[PATCH] implement futimes

[Rob Landley]

On Saturday 21 November 2009 18:48:17 Mike Frysinger wrote:
> On Wednesday 18 November 2009 19:32:14 Rob Landley wrote:
> > User Mode Linux uses futimes(), and thus won't build against uClibc. 
> > This was my quick and dirty first stab at implementing it.  Not very
> > heavily tested.
>
> convince them to convert to futimens() which is in POSIX.

It may be in POSIX, but it's not in any uClibc release.

Once again, it's been more than a year since the last time the uClibc 
development branch got a release.  It's nice there's been exactly one uClibc 
bugfix release since then, but in that time both the Linux kernel and the 
BusyBox project have each had three or four development releases, plus a dozen 
or so bugfix releases.  I've personally cut four releases from my Firmware 
Linux project's development repository since uClibc 0.9.30.0 came out.

In fact, in the time uClibc's been publicly contemplating a 0.9.30.2 bugfix 
release ("preparing 0.9.30.2" posted October 12, still not out yet), the 
kernel's put out three bugfix releases.  Note that the "stable" kernel releases 
are something two guys do in their spare time.

Currently, it's not clear _which_ branch a development release would come 
from, and thus what would be worth testing.  The nptl branch?  The 
inexplicable other development branch which NPTL still hasn't been merged into 
more than 3 years after the OLS presentation on NPTL for uClibc?  The merge of 
these two branches has been "in progress" since Erik was maintainer, but I've 
never even seen an -rc containing NPTL code.

*shrug*  I just ported UML to use futimens() and utimensat(), and maybe 
tomorrow I'll put together another uClibc patch and check the patch into my 
FWL repository.  (People are sending me backports of things like the 
shm_functions because they want to use them on an actual uClibc release that 
_somebody_ supports, and I support the ones I ship.  I'd rather not maintain 
more patches than necessary, but I'm not sure what pushing them into the 
uClibc-dev repository actually accomplishes.  The shm_ functions are already 
_in_ there.)

Whatever happened to "release early, release often"?

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds