In the Windows version of my current personal project, I'm looking to support extended length filepaths. As a result, I'm a little confused with how to use the GetFullPathNameW API to resolve the full name of a long filepath.
According to the MSDN (with regards to the lpFileName parameter):
In the ANSI version of this function, the name is limited to MAX_PATH characters. To extend this limit to 32,767 wide characters, call the Unicode version of the function and prepend "\?\" to the path. For more information, see Naming a File.
If I'm understanding this correctly, in order to use an extended length filepath with GetFullPathNameW, I need to specify a path with the \\?\ prefix attached. Since the \\?\ prefix is only valid before volume letters or UNC paths, this would mean that the API is unusable for resolving the full name of a path relative to the current directory.
If that's the case, is there another API I can use to resolve the full name of a filepath like ..\somedir\somefile.txt if the resulting name's length exceeds MAX_PATH? If not, would I be able to combine GetCurrentDirectory with the relative filepath (\\?\C:\my\cwd\..\somedir\somefile.txt) and use it with GetFullPathNameW, or would I need to handle all of the filepath resolution on my own?
GetFullPathNameA is limited to MAX_PATH characters, because it converts the ANSI name to a UNICODE name beforehand using a hardcoded MAX_PATH-sized (in chars) UNICODE buffer. If the conversion doesn't fail due to the length restrictions, then GetFullPathNameW (or direct GetFullPathName_U[Ex]) is called and the resulting UNICODE name is converted to ANSI.
GetFullPathNameW is a very thin shell over GetFullPathName_U. It is limited to MAXSHORT (0x7fff) length in WCHARs, independent of the \\?\ file prefix. Even without \\?\, it will be work for long (> MAX_PATH) relative names. However, if the lpFileName parameter does not begin with the \\?\ prefix, the result name in the lpBuffer parameter will not begin with \\?\ either.
if you will be use lpBuffer with functions like CreateFileW - this function internally convert Win32Name to NtName. and result will be depended from nape type (RTL_PATH_TYPE). if the name does not begin with \\?\ prefix, the conversion fails because RtlDosPathNameToRelativeNtPathName_U[_WithStatus] fails (because if the path not begin with \\?\ it will be internally call GetFullPathName_U (same function called by GetFullPathNameW) with nBufferLength hardcoded to MAX_PATH (exactly 2*MAX_PATH in bytes – NTDLL functions use buffer size in bytes, not in WCHARs). If name begin with \\?\ prefix, another case in RtlDosPathNameToRelativeNtPathName_U[_WithStatus] is executed – RtlpWin32NtNameToNtPathName, which replaces \\?\ with \??\ and has no MAX_PATH limitation
So the solution may look like this:
if(ULONG len = GetFullPathNameW(FileName, 0, 0, 0))
{
PWSTR buf = (PWSTR)_alloca((4 + len) * sizeof(WCHAR));
buf[0] = L'\\', buf[1] = L'\\', buf[2] = L'?', buf[3] = L'\\';
if (len - 1 == GetFullPathName(FileName, len, buf + 4, &c))
{
CreateFile(buf, ...);
}
}
So we need to specify a path with the \\?\ prefix attached, but not before GetFullPathName - after!
For more info, read this - The Definitive Guide on Win32 to NT Path Conversion
Just to update with the current state:
Starting in Windows 10, version 1607, MAX_PATH limitations have been removed from common Win32 file and directory functions. However, you must opt-in to the new behavior. To enable the new long path behavior, both of the following conditions must be met: ...
For the rest, please see my answer here: https://stackoverflow.com/a/57624626/3736444
Related
I am struggling to understand what may be causing the issue in my case.
I am using
ShellExecuteW('open', 'explorer.exe', '/select,[file_name]', None, win32con.SW_SHOW)
What I am trying to do is open the file in the OS, highlight it, and bring the File Explorer to the foreground. This works fine for most cases, but when I try to open a file that exceeds the MAX_PATH limit (260 characters), the file doesn't open, and instead it takes me to the "My Files" page.
I have tried prepending "\\?\" to the beginning of my file name, because that is what other Stack Overflow posts said to do with regards to overriding the MAX_PATH limit, but it has not changed the situation.
Does the ShellExecuteW function not allow for files that exceed MAX_PATH? And, if so, is there any workaround I could use?
