bpo-45949: Pure Python freeze module for cross builds (GH-29899)

2 files changed, 70 insertions, 0 deletions

diff --git a/Programs/_freeze_module.c b/Programs/_freeze_module.c
index e3f6c11c8b0..d50787666f8 100644
--- a/Programs/_freeze_module.c
+++ b/Programs/_freeze_module.c
@@ -5,6 +5,8 @@
    This is used directly by Tools/scripts/freeze_modules.py, and indirectly by "make regen-frozen".
 
    See Python/frozen.c for more info.
+
+   Keep this file in sync with Programs/_freeze_module.py.
 */
 
 #include <Python.h>
diff --git a/Programs/_freeze_module.py b/Programs/_freeze_module.py
new file mode 100644
index 00000000000..ba638eef6c4
--- /dev/null
+++ b/Programs/_freeze_module.py
@@ -0,0 +1,68 @@
+"""Python implementation of Programs/_freeze_module.c
+
+The pure Python implementation uses same functions and arguments as the C
+implementation.
+
+The generated byte code is slightly different because
+compile() sets the PyCF_SOURCE_IS_UTF8 flag and objects have a
+reference count > 1. Marshal adds the `FLAG_REF` flag and creates a
+reference `hashtable`.
+"""
+
+import marshal
+import sys
+
+header = "/* Auto-generated by Programs/_freeze_module.py */"
+
+
+def read_text(inpath: str) -> bytes:
+    with open(inpath, "rb") as f:
+        return f.read()
+
+
+def compile_and_marshal(name: str, text: bytes) -> bytes:
+    filename = f"<frozen {name}>"
+    # exec == Py_file_input
+    code = compile(text, filename, "exec", optimize=0, dont_inherit=True)
+    return marshal.dumps(code)
+
+
+def get_varname(name: str, prefix: str) -> str:
+    return f"{prefix}{name.replace('.', '_')}"
+
+
+def write_code(outfile, marshalled: bytes, varname: str) -> None:
+    data_size = len(marshalled)
+
+    outfile.write(f"const unsigned char {varname}[] = {{\n")
+
+    for n in range(0, data_size, 16):
+        outfile.write("    ")
+        outfile.write(",".join(str(i) for i in marshalled[n : n + 16]))
+        outfile.write(",\n")
+    outfile.write("};\n")
+
+
+def write_frozen(outpath: str, inpath: str, name: str, marshalled: bytes) -> None:
+    with open(outpath, "w") as outfile:
+        outfile.write(header)
+        outfile.write("\n")
+        arrayname = get_varname(name, "_Py_M__")
+        write_code(outfile, marshalled, arrayname)
+
+
+def main():
+    if len(sys.argv) != 4:
+        sys.exit("need to specify the name, input and output paths\n")
+
+    name = sys.argv[1]
+    inpath = sys.argv[2]
+    outpath = sys.argv[3]
+
+    text = read_text(inpath)
+    marshalled = compile_and_marshal(name, text)
+    write_frozen(outpath, inpath, name, marshalled)
+
+
+if __name__ == "__main__":
+    main()