php-src/ext/standard/proc_open.c

1039 lines
26 KiB
C
Raw Normal View History

/*
+----------------------------------------------------------------------+
2014-09-19 16:33:14 +00:00
| PHP Version 7 |
+----------------------------------------------------------------------+
| Copyright (c) 1997-2016 The PHP Group |
+----------------------------------------------------------------------+
2006-01-01 12:51:34 +00:00
| This source file is subject to version 3.01 of the PHP license, |
| that is bundled with this package in the file LICENSE, and is |
| available through the world-wide-web at the following url: |
2006-01-01 12:51:34 +00:00
| http://www.php.net/license/3_01.txt |
| If you did not receive a copy of the PHP license and are unable to |
| obtain it through the world-wide-web, please send a note to |
| license@php.net so we can mail you a copy immediately. |
+----------------------------------------------------------------------+
| Author: Wez Furlong <wez@thebrainroom.com> |
+----------------------------------------------------------------------+
*/
/* $Id$ */
2004-05-29 09:27:18 +00:00
#if 0 && (defined(__linux__) || defined(sun) || defined(__IRIX__))
# define _BSD_SOURCE /* linux wants this when XOPEN mode is on */
# define _BSD_COMPAT /* irix: uint */
# define _XOPEN_SOURCE 500 /* turn on Unix98 */
# define __EXTENSIONS__ 1 /* Solaris: uint */
#endif
#include "php.h"
#include <stdio.h>
#include <ctype.h>
#include "php_string.h"
#include "ext/standard/head.h"
#include "ext/standard/basic_functions.h"
#include "ext/standard/file.h"
#include "exec.h"
#include "php_globals.h"
#include "SAPI.h"
#include "main/php_network.h"
2004-09-29 06:04:36 +00:00
#ifdef NETWARE
#include <proc.h>
#include <library.h>
#endif
#if HAVE_SYS_WAIT_H
#include <sys/wait.h>
#endif
#if HAVE_SIGNAL_H
#include <signal.h>
#endif
#if HAVE_SYS_STAT_H
#include <sys/stat.h>
#endif
#if HAVE_FCNTL_H
#include <fcntl.h>
#endif
/* This symbol is defined in ext/standard/config.m4.
* Essentially, it is set if you HAVE_FORK || PHP_WIN32
2012-05-29 13:44:04 +00:00
* Other platforms may modify that configure check and add suitable #ifdefs
* around the alternate code.
* */
#ifdef PHP_CAN_SUPPORT_PROC_OPEN
2004-05-29 09:27:18 +00:00
#if 0 && HAVE_PTSNAME && HAVE_GRANTPT && HAVE_UNLOCKPT && HAVE_SYS_IOCTL_H && HAVE_TERMIOS_H
# include <sys/ioctl.h>
# include <termios.h>
2004-05-29 00:30:11 +00:00
# define PHP_CAN_DO_PTS 1
#endif
#include "proc_open.h"
static int le_proc_open;
/* {{{ _php_array_to_envp */
2014-12-13 22:06:14 +00:00
static php_process_env_t _php_array_to_envp(zval *environment, int is_persistent)
{
zval *element;
php_process_env_t env;
zend_string *key, *str;
#ifndef PHP_WIN32
char **ep;
#endif
char *p;
size_t cnt, l, sizeenv = 0;
HashTable *env_hash;
memset(&env, 0, sizeof(env));
2010-08-30 16:52:48 +00:00
if (!environment) {
return env;
}
2010-08-30 16:52:48 +00:00
cnt = zend_hash_num_elements(Z_ARRVAL_P(environment));
2010-08-30 16:52:48 +00:00
if (cnt < 1) {
#ifndef PHP_WIN32
env.envarray = (char **) pecalloc(1, sizeof(char *), is_persistent);
#endif
env.envp = (char *) pecalloc(4, 1, is_persistent);
return env;
}
ALLOC_HASHTABLE(env_hash);
zend_hash_init(env_hash, cnt, NULL, NULL, 0);
/* first, we have to get the size of all the elements in the hash */
ZEND_HASH_FOREACH_STR_KEY_VAL(Z_ARRVAL_P(environment), key, element) {
str = zval_get_string(element);
2014-05-09 11:01:38 +00:00
if (ZSTR_LEN(str) == 0) {
zend_string_release(str);
continue;
}
2010-08-30 16:52:48 +00:00
sizeenv += ZSTR_LEN(str) + 1;
2010-09-08 09:11:18 +00:00
if (key && ZSTR_LEN(key)) {
sizeenv += ZSTR_LEN(key) + 1;
zend_hash_add_ptr(env_hash, key, str);
} else {
zend_hash_next_index_insert_ptr(env_hash, str);
}
} ZEND_HASH_FOREACH_END();
#ifndef PHP_WIN32
ep = env.envarray = (char **) pecalloc(cnt + 1, sizeof(char *), is_persistent);
#endif
p = env.envp = (char *) pecalloc(sizeenv + 4, 1, is_persistent);
ZEND_HASH_FOREACH_STR_KEY_PTR(env_hash, key, str) {
if (key) {
l = ZSTR_LEN(key) + ZSTR_LEN(str) + 2;
memcpy(p, ZSTR_VAL(key), ZSTR_LEN(key));
strncat(p, "=", 1);
strncat(p, ZSTR_VAL(str), ZSTR_LEN(str));
2010-08-30 16:52:48 +00:00
#ifndef PHP_WIN32
*ep = p;
++ep;
#endif
p += l;
} else {
memcpy(p, ZSTR_VAL(str), ZSTR_LEN(str));
#ifndef PHP_WIN32
*ep = p;
++ep;
#endif
