perl namelist should allow both "false" and ".false."

[rljacob]

The MPAS-CICE buildnml is complaining about a logical a user is trying to set using user_nl_mpascice.

config_test_case_diag = false

'ERROR(Build::Namelist::_parse_next): expect a F90 constant for a namelist instead got: false

If they set it instead to

config_test_case_diag = .false.

There there's no error. But both "false" and ".false." should be allowed.

[mvertens]

I assume that this is using the perl old schema with namelist_definition.xml and namelist_defaults.xml. How is config_test_case_diag defined in the namelist_definition.xml?

…

On Mon, Mar 27, 2017 at 1:04 PM, Robert Jacob ***@***.***> wrote: The MPAS-CICE buildnml is complaining about a logical a user is trying to set in using user_nl_mpascice. config_test_case_diag = false 'ERROR(Build::Namelist::_parse_next): expect a F90 constant for a namelist instead got: false If they set it instead to config_test_case_diag = .false. But both "false" and ".false." should be allowed. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1283>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHlxE70z7Td6sYpW_gkNgZtald5qm0jlks5rqAgygaJpZM4MqsF7> .

[rljacob]

Yes that buildnml is still in perl.

In namelist definition:

<entry id="config_test_case_diag" type="logical"
        category="io" group="io">
MISSING DESCRIPTION

Valid values: MISSING POSSIBLE VALUES
Default: Defined in namelist_defaults.xml
</entry>

In namelist_defaults: <config_test_case_diag>false</config_test_case_diag>

[mvertens]

I believe that perl is assuming that a logical will be fortran namelist logical - i.e. .true. or .false. To get the behavior he wants - I would make it a character and then give the valid values as "true" or "false". Hopefully, this will work.

…

On Mon, Mar 27, 2017 at 1:20 PM, Robert Jacob ***@***.***> wrote: Yes that buildnml is still in perl. In namelist definition: <entry id="config_test_case_diag" type="logical" category="io" group="io"> MISSING DESCRIPTION Valid values: MISSING POSSIBLE VALUES Default: Defined in namelist_defaults.xml </entry> In namelist_defaults: <config_test_case_diag>false</config_test_case_diag> — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#1283 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHlxE8jH0WsmcthPNCnpKwZWVAT5BtVCks5rqAvtgaJpZM4MqsF7> .

[akturner]

The fortran standard allows both false and .false. as valid fortran in namelists. The standard actually allows the "f" character (or "t" character for true obviously) followed by anything as a valid value.

[mvertens]

I think this is a limitation of our perl modules for namelist parsing. On Mon, Mar 27, 2017 at 1:33 PM, Mariana Vertenstein <mvertens@ucar.edu> wrote:

…

I believe that perl is assuming that a logical will be fortran namelist logical - i.e. .true. or .false. To get the behavior he wants - I would make it a character and then give the valid values as "true" or "false". Hopefully, this will work. On Mon, Mar 27, 2017 at 1:20 PM, Robert Jacob ***@***.***> wrote: > Yes that buildnml is still in perl. > > In namelist definition: > > <entry id="config_test_case_diag" type="logical" > category="io" group="io"> > MISSING DESCRIPTION > > Valid values: MISSING POSSIBLE VALUES > Default: Defined in namelist_defaults.xml > </entry> > > In namelist_defaults: <config_test_case_diag>false</ > config_test_case_diag> > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#1283 (comment)>, or mute > the thread > <https://github.com/notifications/unsubscribe-auth/AHlxE8jH0WsmcthPNCnpKwZWVAT5BtVCks5rqAvtgaJpZM4MqsF7> > . >

[rljacob]

Does the python parser allow both "false" and ".false."?

[rljacob]

Yes python allows both.

[erichlf]

The issue appears to be in the regex used. I am testing a solution now.