release tarballs should have LAST_UPDATED containing release tag?

[Dan Kegel]

Gerald Pfeifer wrote:
> On Wed, 9 Jul 2003, Dan Kegel wrote:
>   2001-07-27  Richard Henderson  <rth@redhat.com>
> 
>         * test_summary: Copy LAST_UPDATED UTC time to head of summary.
> 
> which unconditionally tries to print LAST_UPDATED instead of first
> checking whether that file exists.
> 
> Richard, my AWK is quite non-existant, else I'd have fixed this right
> away; would you mind having a look? Thanks!
 > ...
>>I see that file is supposed to be created by contrib/gcc_update with
>>the date.  It might be nice for the released gcc tarballs to contain
>>a LAST_UPDATED file containing just the release tag.
> 
> 
> Hmm, we could do that, but I'm not sure whether that's still needed
> once test_summary has been fixed?

Oh, if test_summary is changed to not need that file, that's fine.

Here's a quick hack, not sure if it works yet...
--- gcc-3.3/contrib/test_summary.old    Thu May 16 10:41:58 2002
+++ gcc-3.3/contrib/test_summary        Sun Jul 13 15:38:47 2003
@@ -98,7 +98,7 @@
      srcdir = $1;
      gsub(/\/configure$/, "", srcdir);
      printf "LAST_UPDATED: ";
-    system("tail -1 " srcdir "/LAST_UPDATED");
+    system("if -f " srcdir "/LAST_UPDATED; then tail -1 " srcdir "/LAST_UPDATED ; else echo No timestamp file");
      print "";

      $1 = "configure flags:"; configflags = $0;

simpler fixes are possible... you could just do  || /bin/true after the tail...

(Oddly, when I just tried reproducing the problem in 3.2.3, test_summary didn't seem
to return a nonzero exit status.  I may have been smoking crack before.  I'll
try to get a tested patch here.  I'd rather fix things than just whine.)
- Dan
-- 
Dan Kegel
http://www.kegel.com
http://counter.li.org/cgi-bin/runscript/display-person.cgi?user=78045