On Thu, Aug 05, 1999 at 09:02:51PM -0400, Will Lowe wrote: > > Except for one point. cvs seems to mangle the time stamps, so whenever I > > Check the time on your local box. Are you using ntp to keep your linux > machine in sync with the rest of the world?
No - the ISP for my organisation (APANA) doesn't know or care what ntp is, let alone actively supporting it... My date and time a reasonably accurate though. Look at the date stamp on this message... However, I don't see why that should be a problem. The problem is that when I update the source from CVS, the current date and time are used. When I then run make, it does some comparisons and finds that Makefile.in was updated at the same time as Makefile.am, so it recompiles Makefile.in. For instance, in my current project: -rw-r--r-- 1 bam users 307 Jul 24 17:33 ../acinclude.m4 -rw-r--r-- 1 bam users 39168 Aug 6 10:16 ../aclocal.m4 -rw-r--r-- 1 bam users 20556 Aug 6 10:16 ../configure.in aclocal.m4 and configure.in were just updated by the last upstream source. They both have the same date and time, hence Make wants to rebuild aclocal.m4: [564] [snoopy:bam] ~/cvswork/heimdal/build >make cd .. && aclocal -I ../cf aclocal: couldn't open directory `../cf': No such file or directory make: *** [../aclocal.m4] Error 2 (and that is the problem with the path that I haven't yet investigated). This has nothing to do with my time in relation to other computers on the Internet. It is about comparing the time on different files on my computer. Currently I have hacked around the problem by providing dummy versions of aclocal, automake, and autoheader that do nothing. -- Brian May <[EMAIL PROTECTED]>
pgpdoaFjLKHzD.pgp
Description: PGP signature
