Jeff -- last-minute patch for 2.95, or maybe 2.95.1?

Jeffrey A Law
Thu Jul 8 12:21:00 GMT 1999

  In message < 19990708125857.12445.qmail@deer >you write:
  > Jeff, after (finally) building 2.95 prerelease on my (loaned) UltraSPARC-1
  > and inspecting the output of the u77-test.f test run, I noticed that
  > the date string returned by the DATE intrinsic was wrong:
  >   DATE (dd-mmm-yy): 31-Dec-69
  > Turns out, this problem will likely show up on *any* big-endian system.
  > It is trivial to fix, and the fix is not only "obviously correct", it
  > is easy to reassure oneself that it would be extremely unlikely to cause
  > a problem elsewhere in gcc.  I've verified the fix works on the same
  > UltraSPARC-1, as well as that it doesn't change the correct results
  > obtained on my Pentium II.
  > The bug is not, however, a regression.
  > My concern with not fixing it is that I *hope* GCC 2.95 gets much wider
  > usage vis-a-vis g77 than any previous version of g77, due to the fact
  > that it's the first fully-integrated version of g77 released from the FSF.
  > (OTOH, everyone might wait for 3.0 due to the promise of the rewrite.)
  > So, it's an embarrassing bug to leave around.
  > My choice is to fix it for 2.95, unless there are *zero* other (code-relate
  > d)
  > patches being allowed in at this point, in which case, unless you're willin
  > g
  > to make an exception for this patch, I'd want it put on the list for 2.95.1
  > ,
  > should there be a need for such a patch.
  > What do you think?  I won't apply it to the mainline until I can get the
  > various doc bits all updated per the 2.95 decision, unless otherwise
  > requested.
  >         tq vm, (burley)
  > Wed Jul  7 15:58:16 1999  Craig Burley  <>
  > 	* libU77/date_.c (G77_date_y2kbug_0): G77_time_0 returns
  > 	longint, not integer, and G77_ctime_0 takes longint, not
  > 	integer, argument.
  > 	* libU77/Version.c: Bump version.
I brought this into the gcc-2.95 branch.  Thanks.

Can you please go ahead and install it on the mainline tree too and update the
docs on both trees as you see fit?  Thanks.


More information about the Gcc-patches mailing list