Miscellanous fixups of a miscellaneous nature #171

Don't fork subprocesses in order to handle things in the stdlib (python has a very decent in-process C module for this). Also DRY. Also better chained error messages.

In nearly every case it is better to run subprocess.run() without the burden of forking a bash shell too. All it takes is to turn strings into arrays of strings and *remove* some use of formatting. Aurman specifically, does not use *any* of the fancy shell interpretation that sometimes makes forking a shell to fork the commands, a worthwhile endeavor. So don't do so. This is further cleanup after the removal of subprocesses for file operations. Only changes where this cleanly applies have been done for now... in a followup commit, subprocesses that handle arguments as heavily-quoted strings will be handled.

Some functions pass around lots of string-formatted arguments. Handling this properly would require major refactoring to use lists everywhere. For now, use shlex to convert strings to a proper list-based array that subprocess.run can elegantly handle without running its own shell parser.

Instead of decoding the stdout of subprocess.run and then doing our own writes, just hand subprocess the file object and let it write directly.

Appending is a bit faster than constucting new lists on top of the old ones. It also avoids the need to use two lists.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Miscellanous fixups of a miscellaneous nature #171

Miscellanous fixups of a miscellaneous nature #171

Commits on Jul 1, 2018