From 54cc4b3a742ce2e61a12005236ed9a3a8b8dcd14 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Wed, 3 Sep 2025 08:46:23 +0000
Subject: [PATCH 01/17] Add script to split module based on source paths

This adds a script, `tools/empath-split.py`, which is a wrapper for
Binaryen's `wasm-split`. `wasm-split` has `--multi-split` mode, which
takes a manifest file that lists the name of functions per module.
(Example:
 https://github.com/WebAssembly/binaryen/blob/main/test/lit/wasm-split/multi-split.wast.manifest)

But listing all functions belonging to each module is a tedious process.
`empath-split` takes a wasm file and a text file that has a list of
paths, which can be either directories or functions, and using the
source map information, generates a manifest file, and runs
`wasm-split`.

This makes a small drive-by fix for `emsymbolizer`. Currently when it
takes a 0 address, it returns the location info associated with
offsets[-1], which is the largest offset. This fixes it, and adds an
optional `lower_bound` argument to `find_offset` so that when we want to
get a source info entry, we don't go below the current function start
offset.
---
 test/test_other.py                 |  48 ++++++
 tools/empath-split.py              | 229 +++++++++++++++++++++++++++++
 tools/emsymbolizer.py              |  27 ++--
 tools/link.py                      |   4 +-
 tools/maint/create_entry_points.py |   2 +
 tools/webassembly.py               |  37 +++++
 6 files changed, 335 insertions(+), 12 deletions(-)
 create mode 100755 tools/empath-split.py

