[Patch] PR 31228, Fix close-on-exec race.

[David Daney]

Mark Mitchell wrote:
> David Daney wrote:
>
>   
>> I would like to put the patch for PR 31228, a libgcj bug, on the 4.2
>> branch.  The libgcj maintainer (Tom Tromey) is in agreement.
>>
>> As release manager, do you have an objections?
>>     
>
> Bugzilla doesn't indicate that this is a regression; is there a
> compelling reason that this needs to go on 4.2?
>
>   
It is not a regression.  The bug is a race condition for which there is 
no evidence that it has ever actually caused undesirable behavior or 
even occurred.  It is however a potential security hole in libgcj.

If this were a patch for a Primary GCC language, it might not be 
suitable 4.2 at this point.  However, the libgcj maintainers  sometimes 
like to fix problems even at this late stage.

I have bootstrapped and tested the patch on the 4.2 branch.  So it is 
ready to go.

I guess I don't really have strong feelings about it one way or the 
other.  If the Mark or the java maintainers have a preference that the 
patch be committed or not, I will follow their recommendations.

David Daney