Re: [boost] [filesystem] create_directories not thread-safe?

From: Beman Dawes (bdawes_at_[hidden])
Date: 2004-03-05 12:12:22

At 09:40 AM 3/5/2004, Vladimir Prus wrote:
 >
 >I have an application which sometimes fails saying that it can't create a 

 >directory, because it already exists. The directory is created by
 >
 >     fs::create_directories("run");
 >
 >and the application is multi-threaded. It seems there's a race condition:
 >
 >         if (ph.empty() || exists(ph)) return;
 >
 >         // First create branch, by calling ourself recursively
 >         create_directories(ph.branch_path());
 >         // Now that parent's path exists, create the directory
 >         create_directory(ph);
 >
 >After one thread checks that 'ph' does not exist, another one creates the 

 >same directory, so call to 'create_directory' fails.
 >
 >What can be done? Basically, adding a mutex for MT build and adding a
 >warning to docs are the variants. Opinions?

There is already a general race-condition warning in the docs. See 
www.boost.org/libs/filesystem/doc/index.htm#Race-condition.

Even if the filesystem library itself were made thread-safe, the problem 
could still happen because the state of the external file system may be 
altered by some other process or other machine.

Is there a issue with solving your particular problem by either using your 
own mutex, or catching and discarding the exception, or moving the 
create_directory() call to prolog code that is single-threaded?

--Beman