diff --git a/test/test_other.py b/test/test_other.py
index 82ff8310fcb9a..f510c64ec1e87 100644
--- a/test/test_other.py
+++ b/test/test_other.py
@@ -57,6 +57,7 @@
 emmake = shared.bat_suffix(path_from_root('emmake'))
 emconfig = shared.bat_suffix(path_from_root('em-config'))
 emsize = shared.bat_suffix(path_from_root('emsize'))
+empath_split = shared.bat_suffix(path_from_root('empath-split'))
 emprofile = shared.bat_suffix(path_from_root('emprofile'))
 emstrip = shared.bat_suffix(path_from_root('emstrip'))
 emsymbolizer = shared.bat_suffix(path_from_root('emsymbolizer'))
@@ -16589,3 +16590,50 @@ def test_create_preloaded_file(self):
         return 0;
       }''')
     self.do_runf('main.c', 'done\n', cflags=['-sFORCE_FILESYSTEM', '--post-js=post.js'])
+
+  def test_empath_split(self):
+    create_file('main.cpp', r'''
+      #include <iostream>
+      void foo();
+      int main() {
+        std::cout << "main" << std::endl;
+        foo();
+        return 0;
+      }
+    ''')
+    create_file('foo.cpp', r'''
+      #include <iostream>
+      void foo() { std::cout << "foo" << std::endl; }
+    ''')
+    create_file('path_list', r'''
+      main.cpp
+      foo.cpp
+      /emsdk/emscripten/system
+      /emsdk/emscripten/system/lib/libc/musl
+      /emsdk/emscripten/system/lib/libcxx
+    ''')
+
+    self.run_process([EMCC, 'main.cpp', 'foo.cpp', '-gsource-map', '-g2', '-gsource-map', '-o', 'test.js'])
+    self.run_process([empath_split, 'test.wasm', 'path_list', '-g', '-o', 'test_primary.wasm', '--out-prefix=test_'])
+
+    # Check if functions are correctly assigned and split with the specified
+    # paths. When one path contains another, the inner path should take its
+    # functions first, and the rest is split with the outer path.
+    def has_defined_function(file, func):
+      self.run_process([common.WASM_DIS, file, '-o', 'test.wast'])
+      pattern = re.compile(r'\(\s*func\s+\$' + func + r'[\s\(\)]')
+      with open('test.wast', 'r') as f:
+        return pattern.search(f.read()) is not None
+
+    # main.cpp
+    self.assertTrue(has_defined_function('test_0.wasm', '__original_main'))
+    # foo.cpp
+    self.assertTrue(has_defined_function('test_1.wasm', r'foo\\28\\29'))
+    # /emsdk/emscripten/system
+    self.assertTrue(has_defined_function('test_2.wasm', '__abort_message'))
+    self.assertTrue(has_defined_function('test_2.wasm', 'pthread_cond_wait'))
+    # /emsdk/emscripten/system/lib/libc/musl
+    self.assertTrue(has_defined_function('test_3.wasm', 'strcmp'))
+    # /emsdk/emscripten/system/lib/libcxx
+    self.assertTrue(has_defined_function('test_4.wasm', r'std::__2::ios_base::getloc\\28\\29\\20const'))
+    self.assertTrue(has_defined_function('test_4.wasm', r'std::uncaught_exceptions\\28\\29'))
diff --git a/tools/empath-split.py b/tools/empath-split.py
new file mode 100755
index 0000000000000..90b27eed71a34
--- /dev/null
+++ b/tools/empath-split.py
@@ -0,0 +1,229 @@
+#!/usr/bin/env python3
+# Copyright 2025 The Emscripten Authors.  All rights reserved.
+# Emscripten is available under two separate licenses, the MIT license and the
+# University of Illinois/NCSA Open Source License.  Both these licenses can be
+# found in the LICENSE file.
+
+"""
+Wrapper for 'wasm-split --multi-split' functionality. This script generates a
+.manifest file based on the list of user source paths, using source map
+information.
+
+This assumes the name section exists in the input wasm file, and also assumes
+the sourceMappingURL section exists in the input or a source map file is
+separately supplied with --sourcemap. If we have two files a.c and b.c, to
+generate a source map and the name section, if you compile and link within a
+single command, you can do something like
+$ emcc -g2 -gsrouce-map a.c b.c -o result.js
+If you want to compile and link in separate commands, you can do
+$ emcc -gsource-map a.c -o a.o
+$ emcc -gsource-map b.c -o b.o
+$ emcc -g2 -gsource-map a.o b.o -o result.js
+See https://emscripten.org/docs/porting/Debugging.html for more details.
+
+This takes a wasm file and a paths file, which is a text file containing a list
+of paths as inputs. The paths file should contain a single path per line. A
+single split module will be generated per specified path. If a specified path
+contains another specified path, functions contained in the inner path will be
+split as the inner path's module, and the rest of the functions will be split as
+the outer path's module. Functions that do not belong to any of the specified
+paths will remain in the primary module.
+"""
+
+import argparse
+import os
+import sys
+import tempfile
+from pathlib import PurePath
+
+__scriptdir__ = os.path.dirname(os.path.abspath(__file__))
+__rootdir__ = os.path.dirname(__scriptdir__)
+sys.path.insert(0, __rootdir__)
+
+from tools import config
+from tools import diagnostics
+from tools import emsymbolizer
+from tools import shared
+from tools import webassembly
+from tools.utils import exit_with_error, normalize_path
+
+def parse_args():
+  parser = argparse.ArgumentParser(
+      description='Split a wasm file based on user paths',
+      epilog="""
+This is a wrapper for 'wasm-split --multi-split' functionality, so you should
+add wasm-split's command line options as well. You should or may want to add
+wasm-split options like -o (--output), --out-prefix, -g, and feature
+enabling/disabling options. Run 'wasm-split -h' for the list of options. But you
+should NOT add --manifest, because this will be generated from this script.
+""")
+  parser.add_argument('wasm', help='Path to the input wasm file')
+  parser.add_argument('paths_file', help='Path to the input file containing paths')
+  parser.add_argument('-s', '--sourcemap', help='Force source map file')
+  parser.add_argument('-v', '--verbose', action='store_true',
+                      help='Print verbose info for debugging this script')
+  parser.add_argument('--wasm-split', help='Path to wasm-split executable')
+  parser.add_argument('--preserve-manifest', action='store_true',
+                      help='Preserve generated manifest file. This sets --verbose too.')
