When set, it will conflict with any tests that generate bytecode
because they don't expect it to be set.
Fixable by sanitize the environment before forking, but we can't do
anything about import tests.
This is a pipenv default, and possibly a sane flag to set while doing
Hy development, so lets not let it be a hazard for developers to trip
over.
Instead of just checking that hy2py outputs a nonempty string and doesn't crash, we check that a hy2py-generated Python program works the same as the original Hy program.
This test suggests my plan to make hy2py output real Python has succeeded, so I updated NEWS accordingly.
HyKeywords are no longer an instances of string with a particular
prefix, but a completely separate object.
This means keywords no longer trip isinstance str checks, adding a
little bit of type safety to the compiler.
It also means that HyKeywords evaluate to themselves.
Closes#1352
This commit adds -E support for Hy. Similar to Python, hy will ignore
all PYTHON* environment variables, e.g. PYTHONPATH and PYTHONHOME,
that might be set.
Python Class AST expects a body which is a list of ast.Expr. Force
every entry to be stored as a statement. This means we'll preserve
print statements.
Python also doesn't construct docstrings in classes by setting a
__doc__ attribute, it does it by inspecting the first ast.Expr node of
the class. But this means we can remove the special handler for it.
This means that a HySymbol remembers its original name. That is, `a-b` and `a_b` are different symbols although `(setv a-b 1)` and `(setv a_b 1)` set the same variable (namely, `a_b`).
Most of the edits in this commit are to switch underscores to hyphens in places where mangling hasn't happened yet.
I removed some lexer tests since the lexer no longer does any mangling.
This removes a lot of hy-reprs that were hard-coded into the hy-repr function itself. It also allows you to add a hy-repr for an existing class without monkey-patching the class.
