storep = &u->scheme; + break; + case CURLUPART_USER: + storep = &u->user; + break; + case CURLUPART_PASSWORD: + storep = &u->password; + break; + case CURLUPART_OPTIONS: + storep = &u->options; + break; + case CURLUPART_HOST: + storep = &u->host; + break; + case CURLUPART_PORT: + storep = &u->port; + break; + case CURLUPART_PATH: + storep = &u->path; + break; + case CURLUPART_QUERY: + storep = &u->query; + break; + case CURLUPART_FRAGMENT: + storep = &u->fragment; + break; + default: + return CURLUE_UNKNOWN_PART; + } + if(storep && *storep) { + free(*storep); + *storep = NULL; + } + return CURLUE_OK; + } + + switch(what) { + case CURLUPART_SCHEME: + if(!(flags & CURLU_NON_SUPPORT_SCHEME) && + /* verify that it is a fine scheme */ + !Curl_builtin_scheme(part)) + return CURLUE_UNSUPPORTED_SCHEME; + storep = &u->scheme; + urlencode = FALSE; /* never */ + break; + case CURLUPART_USER: + storep = &u->user; + break; + case CURLUPART_PASSWORD: + storep = &u->password; + break; + case CURLUPART_OPTIONS: + storep = &u->options; + break; + case CURLUPART_HOST: + storep = &u->host; + break; + case CURLUPART_PORT: + urlencode = FALSE; /* never */ + port = strtol(part, NULL, 10); /* Port number must be decimal */ + if((port <= 0) || (port > 0xffff)) + return CURLUE_BAD_PORT_NUMBER; + storep = &u->port; + break; + case CURLUPART_PATH: + urlskipslash = TRUE; + storep = &u->path; + break; + case CURLUPART_QUERY: + plusencode = urlencode; + appendquery = (flags & CURLU_APPENDQUERY)?1:0; + storep = &u->query; + break; + case CURLUPART_FRAGMENT: + storep = &u->fragment; + break; + case CURLUPART_URL: { + /* + * Allow a new URL to replace the existing (if any) contents. + * + * If the existing contents is enough for a URL, allow a relative URL to + * replace it. + */ + CURLUcode result; + char *oldurl; + char *redired_url; + CURLU *handle2; + + if(Curl_is_absolute_url(part, NULL, MAX_SCHEME_LEN)) { + handle2 = curl_url(); + if(!handle2) + return CURLUE_OUT_OF_MEMORY; + result = parseurl(part, handle2, flags); + if(!result) + mv_urlhandle(handle2, u); + else + curl_url_cleanup(handle2); + return result; + } + /* extract the full "old" URL to do the redirect on */ + result = curl_url_get(u, CURLUPART_URL, &oldurl, flags); + if(result) { + /* couldn't get the old URL, just use the new! */ + handle2 = curl_url(); + if(!handle2) + return CURLUE_OUT_OF_MEMORY; + result = parseurl(part, handle2, flags); + if(!result) + mv_urlhandle(handle2, u); + else + curl_url_cleanup(handle2); + return result; + } + + /* apply the relative part to create a new URL */ + redired_url = Curl_concat_url(oldurl, part); + free(oldurl); + if(!redired_url) + return CURLUE_OUT_OF_MEMORY; + + /* now parse the new URL */ + handle2 = curl_url(); + if(!handle2) { + free(redired_url); + return CURLUE_OUT_OF_MEMORY; + } + result = parseurl(redired_url, handle2, flags); + free(redired_url); + if(!result) + mv_urlhandle(handle2, u); + else + curl_url_cleanup(handle2); + return result; + } + default: + return CURLUE_UNKNOWN_PART; + } + if(storep) { + const char *newp = part; + size_t nalloc = strlen(part); + + if(urlencode) { + const char *i; + char *o; + bool free_part = FALSE; + char *enc = malloc(nalloc * 3 + 1); /* for worst case! */ + if(!enc) + return CURLUE_OUT_OF_MEMORY; + if(plusencode) { + /* space to plus */ + i = part; + for(o = enc; *i; ++o, ++i) + *o = (*i == ' ') ? '+' : *i; + *o = 0; /* zero terminate */ + part = strdup(enc); + if(!part) { + free(enc); + return CURLUE_OUT_OF_MEMORY; + } + free_part = TRUE; + } + for(i = part, o = enc; *i; i++) { + if(Curl_isunreserved(*i) || + ((*i == '/') && urlskipslash) || + ((*i == '=') && appendquery) || + ((*i == '+') && plusencode)) { + *o = *i; + o++; + } + else { + snprintf(o, 4, "%%%02x", *i); + o += 3; + } + } + *o = 0; /* zero terminate */ + newp = enc; + if(free_part) + free((char *)part); + } + else { + char *p; + newp = strdup(part); + if(!newp) + return CURLUE_OUT_OF_MEMORY; + p = (char *)newp; + while(*p) { + /* make sure percent encoded are lower case */ + if((*p == '%') && ISXDIGIT(p[1]) && ISXDIGIT(p[2]) && + (ISUPPER(p[1]) || ISUPPER(p[2]))) { + p[1] = (char)TOLOWER(p[1]); + p[2] = (char)TOLOWER(p[2]); + p += 3; + } + else + p++; + } + } + + if(appendquery) { + /* Append the string onto the old query. Add a '&' separator if none is + present at the end of the exsting query already */ + size_t querylen = u->query ? strlen(u->query) : 0; + bool addamperand = querylen && (u->query[querylen -1] != '&'); + if(querylen) { + size_t newplen = strlen(newp); + char *p = malloc(querylen + addamperand + newplen + 1); + if(!p) { + free((char *)newp); + return CURLUE_OUT_OF_MEMORY; + } + strcpy(p, u->query); /* original query */ + if(addamperand) + p[querylen] = '&'; /* ampersand */ + strcpy(&p[querylen + addamperand], newp); /* new suffix */ + free((char *)newp); + free(*storep); + *storep = p; + return CURLUE_OK; + } + } + + free(*storep); + *storep = (char *)newp; + } + /* set after the string, to make it not assigned if the allocation above + fails */ + if(port) + u->portnum = port; + return CURLUE_OK; +} diff --git a/tests/data/Makefile.inc b/tests/data/Makefile.inc index aa5fff0..0f6ac44 100644 --- a/tests/data/Makefile.inc +++ b/tests/data/Makefile.inc @@ -178,6 +178,8 @@ test1533 test1534 test1535 test1536 test1537 test1538 \ test1540 \ test1550 test1551 test1552 test1553 test1554 test1555 test1556 test1557 \ \ +test1560 \ +\ test1590 \ test1600 test1601 test1602 test1603 test1604 test1605 test1606 test1607 \ test1608 test1609 \ diff --git a/tests/data/test1560 b/tests/data/test1560 new file mode 100644 index 0000000..720df03 --- /dev/null +++ b/tests/data/test1560 @@ -0,0 +1,28 @@ + + + +unittest +URL API + + + +# +# Client-side + + +none + + +file +https +http + + +URL API + + +lib1560 + + + + diff --git a/tests/libtest/Makefile.am b/tests/libtest/Makefile.am index d14f37d..dc97e32 100644 --- a/tests/libtest/Makefile.am +++ b/tests/libtest/Makefile.am @@ -133,3 +133,8 @@ lib1521.c: $(top_srcdir)/tests/libtest/mk-lib1521.pl $(top_srcdir)/include/curl/ checksrc: @PERL@ $(top_srcdir)/lib/checksrc.pl $(srcdir)/*.c + +if CURLDEBUG +# for debug builds, we scan the sources on all regular make invokes +all-local: checksrc +endif diff --git a/tests/libtest/Makefile.inc b/tests/libtest/Makefile.inc index 238ef97..7a3cd16 100644 --- a/tests/libtest/Makefile.inc +++ b/tests/libtest/Makefile.inc @@ -30,6 +30,7 @@ noinst_PROGRAMS = chkhostname libauthretry libntlmconnect \ lib1534 lib1535 lib1536 lib1537 lib1538 \ lib1540 \ lib1550 lib1551 lib1552 lib1553 lib1554 lib1555 lib1556 lib1557 \ + lib1560 \ lib1900 \ lib2033 @@ -507,6 +508,9 @@ lib1557_SOURCES = lib1557.c $(SUPPORTFILES) $(TESTUTIL) $(WARNLESS) lib1557_LDADD = $(TESTUTIL_LIBS) lib1557_CPPFLAGS = $(AM_CPPFLAGS) -DLIB1557 +lib1560_SOURCES = lib1560.c $(SUPPORTFILES) $(TESTUTIL) $(WARNLESS) +lib1560_LDADD = $(TESTUTIL_LIBS) + lib1900_SOURCES = lib1900.c $(SUPPORTFILES) $(TESTUTIL) $(WARNLESS) lib1900_LDADD = $(TESTUTIL_LIBS) lib1900_CPPFLAGS = $(AM_CPPFLAGS) diff --git a/tests/libtest/lib1560.c b/tests/libtest/lib1560.c new file mode 100644 index 0000000..669ea9a --- /dev/null +++ b/tests/libtest/lib1560.c @@ -0,0 +1,760 @@ +/*************************************************************************** + * _ _ ____ _ + * Project ___| | | | _ \| | + * / __| | | | |_) | | + * | (__| |_| | _ <| |___ + * \___|\___/|_| \_\_____| + * + * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. + * + * This software is licensed as described in the file COPYING, which + * you should have received as part of this distribution. The terms + * are also available at https://curl.haxx.se/docs/copyright.html. + * + * You may opt to use, copy, modify, merge, publish, distribute and/or sell + * copies of the Software, and permit persons to whom the Software is + * furnished to do so, under the terms of the COPYING file. + * + * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY + * KIND, either express or implied. + * + ***************************************************************************/ + +/* + * Note: + * + * Since the URL parser by default only accepts schemes that *this instance* + * of libcurl supports, make sure that the test1560 file lists all the schemes + * that this test will assume to be present! + */ + +#include "test.h" + +#include "testutil.h" +#include "warnless.h" +#include "memdebug.h" /* LAST include file */ + +struct part { + CURLUPart part; + const char *name; +}; + + +static int checkparts(CURLU *u, const char *in, const char *wanted, + unsigned int getflags) +{ + int i; + CURLUcode rc; + char buf[256]; + char *bufp = &buf[0]; + size_t len = sizeof(buf); + struct part parts[] = { + {CURLUPART_SCHEME, "scheme"}, + {CURLUPART_USER, "user"}, + {CURLUPART_PASSWORD, "password"}, + {CURLUPART_OPTIONS, "options"}, + {CURLUPART_HOST, "host"}, + {CURLUPART_PORT, "port"}, + {CURLUPART_PATH, "path"}, + {CURLUPART_QUERY, "query"}, + {CURLUPART_FRAGMENT, "fragment"}, + {0, NULL} + }; + buf[0] = 0; + + for(i = 0; parts[i].name; i++) { + char *p = NULL; + size_t n; + rc = curl_url_get(u, parts[i].part, &p, getflags); + if(!rc && p) { + snprintf(bufp, len, "%s%s", buf[0]?" | ":"", p); + } + else + snprintf(bufp, len, "%s[%d]", buf[0]?" | ":"", (int)rc); + + n = strlen(bufp); + bufp += n; + len -= n; + curl_free(p); + } + if(strcmp(buf, wanted)) { + fprintf(stderr, "in: %s\nwanted: %s\ngot: %s\n", in, wanted, buf); + return 1; + } + return 0; +} + +struct redircase { + const char *in; + const char *set; + const char *out; + unsigned int urlflags; + unsigned int setflags; + CURLUcode ucode; +}; + +struct setcase { + const char *in; + const char *set; + const char *out; + unsigned int urlflags; + unsigned int setflags; + CURLUcode ucode; +}; + +struct testcase { + const char *in; + const char *out; + unsigned int urlflags; + unsigned int getflags; + CURLUcode ucode; +}; + +struct urltestcase { + const char *in; + const char *out; + unsigned int urlflags; /* pass to curl_url() */ + unsigned int getflags; /* pass to curl_url_get() */ + CURLUcode ucode; +}; + +struct querycase { + const char *in; + const char *q; + const char *out; + unsigned int urlflags; /* pass to curl_url() */ + unsigned int qflags; /* pass to curl_url_get() */ + CURLUcode ucode; +}; + +static struct testcase get_parts_list[] ={ + {"https://127.0.0.1:443", + "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [17] | [18]", + 0, CURLU_NO_DEFAULT_PORT, CURLUE_OK}, + {"http://%3a:%3a@ex%0ample/%3f+?+%3f+%23#+%23%3f%g7", + "http | : | : | [13] | [6] | [15] | /?+ | ? # | +#?%g7", + 0, CURLU_URLDECODE, CURLUE_OK}, + {"http://%3a:%3a@ex%0ample/%3f?%3f%35#%35%3f%g7", + "http | %3a | %3a | [13] | ex%0ample | [15] | /%3f | %3f%35 | %35%3f%g7", + 0, 0, CURLUE_OK}, + {"http://HO0_-st%41/", + "http | [11] | [12] | [13] | HO0_-st%41 | [15] | / | [17] | [18]", + 0, 0, CURLUE_OK}, + {"file://hello.html", + "", + 0, 0, CURLUE_MALFORMED_INPUT}, + {"http://HO0_-st/", + "http | [11] | [12] | [13] | HO0_-st | [15] | / | [17] | [18]", + 0, 0, CURLUE_OK}, + {"imap://user:pass;option@server/path", + "imap | user | pass | option | server | [15] | /path | [17] | [18]", + 0, 0, CURLUE_OK}, + {"http://user:pass;option@server/path", + "http | user | pass;option | [13] | server | [15] | /path | [17] | [18]", + 0, 0, CURLUE_OK}, + {"file:/hello.html", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + 0, 0, CURLUE_OK}, + {"file://127.0.0.1/hello.html", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + 0, 0, CURLUE_OK}, + {"file:////hello.html", + "file | [11] | [12] | [13] | [14] | [15] | //hello.html | [17] | [18]", + 0, 0, CURLUE_OK}, + {"file:///hello.html", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + 0, 0, CURLUE_OK}, + {"https://127.0.0.1", + "https | [11] | [12] | [13] | 127.0.0.1 | 443 | / | [17] | [18]", + 0, CURLU_DEFAULT_PORT, CURLUE_OK}, + {"https://127.0.0.1", + "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [17] | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https://[::1]:1234", + "https | [11] | [12] | [13] | [::1] | 1234 | / | [17] | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https://127abc.com", + "https | [11] | [12] | [13] | 127abc.com | [15] | / | [17] | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https:// example.com?check", + "", + CURLU_DEFAULT_SCHEME, 0, CURLUE_MALFORMED_INPUT}, + {"https://e x a m p l e.com?check", + "", + CURLU_DEFAULT_SCHEME, 0, CURLUE_MALFORMED_INPUT}, + {"https://example.com?check", + "https | [11] | [12] | [13] | example.com | [15] | / | check | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https://example.com:65536", + "", + CURLU_DEFAULT_SCHEME, 0, CURLUE_BAD_PORT_NUMBER}, + {"https://example.com:0#moo", + "", + CURLU_DEFAULT_SCHEME, 0, CURLUE_BAD_PORT_NUMBER}, + {"https://example.com:01#moo", + "https | [11] | [12] | [13] | example.com | 1 | / | " + "[17] | moo", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https://example.com:1#moo", + "https | [11] | [12] | [13] | example.com | 1 | / | " + "[17] | moo", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com#moo", + "http | [11] | [12] | [13] | example.com | [15] | / | " + "[17] | moo", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com", + "http | [11] | [12] | [13] | example.com | [15] | / | " + "[17] | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com/path/html", + "http | [11] | [12] | [13] | example.com | [15] | /path/html | " + "[17] | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com/path/html?query=name", + "http | [11] | [12] | [13] | example.com | [15] | /path/html | " + "query=name | [18]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com/path/html?query=name#anchor", + "http | [11] | [12] | [13] | example.com | [15] | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://example.com:1234/path/html?query=name#anchor", + "http | [11] | [12] | [13] | example.com | 1234 | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http:///user:password@example.com:1234/path/html?query=name#anchor", + "http | user | password | [13] | example.com | 1234 | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"https://user:password@example.com:1234/path/html?query=name#anchor", + "https | user | password | [13] | example.com | 1234 | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://user:password@example.com:1234/path/html?query=name#anchor", + "http | user | password | [13] | example.com | 1234 | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http:/user:password@example.com:1234/path/html?query=name#anchor", + "http | user | password | [13] | example.com | 1234 | /path/html | " + "query=name | anchor", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http:////user:password@example.com:1234/path/html?query=name#anchor", + "", + CURLU_DEFAULT_SCHEME, 0, CURLUE_MALFORMED_INPUT}, + {NULL, NULL, 0, 0, CURLUE_OK}, +}; + +static struct urltestcase get_url_list[] = { + {"HTTP://test/", "http://test/", 0, 0, CURLUE_OK}, + {"http://HO0_-st..~./", "", 0, 0, CURLUE_MALFORMED_INPUT}, + {"http:/@example.com: 123/", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, + {"http:/@example.com:123 /", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, + {"http:/@example.com:123a/", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, + {"http://host/file\r", "", 0, 0, CURLUE_MALFORMED_INPUT}, + {"http://host/file\n\x03", "", 0, 0, CURLUE_MALFORMED_INPUT}, + {"htt\x02://host/file", "", + CURLU_NON_SUPPORT_SCHEME, 0, CURLUE_MALFORMED_INPUT}, + {" http://host/file", "", 0, 0, CURLUE_MALFORMED_INPUT}, + /* here the password ends at the semicolon and options is 'word' */ + {"imap://user:pass;word@host/file", + "imap://user:pass;word@host/file", + 0, 0, CURLUE_OK}, + /* here the password has the semicolon */ + {"http://user:pass;word@host/file", + "http://user:pass;word@host/file", + 0, 0, CURLUE_OK}, + {"file:///file.txt#moo", + "file:///file.txt#moo", + 0, 0, CURLUE_OK}, + {"file:////file.txt", + "file:////file.txt", + 0, 0, CURLUE_OK}, + {"file:///file.txt", + "file:///file.txt", + 0, 0, CURLUE_OK}, + {"http://example.com/hello/../here", + "http://example.com/hello/../here", + CURLU_PATH_AS_IS, 0, CURLUE_OK}, + {"http://example.com/hello/../here", + "http://example.com/here", + 0, 0, CURLUE_OK}, + {"http://example.com:80", + "http://example.com/", + 0, CURLU_NO_DEFAULT_PORT, CURLUE_OK}, + {"tp://example.com/path/html", + "", + 0, 0, CURLUE_UNSUPPORTED_SCHEME}, + {"http://hello:fool@example.com", + "", + CURLU_DISALLOW_USER, 0, CURLUE_USER_NOT_ALLOWED}, + {"http:/@example.com:123", + "http://example.com:123/", + 0, 0, CURLUE_OK}, + {"http:/:password@example.com", + "http://:password@example.com/", + 0, 0, CURLUE_OK}, + {"http://user@example.com?