User: rhaas Date: 2012/07/17 12:08 PM

Modified: /trunk/src/ GRHydro_Con2Prim.F90

Log: GRHydro: remove comparison of logical to .true.

I appreciate the effort to make the code compile with gcc by no longer using integers in place of logicals. I really do. However it seems that F0RTRAN in its infinite wisdom requires you to write

... .and. mustbisect .eqv. .true.

(note the all important 'v') when testing logicals. I have instead opted to remove the whole .eqv. term in favor of

... .and. mustbisect

From: Roland Haas roland.haas@physics.gatech.edu

File Changes:

Directory: /trunk/src/ ======================

File [modified]: GRHydro_Con2Prim.F90 Delta lines: +2 -2 =================================================================== --- trunk/src/GRHydro_Con2Prim.F90 2012-07-17 17:08:33 UTC (rev 401) +++ trunk/src/GRHydro_Con2Prim.F90 2012-07-17 17:08:35 UTC (rev 402) @@ -556,7 +556,7 @@ tmp = (utau + pnew + udens)**2 - s2 plow = max(pmin, sqrt(s2) - utau - udens)

- if (pnew .lt. plow .or. tmp .le. 0.0d0 .or. mustbisect .eq. .true.) then + if (pnew .lt. plow .or. tmp .le. 0.0d0 .or. mustbisect) then

! Ok, Newton-Raphson ended up finding something unphysical. ! Let's try to find our root via bisection (which converges slower but is more robust) @@ -1039,7 +1039,7 @@ tmp = (utau + pnew + udens)**2 - s2 plow = max(pminl, sqrt(s2) - utau - udens)

- if (pnew .lt. plow .or. tmp .le. 0.0d0 .or. mustbisect .eq. .true.) then + if (pnew .lt. plow .or. tmp .le. 0.0d0 .or. mustbisect) then

! Ok, Newton-Raphson ended up finding something unphysical. ! Let's try to find our root via bisection (which converges slower but is more robust)