p += ZSTR_LEN(str) + 1;
}
2014-08-25 17:24:55 +00:00
zend_string_release(str);
} ZEND_HASH_FOREACH_END();
2010-09-06 10:28:04 +00:00
assert((uint)(p - env.envp) <= sizeenv);
2010-08-30 16:52:48 +00:00
zend_hash_destroy(env_hash);
FREE_HASHTABLE(env_hash);
return env;
}
/* }}} */
/* {{{ _php_free_envp */
static void _php_free_envp(php_process_env_t env, int is_persistent)
{
#ifndef PHP_WIN32
if (env.envarray) {
pefree(env.envarray, is_persistent);
}
#endif
if (env.envp) {
pefree(env.envp, is_persistent);
}
}
/* }}} */
2003-07-12 07:01:30 +00:00
/* {{{ proc_open_rsrc_dtor */
2014-12-13 22:06:14 +00:00
static void proc_open_rsrc_dtor(zend_resource *rsrc)
{
struct php_process_handle *proc = (struct php_process_handle*)rsrc->ptr;
int i;
#ifdef PHP_WIN32
DWORD wstatus;
#elif HAVE_SYS_WAIT_H
int wstatus;
int waitpid_options = 0;
pid_t wait_pid;
#endif
/* Close all handles to avoid a deadlock */
for (i = 0; i < proc->npipes; i++) {
if (proc->pipes[i] != 0) {
2014-05-08 22:24:01 +00:00
GC_REFCOUNT(proc->pipes[i])--;
zend_list_close(proc->pipes[i]);
proc->pipes[i] = 0;
}
}
2010-08-30 16:52:48 +00:00
#ifdef PHP_WIN32
if (FG(pclose_wait)) {
WaitForSingleObject(proc->childHandle, INFINITE);
}
GetExitCodeProcess(proc->childHandle, &wstatus);
if (wstatus == STILL_ACTIVE) {
FG(pclose_ret) = -1;
} else {
FG(pclose_ret) = wstatus;
}
CloseHandle(proc->childHandle);
2010-08-30 16:52:48 +00:00
#elif HAVE_SYS_WAIT_H
2010-08-30 16:52:48 +00:00
if (!FG(pclose_wait)) {
waitpid_options = WNOHANG;
}
do {
wait_pid = waitpid(proc->child, &wstatus, waitpid_options);
} while (wait_pid == -1 && errno == EINTR);
2010-08-30 16:52:48 +00:00
if (wait_pid <= 0) {
FG(pclose_ret) = -1;
2009-07-20 04:31:07 +00:00
} else {
if (WIFEXITED(wstatus))
wstatus = WEXITSTATUS(wstatus);
FG(pclose_ret) = wstatus;
}
2010-08-30 16:52:48 +00:00
#else
FG(pclose_ret) = -1;
#endif
_php_free_envp(proc->env, proc->is_persistent);
pefree(proc->pipes, proc->is_persistent);
pefree(proc->command, proc->is_persistent);
pefree(proc, proc->is_persistent);
2010-08-30 16:52:48 +00:00
}
2003-07-12 07:01:30 +00:00
/* }}} */
2003-07-12 07:01:30 +00:00
/* {{{ PHP_MINIT_FUNCTION(proc_open) */
PHP_MINIT_FUNCTION(proc_open)
{
le_proc_open = zend_register_list_destructors_ex(proc_open_rsrc_dtor, NULL, "process", module_number);
return SUCCESS;
}
2003-07-12 07:01:30 +00:00
/* }}} */
/* {{{ proto bool proc_terminate(resource process [, long signal])
kill a process opened by proc_open */
PHP_FUNCTION(proc_terminate)
{
zval *zproc;
struct php_process_handle *proc;
2014-08-25 17:24:55 +00:00
zend_long sig_no = SIGTERM;
2010-08-30 16:52:48 +00:00
2014-12-13 22:06:14 +00:00
if (zend_parse_parameters(ZEND_NUM_ARGS(), "r|l", &zproc, &sig_no) == FAILURE) {
RETURN_FALSE;
}
2015-02-02 05:23:16 +00:00
if ((proc = (struct php_process_handle *)zend_fetch_resource(Z_RES_P(zproc), "process", le_proc_open)) == NULL) {
RETURN_FALSE;
}
2010-08-30 16:52:48 +00:00
#ifdef PHP_WIN32
if (TerminateProcess(proc->childHandle, 255)) {
RETURN_TRUE;
} else {
RETURN_FALSE;
}
#else
if (kill(proc->child, sig_no) == 0) {
RETURN_TRUE;
} else {
RETURN_FALSE;
}
#endif
}
/* }}} */
/* {{{ proto int proc_close(resource process)
close a process opened by proc_open */
PHP_FUNCTION(proc_close)
{
zval *zproc;
struct php_process_handle *proc;
2010-08-30 16:52:48 +00:00
2014-12-13 22:06:14 +00:00
if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &zproc) == FAILURE) {
RETURN_FALSE;
}
2015-02-02 05:23:16 +00:00
if ((proc = (struct php_process_handle *)zend_fetch_resource(Z_RES_P(zproc), "process", le_proc_open)) == NULL) {
RETURN_FALSE;
}
2010-08-30 16:52:48 +00:00
FG(pclose_wait) = 1;
2014-05-08 22:24:01 +00:00
zend_list_close(Z_RES_P(zproc));
FG(pclose_wait) = 0;
2014-08-25 17:24:55 +00:00
RETURN_LONG(FG(pclose_ret));
}
/* }}} */
/* {{{ proto array proc_get_status(resource process)
get information about a process opened by proc_open */
PHP_FUNCTION(proc_get_status)
{
zval *zproc;
struct php_process_handle *proc;
#ifdef PHP_WIN32
DWORD wstatus;
#elif HAVE_SYS_WAIT_H
int wstatus;
pid_t wait_pid;
#endif
int running = 1, signaled = 0, stopped = 0;
int exitcode = -1, termsig = 0, stopsig = 0;
2010-08-30 16:52:48 +00:00
2014-12-13 22:06:14 +00:00
if (zend_parse_parameters(ZEND_NUM_ARGS(), "r", &zproc) == FAILURE) {
RETURN_FALSE;
}
2015-02-02 05:23:16 +00:00
