Re: [PATCH] Sparc tail calls (take 4)

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

  In message <20000324122246.A19852@cygnus.com>you write:
  > On Fri, Mar 24, 2000 at 12:27:16PM -0700, Jeffrey A Law wrote:
  > > A similar check used to be in the code.  Presumably it got lost when
  > > you had to revamp the block note handling:
  > 
  > Hmm.  No, it's still there.
  > 
  >       if (pass == 0)
  >         {
  >           tail_call_insns = insns;
  > 
  >           /* If the current function's argument block is not large enough
  >              to hold the outoing arguments, or we encountered some other
  >              situation we couldn't handle, zero out the sequence.  */
  >           if (current_function_args_size < args_size.constant
  >               || sibcall_failure)
  >             tail_call_insns = NULL_RTX;
  > 
  > So, Jakub, what exactly prompted you to add the other bit,
  > a failure, or just inspection?  I'd probably agree that the
  > check should be moved up there next to the other argument
  > checks, just for clarity.
No idea why I put the test way down there.  Probably because it was
written at a different time than the others.  It does seem rather odd
since we can test it as soon as the know the size of the argument list.

I've got no objections to moving the test elsewhere, though like you I
would like to know why Jakub added the test elsewhere.

jeff