Build a complete sysroot in the cache directory

Rather than adding various include paths, copy any needed headers into
the sysroot along with any libraries.

This means that emscripten can work a lot more like the
traditional cross compiler (e.g. clang -target=xxx --sysroot=yyy),
and we can start to think of the emscripten driver as a seperate
thing to the sysroot.

Fixes: #9353

Author: sbc100

embuilder.py

@@ -129,13 +129,9 @@ def main():
 
   if args.lto:
     shared.Settings.LTO = "full"
-    # Reconfigure the cache dir to reflect the change
-    shared.reconfigure_cache()
 
   if args.pic:
     shared.Settings.RELOCATABLE = 1
-    # Reconfigure the cache dir to reflect the change
-    shared.reconfigure_cache()
 
   if args.force:
     force = True
@@ -177,6 +173,10 @@ def main():
       if force:
         library.erase()
       library.get_path()
+    elif what == 'sysroot':
+      if force:
+        shared.Cache.erase_file('sysroot_install.stamp')
+      system_libs.ensure_sysroot()
     elif what == 'struct_info':
       if force:
         shared.Cache.erase_file('generated_struct_info.json')

emcc.py

@@ -1367,16 +1367,6 @@ def filter_out_duplicate_dynamic_libs(inputs):
                                              '_emscripten_stack_get_end',
                                              '_emscripten_stack_set_limits']
 
-    # Reconfigure the cache now that settings have been applied. Some settings
-    # such as LTO and SIDE_MODULE/MAIN_MODULE effect which cache directory we use.
-    shared.reconfigure_cache()
-
-    if not compile_only and not options.post_link:
-      ldflags = shared.emsdk_ldflags(newargs)
-      for f in ldflags:
-        newargs.append(f)
-        add_link_flag(len(newargs), f)
-
     # SSEx is implemented on top of SIMD128 instruction set, but do not pass SSE flags to LLVM
     # so it won't think about generating native x86 SSE code.
     newargs = [x for x in newargs if x not in shared.SIMD_INTEL_FEATURE_TOWER and x not in shared.SIMD_NEON_FLAGS]
@@ -1984,6 +1974,7 @@ def is_link_flag(flag):
 
     compile_args = [a for a in newargs if a and not is_link_flag(a)]
     cflags = calc_cflags(options)
+    system_libs.ensure_sysroot()
     system_libs.add_ports_cflags(cflags, shared.Settings)
 
     def use_cxx(src):
@@ -2010,9 +2001,8 @@ def get_compiler(cxx):
       return CC
 
     def get_clang_command(src_file):
-      cxx = use_cxx(src_file)
-      per_file_cflags = shared.get_cflags(args, cxx)
-      return get_compiler(cxx) + cflags + per_file_cflags + compile_args + [src_file]
+      per_file_cflags = shared.get_cflags(args)
+      return get_compiler(use_cxx(src_file)) + cflags + per_file_cflags + compile_args + [src_file]
 
     def get_clang_command_asm(src_file):
       asflags = shared.get_asmflags()
@@ -2106,6 +2096,10 @@ def compile_source_file(i, input_file):
   if specified_target and specified_target.startswith('-'):
     exit_with_error('invalid output filename: `%s`' % specified_target)
 
+  ldflags = shared.emsdk_ldflags(newargs)
+  for f in ldflags:
+    add_link_flag(sys.maxsize, f)
+
   using_lld = not (link_to_object and shared.Settings.LTO)
   link_flags = filter_link_flags(link_flags, using_lld)
 

src/struct_info.json

@@ -1525,7 +1525,7 @@
             ]
     },
     {
-        "file": "../lib/libc/musl/src/internal/pthread_impl.h",
+        "file": "system/lib/libc/musl/src/internal/pthread_impl.h",
         "structs": {
             "pthread": [
               "threadStatus",
@@ -1548,7 +1548,7 @@
         "defines": ["__ATTRP_C11_THREAD"]
     },
     {
-        "file": "../lib/libc/musl/src/internal/libc.h",
+        "file": "system/lib/libc/musl/src/internal/libc.h",
         "structs": {
             "libc": [
               "global_locale"

tests/test_sanity.py

@@ -21,7 +21,7 @@
 from tools.shared import EXPECTED_LLVM_VERSION, Cache
 from tools import shared, system_libs, utils
 