if ((proc = (struct php_process_handle *)zend_fetch_resource(Z_RES_P(zproc), "process", le_proc_open)) == NULL) {
RETURN_FALSE;
}
array_init(return_value);
2014-04-15 11:40:40 +00:00
add_assoc_string(return_value, "command", proc->command);
2014-08-25 17:24:55 +00:00
add_assoc_long(return_value, "pid", (zend_long) proc->child);
2010-08-30 16:52:48 +00:00
#ifdef PHP_WIN32
2010-08-30 16:52:48 +00:00
GetExitCodeProcess(proc->childHandle, &wstatus);
running = wstatus == STILL_ACTIVE;
exitcode = running ? -1 : wstatus;
2010-08-30 16:52:48 +00:00
#elif HAVE_SYS_WAIT_H
2010-08-30 16:52:48 +00:00
errno = 0;
wait_pid = waitpid(proc->child, &wstatus, WNOHANG|WUNTRACED);
2010-08-30 16:52:48 +00:00
if (wait_pid == proc->child) {
if (WIFEXITED(wstatus)) {
running = 0;
exitcode = WEXITSTATUS(wstatus);
}
if (WIFSIGNALED(wstatus)) {
running = 0;
signaled = 1;
#ifdef NETWARE
termsig = WIFTERMSIG(wstatus);
#else
termsig = WTERMSIG(wstatus);
#endif
}
if (WIFSTOPPED(wstatus)) {
stopped = 1;
stopsig = WSTOPSIG(wstatus);
}
} else if (wait_pid == -1) {
running = 0;
}
#endif
add_assoc_bool(return_value, "running", running);
add_assoc_bool(return_value, "signaled", signaled);
add_assoc_bool(return_value, "stopped", stopped);
2014-08-25 17:24:55 +00:00
add_assoc_long(return_value, "exitcode", exitcode);
add_assoc_long(return_value, "termsig", termsig);
add_assoc_long(return_value, "stopsig", stopsig);
}
/* }}} */
/* {{{ handy definitions for portability/readability */
#ifdef PHP_WIN32
# define pipe(pair) (CreatePipe(&pair[0], &pair[1], &security, 0) ? 0 : -1)
# define COMSPEC_NT "cmd.exe"
static inline HANDLE dup_handle(HANDLE src, BOOL inherit, BOOL closeorig)
{
HANDLE copy, self = GetCurrentProcess();
2010-08-30 16:52:48 +00:00
if (!DuplicateHandle(self, src, self, &copy, 0, inherit, DUPLICATE_SAME_ACCESS |
(closeorig ? DUPLICATE_CLOSE_SOURCE : 0)))
return NULL;
return copy;
}
static inline HANDLE dup_fd_as_handle(int fd)
{
return dup_handle((HANDLE)_get_osfhandle(fd), TRUE, FALSE);
}
# define close_descriptor(fd) CloseHandle(fd)
#else
# define close_descriptor(fd) close(fd)
#endif
#define DESC_PIPE 1
#define DESC_FILE 2
#define DESC_PARENT_MODE_WRITE 8
struct php_proc_open_descriptor_item {
int index; /* desired fd number in child process */
php_file_descriptor_t parentend, childend; /* fds for pipes in parent/child */
int mode; /* mode for proc_open code */
int mode_flags; /* mode flags for opening fds */
};
/* }}} */
2003-12-06 10:48:56 +00:00
/* {{{ proto resource proc_open(string command, array descriptorspec, array &pipes [, string cwd [, array env [, array other_options]]])
Run a process with more control over it's file descriptors */
PHP_FUNCTION(proc_open)
{
char *command, *cwd=NULL;
size_t command_len, cwd_len = 0;
zval *descriptorspec;
zval *pipes;
zval *environment = NULL;
zval *other_options = NULL;
php_process_env_t env;
int ndesc = 0;
int i;
zval *descitem = NULL;
zend_string *str_index;
2014-08-25 17:24:55 +00:00
zend_ulong nindex;
struct php_proc_open_descriptor_item *descriptors = NULL;
int ndescriptors_array;
#ifdef PHP_WIN32
PROCESS_INFORMATION pi;
HANDLE childHandle;
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
STARTUPINFOW si;
BOOL newprocok;
SECURITY_ATTRIBUTES security;
DWORD dwCreateFlags = 0;
UINT old_error_mode;
2013-01-07 12:03:12 +00:00
char cur_cwd[MAXPATHLEN];
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
wchar_t *cmdw = NULL, *cwdw = NULL, *envpw = NULL;
size_t tmp_len;
2004-09-29 06:04:36 +00:00
#endif
#ifdef NETWARE
2009-07-20 11:48:04 +00:00
char** child_argv = NULL;
2004-09-29 06:04:36 +00:00
char* command_dup = NULL;
char* orig_cwd = NULL;
int command_num_args = 0;
wiring_t channel;
#endif
php_process_id_t child;
struct php_process_handle *proc;
int is_persistent = 0; /* TODO: ensure that persistent procs will work */
#ifdef PHP_WIN32
2003-12-05 13:48:40 +00:00
int suppress_errors = 0;
int bypass_shell = 0;
int blocking_pipes = 0;
2009-07-20 11:48:04 +00:00
#endif
#if PHP_CAN_DO_PTS
php_file_descriptor_t dev_ptmx = -1; /* master */
php_file_descriptor_t slave_pty = -1;
#endif
2014-12-13 22:06:14 +00:00
if (zend_parse_parameters(ZEND_NUM_ARGS(), "saz/|s!a!a!", &command,
2009-07-20 11:48:04 +00:00
&command_len, &descriptorspec, &pipes, &cwd, &cwd_len, &environment,
&other_options) == FAILURE) {
