fix: to have recon-all work with 5.3 on OSX

[satra]

@dgellis90 and @hjmjohnson - could you please take a look at these changes? this did allow running from start to finish.

from nipype import config
config.enable_provenance()
import os
from nipype.workflows.smri.freesurfer import create_reconall_workflow
recon_all = create_reconall_workflow()
recon_all.inputs.inputspec.subject_id = 'time'
recon_all.inputs.inputspec.subjects_dir = os.getcwd()
recon_all.inputs.inputspec.T1_files = os.path.abspath('mprage003.nii.gz')
recon_all.base_dir = os.getcwd()
recon_all.run(plugin='MultiProc', plugin_args={'n_procs': 4})

also subject_id did not seem to get carried along into the sub commands. so i saw commands like this where the --subject is called with subejct_id instead of 'time' as i used in calling the workflow.

Running: mri_segstats --brainmask brainmask.mgz --ctab /Applications/freesurfer/WMParcStatsLUT.txt
 --etiv --excludeid 0 --in /software/temp/share_blake/ReconAll/AutoRecon2/CA_Normalize/norm.mgz
 --in-intensity-name norm --in-intensity-units MR --pv /software/temp/share_blake/ReconAll/AutoRecon2/CA_Normalize/norm.mgz --seg 
/software/temp/share_blake/ReconAll/AutoRecon3/WM_Parcellation/wmparc.mgz --subject subject_id
 --surf-wm-vol --sum ./wmparc.stats```

[coveralls]

Coverage remained the same at 72.528% when pulling bf46eef on satra:fix/reconwf into a8e3730 on nipy:master.

[ellisdg]

I think the changes look good. I made a PR to your branch that adds an explanation how the subject_id is handled in the workflow.

[coveralls]

Coverage remained the same at 72.528% when pulling 6e40eac on satra:fix/reconwf into a8e3730 on nipy:master.