chore(semantics): Remove determine_main (#702)

Vendor `rules_python`'s `determine_main` and remove the need for a
separate `determine_main` target which complicates exec properties.

Replaces #641 with thanks to Keith.
Fixes #605
Fixes #621

Note that `py_venv_*` already mandates that a `main=` label be provided,
so this is for legacy `py_binary` only. Should be removed in 2.0.

### Changes are visible to end-users: no

- Searched for relevant documentation and updated as needed: yes
- Breaking change (forces users to change their own code or config): no
- Suggested release notes appear below: yes

The internal `determine_main` targets have been eliminated, simplifying some advanced usages.

### Test plan

- Covered by existing test cases

---------

Co-authored-by: aspect-marvin[bot] <[email protected]>

Author: arrdem

py/BUILD.bazel

@@ -27,7 +27,6 @@ bzl_library(
     visibility = ["//visibility:public"],
     deps = [
         "//py/private:py_binary",
-        "//py/private:py_executable",
         "//py/private:py_image_layer",
         "//py/private:py_library",
         "//py/private:py_pex_binary",

py/defs.bzl

@@ -35,9 +35,7 @@ python.toolchain(python_version = "3.9", is_default = True)
 ```
 """
 
-load("@aspect_bazel_lib//lib:utils.bzl", "propagate_common_rule_attributes")
 load("//py/private:py_binary.bzl", _py_binary = "py_binary", _py_test = "py_test")
-load("//py/private:py_executable.bzl", "determine_main")
 load("//py/private:py_image_layer.bzl", _py_image_layer = "py_image_layer")
 load("//py/private:py_library.bzl", _py_library = "py_library")
 load("//py/private:py_pex_binary.bzl", _py_pex_binary = "py_pex_binary")
@@ -62,32 +60,13 @@ py_image_layer = _py_image_layer
 
 resolutions = _resolutions
 
-def _py_binary_or_test(name, rule, srcs, main, data = [], deps = [], resolutions = {}, **kwargs):
-    exec_properties = kwargs.pop("exec_properties", {})
-    non_test_exec_properties = {k: v for k, v in exec_properties.items() if not k.startswith("test.")}
-
-    # Compatibility with rules_python, see docs in py_executable.bzl
-    main_target = "{}.find_main".format(name)
-    determine_main(
-        name = main_target,
-        target_name = name,
-        main = main,
-        srcs = srcs,
-        exec_properties = non_test_exec_properties,
-        **propagate_common_rule_attributes(kwargs)
-    )
-
-    package_collisions = kwargs.pop("package_collisions", None)
-
+def _py_binary_or_test(name, rule, srcs, main, data = [], deps = [], **kwargs):
     rule(
         name = name,
         srcs = srcs,
-        main = main_target,
+        main = main,
         data = data,
         deps = deps,
-        resolutions = resolutions,
-        package_collisions = package_collisions,
-        exec_properties = exec_properties,
         **kwargs
     )
 
@@ -97,8 +76,6 @@ def _py_binary_or_test(name, rule, srcs, main, data = [], deps = [], resolutions
         data = data,
         deps = deps,
         imports = kwargs.get("imports"),
-        resolutions = resolutions,
-        package_collisions = package_collisions,
         tags = ["manual"],
         testonly = kwargs.get("testonly", False),
         target_compatible_with = kwargs.get("target_compatible_with", []),

py/private/BUILD.bazel

@@ -92,11 +92,6 @@ bzl_library(
     srcs = ["providers.bzl"],
 )
 
-bzl_library(
-    name = "py_executable",
-    srcs = ["py_executable.bzl"],
-)
-
 bzl_library(
     name = "transitions",
     srcs = ["transitions.bzl"],

py/private/py_binary.bzl

@@ -18,6 +18,9 @@ def _py_binary_rule_impl(ctx):
     venv_toolchain = ctx.toolchains[VENV_TOOLCHAIN]
     py_toolchain = _py_semantics.resolve_toolchain(ctx)
 
+    # Resolve our `main=` to a label, which it isn't
+    main = _py_semantics.determine_main(ctx)
+
     # Check for duplicate virtual dependency names. Those that map to the same resolution target would have been merged by the depset for us.
     virtual_resolution = _py_library.resolve_virtuals(ctx)
