Improve emar wrapper script

.flake8

@@ -3,6 +3,7 @@ ignore = E111,E114,E501,E261,E266,E121,E402,E241
 # TODO(sbc): Switch from the whitelist of a blacklist such as:
 # exclude = ./third_party/,emmaken.py,./tools/scons/,./tools/ports/,./tests/
 filename =
+  */emar.py
   */emcc.py
   */emscons.py
   */emscripten.py

emar.py

@@ -4,80 +4,92 @@
 # University of Illinois/NCSA Open Source License.  Both these licenses can be
 # found in the LICENSE file.
 
-'''
-emar - ar helper script
-=======================
+"""Archive helper script
 
-This script acts as a frontend replacement for ar. See emcc.
-'''
+This script acts as a frontend replacement for `ar`. See emcc.
+This is needed because, unlike a traditional linker, emscripten can't handle
+archive with duplicate member names.  This is because emscripten extracts
+archive to a temporary location and duplicate filenames will clobber each
+other in this case.
+
+"""
+
+# TODO(sbc): Implement `ar x` within emscripten, in python, to avoid this issue
+# and delete this file.
 
 from __future__ import print_function
+import hashlib
+import os
+import shutil
+import sys
+
 from tools.toolchain_profiler import ToolchainProfiler
+from tools import shared
+from tools.response_file import substitute_response_files, create_response_file
+
 if __name__ == '__main__':
   ToolchainProfiler.record_process_start()
 
-import os, sys
-from tools import shared
-from tools.response_file import substitute_response_files, create_response_file
 
 #
 # Main run() function
 #
 def run():
-  DEBUG = os.environ.get('EMCC_DEBUG')
-  if DEBUG == "0":
-    DEBUG = None
-
-  sys.argv = substitute_response_files(sys.argv)
-
-  newargs = [shared.LLVM_AR] + sys.argv[1:]
-
-  if DEBUG:
-    print('emar:', sys.argv, '  ==>  ', newargs, file=sys.stderr)
+  args = substitute_response_files(sys.argv)
+  newargs = [shared.LLVM_AR] + args[1:]
 
   to_delete = []
-  if len(newargs) > 2:
-    if 'r' in newargs[1]:
+
+  # The 3 argment form of ar doesn't involve other files. For example
+  # 'ar x libfoo.a'.
+  if len(newargs) > 3:
+    cmd = newargs[1]
+    if 'r' in cmd:
       # we are adding files to the archive.
-      # find the .a; everything after it is an input file.
+      # normally the output file is then arg 2, except in the case were the
+      # a or b modifiers are used in which case its arg 3.
+      if 'a' in cmd or 'b' in cmd:
+        new_member_args_start = 4
+      else:
+        new_member_args_start = 3
+
       # we add a hash to each input, to make them unique as
       # possible, as llvm-ar cannot extract duplicate names
       # (and only the basename is used!)
-      i = 1
-      while i < len(newargs):
-        if newargs[i].endswith('.a'):
-          import hashlib, shutil
-          for j in range(i+1, len(newargs)):
-            orig_name = newargs[j]
-            full_name = os.path.abspath(orig_name)
-            dir_name = os.path.dirname(full_name)
-            base_name = os.path.basename(full_name)
-            h = hashlib.md5(full_name.encode('utf-8')).hexdigest()[:8]
-            parts = base_name.split('.')
-            parts[0] += '_' + h
-            newname = '.'.join(parts)
-            full_newname = os.path.join(dir_name, newname)
-            if not os.path.exists(full_newname):
-              try: # it is ok to fail here, we just don't get hashing
-                shutil.copyfile(orig_name, full_newname)
-                newargs[j] = full_newname
-                to_delete.append(full_newname)
-              except:
-                pass
-          break
-        i += 1
-
-  response_filename = None
-  if len(newargs) > 3:
+      for j in range(new_member_args_start, len(newargs)):
+        orig_name = newargs[j]
+        full_name = os.path.abspath(orig_name)
+        dir_name = os.path.dirname(full_name)
+        base_name = os.path.basename(full_name)
+        h = hashlib.md5(full_name.encode('utf-8')).hexdigest()[:8]
+        parts = base_name.split('.')
+        parts[0] += '_' + h
+        newname = '.'.join(parts)
+        full_newname = os.path.join(dir_name, newname)
+        if not os.path.exists(full_newname):
+          try: # it is ok to fail here, we just don't get hashing
+            shutil.copyfile(orig_name, full_newname)
+            newargs[j] = full_newname
+            to_delete.append(full_newname)
+          except:
+            pass
+
+    if shared.DEBUG:
+      print('emar:', sys.argv, '  ==>  ', newargs, file=sys.stderr)
+
     response_filename = create_response_file(newargs[3:], shared.get_emscripten_temp_dir())
     to_delete += [response_filename]
     newargs = newargs[:3] + ['@' + response_filename]
 
