pthread_barrier

pthread_barrier

[Reini Urban]

In http://cygwin.com/ml/cygwin/2011-07/msg00406.html
Corinna hinted that pthread_barrier is a bit hard to implement.

I found 2 nice non-GPL implementations, but I'm not sure about the license.
There's one in libuv (which I need it for), which is
https://github.com/joyent/libuv/blob/master/src/unix/pthread-fixes.c
provided by Sony and Google (for Android), which seems to be MIT licensed.

And there's http://www.cs.berkeley.edu/~bvs/cs267_hw2/particles/pthread_barrier.c
without any license, looks it's some berkeley course material.

Doesn't look too hard to implement.
Should I ask the berkeley guy Brian Van Straalen
or is the libuv version good enough for us?

The GPL pthreads-win32 version looks awful in comparison.
-- 
Reini Urban
http://cpanel.net/   http://www.perl-compiler.org/

Re: pthread_barrier

[Christopher Faylor]

On Tue, Oct 08, 2013 at 11:31:24AM -0500, Reini Urban wrote:
>In http://cygwin.com/ml/cygwin/2011-07/msg00406.html
>Corinna hinted that pthread_barrier is a bit hard to implement.
>
>I found 2 nice non-GPL implementations, but I'm not sure about the license.
>There's one in libuv (which I need it for), which is
>https://github.com/joyent/libuv/blob/master/src/unix/pthread-fixes.c
>provided by Sony and Google (for Android), which seems to be MIT licensed.

This one is problematic since it seems like we'd have to include the license
when we distribute the DLL.  (Shhhh...  I know...)

>And there's http://www.cs.berkeley.edu/~bvs/cs267_hw2/particles/pthread_barrier.c
>without any license, looks it's some berkeley course material.
>
>Doesn't look too hard to implement.
>Should I ask the berkeley guy Brian Van Straalen
>or is the libuv version good enough for us?

Code without a license isn't any better than code with a license.  If
you can get someone to assert that the Berkeley code is in the public
domain that would help.  Otherwise, I don't think either of these are
viable options.

Perhaps someone could describe the implementations to someone who could
implement them from scratch.  That would be the safest way to do this
I think.

cgf

Re: pthread_barrier

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1925 bytes --]

On Oct  8 13:14, Christopher Faylor wrote:
> On Tue, Oct 08, 2013 at 11:31:24AM -0500, Reini Urban wrote:
> >In http://cygwin.com/ml/cygwin/2011-07/msg00406.html
> >Corinna hinted that pthread_barrier is a bit hard to implement.
> >
> >I found 2 nice non-GPL implementations, but I'm not sure about the license.
> >There's one in libuv (which I need it for), which is
> >https://github.com/joyent/libuv/blob/master/src/unix/pthread-fixes.c
> >provided by Sony and Google (for Android), which seems to be MIT licensed.
> 
> This one is problematic since it seems like we'd have to include the license
> when we distribute the DLL.  (Shhhh...  I know...)
> 
> >And there's http://www.cs.berkeley.edu/~bvs/cs267_hw2/particles/pthread_barrier.c
> >without any license, looks it's some berkeley course material.
> >
> >Doesn't look too hard to implement.
> >Should I ask the berkeley guy Brian Van Straalen
> >or is the libuv version good enough for us?
> 
> Code without a license isn't any better than code with a license.  If
> you can get someone to assert that the Berkeley code is in the public
> domain that would help.  Otherwise, I don't think either of these are
> viable options.

Isn't Berkeley code always BSD licensed?  I'm just dreaming of a perfect
world, I guess...

> Perhaps someone could describe the implementations to someone who could
> implement them from scratch.  That would be the safest way to do this
> I think.

Indeed, and in fact the problem is not that there isn't code around
which already implements pthread_barrier stuff, the problem is that none
of that matches the existing, C++ for extremists pthread code in Cygwin.
That's what I was silently implying when I wrote the aforementioned
mail.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]