I read some cases, about this issue. Find this article:Long Paths in .NET, Part 1 of 3 [Kim Hamilton]
If you prefix the file name with "\?\" and call the Unicode versions of the Windows APIs, then you can use file names up to 32K characters in length. In other words, the \?\ prefix is a way to enable long paths while working with the Windows file APIs.
and:
Long paths with the \?\ prefix can be used in most of the file-related Windows APIs, but not all Windows APIs.
I also test ShellExcuteW with \\?\,it failed.
Working well with SHOpenFolderAndSelectItems
CoInitialize(NULL);
LPCWSTR file_name ;//Change the path according to your needs
PIDLIST_ABSOLUTE pidl;
if (SUCCEEDED(SHParseDisplayName(file_name, 0, &pidl, 0, 0)))
{
ITEMIDLIST idNull = { 0 };
LPCITEMIDLIST pidlNull[1] = { &idNull };
SHOpenFolderAndSelectItems(pidl, 1, pidlNull, 0);
ILFree(pidl);
}
Note:CoInitialize or CoInitializeEx must be called before using SHOpenFolderAndSelectItems. Not doing so causes SHOpenFolderAndSelectItems to fail.
I'm using the following function to delete a file to the recycle bin: (C++, MFC, Unicode)
bool DeleteFileToPaperbasket (CString filename)
{
TCHAR Buffer[2048+4];
_tcsncpy_s (Buffer, 2048+4, filename, 2048);
Buffer[_tcslen(Buffer)+1]=0; //Double-Null-Termination
SHFILEOPSTRUCT s;
s.hwnd = NULL;
s.wFunc = FO_DELETE;
s.pFrom = Buffer;
s.pTo = NULL;
s.fFlags = FOF_ALLOWUNDO | FOF_SILENT | FOF_NOERRORUI;
s.fAnyOperationsAborted = false;
s.hNameMappings = NULL;
s.lpszProgressTitle = NULL;
int rc = SHFileOperation(&s);
return (rc==0);
}
This works nicely for most files. But if path+filename exceed 255 characters (and still much shorter that 2048 characters), SHFileOperation returns 124. Which is DE_INVALIDFILES.
But what's wrong? I checked everything a million times. The path is double-null terminated, I'm not using \\?\ and it works for short filenames.
I'm totally out of ideas...
I think backwards comparability is biting you in the --- in several ways, and I'd need to actually see the paths your using and implement some error checking code to help. But here are some hints.
You would not get a DE_INVALIDFILES 0x7C "The path in the source or destination or both was invalid." for a max path violation, you'd get a DE_PATHTOODEEP 0x79 "The source or destination path exceeded or would exceed MAX_PATH."
These error codes(return value) do, can, and have changed over time, to be sure what your specific error code means, you need to check it with GetLastError function(msdn)
Also, taken from the SHFileOperation function documentation: "If you do not check fAnyOperationsAborted as well as the return value, you cannot know that the function accomplished the full task you asked of it and you might proceed under incorrect assumptions."
You should not be using this API for extremely long path names, it has been replaced in vista+ by IFileOperation interface
The explanation for why it may work in explorer and not thru this LEGACY api is - Taken from the msdn page on Naming Files, Paths, and Namespaces
The shell and the file system have different requirements. It is
possible to create a path with the Windows API that the shell user
interface is not able to interpret properly.
Hope this was helpful
The recycle bin doesn't support files whose paths exceed MAX_PATH in length. You can verify this for yourself by trying to recycle such a file in Explorer - you will get an error message about the path being too long.
I need normalize file names such that it don't contain any non-portable characters in it. There is portable_file_name but that just checks and returns bool. I need to anyhow convert the given string to a portable name to create files.
Is there any reusable works ?
I assume that you mean some characters (*:;\"?<>/\|) are acceptable as file name or path name characters on some operating systems (Mac OS 9 for instance) but are not acceptable on others (such as Windows XP). Is that correct?
If so, you should probably do the character conversion yourself. I've done this in the past by using a regex to find and replace the unacceptable file name characters with a dash or something that works on all target operating systems. Then, you may safely use the files on both.