#", + "http://user@example.com/", + 0, 0, CURLUE_OK}, + {"http://user@example.com?", + "http://user@example.com/", + 0, 0, CURLUE_OK}, + {"http://user@example.com#anchor", + "http://user@example.com/#anchor", + 0, 0, CURLUE_OK}, + {"example.com/path/html", + "https://example.com/path/html", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"example.com/path/html", + "", + 0, 0, CURLUE_MALFORMED_INPUT}, + {"http://user:password@example.com:1234/path/html?query=name#anchor", + "http://user:password@example.com:1234/path/html?query=name#anchor", + 0, 0, CURLUE_OK}, + {"http://example.com:1234/path/html?query=name#anchor", + "http://example.com:1234/path/html?query=name#anchor", + 0, 0, CURLUE_OK}, + {"http://example.com/path/html?query=name#anchor", + "http://example.com/path/html?query=name#anchor", + 0, 0, CURLUE_OK}, + {"http://example.com/path/html?query=name", + "http://example.com/path/html?query=name", + 0, 0, CURLUE_OK}, + {"http://example.com/path/html", + "http://example.com/path/html", + 0, 0, CURLUE_OK}, + {"tp://example.com/path/html", + "tp://example.com/path/html", + CURLU_NON_SUPPORT_SCHEME, 0, CURLUE_OK}, + {NULL, NULL, 0, 0, 0} +}; + +static int checkurl(const char *url, const char *out) +{ + if(strcmp(out, url)) { + fprintf(stderr, "Wanted: %s\nGot : %s\n", + out, url); + return 1; + } + return 0; +} + +/* !checksrc! disable SPACEBEFORECOMMA 1 */ +static struct setcase set_parts_list[] = { + {"https://host/", + "path=%4A%4B%4C,", + "https://host/%4a%4b%4c", + 0, 0, CURLUE_NO_HOST}, + {"https://host/mooo?q#f", + "path=NULL,query=NULL,fragment=NULL,", + "https://host/", + 0, 0, CURLUE_NO_HOST}, + {"https://user:secret@host/", + "user=NULL,password=NULL,", + "https://host/", + 0, 0, CURLUE_NO_HOST}, + {NULL, + "scheme=https,user= @:,host=foobar,", + "https://%20%20%20%40%3a@foobar/", + 0, CURLU_URLENCODE, CURLUE_OK}, + {NULL, + "scheme=https,host= ,path= ,user= ,password= ,query= ,fragment= ,", + "https://%20:%20@%20%20/%20?+#%20", + 0, CURLU_URLENCODE, CURLUE_OK}, + {NULL, + "scheme=https,host=foobar,path=/this /path /is /here,", + "https://foobar/this%20/path%20/is%20/here", + 0, CURLU_URLENCODE, CURLUE_OK}, + {"imap://user:secret;opt@host/", + "options=updated,scheme=imaps,password=p4ssw0rd,", + "imaps://user:p4ssw0rd;updated@host/", + 0, 0, CURLUE_NO_HOST}, + {"imap://user:secret;optit@host/", + "scheme=https,", + "https://user:secret@host/", + 0, 0, CURLUE_NO_HOST}, + {"file:///file#anchor", + "scheme=https,host=example,", + "https://example/file#anchor", + 0, 0, CURLUE_NO_HOST}, + {NULL, /* start fresh! */ + "scheme=file,host=127.0.0.1,path=/no,user=anonymous,", + "file:///no", + 0, 0, CURLUE_OK}, + {NULL, /* start fresh! */ + "scheme=ftp,host=127.0.0.1,path=/no,user=anonymous,", + "ftp://anonymous@127.0.0.1/no", + 0, 0, CURLUE_OK}, + {NULL, /* start fresh! */ + "scheme=https,host=example.com,", + "https://example.com/", + 0, CURLU_NON_SUPPORT_SCHEME, CURLUE_OK}, + {"http://user:foo@example.com/path?query#frag", + "fragment=changed,", + "http://user:foo@example.com/path?query#changed", + 0, CURLU_NON_SUPPORT_SCHEME, CURLUE_OK}, + {"http://example.com/", + "scheme=foo,", /* not accepted */ + "http://example.com/", + 0, 0, CURLUE_OK}, + {"http://example.com/", + "scheme=https,path=/hello,fragment=snippet,", + "https://example.com/hello#snippet", + 0, 0, CURLUE_OK}, + {"http://example.com:80", + "user=foo,port=1922,", + "http://foo@example.com:1922/", + 0, 0, CURLUE_OK}, + {"http://example.com:80", + "user=foo,password=bar,", + "http://foo:bar@example.com:80/", + 0, 0, CURLUE_OK}, + {"http://example.com:80", + "user=foo,", + "http://foo@example.com:80/", + 0, 0, CURLUE_OK}, + {"http://example.com", + "host=www.example.com,", + "http://www.example.com/", + 0, 0, CURLUE_OK}, + {"http://example.com:80", + "scheme=ftp,", + "ftp://example.com:80/", + 0, 0, CURLUE_OK}, + {NULL, NULL, NULL, 0, 0, 0} +}; + +static CURLUPart part2id(char *part) +{ + if(!strcmp("url", part)) + return CURLUPART_URL; + if(!strcmp("scheme", part)) + return CURLUPART_SCHEME; + if(!strcmp("user", part)) + return CURLUPART_USER; + if(!strcmp("password", part)) + return CURLUPART_PASSWORD; + if(!strcmp("options", part)) + return CURLUPART_OPTIONS; + if(!strcmp("host", part)) + return CURLUPART_HOST; + if(!strcmp("port", part)) + return CURLUPART_PORT; + if(!strcmp("path", part)) + return CURLUPART_PATH; + if(!strcmp("query", part)) + return CURLUPART_QUERY; + if(!strcmp("fragment", part)) + return CURLUPART_FRAGMENT; + return 9999; /* bad input => bad output */ +} + +static void updateurl(CURLU *u, const char *cmd, unsigned int setflags) +{ + const char *p = cmd; + + /* make sure the last command ends with a comma too! */ + while(p) { + char *e = strchr(p, ','); + if(e) { + size_t n = e-p; + char buf[80]; + char part[80]; + char value[80]; + memcpy(buf, p, n); + buf[n] = 0; + if(2 == sscanf(buf, "%79[^=]=%79[^,]", part, value)) { + CURLUPart what = part2id(part); +#if 0 + /* for debugging this */ + fprintf(stderr, "%s = %s [%d]\n", part, value, (int)what); +#endif + if(!strcmp("NULL", value)) + curl_url_set(u, what, NULL, setflags); + else + curl_url_set(u, what, value, setflags); + } + p = e + 1; + continue; + } + break; + } + +} + +static struct redircase set_url_list[] = { + {"file://localhost/path?query#frag", + "foo#another", + "file:///foo#another", + 0, 0, 0}, + {"http://example.com/path?query#frag", + "https://two.example.com/bradnew", + "https://two.example.com/bradnew", + 0, 0, 0}, + {"http://example.com/path?query#frag", + "../../newpage#foo", + "http://example.com/newpage#foo", + 0, 0, 0}, + {"http://user:foo@example.com/path?query#frag", + "../../newpage", + "http://user:foo@example.com/newpage", + 0, 0, 0}, + {"http://user:foo@example.com/path?query#frag", + "../newpage", + "http://user:foo@example.com/newpage", + 0, 0, 0}, + {NULL, NULL, NULL, 0, 0, 0} +}; + +static int set_url(void) +{ + int i; + CURLUcode rc; + CURLU *urlp; + int error = 0; + + for(i = 0; set_url_list[i].in && !error; i++) { + char *url = NULL; + urlp = curl_url(); + if(!urlp) + break; + rc = curl_url_set(urlp, CURLUPART_URL, set_url_list[i].in, + set_url_list[i].urlflags); + if(!rc) { + rc = curl_url_set(urlp, CURLUPART_URL, set_url_list[i].set, + set_url_list[i].setflags); + if(rc) { + fprintf(stderr, "%s:%d Set URL %s returned %d\n", + __FILE__, __LINE__, set_url_list[i].set, + (int)rc); + error++; + } + else { + rc = curl_url_get(urlp, CURLUPART_URL, &url, 0); + if(rc) { + fprintf(stderr, "%s:%d Get URL returned %d\n", + __FILE__, __LINE__, (int)rc); + error++; + } + else { + if(checkurl(url, set_url_list[i].out)) { + error++; + } + } + } + curl_free(url); + } + else if(rc != set_url_list[i].ucode) { + fprintf(stderr, "Set URL\nin: %s\nreturned %d (expected %d)\n", + set_url_list[i].in, (int)rc, set_url_list[i].ucode); + error++; + } + curl_url_cleanup(urlp); + } + return error; +} + +static int set_parts(void) +{ + int i; + CURLUcode rc; + int error = 0; + + for(i = 0; set_parts_list[i].set && !error; i++) { + char *url = NULL; + CURLU *urlp = curl_url(); + if(!urlp) { + error++; + break; + } + if(set_parts_list[i].in) + rc = curl_url_set(urlp, CURLUPART_URL, set_parts_list[i].in, + set_parts_list[i].urlflags); + else + rc = CURLUE_OK; + if(!rc) { + updateurl(urlp, set_parts_list[i].set, set_parts_list[i].setflags); + rc = curl_url_get(urlp, CURLUPART_URL, &url, 0); + + if(rc) { + fprintf(stderr, "%s:%d Get URL returned %d\n", + __FILE__, __LINE__, (int)rc); + error++; + } + else if(checkurl(url, set_parts_list[i].out)) { + error++; + } + } + else if(rc != set_parts_list[i].ucode) { + fprintf(stderr, "Set parts\nin: %s\nreturned %d (expected %d)\n", + set_parts_list[i].in, (int)rc, set_parts_list[i].ucode); + error++; + } + curl_free(url); + curl_url_cleanup(urlp); + } + return error; +} + +static int get_url(void) +{ + int i; + CURLUcode rc; + int error = 0; + for(i = 0; get_url_list[i].in && !error; i++) { + char *url = NULL; + CURLU *urlp = curl_url(); + if(!urlp) { + error++; + break; + } + rc = curl_url_set(urlp, CURLUPART_URL, get_url_list[i].in, + get_url_list[i].urlflags); + if(!rc) { + rc = curl_url_get(urlp, CURLUPART_URL, &url, get_url_list[i].getflags); + + if(rc) { + fprintf(stderr, "%s:%d returned %d\n", + __FILE__, __LINE__, (int)rc); + error++; + } + else { + if(checkurl(url, get_url_list[i].out)) { + error++; + } + } + } + else if(rc != get_url_list[i].ucode) { + fprintf(stderr, "Get URL\nin: %s\nreturned %d (expected %d)\n", + get_url_list[i].in, (int)rc, get_url_list[i].ucode); + error++; + } + curl_free(url); + curl_url_cleanup(urlp); + } + return error; +} + +static int get_parts(void) +{ + int i; + CURLUcode rc; + CURLU *urlp; + int error = 0; + for(i = 0; get_parts_list[i].in && !error; i++) { + urlp = curl_url(); + if(!urlp) { + error++; + break; + } + rc = curl_url_set(urlp, CURLUPART_URL, + get_parts_list[i].in, + get_parts_list[i].urlflags); + if(rc != get_parts_list[i].ucode) { + fprintf(stderr, "Get parts\nin: %s\nreturned %d (expected %d)\n", + get_parts_list[i].in, (int)rc, get_parts_list[i].ucode); + error++; + } + else if(get_parts_list[i].ucode) { + /* the expected error happened */ + } + else if(checkparts(urlp, get_parts_list[i].in, get_parts_list[i].out, + get_parts_list[i].getflags)) + error++; + curl_url_cleanup(urlp); + } + return error; +} + +static struct querycase