[gnu.gcc.bug] gcc-1.35 - constant floating point bug

andrewt@watsnew.waterloo.edu (Andrew Thomas) (06/07/89)

I am running a uVax II with Ultrix 2.0, gcc-1.35.  I have seen reports
for this bug before, and I am wondering if there is a simple fix.

The following program generates the error noted below it:

-----------------------

main()
{
  float	x;
  x=1;
  printf ("%g\n",x);
}

gccbug.c:4: Magnitude of constant too large for 'float'

--------------------------------

The same behaviour occurs if line 4 is changed to x=1.0.  This can be
circumvented if I say:
  int y = 1;
  x = y;

But what happens when I want x=1.5 ?
--

Andrew Thomas
andrewt@watsnew.waterloo.edu	Systems Design Eng.	University of Waterloo

hallvard@IFI.UIO.NO (Hallvard B Furuseth) (06/07/89)

This may be a bug with the -O option.

I have tried to install gcc-1.35 on a vax with ultrix 2.0, but comparing
the stage1 and stage2 files fails.  The comparison succeeds when I
leave out the -O option from the lines

make CC=stage[12]/gcc CFLAGS="-g -O -Bstage[12]/"
during installation.

I can't send details.  The disc is down, and it seems to intend to stay down
for a while.  If someone asks for details, I will mail them when possible.

Hallvard Furuseth
hallvard@ifi.uio.no