add parts of system bitcode libraries when necessary, and refactor library inclusion decision code

kripken · kripken · commit b0b611c9f738 · 2012-12-07T12:32:49.000-08:00
diff --git a/emcc b/emcc
@@ -991,20 +991,32 @@ try:
     for name, create, fix, library_symbols in [('libcxx',    create_libcxx,    fix_libcxx,    libcxx_symbols),
                                                ('libcxxabi', create_libcxxabi, fix_libcxxabi, libcxxabi_symbols),
                                                ('dlmalloc',  create_dlmalloc,  fix_dlmalloc,  dlmalloc_symbols)]:
-      need = []
-      has = []
+      need = set()
+      has = set()
       for temp_file in temp_files:
         symbols = shared.Building.llvm_nm(temp_file)
         for library_symbol in library_symbols:
           if library_symbol in symbols.undefs:
-            need.append(library_symbol)
+            need.add(library_symbol)
           if library_symbol in symbols.defs:
-            has.append(library_symbol)
-      if DEBUG: print >> sys.stderr, 'emcc: considering including %s: we need |%s| and have |%s|' % (name, str(need), str(has))
-      if force or (need and not has):
+            has.add(library_symbol)
+      for haz in has: # remove symbols that are supplied by another of the inputs
+        if haz in need:
+          need.remove(haz)
+      if DEBUG: print >> sys.stderr, 'emcc: considering including %s: we need %s and have %s' % (name, str(need), str(has))
+      if force or len(need) > 0:
         # We need to build and link the library in
         if DEBUG: print >> sys.stderr, 'emcc: including %s' % name
-        extra_files_to_link.append(shared.Cache.get(name, create))
+        libfile = shared.Cache.get(name, create)
+        if len(has) > 0:
+          # remove the symbols we do not need
+          fixed = in_temp(uniquename(libfile)) + '.bc'
+          shutil.copyfile(libfile, fixed)
+          for haz in has:
+            if DEBUG: print >> sys.stderr, 'emcc: including: removing symbol "%s" that we have' % haz
+            shared.Building.remove_symbol(fixed, haz)
+          libfile = fixed
+        extra_files_to_link.append(libfile)
         force = True
         if fix:
           fix()
diff --git a/tests/runner.py b/tests/runner.py
@@ -5779,6 +5779,39 @@ def test_dlmalloc(self):
         ]:
           self.do_run(src.replace('{{{ NEW }}}', new).replace('{{{ DELETE }}}', delete), '*1,0*')
 
+    def test_dlmalloc_partial(self):
+      if self.emcc_args is None: return self.skip('only emcc will link in dlmalloc')
+      # present part of the symbols of dlmalloc, not all
+      src = open(path_from_root('tests', 'new.cpp')).read().replace('{{{ NEW }}}', 'new int').replace('{{{ DELETE }}}', 'delete') + '''    
+void *
+operator new(size_t size)
+{
+  printf("new %d!\\n", size);
+  return malloc(size);
+}
+'''
+      self.do_run(src, 'new 4!\n*1,0*')
+
+    def test_dlmalloc_partial_2(self):
+      if self.emcc_args is None or 'SAFE_HEAP' in str(self.emcc_args): return self.skip('only emcc will link in dlmalloc, and we do unsafe stuff')
+      # present part of the symbols of dlmalloc, not all. malloc is harder to link than new which is weak.
+      src = r'''
+        #include <stdio.h>
+        #include <stdlib.h>
+        void *malloc(size_t size)
+        {
+          return (void*)123;
+        }
+        int main() {
+          void *x = malloc(10);
+          printf("got %p\n", x);
+          free(x);
+          printf("freed the faker\n");
+          return 1;
+        }
+'''
+      self.do_run(src, 'got 0x7b\nfreed')
+
     def test_libcxx(self):
       self.do_run(open(path_from_root('tests', 'hashtest.cpp')).read(),
                    'june -> 30\nPrevious (in alphabetical order) is july\nNext (in alphabetical order) is march')
diff --git a/tools/shared.py b/tools/shared.py
@@ -706,6 +706,11 @@ def open_make_err(i, mode='r'):
       os.chdir(old_dir)
     return generated_libs
 
+  @staticmethod
+  def remove_symbol(filename, symbol):
+    Popen([LLVM_EXTRACT, filename, '-delete', '-glob=' + symbol, '-o', filename], stderr=PIPE).communicate()
+    Popen([LLVM_EXTRACT, filename, '-delete', '-func=' + symbol, '-o', filename], stderr=PIPE).communicate()
+
   @staticmethod
   def link(files, target, remove_duplicates=False):
     actual_files = []
@@ -775,8 +780,7 @@ def link(files, target, remove_duplicates=False):
           print >> sys.stderr, 'emcc: warning: removing duplicates in', actual
           for dupe in dupes:
             print >> sys.stderr, 'emcc: warning: removing duplicate', dupe
-            Popen([LLVM_EXTRACT, actual, '-delete', '-glob=' + dupe, '-o', actual], stderr=PIPE).communicate()
-            Popen([LLVM_EXTRACT, actual, '-delete', '-func=' + dupe, '-o', actual], stderr=PIPE).communicate()
+            Building.remove_symbol(actual, dupe)
           Popen([LLVM_EXTRACT, actual, '-delete', '-glob=.str', '-o', actual], stderr=PIPE).communicate() # garbage that appears here
         seen_symbols = seen_symbols.union(symbols.defs)
 