+  args, forwarded_args = parser.parse_known_args()
+  if '--manifest' in forwarded_args:
+    exit_with_error('manifest file will be generated by this script and should not be given')
+  if args.preserve_manifest:
+    args.verbose = True
+  return args, forwarded_args
+
+
+def get_path_to_functions_map(wasm, sourcemap, paths, verbose):
+  def is_synthesized_func(func):
+    # TODO There can be more
+    synthesized_names = [
+      'main',
+      '__wasm_call_ctors',
+      '__clang_call_terminate',
+    ]
+    synthesized_prefixes = [
+      'legalstub$',
+      'legalfunc$',
+      '__cxx_global_',
+      '_GLOBAL__',
+      'virtual thunk to ',
+    ]
+    if func in synthesized_names:
+      return True
+    return func.startswith(tuple(synthesized_prefixes))
+
+  with webassembly.Module(wasm) as module:
+    if not module.has_name_section():
+      exit_with_error('Name section does not eixst')
+    if not sourcemap:
+      if not emsymbolizer.get_sourceMappingURL_section(module):
+        exit_with_error('sourceMappingURL section does not exist')
+
+    code_section = module.get_section(webassembly.SecType.CODE)
+
+    funcs = module.get_functions()
+    func_names = module.get_function_names()[module.num_imported_funcs():]
+    assert len(funcs) == len(func_names)
+
+    func_to_src = {}
+    src_to_funcs = {}
+
+    if not sourcemap:
+      sourcemap = module.get_sourceMappingURL()
+    sm = emsymbolizer.WasmSourceMap()
+    sm.parse(sourcemap)
+
+    for func_name, func in zip(func_names, funcs):
+      # From the last address, decrement the address by 1 until we find location
+      # info with source file information. The reason we do this is to reduce
+      # the probability of picking an address where another function is inlined
+      # into, picking the inlined function's source.
+      # We start from the end because it is simpler; it is harder to compute the
+      # first instruction's address, because there is a gap for local types
+      # between function offset and the first instruction.
+      addr = func.offset + func.size - 1
+      while addr > func.offset:
+        loc = sm.lookup(addr, func.offset)
+        # This means there is no source map mappings for the entire function
+        # (because we give func.offset as a lower bound). Exit the loop.
+        if not loc:
+          break
+        # Exit the loop only if a location info with source file information is
+        # found. If not, continue the search.
+        if loc.source:
+          break
+        addr -= 1
+
+      if loc and loc.source:
+        func_to_src[func_name] = normalize_path(loc.source)
+      else:
+        if not is_synthesized_func(func_name):
+          diagnostics.warn(f"No source file information found in the source map for function '{func_name}'")
+
+    for func_name, src in func_to_src.items():
+      if src not in src_to_funcs:
+        src_to_funcs[src] = []
+      src_to_funcs[src].append(func_name)
+
+  # Visit paths in the reverse sorting order, so that we can proces inner paths
+  # first.
+  # e.g. If we have /a/b and /a/b/c, /a/b/c will come first, so we can assign
+  # functions contained in /a/b/c to it first and assign the maining functions
+  # to /a/b.
+  visited_funcs = set()
+  path_to_funcs = {}
+  for path in sorted(paths, reverse=True):
+    ppath = PurePath(path)
+    path_to_funcs[path] = []
+    for src, funcs in src_to_funcs.items():
+      psrc = PurePath(src)
+      if ppath == psrc or ppath in psrc.parents:
+        for func in funcs:
+          if func not in visited_funcs:
+            visited_funcs.add(func)
+            path_to_funcs[path].append(func)
+  return path_to_funcs
+
+
+def main():
+  args, forwarded_args = parse_args()
+  wasm = os.path.expanduser(args.wasm)
+  paths_file = os.path.expanduser(args.paths_file)
+  sourcemap = os.path.expanduser(args.sourcemap) if args.sourcemap else None
+  if args.wasm_split:
+    wasm_split = os.path.expanduser(args.wasm_split)
+  else:
+    wasm_split = os.path.join(config.BINARYEN_ROOT, 'bin', 'wasm-split')
+
+  if not os.path.isfile(wasm):
+    exit_with_error(f"'{wasm}' was not found or not a file")
+  if not os.path.isfile(paths_file):
+    exit_with_error(f"'{paths_file}' was not found or not a file")
+  if sourcemap:
+    if not os.path.isfile(sourcemap):
+      exit_with_error(f"'{sourcemap}' was not found or not a file")
+  if not os.path.isfile(wasm_split):
+    exit_with_error(f"'{wasm_split}' was not found or not a file")
+
+  with open(paths_file, 'r', encoding='utf-8') as f:
+    paths = [normalize_path(path.strip()) for path in f if path.strip()]
+    # To make /a/b/c and /a/b/c/ equivalent
+    paths = [path.rstrip(os.sep) for path in paths]
+    # Remove duplicates
+    paths = list(dict.fromkeys(paths))
+
+  path_to_funcs = get_path_to_functions_map(wasm, sourcemap, paths, args.verbose)
+
+  f = tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=False)
+  try:
+    manifest = f.name
+    for i, path in enumerate(paths):
+      f.write(f'{i}\n')
+      if not path_to_funcs[path]:
+        diagnostics.warn(f'{path} does not match any functions')
+      if args.verbose:
+        print(path)
+        for func in path_to_funcs[path]:
+          print('  ' + func)
+        print()
+      for func in path_to_funcs[path]:
+        f.write(func + '\n')
+      if i < len(paths) - 1:
+        f.write('\n')
+    f.close()
+
+    cmd = [wasm_split, '--multi-split', wasm, '--manifest', manifest]
+    if args.verbose:
+      # This option is used both in this script and wasm-split
+      cmd.append('-v')
+    cmd += forwarded_args
+    if args.verbose:
+      print('\n' + ' '.join(cmd))
+    shared.run_process(cmd)
+  finally:
+    if not args.preserve_manifest:
+      os.remove(manifest)
+
+
+if __name__ == '__main__':
+  sys.exit(main())
diff --git a/tools/emsymbolizer.py b/tools/emsymbolizer.py
index b411f49da3d15..6affa7d4c9b1a 100755
--- a/tools/emsymbolizer.py
+++ b/tools/emsymbolizer.py
@@ -178,7 +178,7 @@ def decodeVLQ(string):
       self.offsets.append(offset)
     self.offsets.sort()
 
