Add a method to explicitly request a UTF8 representation of a FilePath. This...

base/file_path.cc

 // Copyright (c) 2008 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/file_path.h"
 #include "base/file_util.h"
 #include "base/logging.h"
 
 // These includes are just for the *Hack functions, and should be removed
 // when those functions are removed.
@@ skipping 45 matching lines @@
   return path.length() > 1 &&
       FilePath::IsSeparator(path[0]) && FilePath::IsSeparator(path[1]);
 #else  // FILE_PATH_USES_DRIVE_LETTERS
   // Look for a separator in the first position.
   return path.length() > 0 && FilePath::IsSeparator(path[0]);
 #endif  // FILE_PATH_USES_DRIVE_LETTERS
 }
 
 }  // namespace
 
+const std::string FilePath::UTF8Value() const {
+#if defined(OS_POSIX)
+  // TODO(erikkay): I'm not sure how to make this code always correct for Linux.

[darin (slow to review), 2008/12/23 20:23:55]

this is precisely why this kind of function is problematic / evil.  why do we
need a UTF8Value accessor?

if you need to store UTF8 representations, then maybe you should convert to a
file:// URL where we have an escaping mechanism to deal with non-UTF8 code
points.

[Avi (use Gerrit), 2008/12/23 22:58:49]

There are cases where we need to have the path in one format across all
platforms. The example I just hit was converting to a WebCore::PluginInfo, which
takes a String.

[Erik does not do reviews, 2008/12/23 23:02:04]

Unfortunately, as Darin educated me here, this method can't actually work on
Linux.  Since from the Linux perspective, it doesn't actually know what the
filesystem encoding is for a given filename, asking it to convert to UTF8 is
nonsensical.  I'm about to revert this CL.

+  return path_;
+#elif defined(OS_WIN)
+  return WideToUTF8(path_);
+#endif
+}
+
 bool FilePath::IsSeparator(CharType character) {
   for (size_t i = 0; i < arraysize(kSeparators) - 1; ++i) {
     if (character == kSeparators[i]) {
       return true;
     }
   }
 
   return false;
 }
 
@@ skipping 173 matching lines @@
        --pos) {
     // If the string only has two separators and they're at the beginning,
     // don't strip them, unless the string began with more than two separators.
     if (pos != start + 1 || last_stripped == start + 2 ||
         !IsSeparator(path_[start - 1])) {
       path_.resize(pos - 1);
       last_stripped = pos;
     }
   }
 }