Additionally, reverted FakeSystemExit subclassing Exception, but added
an explicit __init__ so that we can emulate the SystemExit.code
functionality without having to change the superclass.
Unittest just checks that both `pubs` and `pubs --help` raise a
`SystemExit` exception with error code 0. Due to how argparse
handles the `--help` keyword, this is the best way I could think to
provide test coverage without heavily modifying the parser structure
or the unittest infrastructure.
To ensure the `pubs` matches the behavior of `pubs --help`, it now
raises the same `SystemExit(0)` exception via `sys.exit(0)`. And
in order to catch it in the unittest, I had to modify the
`FakeSystemExit` behavior slightly.
Made subparsers not required in the main parser, but if no commands are parsed,
then the parser prints its help. When a command is present, the normal
functionality preserved.
Removed the single unittest that checks for this exact behavior.
Basically the validation takes care of setting results that
post-processing was relying on. Hence the validation should occur before
the postprocessing. It actually makes more sense to validate within
post-process.
The default behavior from the requests library is to use the guess from
chardet which is not always reliable while doi.org seems to always
return utf8 encoded data. It's unlikely that this changes without also
updating the header.
For now, the editor_input and edit_file functions do not make sense
outside of the context of an ui. Having them in the content module
was confusing. (See issue fixed by a600855.)
Replace the in-house stubbing by the facilities offered
by pyfakefs for running tests. We don't use `io.open`
anymore, because pyfakefs does not stub it correctly.
We rely on a patched version of pyfakefs, as a few bug
remain in the official repo.