     imports_depset = _py_library.make_imports_depset(ctx, extra_imports_depsets = virtual_resolution.imports)
@@ -78,7 +81,7 @@ def _py_binary_rule_impl(ctx):
             "{{ARG_PYTHON}}": to_rlocation_path(ctx, py_toolchain.python) if py_toolchain.runfiles_interpreter else py_toolchain.python.path,
             "{{ARG_VENV_NAME}}": ".{}.venv".format(ctx.attr.name),
             "{{ARG_PTH_FILE}}": to_rlocation_path(ctx, site_packages_pth_file),
-            "{{ENTRYPOINT}}": to_rlocation_path(ctx, ctx.file.main),
+            "{{ENTRYPOINT}}": to_rlocation_path(ctx, main),
             "{{PYTHON_ENV}}": "\n".join(_dict_to_exports(default_env)).strip(),
             "{{EXEC_PYTHON_BIN}}": "python{}".format(
                 py_toolchain.interpreter_version_info.major,
@@ -114,7 +117,7 @@ def _py_binary_rule_impl(ctx):
         DefaultInfo(
             files = depset([
                 executable_launcher,
-                ctx.file.main,
+                main,
                 site_packages_pth_file,
             ]),
             executable = executable_launcher,
@@ -140,9 +143,21 @@ _attrs = dict({
         default = {},
     ),
     "main": attr.label(
-        doc = "Script to execute with the Python interpreter.",
         allow_single_file = True,
-        mandatory = True,
+        doc = """
+Script to execute with the Python interpreter.
+
+Must be a label pointing to a `.py` source file.
+If such a label is provided, it will be honored.
+
+If no label is provided AND there is only one `srcs` file, that `srcs` file will be used.
+
+If there are more than one `srcs`, a file matching `{name}.py` is searched for.
+This is for historical compatibility with the Bazel native `py_binary` and `rules_python`.
+Relying on this behavior is STRONGLY discouraged, may produce warnings and may
+be deprecated in the future.
+
+""",
     ),
     "venv": attr.string(
         doc = """The name of the Python virtual environment within which deps should be resolved.

py/private/py_executable.bzl

@@ -85,7 +85,61 @@ def _resolve_toolchain(ctx):
         flags = _INTERPRETER_FLAGS,
     )
 
+def _csv(values):
+    """Convert a list of strings to comma separated value string."""
+    return ", ".join(sorted(values))
+
+def _path_endswith(path, endswith):
+    # Use slash to anchor each path to prevent e.g.
+    # "ab/c.py".endswith("b/c.py") from incorrectly matching.
+    return ("/" + path).endswith("/" + endswith)
+
+def _determine_main(ctx):
+    """Determine the main entry point .py source file.
+
+    Args:
+        ctx: The rule ctx.
+
+    Returns:
+        Artifact; the main file. If one can't be found, an error is raised.
+    """
+    if ctx.attr.main:
+        if not ctx.attr.main.label.name.endswith(".py"):
+            fail("main must end in '.py'")
+
+        # Short circuit; if the user gave us a label, believe them.
+        return ctx.file.main
+
+    elif len(ctx.files.srcs) == 1:
+        # If the user only provided one src, take that
+        return ctx.files.srcs[0]
+
+    else:
+        # Legacy rule name based searching :/
+        if ctx.label.name.endswith(".py"):
+            fail("name must not end in '.py'")
+        proposed_main = ctx.label.name + ".py"
+
+        print(ctx.files.srcs)
+        main_files = [src for src in ctx.files.srcs if _path_endswith(src.short_path, proposed_main)]
+
+        if len(main_files) > 1:
+            fail(("file name '{}' specified by 'main' attributes matches multiple files. " +
+                  "Matches: {}").format(
+                proposed_main,
+                _csv([f.short_path for f in main_files]),
+            ))
+
+        elif len(main_files) == 1:
+            return main_files[0]
+
+        else:
+            fail("{} does not specify main=, and has multiple sources. Disambiguate the entrypoint".format(
+                ctx.label,
+            ))
+
 semantics = struct(
     interpreter_flags = _INTERPRETER_FLAGS,
     resolve_toolchain = _resolve_toolchain,
+    determine_main = _determine_main,
 )