referrer_charset is a lie. It's really the user's default_charset.

referrer_charset is a lie.  It's really the user's default_charset.

The "referrer_charset" is used as a fallback for decode filenames in the
Content-Disposition header.  This name is a mis-nomer because the value of this
variable is completely unrelated to the referrer.  Instead, the value actually
is the default charset from the user's profile.  This patch renames the
"referrer_charset" concept to "default_charset" in an attempt to restore sanity
to the world.

[abarth-chromium, 2012-02-01 00:01:54]

@asaka for net OWNERS; @darin for contetn OWNERS.

[asanka, 2012-02-01 01:42:58]

Is the charset passed in to DownloadManager::DownloadUrl() from
RenderViewContextMenu::ExecuteCommand() incorrect?  It seems the charset doesn't
make it beyond DownloadUrlToFile(). I think we should either fix that, or not
pass a charset around if we aren't going to use it.

[cbentzel, 2012-02-01 01:57:15]

I tripped over this problem recently, and tracked down when the problem was
introduced.

From a message to jam@chromium.org

"""
I noticed that "referrer_charset" is now a vestigial argument to
DownloadManager::DownloadUrl.

It looks like this used to be used, but a CL of yours removed it as part of
the content work:
http://codereview.chromium.org/6995013/diff/13005/chrome/browser/download/dow...

Was this an accident, or intentional?
"""

My guess is it's an accident. Agree that we either need to pipe it back in
or remove. Sorry for not entering a bug.

On Tue, Jan 31, 2012 at 8:42 PM, <asanka@chromium.org> wrote:

> Is the charset passed in to DownloadManager::DownloadUrl() from
> RenderViewContextMenu::**ExecuteCommand() incorrect?  It seems the
> charset doesn't
> make it beyond DownloadUrlToFile(). I think we should either fix that, or
> not
> pass a charset around if we aren't going to use it.
>
>
>
http://codereview.chromium.**org/9317018/<http://codereview.chromium.org/9317...
>

[abarth-chromium, 2012-02-01 02:02:19]

I'm not sure I quite follow, but maybe I need to look at the code again.

[darin (slow to review), 2012-02-01 05:51:13]

content/  <--  LGTM