This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: NIST Update - test case diff
Jerry DeLisle wrote:
Paul Thomas wrote:
Jerry,
In the process I discovered 4 latent failures had existed that we did
not catch by inspection before when I established my initial
reference set. I manually checked the results of those four tests
and confirmed that we are getting them correct now.
Really, what were they?
See the following, the FM257 was simply because I did not use a driver to give
it the five "go" inputs after pauses. The others were definitely wrong!
$ diff new old
diff new/FM110.ref old/FM110.ref
257c257
< 1.2345 12.345 123.45 1234.5
---
> 1.2345 12.345 1234.50 12345.0
Only in new: fm257.ref
diff new/FM257.ref old/FM257.ref
14,17d13
< 2 PASS
< 3 PASS
< 4 PASS
< 5 PASS
diff new/FM403.ref old/FM403.ref
339c339
< 98.7654 0.9877E+04 987654.00
---
> 98.7654 0.9877E+04 9876.54
345c345
< 987.654 0.8648E+04 8647.860
---
> 987.654 0.8648E+04 86.479
diff new/FM900.ref old/FM900.ref
289c289
< 98.7654 0.9877E+04 987654.00 987.654
---
> 98.7654 0.9877E+04 9876.54 987.654
295c295
< 0.8648D-02 0.8648E+04 8647.860
---
> 0.8648D-02 0.8648E+04 86.479