I am running into an error while trying to enumerate a directory using boost::filesystem.
The error appears when I try to enumerate a directory for which I have no permissions on an SMB shared directory - ie, I have RW on the share but a subdirectory on the share is access denied.
Platform etc:
Windows 10.
C++ 17 code.
Visual Studio 2019.
Unicode VS project (thus the use of std::wstring).
Shared folders from a Samba server running on Linux.
Sample code below, then details of the fail.
int enumerate(const std::wstring dir_to_enumerate) {
if (directory.size() == 0)
return -1;
boost::filesystem::path dir_to_enumerate(dir_to_enumerate);
boost::system::error_code ec;
if ( ! boost::filesystem::exists(dir_to_enumerate, ec))
return -2;
if (ec.value())
return -4;
if ( ! boost::filesystem::is_directory(dir_to_enumerate))
return -3;
boost::filesystem::directory_iterator it{ dir_to_enumerate };
while (it != boost::filesystem::directory_iterator{}) {
// ... do stuff
}
return 0;
}
Consider a folder C:\temp2 with permissions removed for my account. The test
if ( ! boost::filesystem::exists(dir_to_enumerate, ec))
return -2;
reliably detects C:\temp2 is not accessible.
Consider a folder "temp2" located on a Samba share T:. This folder on the Linux box is /data/temp2. So the directory passed into the function is L"T:\temp2", and that corresponds to location /data/temp2 on the Linux filesystem.
Samba itself shares the location R/W.
If on the Linux server the EXT4 permissions of the directory /data/temp2 are R/W (mode 700, or 750, or 755 you get the idea) for my account then everything works as expected, I can enumerate the contents of T:\temp2
The problem shows up detecting the following case reliably:
If I set the EXT4 permissions on the directory /data/temp2 to denied (mode 000) then the above code behaves unpredictably...
Sometimes, most of the time, the test
if ( ! boost::filesystem::exists(dir_to_enumerate, ec))
return -2;
detects T:\temp2 is not accessible.
But occasionally exists() returns true, and we fall thru to testing the boost::filesystem::error_code value.
However the error_code never indicates an error.
In this case the line
boost::filesystem::directory_iterator it{ dir_to_enumerate };
always throws. I can't figure out exactly what it is throwing.
I'm pulling my hair out. Can anyone tell me what I am missing, or even just what the directory_iterator initializer is throwing?
(And... if I may be so forward: please do not tell me to just use std::filesystem. On Visual Studio 2019, and AFAIK C++17 itself, std::filesystem is incomplete. Boost::filesystem works better.)
Related
I want to check if a certain file exists in the specified location. I have been trying multiple solutions for that but seem like none of them work properly, since all of them return false.
There is no doubt that the file exists in the specified location.
Executable is being run as administrator, so I'm having the appropriate permissions.
Code I used:
#include <io.h>
#include <string>
#include <Shlwapi.h>
std::string str = "C:\WINDOWS\System32\iluminated.dll";
unsigned long attrib = GetFileAttributes(str.c_str());
bool exists1 = (attrib != INVALID_FILE_ATTRIBUTES &&
!(attrib & FILE_ATTRIBUTE_DIRECTORY)) &&
GetLastError() != ERROR_FILE_NOT_FOUND; // false
bool exists2 = ( _access( str.c_str(), 0 ) != -1 ); // false
bool exists3 = PathFileExists(str.c_str()) != 0; // false
Is there anything I'm doing wrong?
You should use double back slashes for paths, since if you use single back slashes in a string they are interpreted as command symbols (line \n, for example):
"C:\\WINDOWS\\System32\\iluminated.dll"
Alternatively, you can use forward slashes, they work on most operating systems:
"C:/WINDOWS/System32/iluminated.dll"
I found the answer. Turns out Windows is always redirecting system32 to syswow64 while trying to access in 64-bit Windows. I had to use SysNative directory, even though it doesnt exist - Windows redirects it to the proper system32 directory.
Since Visual Studio 2012, application projects default to “Any CPU
32-bit preferred”. If you run such an executable on a 64-bit Windows
operating system, then it will start as a 32-bit process and be
affected by WOW64 file system redirection.
When a 32-bit process on 64-bit Windows tries to access
"C:\Windows\System32", WOW64 redirects it to "C:\Windows\SysWOW64".
There are several ways to access the real "C:\Windows\System32"
directory:
Use "C:\Windows\SysNative", which WOW64 redirects to "C:\Windows\System32" even though it does not appear in directory
listings. This is an easy way and unlikely to cause problems.
