Saneyaml doesn't support whitespaces/empty lines

[AyanSinhaMahapatra]

Reference: aboutcode-org/scancode-toolkit#3219

We use pyyaml as YAML dumper with unicode support set to True, and pyyaml also fails to load yaml objects which has text with irregular whitespace in them. An example:

Text at path:

                                 Apache License
                           Version 2.0, January 2004

   TERMS AND CONDITIONS

We run the following to create a yaml file:

license_location = test_env.get_test_loc('yaml/simple-license.txt')
with io.open(license_location, encoding='utf-8') as res:
    license_text = res.read()
data = {}
data["license_text"] = license_text
yaml_string = saneyaml.dump(data, indent=4)
yaml_location = test_env.get_test_loc('yaml/simple-license.yaml')
with io.open(yaml_location, 'w') as o:
    o.write(yaml_string)

Now on this, both loaded_yaml = saneyaml.load(yaml_string) and loaded_yaml = yaml.load(yaml_string) and yaml.CSafeLoader(yaml_string).get_data() fails with the same error as the referenced issue above.

We need to update the dump function in Saneyaml to modify the dumper to check for irregular whitespace and only dump valid yaml.

[AyanSinhaMahapatra]

There was two possibilities here:

1. We can either make the code such that all license texts are YAML safe, i.e. add spaces when we have empty newlines (in files which had different indentations).
2. Or we could modify the license texts itself to have this.

We are moving forward with 2. as this makes us consistant and not change the license texts, and we are not having to update a lot of license texts there anyway, only the ones which were producing invalid YAML.