RETURN_FALSE;
}
command = pestrdup(command, is_persistent);
2009-07-20 11:48:04 +00:00
#ifdef PHP_WIN32
if (other_options) {
zval *item = zend_hash_str_find(Z_ARRVAL_P(other_options), "suppress_errors", sizeof("suppress_errors") - 1);
if (item != NULL) {
2014-08-25 17:24:55 +00:00
if (Z_TYPE_P(item) == IS_TRUE || ((Z_TYPE_P(item) == IS_LONG) && Z_LVAL_P(item))) {
suppress_errors = 1;
}
2010-08-30 16:52:48 +00:00
}
item = zend_hash_str_find(Z_ARRVAL_P(other_options), "bypass_shell", sizeof("bypass_shell") - 1);
if (item != NULL) {
2014-08-25 17:24:55 +00:00
if (Z_TYPE_P(item) == IS_TRUE || ((Z_TYPE_P(item) == IS_LONG) && Z_LVAL_P(item))) {
bypass_shell = 1;
}
2010-08-30 16:52:48 +00:00
}
item = zend_hash_str_find(Z_ARRVAL_P(other_options), "blocking_pipes", sizeof("blocking_pipes") - 1);
if (item != NULL) {
if (Z_TYPE_P(item) == IS_TRUE || ((Z_TYPE_P(item) == IS_LONG) && Z_LVAL_P(item))) {
blocking_pipes = 1;
}
}
}
#endif
2010-08-30 16:52:48 +00:00
command_len = strlen(command);
if (environment) {
2014-12-13 22:06:14 +00:00
env = _php_array_to_envp(environment, is_persistent);
} else {
memset(&env, 0, sizeof(env));
}
ndescriptors_array = zend_hash_num_elements(Z_ARRVAL_P(descriptorspec));
descriptors = safe_emalloc(sizeof(struct php_proc_open_descriptor_item), ndescriptors_array, 0);
memset(descriptors, 0, sizeof(struct php_proc_open_descriptor_item) * ndescriptors_array);
#ifdef PHP_WIN32
/* we use this to allow the child to inherit handles */
memset(&security, 0, sizeof(security));
security.nLength = sizeof(security);
security.bInheritHandle = TRUE;
security.lpSecurityDescriptor = NULL;
#endif
2010-08-30 16:52:48 +00:00
/* walk the descriptor spec and set up files/pipes */
ZEND_HASH_FOREACH_KEY_VAL(Z_ARRVAL_P(descriptorspec), nindex, str_index, descitem) {
zval *ztype;
if (str_index) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "descriptor spec must be an integer indexed array");
goto exit_fail;
}
2014-10-23 14:29:02 +00:00
descriptors[ndesc].index = (int)nindex;
if (Z_TYPE_P(descitem) == IS_RESOURCE) {
/* should be a stream - try and dup the descriptor */
php_stream *stream;
php_socket_t fd;
php_stream_from_zval(stream, descitem);
if (FAILURE == php_stream_cast(stream, PHP_STREAM_AS_FD, (void **)&fd, REPORT_ERRORS)) {
goto exit_fail;
}
#ifdef PHP_WIN32
2014-10-23 14:29:02 +00:00
descriptors[ndesc].childend = dup_fd_as_handle((int)fd);
if (descriptors[ndesc].childend == NULL) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "unable to dup File-Handle for descriptor %d", nindex);
goto exit_fail;
}
#else
descriptors[ndesc].childend = dup(fd);
if (descriptors[ndesc].childend < 0) {
php_error_docref(NULL, E_WARNING, "unable to dup File-Handle for descriptor " ZEND_ULONG_FMT " - %s", nindex, strerror(errno));
goto exit_fail;
}
#endif
descriptors[ndesc].mode = DESC_FILE;
} else if (Z_TYPE_P(descitem) != IS_ARRAY) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Descriptor item must be either an array or a File-Handle");
goto exit_fail;
} else {
if ((ztype = zend_hash_index_find(Z_ARRVAL_P(descitem), 0)) != NULL) {
convert_to_string_ex(ztype);
} else {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Missing handle qualifier in array");
goto exit_fail;
}
if (strcmp(Z_STRVAL_P(ztype), "pipe") == 0) {
php_file_descriptor_t newpipe[2];
zval *zmode;
if ((zmode = zend_hash_index_find(Z_ARRVAL_P(descitem), 1)) != NULL) {
convert_to_string_ex(zmode);
} else {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Missing mode parameter for 'pipe'");
goto exit_fail;
}
descriptors[ndesc].mode = DESC_PIPE;
if (0 != pipe(newpipe)) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "unable to create pipe %s", strerror(errno));
goto exit_fail;
}
if (strncmp(Z_STRVAL_P(zmode), "w", 1) != 0) {
descriptors[ndesc].parentend = newpipe[1];
descriptors[ndesc].childend = newpipe[0];
descriptors[ndesc].mode |= DESC_PARENT_MODE_WRITE;
} else {
descriptors[ndesc].parentend = newpipe[0];
descriptors[ndesc].childend = newpipe[1];
}
#ifdef PHP_WIN32
/* don't let the child inherit the parent side of the pipe */
descriptors[ndesc].parentend = dup_handle(descriptors[ndesc].parentend, FALSE, TRUE);
#endif
descriptors[ndesc].mode_flags = descriptors[ndesc].mode & DESC_PARENT_MODE_WRITE ? O_WRONLY : O_RDONLY;
#ifdef PHP_WIN32