Use Wow64DisableWow64FsRedirection and Wow64RevertWow64FsRedirection.
Use a 64-bit process.
Source: https://social.msdn.microsoft.com/Forums/en-US/c54f8368-035e-478e-b988-b180a3c7e3da/file-not-found-for-existing-file-in-system32-directory?forum=csharpgeneral
So I am running a C++ app, built with CMake (not by me). It works fine on other people testing with it, but not found anyone to test on centos7 yet.
The issue seems to arrise at this snippet of code:
struct stat fileStat;
if ( stat( pszFilePath, &fileStat) == -1 )
{
DEBUG_ERR(( "Can't open input dir [%s]\n", pszFilePath ));
return( false );
}
Which is the first part of the ReadFileInfo call here:
time_t dateChange;
DWORD dwSize;
if ( ! CFileList::ReadFileInfo( GetFilePath(), dateChange, dwSize ))
{
DEBUG_ERR(( "Can't get stats info for file '%s'\n", static_cast<LPCTSTR>(GetFilePath()) ));
return false;
}
Now, pszFilePath is many value, a few examples are:
'scripts/sphere_template_vend.scp'
'scripts/sphere_serv_triggers.scp'
The application is owned by root, the whole folder it all sits in is owned by root. The scripts folder is there and has read/write permissions as do all files (all also owned by root)
Running the application triggers both errors in the above code, there is nothing before them that influences anything.
Im not a C++ developer and do not have the tools to compile with debugs for checking the current path and so on, but I see no reason why it throws these errors. The files exist, the files are accessible, no one else seems to have this problem.
I do have cPanel on the server, but it shouldn't be causing any issues as I am using root user and also keeping out of the /home/ directory
Anyone know of any issue this could be because of? I tried with a '/sphere/' prepending the paths but it still has the same issue, it seems the application does not have access to the files (the application oddly reports line errors within the files it says it cannot read, but they do not match, so assuming its not correct).
Issue reported on the Github for the project here: https://github.com/Sphereserver/Source/issues/64
But no one seems to know whats going on
I am MFC guy working on visual studio 2010 create some executables using visual studio!! but on linux and mac my executables are not working as usual windows!!.
So i decided to use "MinGW" compiler to create executables.
Note:-Please give me one suggestion is that," is minGW is best compiler for cross plateform working ??or any thing else is there??"
I successfully install WinGW compiler on my C drive and start working with following program..
#include <iostream>
using namespace std;
int main ()
{
cout << "Hello ";
return 0;
}
I compile it using following command,
g++ -static-libgcc -static-libstdc++ Main.cpp
I found one executable in same folder with name a.exe :).Working fine:)
But after some time i decide to modified same program in following manner like,
int main ()
{
return 0;
}
I compile it with same command but when i execute it using command line it show me error "Access is denied so i goto that folder and run same executable as "Run as Administrator" it show me one messagebox with the message windows cannot access the specified device path or file. you may not have appropriate permissions
---EDIT--
follwing code is NOT WORKING:-
int main ()
{
int k;
return 0;
}
but this program WORKING :-
int main()
{
int k;
k = 0;
return 0;
}
If you are getting this access denied error, then the most likely cause is that the executable file is open in another process, probably the linker or debugger. Try installing Process Explorer and hit Ctrl+F and type in the name of your .exe. This should show the processes that the .exe file open. Kill those processes (or if you are still debugging, then end debugging first). You then should be able to build again.
Note that this has nothing to do with Microsoft APIs, as in any case you're using gcc.
EDIT: If there are no processes holding the .exe then it may be that there is some other kind of permission problem. Does the .exe file exist? Can you delete the file and rebuild? Another thing to try is run Process Monitor and filter for the name of the .exe -- that may show a regular permission denied error, or perhaps another error such as a sharing conflict.
Note:-Please give me one suggestion is that," is minGW is best compiler for cross plateform working ??or any thing else is there??"
No. And there's nothing else out there.
Use whatever compiler is available on target platform, ensure your code compiles on all of them.
Avoid platform-specific and compiler-specific code at all costs (use cross-platform frameworks).
I successfully install WinGW
There are many different versions of mingw provided by different sites. If you install compiler from mingw.org using mingw-get, it'll probably work. If you install mingw from some other site, it may or may not work.
I compile it using following command,
Use a build systems. cmake, qmake or something similar.
it show me error "Access is denied
Launch process monitor and see after which system call it terminates. It is also possible that your antivirus software interferes with your program, or maybe there's some stray dll in your path or something like that.
