gh-95853: WASM: better version and asset handling in scripts (GH-96045)

- support EMSDK tot-upstream and git releases
- allow WASM assents for wasm64-emscripten and WASI. This makes single file distributions on WASI easier.
- decouple WASM assets from browser builds

Author: tiran

Makefile.pre.in

@@ -803,10 +803,11 @@ $(DLLLIBRARY) libpython$(LDVERSION).dll.a: $(LIBRARY_OBJS)
 # wasm assets directory is relative to current build dir, e.g. "./usr/local".
 # --preload-file turns a relative asset path into an absolute path.
 
+.PHONY: wasm_stdlib
+wasm_stdlib: $(WASM_STDLIB)
 $(WASM_STDLIB): $(srcdir)/Lib/*.py $(srcdir)/Lib/*/*.py \
 	    $(srcdir)/Tools/wasm/wasm_assets.py \
-	    Makefile pybuilddir.txt Modules/Setup.local \
-	    python.html python.worker.js
+	    Makefile pybuilddir.txt Modules/Setup.local
 	$(PYTHON_FOR_BUILD) $(srcdir)/Tools/wasm/wasm_assets.py \
 	    --buildroot . --prefix $(prefix)
 

Tools/wasm/wasm_assets.py

@@ -108,6 +108,14 @@
     "_zoneinfo": ["zoneinfo/"],
 }
 
+SYSCONFIG_NAMES = (
+    "_sysconfigdata__emscripten_wasm32-emscripten",
+    "_sysconfigdata__emscripten_wasm32-emscripten",
+    "_sysconfigdata__wasi_wasm32-wasi",
+    "_sysconfigdata__wasi_wasm64-wasi",
+)
+
+
 def get_builddir(args: argparse.Namespace) -> pathlib.Path:
     """Get builddir path from pybuilddir.txt
     """
@@ -120,7 +128,11 @@ def get_sysconfigdata(args: argparse.Namespace) -> pathlib.Path:
     """Get path to sysconfigdata relative to build root
     """
     data_name = sysconfig._get_sysconfigdata_name()
-    assert "emscripten_wasm32" in data_name
+    if not data_name.startswith(SYSCONFIG_NAMES):
+        raise ValueError(
+            f"Invalid sysconfig data name '{data_name}'.",
+            SYSCONFIG_NAMES
+        )
     filename = data_name + ".py"
     return args.builddir / filename
 

Tools/wasm/wasm_build.py

@@ -20,6 +20,7 @@
 import dataclasses
 import os
 import pathlib
+import re
 import shlex
 import shutil
 import subprocess
@@ -99,6 +100,24 @@ def get_emscripten_root(emconfig: pathlib.Path = EM_CONFIG) -> pathlib.PurePath:
 EMSCRIPTEN_ROOT = get_emscripten_root()
 
 
+def read_python_version(configure: pathlib.Path = CONFIGURE) -> str:
+    """Read PACKAGE_VERSION from configure script
+
+    configure and configure.ac are the canonical source for major and
+    minor version number.
+    """
+    version_re = re.compile("^PACKAGE_VERSION='(\d\.\d+)'")
+    with configure.open(encoding="utf-8") as f:
+        for line in f:
+            mo = version_re.match(line)
+            if mo:
+                return mo.group(1)
+    raise ValueError(f"PACKAGE_VERSION not found in {configure}")
+
+
+PYTHON_VERSION = read_python_version()
+
+
 class ConditionError(ValueError):
     def __init__(self, info: str, text: str):
         self.info = info
@@ -174,6 +193,9 @@ def _check_emscripten():
         raise MissingDependency(os.fspath(version_txt), INSTALL_EMSDK)
     with open(version_txt) as f:
         version = f.read().strip().strip('"')
+    if version.endswith("-git"):
+        # git / upstream / tot-upstream installation
+        version = version[:-4]
     version_tuple = tuple(int(v) for v in version.split("."))
     if version_tuple < EMSDK_MIN_VERSION:
         raise MissingDependency(
@@ -221,7 +243,7 @@ def _check_wasi():
         # workaround for https://github.com/python/cpython/issues/95952
         "HOSTRUNNER": (
             "wasmtime run "
-            "--env PYTHONPATH=/{relbuilddir}/build/lib.wasi-wasm32-$(VERSION):/Lib "
+            "--env PYTHONPATH=/{relbuilddir}/build/lib.wasi-wasm32-{version}:/Lib "
             "--mapdir /::{srcdir} --"
         ),
     },
@@ -362,6 +384,7 @@ def getenv(self) -> dict:
                 env[key] = value.format(
                     relbuilddir=self.builddir.relative_to(SRCDIR),
                     srcdir=SRCDIR,
+                    version=PYTHON_VERSION,
                 )
             else:
                 env[key] = value

configure

@@ -1581,7 +1581,7 @@ AC_MSG_RESULT($LDLIBRARY)
 
 # LIBRARY_DEPS, LINK_PYTHON_OBJS and LINK_PYTHON_DEPS variable
 AS_CASE([$ac_sys_system/$ac_sys_emscripten_target],
-  [Emscripten/browser*], [LIBRARY_DEPS='$(PY3LIBRARY) $(WASM_STDLIB)'],
+  [Emscripten/browser*], [LIBRARY_DEPS='$(PY3LIBRARY) $(WASM_STDLIB) python.html python.worker.js'],
   [LIBRARY_DEPS='$(PY3LIBRARY) $(EXPORTSYMS)']
 )
 LINK_PYTHON_DEPS='$(LIBRARY_DEPS)'