-  def find_offset(self, offset):
+  def find_offset(self, offset, lower_bound=None):
     # Find the largest mapped offset <= the search offset
     lo = 0
     hi = len(self.offsets)
@@ -189,11 +189,22 @@ def find_offset(self, offset):
         hi = mid
       else:
         lo = mid + 1
-    return self.offsets[lo - 1]
+    if lo == 0:
+      return None
+    # If lower bound is given, return the offset only if the offset is equal to
+    # or greather than the lower bound
+    if lower_bound:
+      if self.offsets[lo - 1] >= lower_bound:
+        return self.offsets[lo - 1]
+      else:
+        return None
+    else:
+      return self.offsets[lo - 1]
 
-  def lookup(self, offset):
-    nearest = self.find_offset(offset)
-    assert nearest in self.mappings, 'Sourcemap has an offset with no mapping'
+  def lookup(self, offset, lower_bound=None):
+    nearest = self.find_offset(offset, lower_bound)
+    if not nearest:
+      return None
     info = self.mappings[nearest]
     return LocationInfo(
         self.sources[info.source] if info.source is not None else None,
@@ -206,12 +217,8 @@ def symbolize_address_sourcemap(module, address, force_file):
   URL = force_file
   if not URL:
     # If a sourcemap file is not forced, read it from the wasm module
-    section = get_sourceMappingURL_section(module)
-    assert section
-    module.seek(section.offset)
-    assert module.read_string() == 'sourceMappingURL'
     # TODO: support stripping/replacing a prefix from the URL
-    URL = module.read_string()
+    URL = module.get_sourceMappingURL()
 
   if shared.DEBUG:
     print(f'Source Mapping URL: {URL}')
diff --git a/tools/link.py b/tools/link.py
index 65ac42b47bd47..f8adf61cc360c 100644
--- a/tools/link.py
+++ b/tools/link.py
@@ -2301,8 +2301,8 @@ def phase_binaryen(target, options, wasm_target):
       intermediate_debug_info -= 1
     # currently binaryen's DWARF support will limit some optimizations; warn on
     # that. see https://github.com/emscripten-core/emscripten/issues/15269
-    if settings.GENERATE_DWARF:
-      diagnostics.warning('limited-postlink-optimizations', 'running limited binaryen optimizations because DWARF info requested (or indirectly required)')
+    #if settings.GENERATE_DWARF:
+    #  diagnostics.warning('limited-postlink-optimizations', 'running limited binaryen optimizations because DWARF info requested (or indirectly required)')
     with ToolchainProfiler.profile_block('wasm_opt'):
       building.run_wasm_opt(wasm_target,
                             wasm_target,
diff --git a/tools/maint/create_entry_points.py b/tools/maint/create_entry_points.py
index d610960c03349..93693b5496978 100755
--- a/tools/maint/create_entry_points.py
+++ b/tools/maint/create_entry_points.py
@@ -43,6 +43,7 @@
 emstrip
 emsymbolizer
 emscan-deps
+empath-split
 tools/file_packager
 tools/webidl_binder
 test/runner
@@ -56,6 +57,7 @@
   'emdwp': 'tools/emdwp',
   'emnm': 'tools/emnm',
   'emsymbolizer': 'tools/emsymbolizer',
+  'empath-split': 'tools/empath-split',
 }
 
 
diff --git a/tools/webassembly.py b/tools/webassembly.py
index d4c26c0cab15d..ce574f713afbd 100644
--- a/tools/webassembly.py
+++ b/tools/webassembly.py
@@ -521,6 +521,34 @@ def get_function_types(self):
     num_types = self.read_uleb()
     return [self.read_uleb() for _ in range(num_types)]
 
+  @memoize
+  def get_function_names(self):
+    num_funcs = self.num_imported_funcs() + len(self.get_functions())
+    names = [None] * num_funcs
+
+    name_section = self.get_custom_section('name')
+    if not name_section:
+      return names
+
+    self.seek(name_section.offset)
+    self.read_string()  # section name
+    section_end = name_section.offset + name_section.size
+
+    while self.tell() < section_end:
+      subsection_id = self.read_byte()
+      subsection_size = self.read_uleb()
+      if subsection_id == 1:  # function names
+        count = self.read_uleb()
+        for _ in range(count):
+          func_idx = self.read_uleb()
+          func_name = self.read_string()
+          assert func_idx < len(names)
+          names[func_idx] = func_name
+      else:
+        self.skip(subsection_size)
+
+    return names
+
   def has_name_section(self):
     return self.get_custom_section('name') is not None
 
@@ -579,6 +607,15 @@ def get_target_features(self):
       features[feature] = prefix
     return features
 
+  @memoize
+  def get_sourceMappingURL(self):
+    section = self.get_custom_section('sourceMappingURL')
+    if not section:
+      return ''
+    self.seek(section.offset)
+    self.read_string() # 'sourceMappingURL'
+    return self.read_string()
+
 
 def parse_dylink_section(wasm_file):
   with Module(wasm_file) as module:

From ea300b2e8b6c200e80b83714b4ac40e17faa2db5 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 17:09:26 +0000
Subject: [PATCH 02/17] Revert accidental change

---
 tools/link.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/link.py b/tools/link.py
index f8adf61cc360c..65ac42b47bd47 100644
--- a/tools/link.py
+++ b/tools/link.py
@@ -2301,8 +2301,8 @@ def phase_binaryen(target, options, wasm_target):
       intermediate_debug_info -= 1
     # currently binaryen's DWARF support will limit some optimizations; warn on
     # that. see https://github.com/emscripten-core/emscripten/issues/15269
-    #if settings.GENERATE_DWARF:
-    #  diagnostics.warning('limited-postlink-optimizations', 'running limited binaryen optimizations because DWARF info requested (or indirectly required)')
+    if settings.GENERATE_DWARF:
+      diagnostics.warning('limited-postlink-optimizations', 'running limited binaryen optimizations because DWARF info requested (or indirectly required)')
     with ToolchainProfiler.profile_block('wasm_opt'):
       building.run_wasm_opt(wasm_target,
                             wasm_target,

