Implement SysInfo::CPUArchitecture() on Windows.

base/sys_info_posix.cc

 // Copyright (c) 2011 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.
 
 #include "base/sys_info.h"
 
 #include <errno.h>
 #include <string.h>
 #include <sys/param.h>
 #include <sys/utsname.h>
@@ skipping 61 matching lines @@
 }
 #endif
 
 // static
 std::string SysInfo::CPUArchitecture() {
   struct utsname info;
   if (uname(&info) < 0) {
     NOTREACHED();
     return "";
   }
-  return std::string(info.machine);
+  std::string arch(info.machine);
+  if (arch == "i386" || arch == "i486" || arch == "i586" || arch == "i686") {
+    arch = "x86";
+  } else if (arch == "amd64") {
+    arch = "x86_64";

[Mark Mentovai, 2012/09/13 12:40:19]

For Mac and Linux x86_*, this returns the kernel’s bitness, which may not be the
same thing as the underlying hardware’s actual capability (32-bit kernel running
on 64-bit CPU) and may not be the same thing as the running Chrome’s bitness
(32-bit Chrome running on 64-bit kernel, or for that matter on the Mac, 64-bit
Chrome running on 32-bit kernel—except there currently is no 64-bit Mac Chrome).

Is all of this OK? Is it comparable to the behavior on Windows? The comment in
the header doesn’t explain which architecture of (CPU, kernel, app) this is
expected to return.

[Lei Zhang, 2012/09/21 21:10:21]

Right, I think CPUArchitecture() should be renamed to OSArch(). The behavior on
Windows matches the behavior on Linux. I've added additional comments in the
header.

+  }
+  return arch;
 }
 
 // static
 size_t SysInfo::VMAllocationGranularity() {
   return getpagesize();
 }
 
 }  // namespace base