Make rezipping wheels optional and off by default

example/BUILD

@@ -40,6 +40,9 @@ py_test(
     name = "hello_native",
     main = "hello.py",
     srcs = ["hello.py"],
+    deps = [
+        py_requirement("pyyaml"),
+    ]
 )
 
 zapp_test(
@@ -73,7 +76,6 @@ zapp_test(
     ],
 )
 
-
 zapp_test(
     name = "hello_unzipped",
     zip_safe = False,

zapp/BUILD

@@ -35,7 +35,7 @@ toolchain(
 # Zapp plugins used as a runtime library by rules_zapp
 py_library(
     name = "zapp_support",
-    srcs = glob(["support/**/*.py"]),
+    srcs = ["__init__.py"] + glob(["support/**/*.py"]),
     imports = [
         "..",
     ]

zapp/compiler/__main__.py

@@ -6,9 +6,11 @@ import argparse
 import json
 import os
 import pathlib
+import re
 import stat
 import sys
 import zipfile
+from collections import defaultdict
 from email.parser import Parser
 from itertools import chain
 from pathlib import Path
@@ -20,6 +22,9 @@ from zapp.support.unpack import cache_wheel_path
 parser = argparse.ArgumentParser(description="The (bootstrap) Zapp compiler")
 parser.add_argument("-o", "--out", dest="output", help="Output target file")
 parser.add_argument("-d", "--debug", dest="debug", action="store_true", default=False)
+parser.add_argument(
+    "--use-wheels", dest="use_wheels", action="store_true", default=False
+)
 parser.add_argument("manifest", help="The (JSON) manifest")
 
 
@@ -51,6 +56,9 @@ for script in {scripts!r}:
 """
 
 
+whl_workspace_pattern = re.compile(r"^external/(?P<workspace>[^/]*?)/site-packages/")
+
+
 def dsub(d1: dict, d2: dict) -> dict:
     """Dictionary subtraction. Remove k/vs from d1 if they occur in d2."""
 
@@ -76,7 +84,7 @@ def dir_walk_prefixes(path):
         yield os.path.join(*segments)
 
 
-def load_wheel(opts, manifest, path):
+def load_wheel(opts, path):
     """Load a single wheel, returning ..."""
 
     def _parse_email(msg):
@@ -97,17 +105,8 @@ def load_wheel(opts, manifest, path):
     with open(os.path.join(path, "WHEEL")) as wheelf:
         wheel = _parse_email(wheelf.read())
 
-    prefix = os.path.dirname(path)
-
     # Naive glob of sources; note that bazel may hvae inserted empty __init__.py trash
-    sources = [
+    sources = []
-        (
-            dest,
-            spec,
-        )
-        for dest, spec in manifest["sources"].items()
-        if spec["source"].startswith(prefix)
-    ]
 
     # Retain only manifest-listed sources (dealing with __init__.py trash, but maybe not all conflicts)
     with open(os.path.join(path, "RECORD")) as recordf:
@@ -187,13 +186,23 @@ def rezip_wheels(opts, manifest):
 
     Wheels which are unzipped should be re-zipped into the cache, if not present in the cache.
 
-    Files sourced from unzipped wheels should be removed, and a single wheel reference inserted."""
+    Files sourced from unzipped wheels should be removed, and a single wheel reference inserted.
+    """
 
-    wheels = [
+    whl_srcs = defaultdict(dict)
-        load_wheel(opts, manifest, os.path.dirname(s["source"]))
+    for k, s in list(manifest["sources"].items()):
-        for _, s in manifest["sources"].items()
+        src = s["source"]
-        if s["source"].endswith("/WHEEL")
+        m = re.match(whl_workspace_pattern, src)
-    ]
+        if m:
+            whl_srcs[m.group(1)][re.sub(whl_workspace_pattern, "", src)] = s
+            del manifest["sources"][k]
+
+    wheels = []
+    for bundle in whl_srcs.values():
+        whlk = next((k for k in bundle.keys() if k.endswith("WHEEL")), None)
+        whl_manifest = load_wheel(opts, os.path.dirname(bundle[whlk]["source"]))
+        whl_manifest["sources"].update(bundle)
+        wheels.append(whl_manifest)
 
     manifest["requirements"] = {}
 
@@ -201,8 +210,6 @@ def rezip_wheels(opts, manifest):
     for w in wheels:
         # Try to cheat and hit in the local cache first rather than building wheels every time
         wn = wheel_name(w)
-        # Expunge sources available in the wheel
-        manifest["sources"] = dsub(manifest["sources"], w["sources"])
 
         # We may have a double-path dependency.
         # If we DON'T, we have to zip
@@ -302,7 +309,6 @@ def enable_unzipping(opts, manifest):
 
 
 def fix_sources(opts, manifest):
-
     manifest["sources"] = {f: m for f, m in manifest["sources"]}
 
     return manifest
@@ -318,6 +324,7 @@ def main():
         setattr(opts, "tmpdir", d)
 
         manifest = fix_sources(opts, manifest)
+        if opts.use_wheels:
             manifest = rezip_wheels(opts, manifest)
         manifest = ensure_srcs_map(opts, manifest)
         manifest = enable_unzipping(opts, manifest)
@@ -339,11 +346,15 @@ def main():
                     "manifest": manifest,
                 },
                 sys.stdout,
-                indent=2
+                indent=2,
             )
 
         with open(opts.output, "w") as zapp:
-            shebang = "#!" + manifest["shebang"] + "\n"
+            shebang = manifest["shebang"]
+            if not shebang.endswith("\n"):
+                shebang = shebang + "\n"
+            if not shebang.startswith("#!"):
+                shebang = "#!" + shebang
             zapp.write(shebang)
 
         # Now we're gonna build the zapp from the manifest

zapp/zapp.bzl

@@ -141,6 +141,14 @@ def _zapp_impl(ctx):
         is_executable = False,
     )
 
+    args = [
+        "--debug",
+        "-o", ctx.outputs.executable.path,
+        manifest_file.path
+    ]
+    if ctx.attr.use_wheels:
+        args = ["--use-wheels"] + args
+
     # Run compiler
     ctx.actions.run(
         inputs = [
@@ -150,10 +158,7 @@ def _zapp_impl(ctx):
         outputs = [ctx.outputs.executable],
         progress_message = "Building zapp file %s" % ctx.label,
         executable = ctx.executable.compiler,
-        arguments = [
+        arguments = args,
-            "-o", ctx.outputs.executable.path,
-            manifest_file.path
-        ],
         mnemonic = "PythonCompile",
         use_default_shell_env = True,
         execution_requirements = {
@@ -177,13 +182,14 @@ _zapp_attrs = {
     "entry_point": attr.string(),
     "prelude_points": attr.string_list(),
     "zip_safe": attr.bool(default = True),
+    "use_wheels": attr.bool(default = False),
     # FIXME: These are really toolchain parameters, probably.
     "compiler": attr.label(
         default = Label(DEFAULT_COMPILER),
         executable = True,
         cfg = "host",
     ),
-    "shebang": attr.string(default = "/usr/bin/env %py3%"),
+    "shebang": attr.string(default = "#!/usr/bin/env %py3%"),
 }
 
 _zapp = rule(