append_list[] = { + {"HTTP://test/?s", "name=joe\x02", "http://test/?s&name=joe%02", + 0, CURLU_URLENCODE, CURLUE_OK}, + {"HTTP://test/?size=2#f", "name=joe=", "http://test/?size=2&name=joe=#f", + 0, CURLU_URLENCODE, CURLUE_OK}, + {"HTTP://test/?size=2#f", "name=joe doe", + "http://test/?size=2&name=joe+doe#f", + 0, CURLU_URLENCODE, CURLUE_OK}, + {"HTTP://test/", "name=joe", "http://test/?name=joe", 0, 0, CURLUE_OK}, + {"HTTP://test/?size=2", "name=joe", "http://test/?size=2&name=joe", + 0, 0, CURLUE_OK}, + {"HTTP://test/?size=2&", "name=joe", "http://test/?size=2&name=joe", + 0, 0, CURLUE_OK}, + {"HTTP://test/?size=2#f", "name=joe", "http://test/?size=2&name=joe#f", + 0, 0, CURLUE_OK}, + {NULL, NULL, NULL, 0, 0, 0} +}; + +static int append(void) +{ + int i; + CURLUcode rc; + CURLU *urlp; + int error = 0; + for(i = 0; append_list[i].in && !error; i++) { + urlp = curl_url(); + if(!urlp) { + error++; + break; + } + rc = curl_url_set(urlp, CURLUPART_URL, + append_list[i].in, + append_list[i].urlflags); + if(rc) + error++; + else + rc = curl_url_set(urlp, CURLUPART_QUERY, + append_list[i].q, + append_list[i].qflags | CURLU_APPENDQUERY); + if(error) + ; + else if(rc != append_list[i].ucode) { + fprintf(stderr, "Append\nin: %s\nreturned %d (expected %d)\n", + append_list[i].in, (int)rc, append_list[i].ucode); + error++; + } + else if(append_list[i].ucode) { + /* the expected error happened */ + } + else { + char *url; + rc = curl_url_get(urlp, CURLUPART_URL, &url, 0); + if(rc) { + fprintf(stderr, "%s:%d Get URL returned %d\n", + __FILE__, __LINE__, (int)rc); + error++; + } + else { + if(checkurl(url, append_list[i].out)) { + error++; + } + curl_free(url); + } + } + curl_url_cleanup(urlp); + } + return error; +} + +int test(char *URL) +{ + (void)URL; /* not used */ + + if(append()) + return 5; + + if(set_url()) + return 1; + + if(set_parts()) + return 2; + + if(get_url()) + return 3; + + if(get_parts()) + return 4; + + printf("success\n"); + return 0; +} -- 2.17.2 From 581a3b902b949f090776c5295a8aa0786edba773 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Sat, 8 Sep 2018 16:02:25 +0200 Subject: [PATCH 02/14] curl_url-docs: fix AVAILABILITY as Added in curl 7.62.0 Upstream-commit: 890eea5aade0fc4ee167e83948d53351c11dd1ae Signed-off-by: Kamil Dudka --- docs/libcurl/curl_url.3 | 2 +- docs/libcurl/curl_url_cleanup.3 | 2 +- docs/libcurl/curl_url_dup.3 | 2 +- docs/libcurl/curl_url_get.3 | 2 +- docs/libcurl/curl_url_set.3 | 2 +- 5 files changed, 5 insertions(+), 5 deletions(-) diff --git a/docs/libcurl/curl_url.3 b/docs/libcurl/curl_url.3 index 0a56264..a14c45b 100644 --- a/docs/libcurl/curl_url.3 +++ b/docs/libcurl/curl_url.3 @@ -55,7 +55,7 @@ Returns a \fBCURLU *\fP if successful, or NULL if out of memory. } .fi .SH AVAILABILITY -Added in curl 7.63.0 +Added in curl 7.62.0 .SH "SEE ALSO" .BR curl_url_cleanup "(3), " curl_url_get "(3), " curl_url_set "(3), " .BR curl_url_dup "(3), " diff --git a/docs/libcurl/curl_url_cleanup.3 b/docs/libcurl/curl_url_cleanup.3 index a8158b7..4d095a9 100644 --- a/docs/libcurl/curl_url_cleanup.3 +++ b/docs/libcurl/curl_url_cleanup.3 @@ -38,7 +38,7 @@ none curl_url_cleanup(url); .fi .SH AVAILABILITY -Added in curl 7.63.0 +Added in curl 7.62.0 .SH "SEE ALSO" .BR curl_url_dup "(3), " curl_url "(3), " curl_url_set "(3), " .BR curl_url_get "(3), " diff --git a/docs/libcurl/curl_url_dup.3 b/docs/libcurl/curl_url_dup.3 index 4815dbd..c0259e0 100644 --- a/docs/libcurl/curl_url_dup.3 +++ b/docs/libcurl/curl_url_dup.3 @@ -46,7 +46,7 @@ Returns a new handle or NULL if out of memory. curl_url_cleanup(url); .fi .SH AVAILABILITY -Added in curl 7.63.0 +Added in curl 7.62.0 .SH "SEE ALSO" .BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), " .BR curl_url_get "(3), " diff --git a/docs/libcurl/curl_url_get.3 b/docs/libcurl/curl_url_get.3 index 824d496..b1313ea 100644 --- a/docs/libcurl/curl_url_get.3 +++ b/docs/libcurl/curl_url_get.3 @@ -104,7 +104,7 @@ If this function returns an error, no URL part is returned. } .fi .SH AVAILABILITY -Added in curl 7.63.0 +Added in curl 7.62.0 .SH "SEE ALSO" .BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_set "(3), " .BR curl_url_dup "(3), " diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3 index 75fc0d9..79272e8 100644 --- a/docs/libcurl/curl_url_set.3 +++ b/docs/libcurl/curl_url_set.3 @@ -114,7 +114,7 @@ If this function returns an error, no URL part is returned. curl_url_cleanup(url); .fi .SH AVAILABILITY -Added in curl 7.63.0 +Added in curl 7.62.0 .SH "SEE ALSO" .BR curl_url_cleanup "(3), " curl_url "(3), " curl_url_get "(3), " .BR curl_url_dup "(3), " -- 2.17.2 From 9c33cac88a9d94557ba48df7c290afc950895bc4 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Sat, 8 Sep 2018 19:39:57 +0200 Subject: [PATCH 03/14] curl_url_set.3: correct description Upstream-commit: 8b85a3cac516a302a8ce3911cf8b9a229b62a59d Signed-off-by: Kamil Dudka --- docs/libcurl/curl_url_set.3 | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3 index 79272e8..0d6e9aa 100644 --- a/docs/libcurl/curl_url_set.3 +++ b/docs/libcurl/curl_url_set.3 @@ -21,7 +21,7 @@ .\" ************************************************************************** .TH curl_url_set 3 "6 Aug 2018" "libcurl" "libcurl Manual" .SH NAME -curl_url_set - set a part from a URL +curl_url_set - set a URL part .SH SYNOPSIS .B #include -- 2.17.2 From dc2c1d978ec78a5f278d194e1b258015e8bfd664 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Sat, 8 Sep 2018 22:57:36 +0200 Subject: [PATCH 04/14] urlapi: avoid derefencing a possible NULL pointer Coverity CID 1439134 Upstream-commit: 01dedc99fc8d386fe955421ab05a1c4094c9190b Signed-off-by: Kamil Dudka --- lib/urlapi.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/urlapi.c b/lib/urlapi.c index 8287861..3183598 100644 --- a/lib/urlapi.c +++ b/lib/urlapi.c @@ -438,10 +438,10 @@ static CURLUcode parse_hostname_login(struct Curl_URL *u, *hostname = ++ptr; /* We could use the login information in the URL so extract it. Only parse - options if the handler says we should. */ + options if the handler says we should. Note that 'h' might be NULL! */ ccode = Curl_parse_login_details(login, ptr - login - 1, &userp, &passwdp, - h->flags & PROTOPT_URLOPTIONS ? + (h && (h->flags & PROTOPT_URLOPTIONS)) ? &optionsp:NULL); if(ccode) { result = CURLUE_MALFORMED_INPUT; -- 2.17.2 From 6684d372c20609afd21f21399deda6deedea911e Mon Sep 17 00:00:00 2001 From: Daniel Gustafsson Date: Sat, 8 Sep 2018 23:05:21 +0200 Subject: [PATCH 05/14] url.c: fix comment typo and indentation Closes #2960 Upstream-commit: 6e4b8c5073c3985cef98656c3b375981d25a8898 Signed-off-by: Kamil Dudka --- lib/url.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/url.c b/lib/url.c index 4f75f11..dcc6cc8 100644 --- a/lib/url.c +++ b/lib/url.c @@ -1939,7 +1939,7 @@ static struct connectdata *allocate_conn(struct Curl_easy *data) return NULL; } -/* returns the handdler if the given scheme is built-in */ +/* returns the handler if the given scheme is built-in */ const struct Curl_handler *Curl_builtin_scheme(const char *scheme) { const struct Curl_handler * const *pp; @@ -2245,7 +2245,7 @@ static CURLcode parseurlandfillconn(struct Curl_easy *data, the host-name part */ memmove(path + hostlen + 1, path, pathlen + 1); - /* now copy the trailing host part in front of the existing path */ + /* now copy the trailing host part in front of the existing path */ memcpy(path + 1, query, hostlen); path[0]='/'; /* prepend the missing slash */ -- 2.17.2 From 0f8d6ab26abd00459d1364a69d7771a6b3a58ce3 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Mon, 10 Sep 2018 10:09:18 +0200 Subject: [PATCH 06/14] libcurl-url.3: overview man page for the URL API Closes #2967 Upstream-commit: 11e8a43f853b9bf050db58f073e6f2411821ce60 Signed-off-by: Kamil Dudka --- docs/libcurl/Makefile.inc | 1 + docs/libcurl/libcurl-url.3 | 137 +++++++++++++++++++++++++++++++++++++ 2 files changed, 138 insertions(+) create mode 100644 docs/libcurl/libcurl-url.3 diff --git a/docs/libcurl/Makefile.inc b/docs/libcurl/Makefile.inc index 955492c..97cb50c 100644 --- a/docs/libcurl/Makefile.inc +++ b/docs/libcurl/Makefile.inc @@ -23,4 +23,5 @@ man_MANS = curl_easy_cleanup.3 curl_easy_getinfo.3 curl_easy_init.3 \ curl_mime_filename.3 curl_mime_subparts.3 \ curl_mime_type.3 curl_mime_headers.3 curl_mime_encoder.3 libcurl-env.3 \ curl_url.3 curl_url_cleanup.3 curl_url_dup.3 curl_url_get.3 curl_url_set.3 \ + libcurl-url.3 \ libcurl-security.3 diff --git a/docs/libcurl/libcurl-url.3 b/docs/libcurl/libcurl-url.3 new file mode 100644 index 0000000..4ad0a15 --- /dev/null +++ b/docs/libcurl/libcurl-url.3 @@ -0,0 +1,137 @@ +.\" ************************************************************************** +.