+  if shared.DEBUG:
+    print('emar:', sys.argv, '  ==>  ', newargs, file=sys.stderr)
+
   try:
-    return shared.run_process(newargs, stdin=sys.stdin).returncode
+    return shared.run_process(newargs, stdin=sys.stdin, check=False).returncode
   finally:
     for d in to_delete:
       shared.try_delete(d)
 
+
 if __name__ == '__main__':
   sys.exit(run())

tests/test_other.py

@@ -5803,37 +5803,38 @@ def test_call_nonemterpreted_during_sleep(self):
     self.assertContained('cannot have an EM_ASM on the stack when emterpreter pauses/resumes', run_js('a.out.js', stderr=STDOUT, assert_returncode=None))
 
   def test_link_with_a_static(self):
-    for args in [[], ['-O2']]:
-      print(args)
-      self.clear()
-      create_test_file('x.c', r'''
+    create_test_file('x.c', r'''
 int init_weakref(int a, int b) {
-    return a + b;
+  return a + b;
 }
 ''')
-      create_test_file('y.c', r'''
+    create_test_file('y.c', r'''
 static int init_weakref(void) { // inlined in -O2, not in -O0 where it shows up in llvm-nm as 't'
-    return 150;
+  return 150;
 }
 
 int testy(void) {
-    return init_weakref();
+  return init_weakref();
 }
 ''')
-      create_test_file('z.c', r'''
+    create_test_file('z.c', r'''
 extern int init_weakref(int, int);
 extern int testy(void);
 
 int main(void) {
-    return testy() + init_weakref(5, 6);
+  return testy() + init_weakref(5, 6);
 }
 ''')
-      run_process([PYTHON, EMCC, 'x.c', '-o', 'x.o'])
-      run_process([PYTHON, EMCC, 'y.c', '-o', 'y.o'])
-      run_process([PYTHON, EMCC, 'z.c', '-o', 'z.o'])
-      run_process([PYTHON, EMAR, 'rc', 'libtest.a', 'y.o'])
-      run_process([PYTHON, EMAR, 'rc', 'libtest.a', 'x.o'])
-      run_process([PYTHON, EMRANLIB, 'libtest.a'])
+    run_process([PYTHON, EMCC, 'x.c', '-o', 'x.o'])
+    run_process([PYTHON, EMCC, 'y.c', '-o', 'y.o'])
+    run_process([PYTHON, EMCC, 'z.c', '-o', 'z.o'])
+    try_delete('libtest.a')
+    run_process([PYTHON, EMAR, 'rc', 'libtest.a', 'y.o'])
+    run_process([PYTHON, EMAR, 'rc', 'libtest.a', 'x.o'])
+    run_process([PYTHON, EMRANLIB, 'libtest.a'])
+
+    for args in [[], ['-O2']]:
+      print('args:', args)
       run_process([PYTHON, EMCC, 'z.o', 'libtest.a', '-s', 'EXIT_RUNTIME=1'] + args)
       run_js('a.out.js', assert_returncode=161)