Thanks. touch does the trick. What happens is interesting. When I prepare the module I run a script called make_manifest.pl which performs a number of routine things on the distribution namely recurse the tree, remove the Makefile, Makfile.old, /blib, &1 and other extaneous files, does a global s/\015/012/\012 on all the files to de-Windows them, write all the pod into a /html dir using pod2html and rewrite a complete and accurate MANIFEST. This effectively touchs all the files with the current time on the generating system. The whole thing is then tar -cf and gziped.
On upload the originating system time stamps are (mostly!) retained (in GMT) so the times represent the originating system time clock ie my Win32 system. When you run Makefile.PL the resultant Makefile carries the local system timestamp. This was proven this morning when I uploaded a tarball generated 9 hours earlier and the times were out by 8 hours and 55 minutes. This of course caused no problems as the Makefile was now much newer than Makefile.PL. The problem was that my sytem clock is 10 minutes fast.
I had tried vi and then ZZ but this does not change the required timestamp whereas touch does. Thanks
cheers
tachyon
s&&rsenoyhcatreve&&&s&n.+t&"$'$`$\"$\&"&ee&&y&srve&&d&&print
In reply to Re: Re: Makefile time incorrect causing error
by tachyon
in thread Makefile time incorrect causing error
by tachyon
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |