Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming.threads > #1082

Re: String multithreading issue

Show all headers | View raw

On 20.09.2012 09:50, itt ium wrote:
> The following code when checked with helgrind shows possible race conditions. Is there a bug in std::string class?
>
> void *do_work(void *)
> {
>      std::string s;
>      s.erase();
>      return 0;
> }
>
> int main()
> {
>      pthread_t tid[2];
>      pthread_create(&tid[0], NULL, do_work, NULL);
>      pthread_create(&tid[1], NULL, do_work, NULL);
>
>      return 0;
> }

There is nothing wrong with your code. And most likely there is nothing 
wrong with std::string, too.

Most likely helgrind does not understand the way that string::erase is 
implemented. I guess they add a reference to an existing empty string, 
witch is immutable once the runtime has been initialized.

Sometimes small immutable contants are initialized lazy to avoid static 
initialization sequence problems. This lazy initializations might be 
implemented not fully thread-safe because there is no requirement that 
all empty strings must refer to the same storage. In fact, more than one 
instance might be created. But the latter is only wild guess.


Marcel

Back to comp.programming.threads | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

String multithreading issue itt ium <ittium@gmail.com> - 2012-09-20 00:50 -0700
  Re: String multithreading issue Marcel Müller <news.5.maazl@spamgourmet.com> - 2012-09-20 11:54 +0200
  Re: String multithreading issue bartek szurgot <basz@no.spam> - 2012-09-20 17:53 +0200
    Re: String multithreading issue Marcel Müller <news.5.maazl@spamgourmet.com> - 2012-09-21 08:09 +0200
      Re: String multithreading issue Geoff Clare <geoff@clare.See-My-Signature.invalid> - 2012-09-21 13:42 +0100
  Re: String multithreading issue Marc <marc.glisse@gmail.com> - 2012-09-20 16:37 +0000

csiph-web