2014-08-25 17:24:55 +00:00
if (Z_STRLEN_P(zmode) >= 2 && Z_STRVAL_P(zmode)[1] == 'b')
descriptors[ndesc].mode_flags |= O_BINARY;
#endif
} else if (strcmp(Z_STRVAL_P(ztype), "file") == 0) {
zval *zfile, *zmode;
php_socket_t fd;
php_stream *stream;
descriptors[ndesc].mode = DESC_FILE;
if ((zfile = zend_hash_index_find(Z_ARRVAL_P(descitem), 1)) != NULL) {
convert_to_string_ex(zfile);
} else {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Missing file name parameter for 'file'");
goto exit_fail;
}
if ((zmode = zend_hash_index_find(Z_ARRVAL_P(descitem), 2)) != NULL) {
convert_to_string_ex(zmode);
} else {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Missing mode parameter for 'file'");
goto exit_fail;
}
/* try a wrapper */
stream = php_stream_open_wrapper(Z_STRVAL_P(zfile), Z_STRVAL_P(zmode),
REPORT_ERRORS|STREAM_WILL_CAST, NULL);
/* force into an fd */
if (stream == NULL || FAILURE == php_stream_cast(stream,
PHP_STREAM_CAST_RELEASE|PHP_STREAM_AS_FD,
(void **)&fd, REPORT_ERRORS)) {
goto exit_fail;
}
#ifdef PHP_WIN32
2014-10-23 14:29:02 +00:00
descriptors[ndesc].childend = dup_fd_as_handle((int)fd);
_close((int)fd);
/* simulate the append mode by fseeking to the end of the file
this introduces a potential race-condition, but it is the best we can do, though */
if (strchr(Z_STRVAL_P(zmode), 'a')) {
SetFilePointer(descriptors[ndesc].childend, 0, NULL, FILE_END);
}
#else
descriptors[ndesc].childend = fd;
#endif
} else if (strcmp(Z_STRVAL_P(ztype), "pty") == 0) {
#if PHP_CAN_DO_PTS
if (dev_ptmx == -1) {
/* open things up */
dev_ptmx = open("/dev/ptmx", O_RDWR);
if (dev_ptmx == -1) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "failed to open /dev/ptmx, errno %d", errno);
goto exit_fail;
}
grantpt(dev_ptmx);
unlockpt(dev_ptmx);
slave_pty = open(ptsname(dev_ptmx), O_RDWR);
if (slave_pty == -1) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "failed to open slave pty, errno %d", errno);
goto exit_fail;
}
}
descriptors[ndesc].mode = DESC_PIPE;
descriptors[ndesc].childend = dup(slave_pty);
descriptors[ndesc].parentend = dup(dev_ptmx);
descriptors[ndesc].mode_flags = O_RDWR;
#else
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "pty pseudo terminal not supported on this system");
goto exit_fail;
#endif
} else {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "%s is not a valid descriptor spec/mode", Z_STRVAL_P(ztype));
goto exit_fail;
}
}
ndesc++;
} ZEND_HASH_FOREACH_END();
#ifdef PHP_WIN32
if (cwd == NULL) {
char *getcwd_result;
getcwd_result = VCWD_GETCWD(cur_cwd, MAXPATHLEN);
if (!getcwd_result) {
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "Cannot get current directory");
goto exit_fail;
}
2013-01-07 12:03:12 +00:00
cwd = cur_cwd;
}
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
cwdw = php_win32_cp_any_to_w(cwd);
if (!cwdw) {
php_error_docref(NULL, E_WARNING, "CWD conversion failed");
goto exit_fail;
}
memset(&si, 0, sizeof(si));
si.cb = sizeof(si);
si.dwFlags = STARTF_USESTDHANDLES;
2010-08-30 16:52:48 +00:00
si.hStdInput = GetStdHandle(STD_INPUT_HANDLE);
si.hStdOutput = GetStdHandle(STD_OUTPUT_HANDLE);
si.hStdError = GetStdHandle(STD_ERROR_HANDLE);
2010-08-30 16:52:48 +00:00
/* redirect stdin/stdout/stderr if requested */
for (i = 0; i < ndesc; i++) {
switch(descriptors[i].index) {
case 0:
si.hStdInput = descriptors[i].childend;
break;
case 1:
si.hStdOutput = descriptors[i].childend;
break;
case 2:
si.hStdError = descriptors[i].childend;
break;
}
}
2010-08-30 16:52:48 +00:00
memset(&pi, 0, sizeof(pi));
2010-08-30 16:52:48 +00:00
if (suppress_errors) {
old_error_mode = SetErrorMode(SEM_FAILCRITICALERRORS|SEM_NOGPFAULTERRORBOX);
}
2010-08-30 16:52:48 +00:00
dwCreateFlags = NORMAL_PRIORITY_CLASS;
if(strcmp(sapi_module.name, "cli") != 0) {
dwCreateFlags |= CREATE_NO_WINDOW;
}
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
envpw = php_win32_cp_env_any_to_w(env.envp);
if (envpw) {
dwCreateFlags |= CREATE_UNICODE_ENVIRONMENT;
} else {
if (env.envp) {
php_error_docref(NULL, E_WARNING, "ENV conversion failed");
goto exit_fail;
}
}
cmdw = php_win32_cp_conv_any_to_w(command, command_len, &tmp_len);
if (!cmdw) {
php_error_docref(NULL, E_WARNING, "Command conversion failed");
goto exit_fail;
}
if (bypass_shell) {
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