From 355fc6006b06ac259406a36cfc77203cb9310a0e Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 17:36:53 +0000
Subject: [PATCH 03/17] comments

---
 tools/empath-split.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 90b27eed71a34..0aea45530f890 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -92,6 +92,9 @@ def is_synthesized_func(func):
       return True
     return func.startswith(tuple(synthesized_prefixes))
 
+  # Compute {func_name: src file} map, and revert it to get
+  # {src file: list of functions} map, and construct {path: list of functions}
+  # map from it
   with webassembly.Module(wasm) as module:
     if not module.has_name_section():
       exit_with_error('Name section does not eixst')
@@ -192,8 +195,10 @@ def main():
     # Remove duplicates
     paths = list(dict.fromkeys(paths))
 
+  # Compute {path: list of functions} map
   path_to_funcs = get_path_to_functions_map(wasm, sourcemap, paths, args.verbose)
 
+  # Write .manifest file
   f = tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=False)
   try:
     manifest = f.name

From e6d39d6711335d344b857872b497615e2c91ae48 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 18:06:12 +0000
Subject: [PATCH 04/17] Fix regex so that (import .. (func ..)) is not included

---
 test/test_other.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/test_other.py b/test/test_other.py
index f510c64ec1e87..da912530d2623 100644
--- a/test/test_other.py
+++ b/test/test_other.py
@@ -16621,7 +16621,7 @@ def test_empath_split(self):
     # functions first, and the rest is split with the outer path.
     def has_defined_function(file, func):
       self.run_process([common.WASM_DIS, file, '-o', 'test.wast'])
-      pattern = re.compile(r'\(\s*func\s+\$' + func + r'[\s\(\)]')
+      pattern = re.compile(r'^\s*\(\s*func\s+\$' + func + r'[\s\(\)]', flags=re.MULTILINE)
       with open('test.wast', 'r') as f:
         return pattern.search(f.read()) is not None
 

From 1246d2e34cc6c0755b02029332881808eb88cf8b Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 18:11:55 +0000
Subject: [PATCH 05/17] comment typo

---
 tools/empath-split.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 0aea45530f890..0dbd4cd8bfdb4 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -148,7 +148,7 @@ def is_synthesized_func(func):
         src_to_funcs[src] = []
       src_to_funcs[src].append(func_name)
 
-  # Visit paths in the reverse sorting order, so that we can proces inner paths
+  # Visit paths in the reverse sorting order, so that we can process inner paths
   # first.
   # e.g. If we have /a/b and /a/b/c, /a/b/c will come first, so we can assign
   # functions contained in /a/b/c to it first and assign the maining functions

From 9cb04b770feae8f62b5aa2deb07a6ef54066601b Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 16:30:48 -0700
Subject: [PATCH 06/17] Update tools/empath-split.py

Co-authored-by: Derek Schuff <dschuff@chromium.org>
---
 tools/empath-split.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 0dbd4cd8bfdb4..5551007586d9f 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -92,7 +92,7 @@ def is_synthesized_func(func):
       return True
     return func.startswith(tuple(synthesized_prefixes))
 
-  # Compute {func_name: src file} map, and revert it to get
+  # Compute {func_name: src file} map, and invert it to get
   # {src file: list of functions} map, and construct {path: list of functions}
   # map from it
   with webassembly.Module(wasm) as module:

From 4440ebf7ca972aa0f047aa92eae28d1b0651d03f Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 23:58:44 +0000
Subject: [PATCH 07/17] Address comments

---
 test/test_other.py    | 2 +-
 tools/empath-split.py | 4 +---
 tools/webassembly.py  | 4 ++--
 3 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/test/test_other.py b/test/test_other.py
index da912530d2623..3f59c482245af 100644
--- a/test/test_other.py
+++ b/test/test_other.py
@@ -16613,7 +16613,7 @@ def test_empath_split(self):
       /emsdk/emscripten/system/lib/libcxx
     ''')
 
-    self.run_process([EMCC, 'main.cpp', 'foo.cpp', '-gsource-map', '-g2', '-gsource-map', '-o', 'test.js'])
+    self.run_process([EMCC, 'main.cpp', 'foo.cpp', '-gsource-map', '-g2', '-o', 'test.js'])
     self.run_process([empath_split, 'test.wasm', 'path_list', '-g', '-o', 'test_primary.wasm', '--out-prefix=test_'])
 
     # Check if functions are correctly assigned and split with the specified
diff --git a/tools/empath-split.py b/tools/empath-split.py
index 5551007586d9f..58a92894a656f 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -102,10 +102,8 @@ def is_synthesized_func(func):
       if not emsymbolizer.get_sourceMappingURL_section(module):
         exit_with_error('sourceMappingURL section does not exist')
 
-    code_section = module.get_section(webassembly.SecType.CODE)
-
     funcs = module.get_functions()
-    func_names = module.get_function_names()[module.num_imported_funcs():]
+    func_names = module.get_function_names()
     assert len(funcs) == len(func_names)
 
     func_to_src = {}
diff --git a/tools/webassembly.py b/tools/webassembly.py
index ce574f713afbd..192bfdf850432 100644
--- a/tools/webassembly.py
+++ b/tools/webassembly.py
@@ -522,7 +522,7 @@ def get_function_types(self):
     return [self.read_uleb() for _ in range(num_types)]
 
   @memoize
-  def get_function_names(self):
+  def get_function_names(self, remove_imports=True):
     num_funcs = self.num_imported_funcs() + len(self.get_functions())
     names = [None] * num_funcs
 
@@ -547,7 +547,7 @@ def get_function_names(self):
       else:
         self.skip(subsection_size)
 
-    return names
+    return names[self.num_imported_funcs():] if remove_imports else names
 
   def has_name_section(self):
     return self.get_custom_section('name') is not None

From 829e697771b782edc955cdd3a8684c5d5d957989 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Mon, 15 Sep 2025 23:58:57 +0000
Subject: [PATCH 08/17] ruff fixes

---
 test/test_other.py    | 2 +-
 tools/empath-split.py | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/test/test_other.py b/test/test_other.py
index 3f59c482245af..10b995a6e85be 100644
--- a/test/test_other.py
+++ b/test/test_other.py
@@ -16622,7 +16622,7 @@ def test_empath_split(self):
     def has_defined_function(file, func):
       self.run_process([common.WASM_DIS, file, '-o', 'test.wast'])
       pattern = re.compile(r'^\s*\(\s*func\s+\$' + func + r'[\s\(\)]', flags=re.MULTILINE)
-      with open('test.wast', 'r') as f:
+      with open('test.wast') as f:
         return pattern.search(f.read()) is not None
 
     # main.cpp
diff --git a/tools/empath-split.py b/tools/empath-split.py
index 58a92894a656f..0d1c10c104ce3 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -73,7 +73,7 @@ def parse_args():
   return args, forwarded_args
 
 
-def get_path_to_functions_map(wasm, sourcemap, paths, verbose):
+def get_path_to_functions_map(wasm, sourcemap, paths):
   def is_synthesized_func(func):
     # TODO There can be more
     synthesized_names = [
@@ -186,7 +186,7 @@ def main():
   if not os.path.isfile(wasm_split):
     exit_with_error(f"'{wasm_split}' was not found or not a file")
 
-  with open(paths_file, 'r', encoding='utf-8') as f:
+  with open(paths_file, encoding='utf-8') as f:
     paths = [normalize_path(path.strip()) for path in f if path.strip()]
     # To make /a/b/c and /a/b/c/ equivalent
     paths = [path.rstrip(os.sep) for path in paths]

From 757253d7e9d9d9e971fe390ec194767e0fe49a11 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Tue, 16 Sep 2025 00:03:56 +0000
Subject: [PATCH 09/17] More ruff fix

---
 tools/empath-split.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 0d1c10c104ce3..6712f1204def6 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -47,6 +47,7 @@
 from tools import webassembly
 from tools.utils import exit_with_error, normalize_path
 
+
 def parse_args():
   parser = argparse.ArgumentParser(
       description='Split a wasm file based on user paths',

From 272885c47c247c1f335136eeb08bce760338d2a1 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Tue, 16 Sep 2025 01:01:02 +0000
Subject: [PATCH 10/17] Add generated empath-split

---
 empath-split | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)
 create mode 100755 empath-split

diff --git a/empath-split b/empath-split
new file mode 100755
index 0000000000000..c8fb10c8e5917
--- /dev/null
+++ b/empath-split
@@ -0,0 +1,35 @@
+#!/bin/sh
+# Copyright 2020 The Emscripten Authors.  All rights reserved.
+# Emscripten is available under two separate licenses, the MIT license and the
+# University of Illinois/NCSA Open Source License.  Both these licenses can be
+# found in the LICENSE file.
+#
+# Entry point for running python scripts on UNIX systems.
+#
+# Automatically generated by `create_entry_points.py`; DO NOT EDIT.
+#
+# To make modifications to this file, edit `tools/run_python.sh` and then run
+# `tools/maint/create_entry_points.py`
+
+# $PYTHON -E does not ignore _PYTHON_SYSCONFIGDATA_NAME, an internal of cpython
+# used in cross compilation via setup.py, so we unset it explicitly here.
+unset _PYTHON_SYSCONFIGDATA_NAME
+
+if [ -z "$PYTHON" ]; then
+  PYTHON=$EMSDK_PYTHON
+fi
+
+if [ -z "$PYTHON" ]; then
+  PYTHON=$(command -v python3 2> /dev/null)
+fi
+
+if [ -z "$PYTHON" ]; then
+  PYTHON=$(command -v python 2> /dev/null)
+fi
+
+if [ -z "$PYTHON" ]; then
+  echo 'unable to find python in $PATH'
+  exit 1
+fi
+
+exec "$PYTHON" -E "$(dirname $0)/tools/empath-split.py" "$@"