\" * _ _ ____ _ +.\" * Project ___| | | | _ \| | +.\" * / __| | | | |_) | | +.\" * | (__| |_| | _ <| |___ +.\" * \___|\___/|_| \_\_____| +.\" * +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. +.\" * +.\" * This software is licensed as described in the file COPYING, which +.\" * you should have received as part of this distribution. The terms +.\" * are also available at https://curl.haxx.se/docs/copyright.html. +.\" * +.\" * You may opt to use, copy, modify, merge, publish, distribute and/or sell +.\" * copies of the Software, and permit persons to whom the Software is +.\" * furnished to do so, under the terms of the COPYING file. +.\" * +.\" * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY +.\" * KIND, either express or implied. +.\" * +.\" ************************************************************************** +.TH libcurl 3 "10 Sep 2018" "libcurl" "libcurl url interface" +.SH NAME +libcurl-url \- URL interface overview +.SH DESCRIPTION +The URL interface provides a set of functions for parsing and generating URLs. +.SH INCLUDE +You still only include in your code. Note that the URL API was +introduced in 7.62.0. +.SH CREATE +Create a handle that holds URL info and resources with \fIcurl_url(3)\fP: + + CURLU *h = curl_url(); +.SH CLEANUP +When done with it, clean it up with \fIcurl_url_cleanup(3)\fP: + + curl_url_cleanup(h); +.SH DUPLICATE +When you need a copy of a handle, just duplicate it with \fIcurl_url_dup(3)\fP: + + CURLU *nh = curl_url_dup(h); +.SH PARSING +By "setting" a URL to the handle with \fIcurl_url_set(3)\fP, the URL is parsed +and stored in the handle. If the URL is not syntactically correct it will +return an error instead. + +.nf + rc = curl_url_set(h, CURLUPART_URL, + "https://example.com:449/foo/bar?name=moo", 0); +.fi + +The zero in the fourth argument is a bitmask for changing specific features. + +If successful, this stores the URL in its individual parts within the handle. +.SH REDIRECT +When a handle already contains info about a URL, setting a relative URL will +make it "redirect" to adapt to it. + + rc = curl_url_set(h, CURLUPART_URL, "../test?another", 0); +.SH "GET URL" +The `CURLU` handle represents a URL and you can easily extract that with +\fIcurl_url_get(3)\fP: + + char *url; + rc = curl_url_get(h, CURLUPART_URL, &url, 0); + curl_free(url); + +The zero in the fourth argument is a bitmask for changing specific features. +.SH "GET PARTS" +When a URL has been parsed or parts have been set, you can extract those +pieces from the handle at any time. + +.nf + rc = curl_url_get(h, CURLUPART_HOST, &host, 0); + rc = curl_url_get(h, CURLUPART_SCHEME, &scheme, 0); + rc = curl_url_get(h, CURLUPART_USER, &user, 0); + rc = curl_url_get(h, CURLUPART_PASSWORD, &password, 0); + rc = curl_url_get(h, CURLUPART_PORT, &port, 0); + rc = curl_url_get(h, CURLUPART_PATH, &path, 0); + rc = curl_url_get(h, CURLUPART_QUERY, &query, 0); + rc = curl_url_get(h, CURLUPART_FRAGMENT, &fragment, 0); +.fi + +Extracted parts are not URL decoded unless the user also asks for it with the +CURLU_URLDECODE flag set in the fourth bitmask argument. + +Remember to free the returned string with \fIcurl_free(3)\fP when you're done +with it! +.SH "SET PARTS" +A user set individual URL parts, either after having parsed a full URL or +instead of parsing such. + +.nf + rc = curl_url_set(urlp, CURLUPART_HOST, "www.example.com", 0); + rc = curl_url_set(urlp, CURLUPART_SCHEME, "https", 0); + rc = curl_url_set(urlp, CURLUPART_USER, "john", 0); + rc = curl_url_set(urlp, CURLUPART_PASSWORD, "doe", 0); + rc = curl_url_set(urlp, CURLUPART_PORT, "443", 0); + rc = curl_url_set(urlp, CURLUPART_PATH, "/index.html", 0); + rc = curl_url_set(urlp, CURLUPART_QUERY, "name=john", 0); + rc = curl_url_set(urlp, CURLUPART_FRAGMENT, "anchor", 0); +.fi + +Set parts are not URL encoded unless the user asks for it with the +`CURLU_URLENCODE` flag. +.SH "APPENDQUERY" +An application can append a string to the right end of the query part with the +`CURLU_APPENDQUERY` flag to \fIcurl_url_set(3)\fP. + +Imagine a handle that holds the URL `https://example.com/?shoes=2`. An +application can then add the string `hat=1` to the query part like this: + +.nf + rc = curl_url_set(urlp, CURLUPART_QUERY, "hat=1", CURLU_APPENDQUERY); +.fi + +It will even notice the lack of an ampersand (`&`) separator so it will inject +one too, and the handle's full URL will then equal +`https://example.com/?shoes=2&hat=1`. + +The appended string can of course also get URL encoded on add, and if asked to +URL encode, the encoding process will skip the '=' character. For example, +append `candy=N&N` to what we already have, and URL encode it to deal with the +ampersand in the data: + +.nf + rc = curl_url_set(urlp, CURLUPART_QUERY, "candy=N&N", + CURLU_APPENDQUERY | CURLU_URLENCODE); +.fi + +Now the URL looks like +.nf + https://example.com/?shoes=2&hat=1&candy=N%26N` +.fi +.SH "SEE ALSO" +.BR curl_url "(3), " curl_url_cleanup "(3), " curl_url_get "(3), " +.BR curl_url_dup "(3), " curl_url_set "(3), " CURLOPT_URL "(3), " -- 2.17.2 From 4c235b460cf40f8ce0c6ad06b44ecb4dddc128e4 Mon Sep 17 00:00:00 2001 From: Dave Reisner Date: Mon, 10 Sep 2018 09:38:46 -0400 Subject: [PATCH 07/14] curl_url_set.3: fix typo in reference to CURLU_APPENDQUERY Upstream-commit: 04110573801feb2f278e2f774087a0525d5e8d0a Signed-off-by: Kamil Dudka --- docs/libcurl/curl_url_set.3 | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3 index 0d6e9aa..b2b273f 100644 --- a/docs/libcurl/curl_url_set.3 +++ b/docs/libcurl/curl_url_set.3 @@ -75,7 +75,7 @@ If used in with \fICURLU_APPENDQUERY\fP, the provided part will be appended on the end of the existing query - and if the previous part didn't end with an ampersand (&), an ampersand will be inserted before the new appended part. -When \fCURLU_APPENDQUERY\fP is used together with \fICURLU_URLENCODE\fP, +When \fICURLU_APPENDQUERY\fP is used together with \fICURLU_URLENCODE\fP, the '=' symbols will not be URL encoded. The question mark in the URL is not part of the actual query contents. -- 2.17.2 From fb07ea0cf9c612b2fad6a113b1d40aa7896fe43a Mon Sep 17 00:00:00 2001 From: Dave Reisner Date: Mon, 10 Sep 2018 09:39:33 -0400 Subject: [PATCH 08/14] curl_url_set.3: properly escape \n in example code This yields "the scheme is %s\n" instead of "the scheme is %s0 Closes #2970 Upstream-commit: c1e5980f6672a2bd2d26894f093b435f2deb04e0 Signed-off-by: Kamil Dudka --- docs/libcurl/curl_url_get.3 | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/libcurl/curl_url_get.3 b/docs/libcurl/curl_url_get.3 index b1313ea..53f7954 100644 --- a/docs/libcurl/curl_url_get.3 +++ b/docs/libcurl/curl_url_get.3 @@ -97,7 +97,7 @@ If this function returns an error, no URL part is returned. char *scheme; rc = curl_url_get(url, CURLUPART_SCHEME, &scheme, 0); if(!rc) { - printf("the scheme is %s\n", scheme); + printf("the scheme is %s\\n", scheme); curl_free(scheme); } curl_url_cleanup(url); -- 2.17.2 From 376ae7de5a5a5f5b5513e6055700d010f21d4da3 Mon Sep 17 00:00:00 2001 From: Daniel Gustafsson Date: Wed, 19 Sep 2018 13:44:10 +0200 Subject: [PATCH 09/14] urlapi: don't set value which is never read In the CURLUPART_URL case, there is no codepath which invokes url decoding so remove the assignment of the urldecode variable. This fixes the deadstore bug-report from clang static analysis. Closes #3015 Reviewed-by: Daniel Stenberg Upstream-commit: 522e647cc52c45ebdb58d57f242204f9a72c45dd Signed-off-by: Kamil Dudka --- lib/urlapi.c | 1 - 1 file changed, 1 deletion(-) diff --git a/lib/urlapi.c b/lib/urlapi.c index 3183598..127f390 100644 --- a/lib/urlapi.c +++ b/lib/urlapi.c @@ -970,7 +970,6 @@ CURLUcode curl_url_get(CURLU *u, CURLUPart what, char *scheme; char *options = u->options; char *port = u->port; - urldecode = FALSE; /* not for the whole thing */ if(u->scheme && strcasecompare("file", u->scheme)) { url = aprintf("file://%s%s%s", u->path, -- 2.17.2 From 26dd137f3ca894e6402a98889d3b182f608d3c7f Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Wed, 19 Sep 2018 10:17:03 +0200 Subject: [PATCH 10/14] urlapi: add CURLU_GUESS_SCHEME and fix hostname acceptance In order for this API to fully work for libcurl itself, it now offers a CURLU_GUESS_SCHEME flag that makes it "guess" scheme based on the host name prefix just like libcurl always did. If there's no known prefix, it will guess "http://". Separately, it relaxes the check of the host name so that IDN host names can be passed in as well. Both these changes are necessary for libcurl itself to use this API. Assisted-by: Daniel Gustafsson Closes #3018 Upstream-commit: 9307c219ad4741db860b864c860ac2f8bf9fad9d Signed-off-by: Kamil Dudka --- docs/libcurl/curl_url_set.3 | 10 ++++++++ include/curl/urlapi.h | 1 + lib/urlapi.c | 48 ++++++++++++++++++++++++++++--------- tests/data/test1560 | 6 +++++ tests/libtest/lib1560.c | 26 +++++++++++++++++++- 5 files changed, 79 insertions(+), 12 deletions(-) diff --git a/docs/libcurl/curl_url_set.3 b/docs/libcurl/curl_url_set.3 index b2b273f..95b76bd 100644 --- a/docs/libcurl/curl_url_set.3 +++ b/docs/libcurl/curl_url_set.3 @@ -96,6 +96,16 @@ The query part gets space-to-plus conversion before the URL conversion. This URL encoding is charset unaware and will convert the input on a byte-by-byte manner. +.IP CURLU_DEFAULT_SCHEME +If set, will make libcurl allow the URL to be set without a scheme and then +sets that to the default scheme: HTTPS. Overrides the \fICURLU_GUESS_SCHEME\fP +option if both are set. +.IP CURLU_GUESS_SCHEME +If set, will make libcurl allow the URL to be set without a scheme and it +instead "guesses" which scheme that was intended based on the host name. If +the outermost sub-domain name matches DICT, FTP, IMAP, LDAP, POP3 or SMTP then +that scheme will be used, otherwise it picks HTTP. Conflicts with the +\fICURLU_DEFAULT_SCHEME\fP option which takes precendence if both are set. .SH RETURN VALUE Returns a CURLUcode error value, which is CURLUE_OK (0) if everything went fine. diff --git a/include/curl/urlapi.h b/include/curl/urlapi.h index b16cfce..319de35 100644 --- a/include/curl/urlapi.h +++ b/include/curl/urlapi.h @@ -75,6 +75,7 @@ typedef enum { #define CURLU_URLDECODE (1<<6) /* URL decode on get */ #define CURLU_URLENCODE (1<<7) /* URL encode on set */ #define CURLU_APPENDQUERY (1<<8) /* append a form style part */ +#define CURLU_GUESS_SCHEME (1<<9) /* legacy curl-style guessing */ typedef struct Curl_URL CURLU; diff --git a/lib/urlapi.c b/lib/urlapi.c index 127f390..45f1e14 100644 --- a/lib/urlapi.c +++ b/lib/urlapi.c @@ -554,7 +554,7 @@ static CURLUcode junkscan(char *part) static CURLUcode hostname_check(char *hostname, unsigned int flags) { - const char *l; /* accepted characters */ + const char *l = NULL; /* accepted characters */ size_t len; size_t hlen = strlen(hostname); (void)flags; @@ -564,14 +564,21 @@ static CURLUcode hostname_check(char *hostname, unsigned int flags) l = "0123456789abcdefABCDEF::."; hlen -= 2; } - else /* % for URL escaped letters */ - l = "0123456789abcdefghijklimnopqrstuvwxyz-_.ABCDEFGHIJKLIMNOPQRSTUVWXYZ%"; - - len = strspn(hostname, l); - if(hlen != len) - /* hostname with bad content */ - return CURLUE_MALFORMED_INPUT; + if(l) { + /* only valid letters are ok */ + len = strspn(hostname, l); + if(hlen != len) + /* hostname with bad content */ + return CURLUE_MALFORMED_INPUT; + } + else { + /* letters from the second string is not ok */ + len = strcspn(hostname, " "); + if(hlen != len) + /* hostname with bad content */ + return CURLUE_MALFORMED_INPUT; + } return CURLUE_OK; } @@ -587,7 +594,7 @@ static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags) CURLUcode result; bool url_has_scheme = FALSE; char schemebuf[MAX_SCHEME_LEN]; - char *schemep; + char *schemep = NULL; size_t schemelen = 0; size_t urllen; const struct Curl_handler *h = NULL; @@ -723,9 +730,10 @@ static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags) else { /* no scheme! */ - if(!(flags & CURLU_DEFAULT_SCHEME)) + if(!(flags & (CURLU_DEFAULT_SCHEME|CURLU_GUESS_SCHEME))) return CURLUE_MALFORMED_INPUT; - schemep = (char *) DEFAULT_SCHEME; + if(flags & CURLU_DEFAULT_SCHEME) + schemep = (char *) DEFAULT_SCHEME; /* * The URL was badly formatted, let's try without scheme specified. @@ -744,6 +752,24 @@ static CURLUcode seturl(const char *url, CURLU *u, unsigned int flags) memcpy(hostname, hostp, len); hostname[len] = 0; + if((flags & CURLU_GUESS_SCHEME) && !schemep) { + /* legacy curl-style guess based on host name */ + if(checkprefix("ftp.", hostname)) + schemep = (char *)"ftp"; + else if(checkprefix("dict.", hostname)) + schemep = (char *)"dict"; + else if(checkprefix("ldap.", hostname)) + schemep = (char *)"ldap"; + else if(checkprefix("imap.", hostname)) + schemep = (char *)"imap"; + else if(checkprefix("smtp.", hostname)) + schemep = (char *)"smtp"; + else if(checkprefix("pop3.", hostname)) + schemep = (char *)"pop3"; + else + schemep = (char *)"http"; + } + len = strlen(p); memcpy(path, p, len); path[len] = 0; diff --git a/tests/data/test1560 b/tests/data/test1560 index 720df03..4b6c97a 100644 --- a/tests/data/test1560 +++ b/tests/data/test1560 @@ -16,6 +16,12 @@ none file https http +pop3 +smtp +imap +ldap +dict +ftp URL API diff --git a/tests/libtest/lib1560.c b/tests/libtest/lib1560.c index 669ea9a..30fb582 100644 --- a/tests/libtest/lib1560.c +++ b/tests/libtest/lib1560.c @@ -246,8 +246,32 @@ static struct testcase get_parts_list[] ={ }; static struct urltestcase get_url_list[] = { + {"smtp.example.com/path/html", + "smtp://smtp.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"https.example.com/path/html", + "http://https.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"dict.example.com/path/html", + "dict://dict.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"pop3.example.com/path/html", + "pop3://pop3.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"ldap.example.com/path/html", + "ldap://ldap.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"imap.example.com/path/html", + "imap://imap.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"ftp.example.com/path/html", + "ftp://ftp.example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, + {"example.com/path/html", + "http://example.com/path/html", + CURLU_GUESS_SCHEME, 0, CURLUE_OK}, {"HTTP://test/", "http://test/", 0, 0, CURLUE_OK}, - {"http://HO0_-st..~./", "", 0, 0, CURLUE_MALFORMED_INPUT}, + {"http://HO0_-st..~./", "http://HO0_-st..~./", 0, 0, CURLUE_OK}, {"http:/@example.com: 123/", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, {"http:/@example.com:123 /", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, {"http:/@example.com:123a/", "", 0, 0, CURLUE_BAD_PORT_NUMBER}, -- 2.17.2 From 4e335817d4ac0ee5596363004bfcaaad15bc6127 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Wed, 19 Sep 2018 11:28:40 +0200 Subject: [PATCH 11/14] urlapi: document the error codes, remove two unused ones Assisted-by: Daniel Gustafsson Closes #3019 Upstream-commit: 5c73093edb3bd527db9c8abdee53d0f18e6a4cc1 Signed-off-by: Kamil Dudka --- docs/libcurl/libcurl-errors.3 | 37 ++++++++++++++++++++++++++++- docs/libcurl/symbols-in-versions | 2 -- include/curl/urlapi.h | 8 +++---- tests/libtest/lib1560.c | 40 ++++++++++++++++---------------- 4 files changed, 59 insertions(+), 28 deletions(-) diff --git a/docs/libcurl/libcurl-errors.3 b/docs/libcurl/libcurl-errors.3 index 30c57b3..411a272 100644 --- a/docs/libcurl/libcurl-errors.3 +++ b/docs/libcurl/libcurl-errors.3 @@ -5,7 +5,7 @@ .\" * | (__| |_| | _ <| |___ .\" * \___|\___/|_| \_\_____| .\" * -.\" * Copyright (C) 1998 - 2015, Daniel Stenberg, , et al. +.\" * Copyright (C) 1998 - 2018, Daniel Stenberg, , et al. .\" * .\" * This software is licensed as described in the file COPYING, which .\" * you should have received as part of this distribution. The terms @@ -307,6 +307,41 @@ Not enough memory was available. .IP "CURLSHE_NOT_BUILT_IN (5)" The requested sharing could not be done because the library you use don't have that particular feature enabled. (Added in 7.23.0) +.SH "CURLUcode" +.IP "CURLUE_BAD_HANDLE (1)" +An argument that should be a CURLU pointer was passed in as a NULL. +.IP "CURLUE_BAD_PARTPOINTER (2)" +A NULL pointer was passed to the 'part' argument of \fIcurl_url_get(3)\fP. +.IP "CURLUE_MALFORMED_INPUT (3)" +A malformed input was passed to a URL API function. +.IP "CURLUE_BAD_PORT_NUMBER (4)" +The port number was not a decimal number between 0 and 65535. +.IP "CURLUE_UNSUPPORTED_SCHEME (5)" +This libcurl build doesn't support the given URL scheme. +.IP "CURLUE_URLDECODE (6)" +URL decode error, most likely because of rubbish in the input. +.IP "CURLUE_OUT_OF_MEMORY (7)" +A memory function failed. +.IP "CURLUE_USER_NOT_ALLOWED (8)" +Credentials was passed in the URL when prohibited. +.IP "CURLUE_UNKNOWN_PART (9)" +An unknown part ID was passed to a URL API function. +.IP "CURLUE_NO_SCHEME (10)" +There is no scheme part in the URL. +.IP "CURLUE_NO_USER (11)" +There is no user part in the URL. +.IP "CURLUE_NO_PASSWORD (12)" +There is no password part in the URL. +.IP "CURLUE_NO_OPTIONS (13)" +There is no options part in the URL. +.IP "CURLUE_NO_HOST (14)" +There is no host part in the URL. +.IP "CURLUE_NO_PORT (15)" +There is no port part in the URL. +.IP "CURLUE_NO_QUERY (16)" +There is no query part in the URL. +.IP "CURLUE_NO_FRAGMENT (17)" +There is no fragment part in