newprocok = CreateProcessW(NULL, cmdw, &security, &security, TRUE, dwCreateFlags, envpw, cwdw, &si, &pi);
} else {
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
int ret;
size_t len;
wchar_t *cmdw2;
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
len = (sizeof(COMSPEC_NT) + sizeof(" /c ") + tmp_len + 1);
cmdw2 = (wchar_t *)malloc(len * sizeof(wchar_t));
2016-08-29 15:25:46 +00:00
if (!cmdw2) {
php_error_docref(NULL, E_WARNING, "Command conversion failed");
goto exit_fail;
}
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
ret = _snwprintf(cmdw2, len, L"%hs /c %s", COMSPEC_NT, cmdw);
if (-1 == ret) {
2016-08-29 15:25:46 +00:00
free(cmdw2);
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
php_error_docref(NULL, E_WARNING, "Command conversion failed");
goto exit_fail;
}
2010-08-30 16:52:48 +00:00
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
newprocok = CreateProcessW(NULL, cmdw2, &security, &security, TRUE, dwCreateFlags, envpw, cwdw, &si, &pi);
free(cmdw2);
}
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
free(cwdw);
cwdw = NULL;
free(cmdw);
cmdw = NULL;
free(envpw);
envpw = NULL;
if (suppress_errors) {
SetErrorMode(old_error_mode);
}
2010-08-30 16:52:48 +00:00
if (FALSE == newprocok) {
2008-08-21 11:27:57 +00:00
DWORD dw = GetLastError();
/* clean up all the descriptors */
for (i = 0; i < ndesc; i++) {
CloseHandle(descriptors[i].childend);
if (descriptors[i].parentend) {
CloseHandle(descriptors[i].parentend);
}
}
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "CreateProcess failed, error code - %u", dw);
goto exit_fail;
}
childHandle = pi.hProcess;
child = pi.dwProcessId;
CloseHandle(pi.hThread);
#elif defined(NETWARE)
if (cwd) {
orig_cwd = getcwd(NULL, PATH_MAX);
chdir2(cwd);
}
channel.infd = descriptors[0].childend;
channel.outfd = descriptors[1].childend;
channel.errfd = -1;
/* Duplicate the command as processing downwards will modify it*/
command_dup = strdup(command);
2007-02-24 03:24:40 +00:00
if (!command_dup) {
goto exit_fail;
}
/* get a number of args */
construct_argc_argv(command_dup, NULL, &command_num_args, NULL);
child_argv = (char**) malloc((command_num_args + 1) * sizeof(char*));
if(!child_argv) {
free(command_dup);
if (cwd && orig_cwd) {
chdir2(orig_cwd);
free(orig_cwd);
}
}
/* fill the child arg vector */
construct_argc_argv(command_dup, NULL, &command_num_args, child_argv);
child_argv[command_num_args] = NULL;
child = procve(child_argv[0], PROC_DETACHED|PROC_INHERIT_CWD, NULL, &channel, NULL, NULL, 0, NULL, (const char**)child_argv);
free(child_argv);
free(command_dup);
if (cwd && orig_cwd) {
chdir2(orig_cwd);
free(orig_cwd);
}
if (child < 0) {
/* failed to fork() */
/* clean up all the descriptors */
for (i = 0; i < ndesc; i++) {
close(descriptors[i].childend);
if (descriptors[i].parentend)
close(descriptors[i].parentend);
}
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "procve failed - %s", strerror(errno));
goto exit_fail;
}
#elif HAVE_FORK
/* the unix way */
child = fork();
if (child == 0) {
/* this is the child process */
#if PHP_CAN_DO_PTS
if (dev_ptmx >= 0) {
int my_pid = getpid();
2004-05-29 00:30:11 +00:00
#ifdef TIOCNOTTY
/* detach from original tty. Might only need this if isatty(0) is true */
ioctl(0,TIOCNOTTY,NULL);
2004-05-29 00:30:11 +00:00
#else
setsid();
#endif
/* become process group leader */
setpgid(my_pid, my_pid);
tcsetpgrp(0, my_pid);
}
#endif
2010-08-30 16:52:48 +00:00
/* close those descriptors that we just opened for the parent stuff,
* dup new descriptors into required descriptors and close the original
* cruft */
for (i = 0; i < ndesc; i++) {
switch (descriptors[i].mode & ~DESC_PARENT_MODE_WRITE) {
case DESC_PIPE:
close(descriptors[i].parentend);
break;
}
if (dup2(descriptors[i].childend, descriptors[i].index) < 0)
perror("dup2");
if (descriptors[i].childend != descriptors[i].index)
close(descriptors[i].childend);
}
#if PHP_CAN_DO_PTS
if (dev_ptmx >= 0) {
close(dev_ptmx);
close(slave_pty);
}
#endif
2010-08-30 16:52:48 +00:00
if (cwd) {
php_ignore_value(chdir(cwd));
}
2010-08-30 16:52:48 +00:00
2009-07-20 11:48:04 +00:00
if (env.envarray) {
execle("/bin/sh", "sh", "-c", command, NULL, env.envarray);
} else {
execl("/bin/sh", "sh", "-c", command, NULL);
}
_exit(127);
} else if (child < 0) {
/* failed to fork() */
/* clean up all the descriptors */