From e47d6a44643a75bc02bb07783c18ec9d2afe783a Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Tue, 16 Sep 2025 01:36:33 +0000
Subject: [PATCH 11/17] Maybe I shouldn't add this after all

---
 empath-split | 35 -----------------------------------
 1 file changed, 35 deletions(-)
 delete mode 100755 empath-split

diff --git a/empath-split b/empath-split
deleted file mode 100755
index c8fb10c8e5917..0000000000000
--- a/empath-split
+++ /dev/null
@@ -1,35 +0,0 @@
-#!/bin/sh
-# Copyright 2020 The Emscripten Authors.  All rights reserved.
-# Emscripten is available under two separate licenses, the MIT license and the
-# University of Illinois/NCSA Open Source License.  Both these licenses can be
-# found in the LICENSE file.
-#
-# Entry point for running python scripts on UNIX systems.
-#
-# Automatically generated by `create_entry_points.py`; DO NOT EDIT.
-#
-# To make modifications to this file, edit `tools/run_python.sh` and then run
-# `tools/maint/create_entry_points.py`
-
-# $PYTHON -E does not ignore _PYTHON_SYSCONFIGDATA_NAME, an internal of cpython
-# used in cross compilation via setup.py, so we unset it explicitly here.
-unset _PYTHON_SYSCONFIGDATA_NAME
-
-if [ -z "$PYTHON" ]; then
-  PYTHON=$EMSDK_PYTHON
-fi
-
-if [ -z "$PYTHON" ]; then
-  PYTHON=$(command -v python3 2> /dev/null)
-fi
-
-if [ -z "$PYTHON" ]; then
-  PYTHON=$(command -v python 2> /dev/null)
-fi
-
-if [ -z "$PYTHON" ]; then
-  echo 'unable to find python in $PATH'
-  exit 1
-fi
-
-exec "$PYTHON" -E "$(dirname $0)/tools/empath-split.py" "$@"

From d20c87bc17274cc687db8a779609d11e942f9c7c Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Tue, 16 Sep 2025 01:38:03 +0000
Subject: [PATCH 12/17] Add the scripts to .gitignore

---
 .gitignore | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/.gitignore b/.gitignore
index 6700e4172e80d..558e31c0b3b46 100644
--- a/.gitignore
+++ b/.gitignore
@@ -46,6 +46,7 @@ emdump
 emdwp
 emmake
 emnm
+empath-split
 emprofile
 emranlib
 emrun
@@ -67,6 +68,7 @@ emdump.bat
 emdwp.bat
 emmake.bat
 emnm.bat
+empath-split.bat
 emprofile.bat
 emranlib.bat
 emrun.bat

From e36b0f59745f9aef43d4fdc9b965fd56eb0b76bc Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Tue, 16 Sep 2025 02:36:11 +0000
Subject: [PATCH 13/17] fix

---
 tools/empath-split.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 6712f1204def6..383e21dcb5c00 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -195,7 +195,7 @@ def main():
     paths = list(dict.fromkeys(paths))
 
   # Compute {path: list of functions} map
-  path_to_funcs = get_path_to_functions_map(wasm, sourcemap, paths, args.verbose)
+  path_to_funcs = get_path_to_functions_map(wasm, sourcemap, paths)
 
   # Write .manifest file
   f = tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=False)

From 3743aa73f948b175406ea826ca89e209309d212c Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Wed, 17 Sep 2025 01:01:06 +0000
Subject: [PATCH 14/17] Address comments

---
 tools/empath-split.py | 45 +++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 23 deletions(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 383e21dcb5c00..d7f4cb7448769 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -40,12 +40,14 @@
 __rootdir__ = os.path.dirname(__scriptdir__)
 sys.path.insert(0, __rootdir__)
 
+from tools import building
 from tools import config
 from tools import diagnostics
 from tools import emsymbolizer
 from tools import shared
+from tools import utils
 from tools import webassembly
-from tools.utils import exit_with_error, normalize_path
+from tools.utils import exit_with_error
 
 
 def parse_args():
@@ -137,7 +139,7 @@ def is_synthesized_func(func):
         addr -= 1
 
       if loc and loc.source:
-        func_to_src[func_name] = normalize_path(loc.source)
+        func_to_src[func_name] = utils.normalize_path(loc.source)
       else:
         if not is_synthesized_func(func_name):
           diagnostics.warn(f"No source file information found in the source map for function '{func_name}'")
@@ -169,33 +171,30 @@ def is_synthesized_func(func):
 
