Task factory "task per file" fails with source "containerUrl"

[tojozefi]

Using task factory "task per file" with source defined as container URL fails with error:
AttributeError: 'SplitResult' object has no attribute 'pathname'

Here is extract from the job template:
{
(...)
"parameters":{
(...)
"inputContainerUrl":{
"type":"string",
"defaultValue":"https://tojobatch.blob.core.windows.net/fgrp-sequencer-input?st=2020-04-29T20%3A20%3A06Z&se=2020-05-30T20%3A20%3A00Z&sp=rl&sv=2018-03-28&sr=c&sig=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
},
(...)
"job":{
(...)
"taskFactory":{
"type":"taskPerFile",
"source":{
"containerUrl": "[parameters('inputContainerUrl')]"
},
(...)

The same template works ok when using source with "fileGroup" parameter.

Software level:
$ az --version
azure-cli 2.4.0 *
command-modules-nspkg 2.0.3
core 2.4.0 *
nspkg 3.0.4
telemetry 1.0.4
Extensions:
image-copy-extension 0.0.9
azure-batch-cli-extensions 5.0.1

[bgklein]

I believe this section of code was never actually functional:

azure-batch-cli-extensions/azext/batch/_file_utils.py

Lines 346 to 351 in 14913ee

	if resource_file.source.container_url:
	# Input data storage in arbitrary container
	uri = urlsplit(resource_file.source.container_url)
	container = uri.pathname.split('/')[1]
	blobs = self.list_container_contents(resource_file.source, container, storage_client)
	return convert_blobs_to_resource_files(blobs, resource_file)

(or at least not for the last three years). According to urllib.parse.urlsplit documentation for both py2 and py3 and manually executing these 5 lines on your url it looks to be path instead of pathname.

I'll look to update this in our next version as it really should be using our native support for containers now anyways rather than manually listing the contents (https://docs.microsoft.com/en-us/python/api/azure-batch/azure.batch.models.resourcefile?view=azure-python).

[tojozefi]

Any plans on solving this issue?