the URL. .SH "SEE ALSO" .BR curl_easy_strerror "(3), " curl_multi_strerror "(3), " .BR curl_share_strerror "(3), " CURLOPT_ERRORBUFFER "(3), " diff --git a/docs/libcurl/symbols-in-versions b/docs/libcurl/symbols-in-versions index c797cb7..3b3861f 100644 --- a/docs/libcurl/symbols-in-versions +++ b/docs/libcurl/symbols-in-versions @@ -736,14 +736,12 @@ CURLUE_NO_FRAGMENT 7.62.0 CURLUE_NO_HOST 7.62.0 CURLUE_NO_OPTIONS 7.62.0 CURLUE_NO_PASSWORD 7.62.0 -CURLUE_NO_PATH 7.62.0 CURLUE_NO_PORT 7.62.0 CURLUE_NO_QUERY 7.62.0 CURLUE_NO_SCHEME 7.62.0 CURLUE_NO_USER 7.62.0 CURLUE_OK 7.62.0 CURLUE_OUT_OF_MEMORY 7.62.0 -CURLUE_RELATIVE 7.62.0 CURLUE_UNKNOWN_PART 7.62.0 CURLUE_UNSUPPORTED_SCHEME 7.62.0 CURLUE_URLDECODE 7.62.0 diff --git a/include/curl/urlapi.h b/include/curl/urlapi.h index 319de35..90dd56c 100644 --- a/include/curl/urlapi.h +++ b/include/curl/urlapi.h @@ -35,7 +35,7 @@ typedef enum { CURLUE_BAD_PORT_NUMBER, /* 4 */ CURLUE_UNSUPPORTED_SCHEME, /* 5 */ CURLUE_URLDECODE, /* 6 */ - CURLUE_RELATIVE, /* 7 */ + CURLUE_OUT_OF_MEMORY, /* 7 */ CURLUE_USER_NOT_ALLOWED, /* 8 */ CURLUE_UNKNOWN_PART, /* 9 */ CURLUE_NO_SCHEME, /* 10 */ @@ -44,10 +44,8 @@ typedef enum { CURLUE_NO_OPTIONS, /* 13 */ CURLUE_NO_HOST, /* 14 */ CURLUE_NO_PORT, /* 15 */ - CURLUE_NO_PATH, /* 16 */ - CURLUE_NO_QUERY, /* 17 */ - CURLUE_NO_FRAGMENT, /* 18 */ - CURLUE_OUT_OF_MEMORY /* 19 */ + CURLUE_NO_QUERY, /* 16 */ + CURLUE_NO_FRAGMENT /* 17 */ } CURLUcode; typedef enum { diff --git a/tests/libtest/lib1560.c b/tests/libtest/lib1560.c index 30fb582..224cb88 100644 --- a/tests/libtest/lib1560.c +++ b/tests/libtest/lib1560.c @@ -129,7 +129,7 @@ struct querycase { static struct testcase get_parts_list[] ={ {"https://127.0.0.1:443", - "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [17] | [18]", + "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [16] | [17]", 0, CURLU_NO_DEFAULT_PORT, CURLUE_OK}, {"http://%3a:%3a@ex%0ample/%3f+?+%3f+%23#+%23%3f%g7", "http | : | : | [13] | [6] | [15] | /?+ | ? # | +#?%g7", @@ -138,43 +138,43 @@ static struct testcase get_parts_list[] ={ "http | %3a | %3a | [13] | ex%0ample | [15] | /%3f | %3f%35 | %35%3f%g7", 0, 0, CURLUE_OK}, {"http://HO0_-st%41/", - "http | [11] | [12] | [13] | HO0_-st%41 | [15] | / | [17] | [18]", + "http | [11] | [12] | [13] | HO0_-st%41 | [15] | / | [16] | [17]", 0, 0, CURLUE_OK}, {"file://hello.html", "", 0, 0, CURLUE_MALFORMED_INPUT}, {"http://HO0_-st/", - "http | [11] | [12] | [13] | HO0_-st | [15] | / | [17] | [18]", + "http | [11] | [12] | [13] | HO0_-st | [15] | / | [16] | [17]", 0, 0, CURLUE_OK}, {"imap://user:pass;option@server/path", - "imap | user | pass | option | server | [15] | /path | [17] | [18]", + "imap | user | pass | option | server | [15] | /path | [16] | [17]", 0, 0, CURLUE_OK}, {"http://user:pass;option@server/path", - "http | user | pass;option | [13] | server | [15] | /path | [17] | [18]", + "http | user | pass;option | [13] | server | [15] | /path | [16] | [17]", 0, 0, CURLUE_OK}, {"file:/hello.html", - "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [16] | [17]", 0, 0, CURLUE_OK}, {"file://127.0.0.1/hello.html", - "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [16] | [17]", 0, 0, CURLUE_OK}, {"file:////hello.html", - "file | [11] | [12] | [13] | [14] | [15] | //hello.html | [17] | [18]", + "file | [11] | [12] | [13] | [14] | [15] | //hello.html | [16] | [17]", 0, 0, CURLUE_OK}, {"file:///hello.html", - "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [17] | [18]", + "file | [11] | [12] | [13] | [14] | [15] | /hello.html | [16] | [17]", 0, 0, CURLUE_OK}, {"https://127.0.0.1", - "https | [11] | [12] | [13] | 127.0.0.1 | 443 | / | [17] | [18]", + "https | [11] | [12] | [13] | 127.0.0.1 | 443 | / | [16] | [17]", 0, CURLU_DEFAULT_PORT, CURLUE_OK}, {"https://127.0.0.1", - "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [17] | [18]", + "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://[::1]:1234", - "https | [11] | [12] | [13] | [::1] | 1234 | / | [17] | [18]", + "https | [11] | [12] | [13] | [::1] | 1234 | / | [16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://127abc.com", - "https | [11] | [12] | [13] | 127abc.com | [15] | / | [17] | [18]", + "https | [11] | [12] | [13] | 127abc.com | [15] | / | [16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https:// example.com?check", "", @@ -183,7 +183,7 @@ static struct testcase get_parts_list[] ={ "", CURLU_DEFAULT_SCHEME, 0, CURLUE_MALFORMED_INPUT}, {"https://example.com?check", - "https | [11] | [12] | [13] | example.com | [15] | / | check | [18]", + "https | [11] | [12] | [13] | example.com | [15] | / | check | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://example.com:65536", "", @@ -193,27 +193,27 @@ static struct testcase get_parts_list[] ={ CURLU_DEFAULT_SCHEME, 0, CURLUE_BAD_PORT_NUMBER}, {"https://example.com:01#moo", "https | [11] | [12] | [13] | example.com | 1 | / | " - "[17] | moo", + "[16] | moo", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://example.com:1#moo", "https | [11] | [12] | [13] | example.com | 1 | / | " - "[17] | moo", + "[16] | moo", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"http://example.com#moo", "http | [11] | [12] | [13] | example.com | [15] | / | " - "[17] | moo", + "[16] | moo", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"http://example.com", "http | [11] | [12] | [13] | example.com | [15] | / | " - "[17] | [18]", + "[16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"http://example.com/path/html", "http | [11] | [12] | [13] | example.com | [15] | /path/html | " - "[17] | [18]", + "[16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"http://example.com/path/html?query=name", "http | [11] | [12] | [13] | example.com | [15] | /path/html | " - "query=name | [18]", + "query=name | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"http://example.com/path/html?query=name#anchor", "http | [11] | [12] | [13] | example.com | [15] | /path/html | " -- 2.17.2 From 88dfdac2fc1b34a321a323868ea06116c72fe6d2 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Fri, 21 Sep 2018 08:17:39 +0200 Subject: [PATCH 12/14] urlapi: fix support for address scope in IPv6 numerical addresses Closes #3024 Upstream-commit: 2097cd515289581df5dfb6eeb5942d083a871fa4 Signed-off-by: Kamil Dudka --- lib/urlapi-int.h | 4 ++++ lib/urlapi.c | 8 ++------ tests/libtest/lib1560.c | 3 +++ 3 files changed, 9 insertions(+), 6 deletions(-) diff --git a/lib/urlapi-int.h b/lib/urlapi-int.h index 7ac09fd..a5bb8ea 100644 --- a/lib/urlapi-int.h +++ b/lib/urlapi-int.h @@ -22,6 +22,10 @@ * ***************************************************************************/ #include "curl_setup.h" +/* scheme is not URL encoded, the longest libcurl supported ones are 6 + letters */ +#define MAX_SCHEME_LEN 8 + bool Curl_is_absolute_url(const char *url, char *scheme, size_t buflen); char *Curl_concat_url(const char *base, const char *relurl); size_t Curl_strlen_url(const char *url, bool relative); diff --git a/lib/urlapi.c b/lib/urlapi.c index 45f1e14..a12112e 100644 --- a/lib/urlapi.c +++ b/lib/urlapi.c @@ -53,10 +53,6 @@ struct Curl_URL { #define DEFAULT_SCHEME "https" -/* scheme is not URL encoded, the longest libcurl supported ones are 6 - letters */ -#define MAX_SCHEME_LEN 8 - static void free_urlhandle(struct Curl_URL *u) { free(u->scheme); @@ -480,7 +476,7 @@ static CURLUcode parse_port(struct Curl_URL *u, char *hostname) char endbracket; int len; - if((1 == sscanf(hostname, "[%*45[0123456789abcdefABCDEF:.]%c%n", + if((1 == sscanf(hostname, "[%*45[0123456789abcdefABCDEF:.%%]%c%n", &endbracket, &len)) && (']' == endbracket)) { /* this is a RFC2732-style specified IP-address */ @@ -561,7 +557,7 @@ static CURLUcode hostname_check(char *hostname, unsigned int flags) if(hostname[0] == '[') { hostname++; - l = "0123456789abcdefABCDEF::."; + l = "0123456789abcdefABCDEF::.