 def main():
   args, forwarded_args = parse_args()
-  wasm = os.path.expanduser(args.wasm)
-  paths_file = os.path.expanduser(args.paths_file)
-  sourcemap = os.path.expanduser(args.sourcemap) if args.sourcemap else None
   if args.wasm_split:
-    wasm_split = os.path.expanduser(args.wasm_split)
+    wasm_split = args.wasm_split
   else:
-    wasm_split = os.path.join(config.BINARYEN_ROOT, 'bin', 'wasm-split')
-
-  if not os.path.isfile(wasm):
-    exit_with_error(f"'{wasm}' was not found or not a file")
-  if not os.path.isfile(paths_file):
-    exit_with_error(f"'{paths_file}' was not found or not a file")
-  if sourcemap:
-    if not os.path.isfile(sourcemap):
-      exit_with_error(f"'{sourcemap}' was not found or not a file")
+    wasm_split = os.path.join(building.get_binaryen_bin(), 'wasm-split')
+
+  if not os.path.isfile(args.wasm):
+    exit_with_error(f"'{args.wasm}' was not found or not a file")
+  if not os.path.isfile(args.paths_file):
+    exit_with_error(f"'{args.paths_file}' was not found or not a file")
+  if args.sourcemap:
+    if not os.path.isfile(args.sourcemap):
+      exit_with_error(f"'{args.sourcemap}' was not found or not a file")
   if not os.path.isfile(wasm_split):
     exit_with_error(f"'{wasm_split}' was not found or not a file")
 
-  with open(paths_file, encoding='utf-8') as f:
-    paths = [normalize_path(path.strip()) for path in f if path.strip()]
-    # To make /a/b/c and /a/b/c/ equivalent
-    paths = [path.rstrip(os.sep) for path in paths]
-    # Remove duplicates
-    paths = list(dict.fromkeys(paths))
+  paths = utils.read_file(args.paths_file).splitlines()
+  paths = [utils.normalize_path(path.strip()) for path in paths if path.strip()]
+  # To make /a/b/c and /a/b/c/ equivalent
+  paths = [path.rstrip(os.sep) for path in paths]
+  # Remove duplicates
+  paths = list(dict.fromkeys(paths))
 
   # Compute {path: list of functions} map
-  path_to_funcs = get_path_to_functions_map(wasm, sourcemap, paths)
+  path_to_funcs = get_path_to_functions_map(args.wasm, args.sourcemap, paths)
 
   # Write .manifest file
   f = tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=False)
@@ -216,7 +215,7 @@ def main():
         f.write('\n')
     f.close()
 
-    cmd = [wasm_split, '--multi-split', wasm, '--manifest', manifest]
+    cmd = [wasm_split, '--multi-split', args.wasm, '--manifest', manifest]
     if args.verbose:
       # This option is used both in this script and wasm-split
       cmd.append('-v')

From f5f246833e8f06589519248b625b2c0811dd38a5 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Wed, 17 Sep 2025 01:14:26 +0000
Subject: [PATCH 15/17] Remove try-finally

---
 tools/empath-split.py | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index d7f4cb7448769..f12705552caa9 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -197,8 +197,7 @@ def main():
   path_to_funcs = get_path_to_functions_map(args.wasm, args.sourcemap, paths)
 
   # Write .manifest file
-  f = tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=False)
-  try:
+  with tempfile.NamedTemporaryFile(suffix=".manifest", mode='w+', delete=args.preserve_manifest) as f:
     manifest = f.name
     for i, path in enumerate(paths):
       f.write(f'{i}\n')
@@ -213,7 +212,7 @@ def main():
         f.write(func + '\n')
       if i < len(paths) - 1:
         f.write('\n')
-    f.close()
+    f.flush()
 
     cmd = [wasm_split, '--multi-split', args.wasm, '--manifest', manifest]
     if args.verbose:
@@ -223,9 +222,6 @@ def main():
     if args.verbose:
       print('\n' + ' '.join(cmd))
     shared.run_process(cmd)
-  finally:
-    if not args.preserve_manifest:
-      os.remove(manifest)
 
 
 if __name__ == '__main__':

From be1de651eaad02524ded7fe4f65533a0bb6d3703 Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Wed, 17 Sep 2025 02:34:34 +0000
Subject: [PATCH 16/17] Ruff fix

---
 tools/empath-split.py | 1 -
 1 file changed, 1 deletion(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index f12705552caa9..9b2808dd0b77b 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -41,7 +41,6 @@
 sys.path.insert(0, __rootdir__)
 
 from tools import building
-from tools import config
 from tools import diagnostics
 from tools import emsymbolizer
 from tools import shared

From a727aa1866a05e083bedd48b03db3c543bce686a Mon Sep 17 00:00:00 2001
From: Heejin Ahn <aheejin@gmail.com>
Date: Thu, 18 Sep 2025 12:08:19 -0700
Subject: [PATCH 17/17] Update tools/empath-split.py

Co-authored-by: Derek Schuff <dschuff@chromium.org>
---
 tools/empath-split.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/empath-split.py b/tools/empath-split.py
index 9b2808dd0b77b..218aba735e3d0 100755
--- a/tools/empath-split.py
+++ b/tools/empath-split.py
@@ -151,7 +151,7 @@ def is_synthesized_func(func):
   # Visit paths in the reverse sorting order, so that we can process inner paths
   # first.
   # e.g. If we have /a/b and /a/b/c, /a/b/c will come first, so we can assign
-  # functions contained in /a/b/c to it first and assign the maining functions
+  # functions contained in /a/b/c to it first and assign the remaining functions
   # to /a/b.
   visited_funcs = set()
   path_to_funcs = {}