[gfortran] I/O of large integer and real kinds, round 2.1

[Steve Kargl]

On Mon, Jun 20, 2005 at 11:31:46PM +0200, FX Coudert wrote:
> >>Built, tested and regtested on:
> >>   - i686-linux, which has real(10)
> >>   - i386-freebsd
> >>   - sparc-solaris, which has real(16)
> >>   - x86_64-linux, which has real(10) and integer(16)
> >>   - alpha-linux, which has integer(16)
> 
> Well, the patch doesn't apply cleanly on CVS anymore (due to another of 
> my patches), so here is the new diff (well, I didn't do all the 
> regtesting again; i686-linux is under way, though).
> 
> OK for mainline? (and should it be backported to 4.0, too?)
> 
> FX


I sprinkled some debugging printfs into write.c around line 385.
There is some type of problem with the implementation of log10l()
narrow HUGE() and TINY() for REAL*10 on amd64-*-freebsd.

abslog  = inf
edigits = -2147483647
buffer = +1.189731e+4932                
ndigits = -2147483622

-- 
steve