for (i = 0; i < ndesc; i++) {
close(descriptors[i].childend);
if (descriptors[i].parentend)
close(descriptors[i].parentend);
}
2014-12-13 22:06:14 +00:00
php_error_docref(NULL, E_WARNING, "fork failed - %s", strerror(errno));
goto exit_fail;
}
#else
# error You lose (configure should not have let you get here)
#endif
/* we forked/spawned and this is the parent */
proc = (struct php_process_handle*)pemalloc(sizeof(struct php_process_handle), is_persistent);
proc->is_persistent = is_persistent;
proc->command = command;
proc->pipes = pemalloc(sizeof(zend_resource *) * ndesc, is_persistent);
proc->npipes = ndesc;
proc->child = child;
#ifdef PHP_WIN32
proc->childHandle = childHandle;
#endif
proc->env = env;
if (pipes != NULL) {
zval_dtor(pipes);
2015-01-03 09:22:58 +00:00
}
2014-03-03 08:50:35 +00:00
array_init(pipes);
#if PHP_CAN_DO_PTS
if (dev_ptmx >= 0) {
close(dev_ptmx);
close(slave_pty);
}
#endif
/* clean up all the child ends and then open streams on the parent
* ends, where appropriate */
for (i = 0; i < ndesc; i++) {
char *mode_string=NULL;
php_stream *stream = NULL;
close_descriptor(descriptors[i].childend);
switch (descriptors[i].mode & ~DESC_PARENT_MODE_WRITE) {
case DESC_PIPE:
switch(descriptors[i].mode_flags) {
#ifdef PHP_WIN32
case O_WRONLY|O_BINARY:
mode_string = "wb";
break;
case O_RDONLY|O_BINARY:
mode_string = "rb";
break;
#endif
case O_WRONLY:
mode_string = "w";
break;
case O_RDONLY:
mode_string = "r";
break;
case O_RDWR:
mode_string = "r+";
break;
}
#ifdef PHP_WIN32
2007-04-16 08:09:56 +00:00
stream = php_stream_fopen_from_fd(_open_osfhandle((zend_intptr_t)descriptors[i].parentend,
descriptors[i].mode_flags), mode_string, NULL);
2015-06-30 15:21:21 +00:00
php_stream_set_option(stream, PHP_STREAM_OPTION_PIPE_BLOCKING, blocking_pipes, NULL);
#else
stream = php_stream_fopen_from_fd(descriptors[i].parentend, mode_string, NULL);
# if defined(F_SETFD) && defined(FD_CLOEXEC)
/* mark the descriptor close-on-exec, so that it won't be inherited by potential other children */
fcntl(descriptors[i].parentend, F_SETFD, FD_CLOEXEC);
# endif
#endif
if (stream) {
zval retfp;
/* nasty hack; don't copy it */
stream->flags |= PHP_STREAM_FLAG_NO_SEEK;
2010-08-30 16:52:48 +00:00
php_stream_to_zval(stream, &retfp);
add_index_zval(pipes, descriptors[i].index, &retfp);
2014-04-09 11:05:22 +00:00
proc->pipes[i] = Z_RES(retfp);
Z_ADDREF(retfp);
}
break;
default:
proc->pipes[i] = NULL;
}
}
efree(descriptors);
2015-02-02 05:23:16 +00:00
ZVAL_RES(return_value, zend_register_resource(proc, le_proc_open));
return;
exit_fail:
efree(descriptors);
_php_free_envp(env, is_persistent);
pefree(command, is_persistent);
Fixed the UTF-8 and long path support in the streams on Windows. Since long the default PHP charset is UTF-8, however the Windows part is out of step with this important point. The current implementation in PHP doesn't technically permit to handle UTF-8 filepath and several other things. Till now, only the ANSI compatible APIs are being used. Here is more about it https://msdn.microsoft.com/en-us/library/windows/desktop/dd317752%28v=vs.85%29.aspx The patch fixes not only issues with multibyte filenames under incompatible codepages, but indirectly also issues with some other multibyte encodings like BIG5, Shift-JIS, etc. by providing a clean way to access filenames in UTF-8. Below is a small list of issues from the bug tracker, that are getting fixed: https://bugs.php.net/63401 https://bugs.php.net/41199 https://bugs.php.net/50203 https://bugs.php.net/71509 https://bugs.php.net/64699 https://bugs.php.net/64506 https://bugs.php.net/30195 https://bugs.php.net/65358 https://bugs.php.net/61315 https://bugs.php.net/70943 https://bugs.php.net/70903 https://bugs.php.net/63593 https://bugs.php.net/54977 https://bugs.php.net/54028 https://bugs.php.net/43148 https://bugs.php.net/30730 https://bugs.php.net/33350 https://bugs.php.net/35300 https://bugs.php.net/46990 https://bugs.php.net/61309 https://bugs.php.net/69333 https://bugs.php.net/45517 https://bugs.php.net/70551 https://bugs.php.net/50197 https://bugs.php.net/72200 https://bugs.php.net/37672 Yet more related tickets can for sure be found - on bugs.php.net, Stackoverflow and Github. Some of the bugs are pretty recent, some descend to early 2000th, but the user comments in there last even till today. Just for example, bug #30195 was opened in 2004, the latest comment in there was made in 2014. It is certain, that these bugs descend not only to pure PHP use cases, but get also redirected from the popular PHP based projects. Given the modern systems (and those supported by PHP) are always based on NTFS, there is no excuse to keep these issues unresolved. The internalization approach on Windows is in many ways different from UNIX and Linux, while it supports and is based on Unicode. It depends on the current system code page, APIs used and exact kind how the binary was compiled The locale doesn't affect the way Unicode or ANSI API work. PHP in particular is being compiled without _UNICODE defined and this is conditioned by the way we handle strings. Here is more about it https://msdn.microsoft.com/en-us/library/tsbaswba.aspx However, with any system code page ANSI functions automatically convert paths to UTF-16. Paths in some encodings incompatible with the current system code page, won't work correctly with ANSI APIs. PHP till now only uses the ANSI Windows APIs. For example, on a system with the current code page 1252, the paths in cp1252 are supported and transparently converted to UTF-16 by the ANSI functions. Once one wants to handle a filepath encoded with cp932 on that particular system, an ANSI or a POSIX compatible function used in PHP will produce an erroneous result. When trying to convert that cp932 path to UTF-8 and passing to the ANSI functions, an ANSI function would likely interpret the UTF-8 string as some string in the current code page and create a filepath that represents every single byte of the UTF-8 string. These behaviors are not only broken but also disregard the documented INI settings. This patch solves the issies with the multibyte paths on Windows by intelligently enforcing the usage of the Unicode aware APIs. For functions expect Unicode (fe CreateFileW, FindFirstFileW, etc.), arguments will be converted to UTF-16 wide chars. For functions returning Unicode aware data (fe GetCurrentDirectoryW, etc.), resulting wide string is converted back to char's depending on the current PHP charset settings, either to the current ANSI codepage (this is the behavior prior to this patch) or to UTF-8 (the default behavior). In a particular case, users might have to explicitly set internal_encoding or default_charset, if filenames in ANSI codepage are necessary. Current tests show no regressions and witness that this will be an exotic case, the current default UTF-8 encoding is compatible with any supported system. The dependency libraries are long switching to Unicode APIs, so some tests were also added for extensions not directly related to streams. At large, the patch brings over 150 related tests into the core. Those target and was run on various environments with European, Asian, etc. codepages. General PHP frameworks was tested and showed no regressions. The impact on the current C code base is low, the most places affected are the Windows only places in the three files tsrm_win32.c, zend_virtual_cwd.c and plain_wrapper.c. The actual implementation of the most of the wide char supporting functionality is in win32/ioutil.* and win32/codepage.*, several low level functionsare extended in place to avoid reimplementation for now. No performance impact was sighted. As previously mentioned, the ANSI APIs used prior the patch perform Unicode conversions internally. Using the Unicode APIs directly while doing custom conversions just retains the status quo. The ways to optimize it are open (fe. by implementing caching for the strings converted to wide variants). The long path implementation is user transparent. If a path exceeds the length of _MAX_PATH, it'll be automatically prefixed with \\?\. The MAXPATHLEN is set to 2048 bytes. Appreciation to Pierre Joye, Matt Ficken, @algo13 and others for tips, ideas and testing. Thanks.
2016-06-20 07:32:19 +00:00
#ifdef PHP_WIN32
free(cwdw);
free(cmdw);
free(envpw);
#endif
#if PHP_CAN_DO_PTS
if (dev_ptmx >= 0) {
close(dev_ptmx);
}
if (slave_pty >= 0) {
close(slave_pty);
}
#endif
RETURN_FALSE;
}
/* }}} */
#endif /* PHP_CAN_SUPPORT_PROC_OPEN */
/*
* Local variables:
* tab-width: 4
* c-basic-offset: 4
* End:
* vim600: sw=4 ts=4 fdm=marker
* vim<600: sw=4 ts=4
*/