Check the permissions for the entire folder in which the executable resides. Trying to 'Run as Administrator' doesn't have any effect if the folder doesn't allow the permissions.
It doesn't have anything to do with your code. This is an environmental problem, something is pretty messed up about the permissions your user account has for one or more of the directories on your hard disk. The generic diagnostic is that the default working directory for the program does not permit read or list access.
A possible starting point would be to use Explorer and right-click the directory where MinGW is installed. Use the Security tab and ensure that your user account has all permissions enabled. Further narrow it down to trying to run the program from the command prompt, using different directories as the default directory.
On a Windows XP, 32-bit machine I've cloned the libgit2 repository using
git clone git://github.com/libgit2/libgit2.git trunk
Then I configured from trunk_build_debug for mingw using:
cmake ../trunk -DCMAKE_BUILD_TYPE=Debug -DCMAKE_C_FLAGS=-g
and build-ed it with
make
With this newly build library I'm trying to create a new repository. So I'm using
error_code = git_repository_init( &repo, path, 0 );
if ( error_code != 0 )
{
const git_error * error = giterr_last();
/* the text representation:
( error == NULL || error->message == NULL ?
"(no description)" :
error->message )
*/
}
On first attempt this fails with the error message:
Git error -1: The global file '.config/git/config' doesn't exist: The
system cannot find the file specified.
The .git directory is created and I can use git status to query it. Also, I have git installed in Program Files/git and it works just fine. The file that is mentioned does not exist in user directory, that is true. I'm not happy because I can't check the returned error code (the error may be "real").
So I've created an empty file in D:\Documents and Settings\User Name.config\git\config
The .git directory is again created but I get the same -1 error code. This time giterr_last() returns NULL.
Tracing the code in debugger seems to indicate that the library is not happy with my empty file.
However, I think this indicates that I'm doing something wrong. Is there an initialization method? Do I need to create the file and use the git_config_ functions to point the library to it?
BTW, is this the wrong place to ask this? Is there a dedicated forum where I should ask this question?
Thanks
Here is the back trace where the call fails: pastebin
It seems that the configure file inside .git directory can't be parsed (not shown in in trace is the 'config_parse()' that fails).
I've opened an issue in libgit2.
To create a new repository, one should rather rely on git_repository_init(). For more information, you can peek at the header and the tests.
git_repository_open() should rather be used when the repository already exists.
Is there a dedicated forum where I should ask this question?
It's ok to ask programming related question on StackOverflow. Beside this, some libgit2 developers hang out in the #libgit2 channel on irc.freenode.net. However, when you encounter bugs, it's recommended to submit them to the issue tracker.
Update
There had been some issues in the past with the way the configuration files are being probbed on Windows. Along with some attempts to fix them. See this issue for some background about this.
Beside the complexity of the probbing, some users may want to explictly provide alternate locations for their files. This pull request works this angle.
Considering what you discovered while debugging the libgit2, maybe now would be a good time to open a bug in the issue tracker. Beware that two issues may be at play: the probbing one and and the empty config issue.
I'm trying to compile and run the app, which was created 4 years ago. It was developed for Windows in Embarcadero RAD Studio C++ builder. Now I try to compile and run it in Windows in VirtualBox using the latest version of RAD Studio. I have a system call to another app in my app:
system("dot.exe -Tjpg -o nfa.jpg NFA_graph.txt");
It keeps returning 1 and the file is not created. I also tried
system("Echo %CD% >> z:\log.txt");
and the file is not created. I also tried like this:
FILE *fpipe;
char *command = "Echo %CD% >> z:\log.txt";
char line[256];
if (0 == (fpipe = (FILE*)_popen(command, "r")))
{
perror("popen() failed.");
exit(1);
}
while (fread(line, sizeof line, 1, fpipe))
{
ShowMessage(line);
}
_pclose(fpipe);
And nothing I get. I wonder if the reason of such strange behaviour is that I'm running this all in VirtualBox?
You're not escaping your \ characters. You should use / in file paths, or \\ if you must. In addition, Windows 7 won't let you write to the root directory of a hard drive w/o administrator access.
To determine if a command environment is available, first do this:
if (!system(NULL)) {
// Can't make any system() calls
}
If your command environment is available, then you need to fully specify the path, making sure to escape the \'s like I mentioned above, and don't write anything to a drive's root directory. Also make note that opening files does not default create directories.
No, it's very unlikely. I see few issues with your code: you did not check errno if system() returns 1. It can help you to spot a real problem. Also, all backslashes must be Esc'ed.
I'm suggesting that dot.exe is not in PATH environment variable, that's the reason of the system() failure.