%"; hlen -= 2; } diff --git a/tests/libtest/lib1560.c b/tests/libtest/lib1560.c index 224cb88..7a5be81 100644 --- a/tests/libtest/lib1560.c +++ b/tests/libtest/lib1560.c @@ -128,6 +128,9 @@ struct querycase { }; static struct testcase get_parts_list[] ={ + {"https://[::1%252]:1234", + "https | [11] | [12] | [13] | [::1%252] | 1234 | / | [16] | [17]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://127.0.0.1:443", "https | [11] | [12] | [13] | 127.0.0.1 | [15] | / | [16] | [17]", 0, CURLU_NO_DEFAULT_PORT, CURLUE_OK}, -- 2.17.2 From 6c9f3f4bc604ba06a4f43807ace9189503a5e9fc Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Fri, 2 Nov 2018 15:11:16 +0100 Subject: [PATCH 13/14] URL: fix IPv6 numeral address parser Regression from 46e164069d1a52. Extended test 1560 to verify. Reported-by: tpaukrt on github Fixes #3218 Closes #3219 Upstream-commit: b28094833a971870fd8c07960b3b12bf6fbbaad3 Signed-off-by: Kamil Dudka --- lib/urlapi.c | 8 ++++++-- tests/libtest/lib1560.c | 9 +++++++++ 2 files changed, 15 insertions(+), 2 deletions(-) diff --git a/lib/urlapi.c b/lib/urlapi.c index a12112e..8626052 100644 --- a/lib/urlapi.c +++ b/lib/urlapi.c @@ -481,8 +481,12 @@ static CURLUcode parse_port(struct Curl_URL *u, char *hostname) (']' == endbracket)) { /* this is a RFC2732-style specified IP-address */ portptr = &hostname[len]; - if (*portptr != ':') - return CURLUE_MALFORMED_INPUT; + if(*portptr) { + if(*portptr != ':') + return CURLUE_MALFORMED_INPUT; + } + else + portptr = NULL; } else portptr = strchr(hostname, ':'); diff --git a/tests/libtest/lib1560.c b/tests/libtest/lib1560.c index 7a5be81..483035c 100644 --- a/tests/libtest/lib1560.c +++ b/tests/libtest/lib1560.c @@ -128,6 +128,15 @@ struct querycase { }; static struct testcase get_parts_list[] ={ + {"http://[fd00:a41::50]:8080", + "http | [11] | [12] | [13] | [fd00:a41::50] | 8080 | / | [16] | [17]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://[fd00:a41::50]/", + "http | [11] | [12] | [13] | [fd00:a41::50] | [15] | / | [16] | [17]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, + {"http://[fd00:a41::50]", + "http | [11] | [12] | [13] | [fd00:a41::50] | [15] | / | [16] | [17]", + CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, {"https://[::1%252]:1234", "https | [11] | [12] | [13] | [::1%252] | 1234 | / | [16] | [17]", CURLU_DEFAULT_SCHEME, 0, CURLUE_OK}, -- 2.17.2 From 9fa7298750c1d66331dc55a202277b131868c048 Mon Sep 17 00:00:00 2001 From: Daniel Stenberg Date: Wed, 2 Jan 2019 20:18:27 +0100 Subject: [PATCH 14/14] xattr: strip credentials from any URL that is stored Both user and password are cleared uncondtitionally. Added unit test 1621 to verify. Fixes #3423 Closes #3433 Upstream-commit: 98e6629154044e4ab1ee7cff8351c7ebcb131e88 Signed-off-by: Kamil Dudka --- src/tool_xattr.c | 63 +++++++++++++++++++++++++---- tests/data/Makefile.inc | 2 +- tests/data/test1621 | 27 +++++++++++++ tests/unit/Makefile.inc | 6 ++- tests/unit/unit1621.c | 89 +++++++++++++++++++++++++++++++++++++++++ 5 files changed, 177 insertions(+), 10 deletions(-) create mode 100644 tests/data/test1621 create mode 100644 tests/unit/unit1621.c diff --git a/src/tool_xattr.c b/src/tool_xattr.c index 92b99db..730381b 100644 --- a/src/tool_xattr.c +++ b/src/tool_xattr.c @@ -5,7 +5,7 @@ * | (__| |_| | _ <| |___ * \___|\___/|_| \_\_____| * - * Copyright (C) 1998 - 2014, Daniel Stenberg, , et al. + * Copyright (C) 1998 - 2019, Daniel Stenberg, , et al. * * This software is licensed as described in the file COPYING, which * you should have received as part of this distribution. The terms @@ -49,6 +49,46 @@ static const struct xattr_mapping { { NULL, CURLINFO_NONE } /* last element, abort loop here */ }; +/* returns TRUE if a new URL is returned, that then needs to be freed */ +/* @unittest: 1621 */ +#ifdef UNITTESTS +bool stripcredentials(char **url); +#else +static +#endif +bool stripcredentials(char **url) +{ + CURLU *u; + CURLUcode uc; + char *nurl; + u = curl_url(); + if(u) { + uc = curl_url_set(u, CURLUPART_URL, *url, 0); + if(uc) + goto error; + + uc = curl_url_set(u, CURLUPART_USER, NULL, 0); + if(uc) + goto error; + + uc = curl_url_set(u, CURLUPART_PASSWORD, NULL, 0); + if(uc) + goto error; + + uc = curl_url_get(u, CURLUPART_URL, &nurl, 0); + if(uc) + goto error; + + curl_url_cleanup(u); + + *url = nurl; + return TRUE; + } + error: + curl_url_cleanup(u); + return FALSE; +} + /* store metadata from the curl request alongside the downloaded * file using extended attributes */ @@ -62,17 +102,24 @@ int fwrite_xattr(CURL *curl, int fd) char *value = NULL; CURLcode result = curl_easy_getinfo(curl, mappings[i].info, &value); if(!result && value) { + bool freeptr = FALSE; + if(CURLINFO_EFFECTIVE_URL == mappings[i].info) + freeptr = stripcredentials(&value); + if(value) { #ifdef HAVE_FSETXATTR_6 - err = fsetxattr(fd, mappings[i].attr, value, strlen(value), 0, 0); + err = fsetxattr(fd, mappings[i].attr, value, strlen(value), 0, 0); #elif defined(HAVE_FSETXATTR_5) - err = fsetxattr(fd, mappings[i].attr, value, strlen(value), 0); + err = fsetxattr(fd, mappings[i].attr, value, strlen(value), 0); #elif defined(__FreeBSD_version) - err = extattr_set_fd(fd, EXTATTR_NAMESPACE_USER, mappings[i].attr, value, - strlen(value)); - /* FreeBSD's extattr_set_fd returns the length of the extended attribute - */ - err = err < 0 ? err : 0; + err = extattr_set_fd(fd, EXTATTR_NAMESPACE_USER, mappings[i].attr, + value, strlen(value)); + /* FreeBSD's extattr_set_fd returns the length of the extended + attribute */ + err = err < 0 ? err : 0; #endif + if(freeptr) + curl_free(value); + } } i++; } diff --git a/tests/data/Makefile.inc b/tests/data/Makefile.inc index dd38f89..6172b77 100644 --- a/tests/data/Makefile.inc +++ b/tests/data/Makefile.inc @@ -182,7 +182,7 @@ test1560 \ \ test1590 \ test1600 test1601 test1602 test1603 test1604 test1605 test1606 test1607 \ -test1608 test1609 \ +test1608 test1609 test1621 \ \ test1700 test1701 test1702 \ \ diff --git a/tests/data/test1621 b/tests/data/test1621 new file mode 100644 index 0000000..1117d1b --- /dev/null +++ b/tests/data/test1621 @@ -0,0 +1,27 @@ + + + +unittest +stripcredentials + + + +# +# Client-side + + +none + + +unittest +https + + +unit tests for stripcredentials from URL + + +unit1621 + + + + diff --git a/tests/unit/Makefile.inc b/tests/unit/Makefile.inc index 8b1a607..82eaec7 100644 --- a/tests/unit/Makefile.inc +++ b/tests/unit/Makefile.inc @@ -10,7 +10,7 @@ UNITPROGS = unit1300 unit1301 unit1302 unit1303 unit1304 unit1305 unit1307 \ unit1330 unit1394 unit1395 unit1396 unit1397 unit1398 \ unit1399 \ unit1600 unit1601 unit1602 unit1603 unit1604 unit1605 unit1606 unit1607 \ - unit1608 unit1609 + unit1608 unit1609 unit1621 unit1300_SOURCES = unit1300.c $(UNITFILES) unit1300_CPPFLAGS = $(AM_CPPFLAGS) @@ -95,3 +95,7 @@ unit1608_CPPFLAGS = $(AM_CPPFLAGS) unit1609_SOURCES = unit1609.c $(UNITFILES) unit1609_CPPFLAGS = $(AM_CPPFLAGS) + +unit1621_SOURCES = unit1621.c $(UNITFILES) +unit1621_CPPFLAGS = $(AM_CPPFLAGS) +unit1621_LDADD = $(top_builddir)/src/libcurltool.la $(top_builddir)/lib/libcurl.la diff --git a/tests/unit/unit1621.c b/tests/unit/unit1621.c new file mode 100644 index 0000000..6e07b6e --- /dev/null +++ b/tests/unit/unit1621.c @@ -0,0 +1,89 @@ +/*************************************************************************** + * _ _ ____ _ + * Project ___| | | | _ \| | + * / __| | | | |_) | | + * | (__| |_| | _ <| |___ + * \___|\___/|_| \_\_____| + * + * Copyright (C) 1998 - 2019, Daniel Stenberg, , et al. + * + * This software is licensed as described in the file COPYING, which + * you should have received as part of this distribution. The terms + * are also available at https://curl.haxx.se/docs/copyright.html. + * + * You may opt to use, copy, modify, merge, publish, distribute and/or sell + * copies of the Software, and permit persons to whom the Software is + * furnished to do so, under the terms of the COPYING file. + * + * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY + * KIND, either express or implied. + * + ***************************************************************************/ +#include "curlcheck.h" + +#include "urldata.h" +#include "url.h" + +#include "memdebug.h" /* LAST include file */ + +static CURLcode unit_setup(void) +{ + return CURLE_OK; +} + +static void unit_stop(void) +{ +} + +#ifdef __MINGW32__ +UNITTEST_START +{ + return 0; +} +UNITTEST_STOP +#else + +bool stripcredentials(char **url); + +struct checkthis { + const char *input; + const char *output; +}; + +static struct checkthis tests[] = { + { "ninja://foo@example.com", "ninja://foo@example.com" }, + { "https://foo@example.com", "https://example.com/" }, + { "https://localhost:45", "https://localhost:45/" }, + { "https://foo@localhost:45", "https://localhost:45/" }, + { "http://daniel:password@localhost", "http://localhost/" }, + { "http://daniel@localhost", "http://localhost/" }, + { "http://localhost/", "http://localhost/" }, + { NULL, NULL } /* end marker */ +}; + +UNITTEST_START +{ + bool cleanup; + char *url; + int i; + int rc = 0; + + for(i = 0; tests[i].input; i++) { + url = (char *)tests[i].input; + cleanup = stripcredentials(&url); + printf("Test %u got input \"%s\", output: \"%s\"\n", + i, tests[i].input, url); + + if(strcmp(tests[i].output, url)) { + fprintf(stderr, "Test %u got input \"%s\", expected output \"%s\"\n" + " Actual output: \"%s\"\n", i, tests[i].input, tests[i].output, + url); + rc++; + } + if(cleanup) + curl_free(url); + } + return rc; +} +UNITTEST_STOP +#endif -- 2.17.2