CWL fails to reconginze the existing files to generate new outputs using inplace editing

I am using CWL to generate some output based on certain existing output files in outdir.

---

## Expected Behavior

The CWL should take in the existing output files in the designated output directory and generate new outputs. The designated output dir consists of some .pickle files. The CWL should take them and generate single merged hdf5 file.

## Actual Behavior
Instead CWL scripts starts from the scratch. Which is processing wise time consuming

## Workflow Code
```
cwlVersion: v1.0
class: CommandLineTool
doc: Spladder

baseCommand: [python,-m,spladder.spladder]

hints:
cwltool:InplaceUpdateRequirement:
inplaceUpdate: true

requirements:
EnvVarRequirement:
envDef:
PYTHONPATH: /cluster/home/aalva/spladder/python
InlineJavascriptRequirement: {}
InitialWorkDirRequirement:
listing:
- entry: $(inputs.spladder_dir)
writable: true

inputs:
spladder_gtf:
type: File
inputBinding:
position: 3
prefix: -a
spladder_bams:
type: File[]
inputBinding:
position: 1
prefix: -b
itemSeparator: ","
secondaryFiles: .bai
spladder_dir:
type: Directory
inputBinding:
position: 2
prefix: -o
spladder_phase2:
type: string
inputBinding:
position: 4
prefix: -T
spladder_merge_graphs:
type: string
inputBinding:
position: 5
prefix: -M
spladder_qmode:
type: string
inputBinding:
position: 7
prefix: --qmode
spladder_quantify_graph:
type: string
inputBinding:
position: 6
prefix: --quantify_graph

outputs:
spladder_out:
type: Directory
outputBinding:
glob: $(inputs.spladder_dir.basename)

$namespaces:
cwltool: http://commonwl.org/cwltool#
```

## Full Traceback
```
Strted to run from scratch in temp file
```

## Your Environment
* cwltool version: v1.0

I ran it as `cwltool --enable-ext spl.cwl sp.yml`

Start by reviewing the provided CWL workflow, especially cwltool:InplaceUpdateRequirement and InitialWorkDirRequirement, then reproduce it with `cwltool --enable-ext spl.cwl sp.yml` using the reported cwltool v1.0 environment. Determine why the existing `.pickle` files in the designated output directory are not reused; done means the workflow recognizes them and produces the merged HDF5 output without restarting from scratch.

Written by the indexing model from the issue text.