Introduce the File tool and tweak stub binary implementation

[rmartin16]

Changes

• Introduce the File tool
  • Encapsulates complex file operations such as archive unpacking
  • Incorporates the (now removed) Download tool's functionality
• Some updates to the stub binary error handling

Notes

• I was making the stub binary updates while rebasing Implement persistent build tracking for more intuitive behavior #1714....but it seems more appropriate on its own

PR Checklist:

• All new features have been tested
• All new features have been documented
• I have read the CONTRIBUTING.md file
• I will abide by the code of conduct

[rmartin16]

If you just call repr() here and filename is a Path, it looks like this:

Unable to unpack support package PosixPath("/asdf")

so, this strategy gets what we actually want.

Unable to unpack support package '/asdf'

[freakboy3742]

Torn between whether we should use the explicit ', or {str(filename)!r} - that should ensure that any inline quotes are correctly escaped (I think...)

[rmartin16]

You mean like these differences?

>>> file_path=Path("/pa't'h/to/file")
>>> 
>>> print(f"File is {file_path}")
File is /pa't'h/to/file
>>> 
>>> print(f"File is {str(file_path)!r}")
File is "/pa't'h/to/file"
>>> 
>>> print(f"File is '{file_path}'")
File is '/pa't'h/to/file'

Manually calling str() does seem to handle that edge case best.

[freakboy3742]

That's what I had in mind, yes.

[rmartin16]

ive nerdsniped myself in to implementing a File tool....

[rmartin16]

The updates for the File tool were relatively extensive but also pretty banal. Reviewing the changes for each commit will help discern what I changed for the stub binary versus the File tool.

[freakboy3742]

A large PR, but largely straightforward, with some nice cleanups along the way. One minor suggestion about some possibly overenthusiastic modularity; weak preference would be to remove it, but if you feel particularly strongly to the contrary, I'm not going to fight you too hard over it.

[freakboy3742]

Possibly a little over-enthusiastic refactoring here; function calls aren't free, and in practice, this method can't be overwritten by a subclass. Given it's wrapping a single if clause that won't be needed in a couple of Python releases, I'm not sure the extra method is worth it.

[rmartin16]

That makes sense. I also tweaked it a bit so (in theory, anyway) pyupgrade should have an easier time trimming it down later on.

[rmartin16]

A new problem from this is a corrupted build directory if the stub can't be installed.

Wasn't there work being done to ensure the bundle was deleted if the CreateCommand goes bad?

[freakboy3742]

👍 LGTM.