zero**zero in ElementDescription.F90 [FIXED]

Clearly defined bug reports and their fixes
Post Reply
Victor
Posts: 4
Joined: 18 Jun 2020, 18:47
Antispam: Yes

zero**zero in ElementDescription.F90 [FIXED]

Post by Victor »

Running several test cases, one of them is AdvDiffFluxes, at ElementDescription.F90 lines 789 and 826

s = s + Coeff(i) * u**p(i)

both u and p(i) are zero, thus trying to compute zero**zero.

Easy to check by putting

if(u==0.and.p(i)==0) call Fatal('ElementDescription','Attempted computation of 0**0')

just before.

Can anybody reproduce the bug?
What is the fix?
mika
Posts: 230
Joined: 15 Sep 2009, 07:44

Re: zero**zero in ElementDescription.F90

Post by mika »

The devel version has now been updated to fix this.

Thanks for reporting,
Mika
Victor
Posts: 4
Joined: 18 Jun 2020, 18:47
Antispam: Yes

Re: zero**zero in ElementDescription.F90 [FIXED]

Post by Victor »

There is at least another place in ElementDescription.F90 computing 0**0 at line 1008.
But by now I know how to fix it.
However running AdvDiffFluxes I run into another bug, that I am describing in another issue
concerning referencing a pointer not allocated.
Post Reply