[Bug fortran/70235] [4.9/5/6 Regression] Incorrect output with PF format

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70235

--- Comment #11 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Created attachment 38075 [details]
> A patch for testing

With the patch and using the test attached to comment 5 with y = 1.0 and d=8, I
get the following output

-8pf18.8 y=        0.0000000.
-7pf18.8 y=        0.000000.0
-6pf18.8 y=        0.00000.00
-5pf18.8 y=        0.0000.000
-4pf18.8 y=        0.000.0000
-3pf18.8 y=        0.00.00000
-2pf18.8 y=        0.0.000000
-1pf18.8 y=        0.10000000
0pf18.8 y=        1.00000000
1pf18.8 y=       10.00000000
2pf18.8 y=      100.00000000
3pf18.8 y=     1000.00000000
4pf18.8 y=    10000.00000000
5pf18.8 y=   100000.00000000
6pf18.8 y=  1000000.00000000
7pf18.8 y= 10000000.00000000

which is wrong for P<=-2.