noresm2_3_develop : CAM_CONFIG_OPTS obtains double arguments in case of full-chemistry

[DirkOlivie]

What happened?

When making a full-chemistry compset, CAM_CONFIG_OPTS (seen in env_build.xml) obtains certain arguments double. It looks like :

<entry id="CAM_CONFIG_OPTS" value="-phys cam6 -chem trop_mam_oslo -camnor  -chem trop_mam_oslo -camnor -cosp -chem tropstrat_mam_oslo"> 

So -chem trop_mam_oslo appears twice (it should not be there at all) and -camnor appears twice (it should appear only once). At the end -chem tropstrat_mam_oslo appears once (and that is fine) : I assume that for -chem the last appearance is finally taken into account during the building process, and therefore the compset is actually build as a full-chemistry compset.

These double appearances find their origin probably in the additive choice in config_component.xml :

<entry id="CAM_CONFIG_OPTS">
    ...
    <values match="last" modifier='additive'>
    ...
    <value compset="_CAM60%NORESM"            >-chem trop_mam_oslo -camnor </value>
    <value compset="_CAM60%NORESM%NORBC"      >-chem trop_mam_oslo -camnor -cosp</value>
    <value compset="_CAM60%NORESM%FSST"       >-chem trop_mam_oslo -camnor -cosp</value>
    <value compset="_CAM60%NORESM%NORPDDMSBC" >-chem trop_mam_oslo -camnor -cosp</value>
    <value compset="_CAM60%NORESM%NORPIBC"    >-chem trop_mam_oslo -camnor -cosp</value>
    <value compset="_CAM60%NORESM%NUDGE"      >-chem trop_mam_oslo -camnor -offline_dyn</value>
    <value compset="_CAM60%NORESM%TROPSTRATCHEM"       >-chem tropstrat_mam_oslo </value>
    <value compset="_CAM60%NORESM%NORBC%TROPSTRATCHEM" >-chem tropstrat_mam_oslo </value>

So when choosing a full-chemistry compset (e.g., with NORESM%NORBC%TROPSTRATCHEM) , it will find agreement for all 3 lines :

<value compset="_CAM60%NORESM"            >-chem trop_mam_oslo -camnor </value>
<value compset="_CAM60%NORESM%NORBC"      >-chem trop_mam_oslo -camnor -cosp</value>
<value compset="_CAM60%NORESM%NORBC%TROPSTRATCHEM" >-chem tropstrat_mam_oslo </value>

leading to the double (or unintented) appearance of certain arguments.

I think the model is running fine, but maybe we should try to find a better and cleaner solution. Suggestions are welcome ...

What are the steps to reproduce the bug?

Run the case in /cluster/projects/nn9560k/olivie/cases-noresm2.3-test/NF1850norbc_tropstratchem_f19_f19_20241127_test16 on betzy.

What CAM-Nor tag were you using?

cam_noresm2_3_v1.0.0d

What machine were you running CAM (or NorESM) on?

Betzy

What compiler were you using?

Intel

Path to a case directory, if applicable

/cluster/projects/nn9560k/olivie/cases-noresm2.3-test/NF1850norbc_tropstratchem_f19_f19_20241127_test16 (on Betzy)

Will you be addressing this bug yourself?

Yes, but I will need some help

Extra info

No response