Hi Michael,

2015-01-27 8:56 GMT+01:00 Michael Olbrich <[email protected]>:

> On Mon, Jan 26, 2015 at 05:24:55PM +0100, Guillermo Rodriguez Garcia wrote:
> > Hello all,
> >
> > I am using the ntpd applet in busybox but also wanted to enable the
> ntpdate
> > program (yes I know use of ntpdate is not recommended in general but this
> > is for a very specific use case). However I see that ntpdate is part of
> the
> > "ntp" package and that the ntp package is blocked by BUSYBOX_NTPD. From
> > rules/ntp.in:
> >
> > menuconfig NTP
> >         tristate
> >         prompt "ntp                           "
> >         select OPENSSL          if NTP_CRYPTO
> >         select ROOTFS_VAR_CACHE if NTP_NTPD
> >         select LIBC_M
> >         select LIBC_RT
> >         select GCCLIBS_GCC_S
> >         select LIBMD
> >         depends on !BUSYBOX_NTPD || ALLYES
> >
> > [...]
> > config NTP_NTPDATE
> >         bool
> >         prompt "install ntpdate"
> >         depends on NTP
> >
> >
> > Obviously one would not want to run two ntp services at the same time,
> but
> > since busybox lacks an ntpdate command, there are cases where one would
> > want to use ntpdate from the NTP package while still using Busybox's ntpd
> > for everything else.
> >
> > My question is: Is there any reason why this wouldn't work? Was the
> > "depends on !BUSYBOX_NTPD" line added just to avoid the possibility of
> the
> > two ntp services running at the same time? Or is it there to guard from
> > other possible problems/issues?
>
> We usually add those lines only to prevent two packages from installing the
> same file. In this case we have the !BUSYBOX_NTPD check twice, once for the
> package and once for ntpd. This is clearly wrong. I don't see a reason why
> the check for the packages is needed and I don't remember why I added it. I
> suggest you remove the check (and the comment at the end of the file) and
> see what happens.
>

OK, tried just that: Removed the first (package-level) check, then selected
ntpdate from the ntp package and ntpd from busybox, rebuilt, tested on the
target, and I can confirm that everything works as expected. So I would say
that this check can be removed.

Thank you,

Guillermo Rodriguez Garcia
[email protected]
-- 
ptxdist mailing list
[email protected]

Reply via email to