Update min python version from 3.6 to 3.8

The reason for picking 3.8 here is that it provides all the features
we currently have need of, and it available in the places we care about:

- debian/stable (bookworm): 3.11
- ubuntu/LTS (focal): 3.8
- emsdk: 3.9.2

Replaces: #23378
Fixes: #23387

ChangeLog.md

@@ -23,6 +23,8 @@ See docs/process.md for more on how version tagging works.
 - The minimum version of node required to run emscripten was bumped from v16.20
   to v18.  Version 4.0 was mistakenly shipped with a change that required v20,
   but that was reverted. (#23410)
+- The version of python required to run emscripten was bumped from 3.6 to 3.8.
+  (#23417)
 
 4.0.0 - 01/14/25
 ----------------

pyproject.toml

@@ -33,6 +33,7 @@ lint.ignore = [
   "B006",
   "B011",
   "B018",
+  "B019",
   "B023",
   "B026",
   "B904",

tools/building.py

@@ -476,17 +476,6 @@ def check_closure_compiler(cmd, args, env, allowed_to_fail):
   return True
 
 
-# Remove this once we require python3.7 and can use std.isascii.
-# See: https://docs.python.org/3/library/stdtypes.html#str.isascii
-def isascii(s):
-  try:
-    s.encode('ascii')
-  except UnicodeEncodeError:
-    return False
-  else:
-    return True
-
-
 def get_closure_compiler_and_env(user_args):
   env = shared.env_with_node_in_path()
   closure_cmd = get_closure_compiler()
@@ -623,7 +612,7 @@ def run_closure_cmd(cmd, filename, env):
   tempfiles = shared.get_temp_files()
 
   def move_to_safe_7bit_ascii_filename(filename):
-    if isascii(filename):
+    if filename.isascii():
       return os.path.abspath(filename)
     safe_filename = tempfiles.get('.js').name  # Safe 7-bit filename
     shutil.copyfile(filename, safe_filename)

tools/extract_metadata.py

@@ -4,6 +4,7 @@
 # found in the LICENSE file.
 
 import logging
+from functools import cache
 from typing import List, Dict
 
 from . import webassembly, utils
@@ -140,7 +141,7 @@ def parse_function_for_memory_inits(module, func_index, offset_map):
     parse_function_for_memory_inits(module, t, offset_map)
 
 
-@webassembly.memoize
+@cache
 def get_passive_segment_offsets(module):
   start_func_index = module.get_start()
   assert start_func_index is not None

tools/shared.py

@@ -6,8 +6,8 @@
 from .toolchain_profiler import ToolchainProfiler
 
 from enum import Enum, unique, auto
-from functools import wraps
 from subprocess import PIPE
+import functools
 import atexit
 import json
 import logging
@@ -20,9 +20,9 @@
 import sys
 import tempfile
 
-# We depend on python 3.6 for fstring support
-if sys.version_info < (3, 6):
-  print('error: emscripten requires python 3.6 or above', file=sys.stderr)
+# We depend on python 3.8 features
+if sys.version_info < (3, 8):
+  print(f'error: emscripten requires python 3.8 or above ({sys.executable} {sys.version})', file=sys.stderr)
   sys.exit(1)
 
 from . import colored_logger
@@ -67,6 +67,7 @@
 EMSCRIPTEN_TEMP_DIR = None
 
 logger = logging.getLogger('shared')
+memoize = functools.cache
 
 # warning about absolute-paths is disabled by default, and not enabled by -Wall
 diagnostics.add_warning('absolute-paths', enabled=False, part_of_all=False)
@@ -273,22 +274,6 @@ def get_npm_cmd(name):
   return cmd
 
 
-# TODO(sbc): Replace with functools.cache, once we update to python 3.7
-def memoize(func):
-  called = False
-  result = None
-
-  @wraps(func)
-  def helper():
-    nonlocal called, result
-    if not called:
-      result = func()
-      called = True
-    return result
-
-  return helper
-
-
 @memoize
 def get_clang_version():
   if not os.path.exists(CLANG_CC):

tools/system_libs.py

@@ -2442,17 +2442,8 @@ def calculate(args):
   return ret
 
 
-# Once we require python 3.8 we can use shutil.copytree with
-# dirs_exist_ok=True and remove this function.
 def copytree_exist_ok(src, dst):
-  os.makedirs(dst, exist_ok=True)
-  for entry in os.scandir(src):
-    srcname = os.path.join(src, entry.name)
-    dstname = os.path.join(dst, entry.name)
-    if entry.is_dir():
-      copytree_exist_ok(srcname, dstname)
-    else:
-      shared.safe_copy(srcname, dstname)
+  shutil.copytree(src, dst, dirs_exist_ok=True)
 
 
 def install_system_headers(stamp):

tools/webassembly.py

@@ -6,9 +6,9 @@
 """Utilities for manipulating WebAssembly binaries from python.
 """
 