-SANITY_FILE = shared.Cache.get_path('sanity.txt', root=True)
+SANITY_FILE = shared.Cache.get_path('sanity.txt')
 commands = [[EMCC], [PYTHON, path_from_root('tests', 'runner.py'), 'blahblah']]
 
 

tools/cache.py

@@ -23,22 +23,9 @@ class Cache(object):
   # acquired.
   EM_EXCLUSIVE_CACHE_ACCESS = int(os.environ.get('EM_EXCLUSIVE_CACHE_ACCESS', '0'))
 
-  def __init__(self, dirname, use_subdir=True):
+  def __init__(self, dirname):
     # figure out the root directory for all caching
     dirname = os.path.normpath(dirname)
-    self.root_dirname = dirname
-
-    # if relevant, use a subdir of the cache
-    if use_subdir:
-      subdir = 'wasm'
-      if shared.Settings.LTO:
-        subdir += '-lto'
-      if shared.Settings.RELOCATABLE:
-        subdir += '-pic'
-      if shared.Settings.MEMORY64:
-        subdir += '-memory64'
-      dirname = os.path.join(dirname, subdir)
-
     self.dirname = dirname
     self.acquired_count = 0
 
@@ -81,30 +68,56 @@ def ensure(self):
   def erase(self):
     self.acquire_cache_lock()
     try:
-      if os.path.exists(self.root_dirname):
-        for f in os.listdir(self.root_dirname):
-          tempfiles.try_delete(os.path.join(self.root_dirname, f))
+      if os.path.exists(self.dirname):
+        for f in os.listdir(self.dirname):
+          tempfiles.try_delete(os.path.join(self.dirname, f))
     finally:
       self.release_cache_lock()
 
-  def get_path(self, shortname, root=False):
-    if root:
-      return os.path.join(self.root_dirname, shortname)
-    return os.path.join(self.dirname, shortname)
+  def get_path(self, name):
+    return os.path.join(self.dirname, name)
+
+  def get_sysroot_dir(self, absolute):
+    if absolute:
+      return os.path.join(self.dirname, 'sysroot')
+    return 'sysroot'
+
+  def get_include_dir(self):
+    return os.path.join(self.get_sysroot_dir(absolute=True), 'include')
+
+  def get_lib_dir(self, absolute):
+    path = os.path.join(self.get_sysroot_dir(absolute=absolute), 'lib')
+    if shared.Settings.MEMORY64:
+      path = os.path.join(path, 'wasm64-emscripten')
+    else:
+      path = os.path.join(path, 'wasm32-emscripten')
+    # if relevant, use a subdir of the cache
+    subdir = []
+    if shared.Settings.LTO:
+      subdir.append('lto')
+    if shared.Settings.RELOCATABLE:
+      subdir.append('pic')
+    if subdir:
+      path = os.path.join(path, '-'.join(subdir))
+    return path
+
+  def get_lib_name(self, name):
+    return os.path.join(self.get_lib_dir(absolute=False), name)
 
   def erase_file(self, shortname):
     name = os.path.join(self.dirname, shortname)
     if os.path.exists(name):
       logging.info('Cache: deleting cached file: %s', name)
       tempfiles.try_delete(name)
 
+  def getlib(self, libname, *args, **kwargs):
+    name = self.get_lib_name(libname)
+    return self.get(name, *args, **kwargs)
+
   # Request a cached file. If it isn't in the cache, it will be created with
   # the given creator function
-  def get(self, shortname, creator, what=None, force=False, root=False):
-    if root:
-      cachename = os.path.join(self.root_dirname, shortname)
-    else:
-      cachename = os.path.join(self.dirname, shortname)
+  def get(self, shortname, creator, what=None, force=False):
+    cachename = os.path.join(self.dirname, shortname)
     cachename = os.path.abspath(cachename)
     # Check for existence before taking the lock in case we can avoid the
     # lock completely.

