[Haskell-cafe] Discussion: The CLOEXEC problem

[Niklas Hambüchen]

On 25/07/15 21:09, Donn Cave wrote:
> But there are other things that could turn up.
> For example, you could use flock(2) (Berkeley, not POSIX fcntl
> lock) to keep an advisory file lock until the exec exits.  If the
> file is closed prematurely, you lose the lock, and ... whatever
> happens then.

This is a very valid point. Applications that rely on this will break by
changing the default here.

I'm wondering though whether this is an acceptable price to pay for
better (in my opinion) defaults. Given enough announcement and time, it
should not be too difficult to find Berkeley flock() invocations, and
explicitly fnctl their FDs to CLOEXEC=False, or open() them with
CLOEXEC=False.

I would even be surprised if there is a single Haskell program out there
that uses this; I know of one that uses file locking, bu that's using
fnctl style locks.