Update CI

[zmoon]

• moving to the new micromamba Action
• add more debug flags for GFortran
• similar for Intel
• DEBUG=2 option and document (29863dc)

[zmoon]

@drnimbusrain how do we want to handle FPEs? First one is here (case that hcmref is 0):

--- a/src/canopy_calcs.F90
+++ b/src/canopy_calcs.F90
@@ -78,8 +78,12 @@ SUBROUTINE canopy_calcs
 ! ... calculate wind speed from u and v
                 ubzref   = sqrt((uref**2.0) + (vref**2.0))

-! ... get scaled canopy model profile and sub-canopy layers
-                zhc         = zk/hcmref
+                ! ... get scaled canopy model profile and sub-canopy layers
+                if (hcmref > 0) then
+                    zhc = zk / hcmref
+                else
+                    zhc = 0
+                end if
                 cansublays  = floor(hcmref/modres)

 ! ... check for valid model vegetation types

[drnimbusrain]

I agree with your first FPE change here. Do you want to make this (and other) change to your branch as the FPEs come up with your improved debugging options?

[drnimbusrain]

@zmoon Do you think you will be able to work on this again soon? Thank you!

[zmoon]

@drnimbusrain I'll try to work on it again soon.

[drnimbusrain]

Thanks Zach...I added the additional debug flags and checked/solved the FPEs in latest develop.

…

On Wed, Jan 24, 2024, 1:36 PM Zachary Moon ***@***.***> wrote: @drnimbusrain <https://github.com/drnimbusrain> I'll try to work on it again soon. — Reply to this email directly, view it on GitHub <#85 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGLFYNQFPNHD3JUWKZGQXHLYQFIBDAVCNFSM6AAAAAA3HJFUT2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBYG4YDOMJYG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[zmoon]

That's good, I'll merge that into here before working on it more then.

[zmoon]

@drnimbusrain some notes:

• CI debug build still has DEBUG=1 not 2
• we're currently not testing all code paths, just the default namelist, so there may be other FPEs
• "array temporary was created" warnings we could alleviate by explicitly creating the temporary arrays. This is mainly for going from an array of floats to assigning to a float member of an array of derived type in the ncf read/write.

[drnimbusrain]

@zmoon Thank you for working on this, and agree that more comprehensive checks of the different code pathways/options are needed to check for all FPEs. I will try and work on this with your new debug options more thoroughly as we go along with more robust testing. I think that @btang1 @MaggieMarvin @angehung5 could help with this DEBUG=2 testing of the different options for FPEs as well, since they are going to be focusing on different parts of code. For now I approve, and we can improve later.

[drnimbusrain]

I can also work on the "array temporary was created" issues too. Thanks!

[drnimbusrain]

Looks OK, but just want to know about the leafage option changes here.

[drnimbusrain]

@zmoon Although, what has changed in the leafage option with your PR?

[zmoon]

Should be nothing. I just moved the calculation under the if statement to avoid FPE related to the LAI being unitialized. Little note in d241bea

[drnimbusrain]

OK, I will merge then and work with the DEBUG=2 option some to fix some other potential FPEs and the warnings...