tools/gen_struct_info.py

@@ -246,6 +246,7 @@ def inspect_code(headers, cpp_opts, structs, defines):
                                     '-O0',
                                     '-Werror',
                                     '-Wno-format',
+                                    '-I', shared.path_from_root(),
                                     '-s', 'BOOTSTRAPPING_STRUCT_INFO=1',
                                     '-s', 'WARN_ON_UNDEFINED_SYMBOLS=0',
                                     '-s', 'STRICT=1',

tools/ports/sdl2.py

@@ -82,7 +82,7 @@ def create():
     ports.create_lib(final, o_s)
     return final
 
-  return [shared.Cache.get(libname, create, what='port')]
+  return [shared.Cache.getlib(libname, create, what='port')]
 
 
 def clear(ports, settings, shared):

tools/shared.py

@@ -273,7 +273,7 @@ def check_sanity(force=False):
       return # config stored directly in EM_CONFIG => skip sanity checks
     expected = generate_sanity()
 
-    sanity_file = Cache.get_path('sanity.txt', root=True)
+    sanity_file = Cache.get_path('sanity.txt')
     if os.path.exists(sanity_file):
       sanity_data = open(sanity_file).read()
       if sanity_data != expected:
@@ -420,9 +420,7 @@ def emsdk_ldflags(user_args):
     return []
 
   library_paths = [
-      path_from_root('system', 'local', 'lib'),
-      path_from_root('system', 'lib'),
-      Cache.dirname
+      Cache.get_lib_dir(absolute=True)
   ]
   ldflags = ['-L' + l for l in library_paths]
 
@@ -437,12 +435,12 @@ def emsdk_ldflags(user_args):
   return ldflags
 
 
-def emsdk_cflags(user_args, cxx):
+def emsdk_cflags(user_args):
   # Disable system C and C++ include directories, and add our own (using
   # -isystem so they are last, like system dirs, which allows projects to
   # override them)
 
-  c_opts = ['--sysroot=' + path_from_root('system')]
+  c_opts = ['--sysroot=' + Cache.get_sysroot_dir(absolute=True)]
 
   def array_contains_any_of(hay, needles):
     for n in needles:
@@ -477,24 +475,8 @@ def array_contains_any_of(hay, needles):
 
   sysroot_include_paths = []
 
-  if cxx:
-    sysroot_include_paths += [
-      os.path.join('/include', 'libcxx'),
-      os.path.join('/lib', 'libcxxabi', 'include'),
-    ]
-
-  # TODO: Merge the cache into the sysroot.
-  c_opts += ['-Xclang', '-isystem' + Cache.get_path('include')]
-
   sysroot_include_paths += [
     os.path.join('/include', 'compat'),
-    os.path.join('/include', 'libc'),
-    os.path.join('/lib', 'libc', 'musl', 'arch', 'emscripten'),
-    os.path.join('/local', 'include'),
-    os.path.join('/include', 'SSE'),
-    os.path.join('/include', 'neon'),
-    os.path.join('/lib', 'compiler-rt', 'include'),
-    os.path.join('/lib', 'libunwind', 'include'),
   ]
 
   def include_directive(paths):
@@ -511,7 +493,7 @@ def get_asmflags():
   return ['-target', get_llvm_target()]
 
 
-def get_cflags(user_args, cxx):
+def get_cflags(user_args):
   # Set the LIBCPP ABI version to at least 2 so that we get nicely aligned string
   # data and other nice fixes.
   c_opts = [# '-fno-threadsafe-statics', # disabled due to issue 1289
@@ -538,7 +520,7 @@ def get_cflags(user_args, cxx):
   if os.environ.get('EMMAKEN_NO_SDK') or '-nostdinc' in user_args:
     return c_opts
 
-  return c_opts + emsdk_cflags(user_args, cxx)
+  return c_opts + emsdk_cflags(user_args)
 
 
 # Settings. A global singleton. Not pretty, but nicer than passing |, settings| everywhere
@@ -747,11 +729,6 @@ def asmjs_mangle(name):
     return name
 
 
-def reconfigure_cache():
-  global Cache
-  Cache = cache.Cache(config.CACHE)
-
-
 class JS(object):
   emscripten_license = '''\
 /**