Bug #114: expand setclock script

Greg Schafer gschafer at zip.com.au
Sat Jan 24 19:53:48 PST 2004


On Sat, Jan 24, 2004 at 07:44:04AM -0700, Gerard Beekmans wrote:
> On Mon, 2003-11-03 at 21:24, Dustin D. Cook wrote:
> Taking all comments into consideration I say this is better served in
> BLFS, if they chose to deal with it. There's too many variables to
> determine which clock is more accurate and it would be wrong not to take
> them all into account else it defeats the whole purpose to begin with.

Sorry for making a late entry into this thread :-/ But I think that
Alexander's suggestion of "principle of the least astonishment" is the
sanest thing to do here.

Whatever we do will not please everyone. Too bad. But after doing a bit of
reading up on the matter, I'm convinced that saving the system clock to
hwclock on shutdown is the least problematic thing to do. It's certainly
less problematic than what we do now! I've been caught out a couple of times
myself by this lack, and could of avoided those situations had the default
bootscripts been saving the clock on shutdown.

A change like this shouldn't affect folks who sync to external time sources.
Folks who do fancy stuff with "hwclock --adjust" will be affected but
they'll just have to be aware and remove the symlink as needed.


PS - Gerard, while we're talking about bootscripts, are you still
comfortable being their maintainer? There have been a fair few suggestions
in recent months by various folks that have gone largely ignored. I'm not
very much interested in bootscripts so have mostly avoided commenting on
them. Some of the suggestions have been good but others have been a bit
iffy and I wouldn't support some of those. Anyway, the only reason I raise
this is just in case you might want to free yourself up to offload the
maintainership onto someone else. I don't want to do it but I'm sure we
could find some suitably qulified volunteers if needed.



More information about the lfs-dev mailing list