Try this:
boost::filesystem3::path portable_file_name;
portable_file_name.normalize();
The best I could come up with so far is:
for (auto &c:name)
{
char test[] = { c,0 };
if (!boost::filesystem::portable_file_name(test))
{
c = '_';
}
}
One of the important steps in doing this is converting file names like ./file or ones pointing to symbolic links into filenames that work on other platforms which might not have these concepts. Boost v1.48.0+ actually has the following functions to do so:
path canonical(const path& p, const path& base = current_path());
path canonical(const path& p, system::error_code& ec);
path canonical(const path& p, const path& base, system::error_code& ec);
This usually involves converting relative paths into absolute ones. These functions are also often called before performing security checks (e.g. is the requested file within the web-root directory of a web server?).
Note that cannonical() requires the file to exist.
I'm trying to hide some values in the registry (such as serial numbers) with C++/windows
so I've been looking at this article http://www.codeproject.com/KB/system/NtRegistry.aspx
which says:
How is this possible? The answer is
that a name which is a counted as a
Unicode string can explicitly include
NULL characters (0) as part of the
name. For example, "Key\0". To include
the NULL at the end, the length of the
Unicode string is specified as 4.
There is absolutely no way to specify
this name using the Win32 API since if
"Key\0" is passed as a name, the API
will determine that the name is "Key"
(3 characters in length) because the
"\0" indicates the end of the name.
When a key (or any other object with a
name such as a named Event, Semaphore,
or Mutex) is created with such a name,
any application using the Win32 API
will be unable to open the name, even
though they might seem to see it.
so I tried doing something similar:
HKEY keyHandle;
PHKEY key;
unsigned long status = 0;
wchar_t *wKeyName = new wchar_t[m_keyLength];
MultiByteToWideChar(CP_ACP, 0, m_keyName, m_keyLength, wKeyName, m_keyLength);
wKeyName[18] = '\0';
long result = RegCreateKeyExW(HKEY_LOCAL_MACHINE,
wKeyName,
0,
NULL,
0,
KEY_ALL_ACCESS,
NULL,
&keyHandle,
&status);
where m_keyName is the ASCII text and wKeyName is the wide char text, but in regedit I see that it is treated the same and the key is just cut where I put the '\0'.
what is wrong with it?
The problem is that you are using the Win32 API and not the NT Native API. There is a table about 1/2 way through the article that you referenced that contains the list of Native APIs. For example, you would use NtCreateKey or ZwCreateKey instead of RegCreateKeyExW. The Win32 API assumes that alls strings are terminated by a NUL character whereas the Native API counterparts use a UNICODE_STRING structure for the name.
I'll take a stab in the dark, as I have never tried to do this.
It appears that you are using the wrong function to create your registry key. You should be using the NtCreateKey method because RegCreateKeyEx[AW] will notice your '\0' and chop off past it.
Why not use the class provided in the example? It provides a method called CreateHiddenKey. To use it, simply call SetKey before it. It would be much cleaner.
I have to convert the encoding of a string output of a VB6 application to a specific encoding.
The problem is, I don't know the encoding of the string, because of that:
According to the VB6 documentation when accessing certain API functions the internal Unicode strings are converted to ANSI strings using the default codepage of Windows.
Because of that, the encoding of the string output can be different on different systems, but I have to know it to perform the conversion.
How can I read the default codepage using the Win32 API or - if there's no other way - by reading the registry?
It could be even more succinct by using GetACP - the Win32 API call for returning the default code page! (Default code page is often called "ANSI")
int nCodePage = GetACP();
Also many API calls (such as MultiByteToWideChar) accept the constant value CP_ACP (zero) which always means "use the system code page". So you may not actually need to know the current code page, depending on what you want to do with it.
GetSystemDefaultLCID() gives you the system locale.
If the LCID is not enough and you truly need the codepage, use this code:
TCHAR szCodePage[10];
int cch= GetLocaleInfo(
GetSystemDefaultLCID(), // or any LCID you may be interested in
LOCALE_IDEFAULTANSICODEPAGE,
szCodePage,
countof(szCodePage));
nCodePage= cch>0 ? _ttoi(szCodePage) : 0;
That worked for me, thanks, but can be written more succinctly as:
UINT nCodePage = CP_ACP;
const int cch = ::GetLocaleInfo(LOCALE_SYSTEM_DEFAULT,
LOCALE_RETURN_NUMBER|LOCALE_IDEFAULTANSICODEPAGE,
(LPTSTR)&nCodePage, sizeof(nCodePage) / sizeof(_TCHAR) );