Re: New g77 failure on Alpha with -Os

[Jeffrey A Law <law at cygnus dot com>]

  In message <199807170421.AAA15246@melange.gnu.org>you write:
  > Note that I wrote auto0.f (and auto1.f) earlier this week to stress
  > g77's automatic arrays.  (Basically these are like `foo(n) { int a[n];
  > ... }' in C, or I should say GNU [extended] C.)  The idea was to
  > get one or both of them to fail on HP-UX (10, maybe 9 too) to address
  > a longstanding bug report from Tim Prince concerning automatic
  > arrays on that system.  That failed, though I still intend to get back
  > to that issue soon, but, lo and behold, one of the two tests I wrote
  > fails on one machine using one particular (obscure) optimization
  > option.
  > 
  > So, who knows, maybe when this is investigated, it'll turn out to
  > expose a sufficiently high-level bug in gcc that the fix will make
  > Tim's problems on HP-UX go away!
This kind of thing has happened several times with c-torture.  I think
Tege even mentions it in the c-torture README :-)

jeff