Re: pwrite() and threads

[Giles Lean <giles.lean%pobox.com@localhost>]

Wouter Klouwen <dublet%acm.org@localhost> wrote:

> Giles Lean(giles.lean%pobox.com@localhost) said 2010.05.26 23:26:21 +0000:
> > POSIX doesn't agree with you, I'm afraid.
> 
> Fair enough. I was looking at a UN*X spec that was rather sparse on
> details.

In another lifetime, I got to play standards lawyer (and language
lawyer) from time to time.  ("Your C library is broken!" "No, your
application's expections are wrong. Sorry.") So I've had practice.

FYI the standard is available freely (you have to register) at
http://www.opengroup.org.  (I won't try to give a direct URL; I usually
find my way into their web site via Google.)

> Avoiding threading safety issues is better than dealing with them.

As one who spent uncounted hours dealing with customer threads problems,
some real and bugs on our (I worked for a vendor) part but many being
application problems, I can only heartily endorse this!

The Go language has some interesting alterntives for concurrency; I've
started on a NetBSD port but got thoroughly distracted. I'll try to get
myself undistracted again.

Rob Pike says porting Go is "easy", but I'm not sure Rob Pike's idea of
"easy" will match mine.  Still, at least I'm porting an open source
language to an open source operating system, and it runs on OS X and
FreeBSD already.  Of course their threads support (at OS level, below
pthreads) looks quite different to NetBSD's.

Giles