Introduce build target for bundled data on Windows

On non-Windows platforms, we generate assembly to bundle ICU data.
On Windows however, we previously use a checked in icudt.dll, which
is no longer maintained.

We could use the same assembly approach if we wrap in inline assembly.

Bug: chromium:938223
Change-Id: Ieced205ae2e082a5290bd45499216c6c86751ca1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/deps/icu/+/1503032


Reviewed-by: Jungshik Shin <jshin@chromium.org>

BUILD.gn

@@ -49,11 +49,7 @@ config("icu_config") {
   if (icu_use_data_file) {
     defines += [ "ICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_FILE" ]
   } else {
-    if (is_win) {
-      defines += [ "ICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_SHARED" ]
-    } else {
-      defines += [ "ICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_STATIC" ]
-    }
+    defines += [ "ICU_UTIL_DATA_IMPL=ICU_UTIL_DATA_STATIC" ]
   }
 
   if (is_win) {
@@ -1106,37 +1102,47 @@ if (icu_use_data_file) {
     }
   }
 } else {
+  data_assembly = "$target_gen_dir/${data_bundle_prefix}_dat.S"
+  inline_data_assembly = "$target_gen_dir/${data_bundle_prefix}_dat.cc"
+  action("make_data_assembly") {
+    script = "scripts/make_data_assembly.py"
+    inputs = [ "$data_dir/$data_bundle" ]
+    outputs = [ data_assembly ]
+    args = [
+      rebase_path(inputs[0], root_build_dir),
+      rebase_path(data_assembly, root_build_dir),
+    ]
+    # TODO(GYP): Gyp has considerations here for QNX and for the host
+    # toolchain that have not been ported over.
+    if (is_mac || is_ios) {
+      args += [ "--mac" ]
+    } else if (is_win) {
+      args += [ "--win" ]
+    }
+  }
+
   if (is_win) {
-    # On Windows the target DLL is pre-built so just use a copy rule.
-    # data_bundle and data_dir have to be used to avoid 'unused variable'
-    # error.
-    data_bundle = "icudt.dll"
-    data_dir = "windows"
-    copy("icudata") {
-      sources = [ "$data_dir/$data_bundle" ]
-      outputs = [ "$root_out_dir/$data_bundle" ]
-      data = outputs
+    action("make_inline_data_assembly") {
+      deps = [ ":make_data_assembly" ]
+      script = "scripts/asm_to_inline_asm.py"
+      inputs = [ data_assembly ]
+      outputs =  [ inline_data_assembly ]
+      args = rebase_path([
+        data_assembly,
+        inline_data_assembly,
+      ], root_build_dir)
     }
   } else {
-    data_assembly = "$target_gen_dir/${data_bundle_prefix}_dat.S"
-    action("make_data_assembly") {
-      script = "scripts/make_data_assembly.py"
-      inputs = [ "$data_dir/$data_bundle" ]
-      outputs = [ "$data_assembly" ]
-      args = [
-        rebase_path(inputs[0], root_build_dir),
-        rebase_path(data_assembly, root_build_dir),
-      ]
-      # TODO(GYP): Gyp has considerations here for QNX and for the host
-      # toolchain that have not been ported over.
-      if (is_mac || is_ios) {
-        args += [ "--mac" ]
-      }
-    }
+    not_needed([ "inline_data_assembly" ])
+  }
 
-    source_set("icudata") {
-      sources = [ "$data_assembly" ]
-      defines = [ "U_HIDE_DATA_SYMBOL" ]
+  source_set("icudata") {
+    defines = [ "U_HIDE_DATA_SYMBOL" ]
+    if (is_win) {
+      sources = [ inline_data_assembly ]
+      deps = [ ":make_inline_data_assembly", ]
+    } else {
+      sources = [ data_assembly ]
       deps = [ ":make_data_assembly", ]
     }
   }

scripts/asm_to_inline_asm.py

+#!/usr/bin/env python
+
+# Copyright 2019 The Chromium Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style license that can be
+# found in the LICENSE file.
+
+'''
+Converts a given file in clang assembly syntax to a corresponding
+representation in inline assembly. Specifically, this is used for
+Windows clang builds.
+'''
+
+import argparse
+import sys
+
+def asm_to_inl_asm(in_filename, out_filename):
+  with open(in_filename, 'r') as infile, open(out_filename, 'wb') as outfile:
+    outfile.write('__asm__(\n')
+    for line in infile:
+      # Escape " in .S file before outputing it to inline asm file.
+      line = line.replace('"', '\\"')
+      outfile.write('  "%s\\n"\n' % line.rstrip())
+    outfile.write(');\n')
+  return 0
+
+if __name__ == '__main__':
+  parser = argparse.ArgumentParser(description=__doc__)
+  parser.add_argument('input', help='Name of the input assembly file')
+  parser.add_argument('output', help='Name of the target CC file')
+  args = parser.parse_args()
+  sys.exit(asm_to_inl_asm(args.input, args.output))

scripts/make_data_assembly.py

@@ -12,7 +12,10 @@ parser = optparse.OptionParser()
 parser.add_option("--mac",
                   help="generate assembly file for Mac/iOS (default: False)",
                   action="store_true", default=False)
-parser.set_usage("""make_data_assembly  icu_data [assembly_file] [--mac]
+parser.add_option("--win",
+                  help="generate assembly file for Windows (default: False)",
+                  action="store_true", default=False)
+parser.set_usage("""make_data_assembly  icu_data [assembly_file] [--mac] [--win]
     icu_data: ICU data file to generate assembly from.
     assembly_file: Output file converted from icu_data file.""")
 (options, args) = parser.parse_args()
@@ -56,6 +59,11 @@ if options.mac:
                "\t.const\n"
                "\t.align 4\n"
                "_icudt%s_dat:\n" %tuple([version_number] * 3))
+elif options.win:
+  output.write(".globl icudt%s_dat\n"
+               "\t.section .rdata\n"
+               "\t.balign 16\n"
+               "icudt%s_dat:\n" % tuple([version_number] * 2))
 else:
   output.write(".globl icudt%s_dat\n"
                "\t.section .note.GNU-stack,\"\",%%progbits\n"