+from functools import cache
 from collections import namedtuple
 from enum import IntEnum
-from functools import wraps
 import logging
 import os
 import sys
@@ -55,30 +55,6 @@ def read_sleb(iobuf):
   return leb128.i.decode_reader(iobuf)[0]
 
 
-def memoize(method):
-
-  @wraps(method)
-  def wrapper(self, *args, **kwargs):
-    assert not kwargs
-    key = (method.__name__, args)
-    if key not in self._cache:
-      self._cache[key] = method(self, *args, **kwargs)
-    return self._cache[key]
-
-  return wrapper
-
-
-def once(method):
-
-  @wraps(method)
-  def helper(self, *args, **kwargs):
-    key = method
-    if key not in self._cache:
-      self._cache[key] = method(self, *args, **kwargs)
-
-  return helper
-
-
 class Type(IntEnum):
   I32 = 0x7f # -0x1
   I64 = 0x7e # -0x2
@@ -280,7 +256,7 @@ def sections(self):
       yield Section(section_type, section_size, section_offset, name)
       offset = section_offset + section_size
 
-  @memoize
+  @cache
   def get_types(self):
     type_section = self.get_section(SecType.TYPE)
     if not type_section:
@@ -315,7 +291,7 @@ def parse_features_section(self):
         feature_count -= 1
     return features
 
-  @memoize
+  @cache
   def parse_dylink_section(self):
     dylink_section = next(self.sections())
     assert dylink_section.type == SecType.CUSTOM
@@ -380,7 +356,7 @@ def parse_dylink_section(self):
 
     return Dylink(mem_size, mem_align, table_size, table_align, needed, export_info, import_info)
 
-  @memoize
+  @cache
   def get_exports(self):
     export_section = self.get_section(SecType.EXPORT)
     if not export_section:
@@ -397,7 +373,7 @@ def get_exports(self):
 
     return exports
 
-  @memoize
+  @cache
   def get_imports(self):
     import_section = self.get_section(SecType.IMPORT)
     if not import_section:
@@ -430,7 +406,7 @@ def get_imports(self):
 
     return imports
 
-  @memoize
+  @cache
   def get_globals(self):
     global_section = self.get_section(SecType.GLOBAL)
     if not global_section:
@@ -445,15 +421,15 @@ def get_globals(self):
       globls.append(Global(global_type, mutable, init))
     return globls
 
-  @memoize
+  @cache
   def get_start(self):
     start_section = self.get_section(SecType.START)
     if not start_section:
       return None
     self.seek(start_section.offset)
     return self.read_uleb()
 
-  @memoize
+  @cache
   def get_functions(self):
     code_section = self.get_section(SecType.CODE)
     if not code_section:
@@ -471,14 +447,14 @@ def get_functions(self):
   def get_section(self, section_code):
     return next((s for s in self.sections() if s.type == section_code), None)
 
-  @memoize
+  @cache
   def get_custom_section(self, name):
     for section in self.sections():
       if section.type == SecType.CUSTOM and section.name == name:
         return section
     return None
 
-  @memoize
+  @cache
   def get_segments(self):
     segments = []
     data_section = self.get_section(SecType.DATA)
@@ -496,7 +472,7 @@ def get_segments(self):
       self.seek(offset + size)
     return segments
 
-  @memoize
+  @cache
   def get_tables(self):
     table_section = self.get_section(SecType.TABLE)
     if not table_section:
@@ -512,7 +488,7 @@ def get_tables(self):
 
     return tables
 
-  @memoize
+  @cache
   def get_function_types(self):
     function_section = self.get_section(SecType.FUNCTION)
     if not function_section:
@@ -525,7 +501,7 @@ def get_function_types(self):
   def has_name_section(self):
     return self.get_custom_section('name') is not None
 
-  @once
+  @cache
   def _calc_indexes(self):
     self.imports_by_kind = {}
     for i in self.get_imports():