On Mar 6, 2016, at 2:53 PM, Ralph Corderoy
<ralph(_at_)inputplus(_dot_)co(_dot_)uk> wrote:
To give them some credit, fcntl(3p) here says there were too many
existing brain-damaged implementations around for POSIX.1‐2008 to
deviate. I guess sometimes they're stuck documenting life.
They were (are) a disingenuous bunch of bastards, when it suits their needs ;-)
Witness utf8 creeping in at the last second. And I was part of the "Austin
Group" networking updates mob. There were shenanigans going on there that I
have thankfully long forgotten.
Dot-locking FTW! flock(2) starts off sounding sane on Linux these days,
but then you get into the NOTES with all the NFS exceptions, and it
spirals down. https://manned.org/flock.2
That's not the fault of flock() or lockf(). They define sane interfaces; it's
NFS that's failing here. You *can* get network locking right – as 9P proved
ages ago.
--lyndon
_______________________________________________
Nmh-workers mailing list
Nmh-workers(_at_)nongnu(_dot_)org
https://lists.nongnu.org/mailman/listinfo/nmh-workers