Remove uninitialized const global g_GbFontNameMap.

Remove uninitialized const global g_GbFontNameMap.

C++11 makes uninitialized const PODs an error, because they contain
uninitialized memory (they're uninitialized that can never be initialized
(because they're const).

In this case, the memory was only used by _GetSubFontName() if the lang
parameter was 1, but _GetSubFontName() is only called from one place, with
a lang parameter of 0. So remove _GetSubFontName()'s lang parameter too.

(Using bsearch for searching an array that always has exactly 2 entries is
overkill too, but I'm trying to keep the diff small.)

No intended behavior change. Fixes this error on the clang/win bot:
..\..\third_party\pdfium\core\src\fxge\win32\fx_win32_device.cpp(207,20) :  error(clang): default initialization of an object of const type 'const _FontNameMap [1]'
const _FontNameMap g_GbFontNameMap[1];
                   ^

BUG=chromium:82385
R=bo_xu@foxitsoftware.com

Committed: https://pdfium.googlesource.com/pdfium/+/27e35a8

[jam, 2014-07-14 22:48:20]

Nico: for changes to pdfium, Bo/Jun are better suited to review this since
they're more familiar with the code, and can ensure that the relevant fixes are
also committed in their internal tree.

[Nico, 2014-07-15 21:45:15]

Bo/Jun: Ping.

[Bo Xu, 2014-07-15 22:10:10]

On 2014/07/15 21:45:15, Nico (away) wrote:
> Bo/Jun: Ping.

I looked at this issue. Although the g_GbFontNameMap is not used for now. The
function prototype is reserved to deal with different font maps. Just removing
it does not seem to be a good option. Let me check our developer for this module
and see what they say. My thinking is to give an initial value for
g_GbFontNameMap.

[Nico, 2014-07-15 22:17:58]

On 2014/07/15 22:10:10, Bo Xu wrote:
> On 2014/07/15 21:45:15, Nico (away) wrote:
> > Bo/Jun: Ping.
> 
> I looked at this issue. Although the g_GbFontNameMap is not used for now. The
> function prototype is reserved to deal with different font maps.

The guideline we use for this is to not have things just because they might be
needed one day – that keeps the code simpler, and maybe it turns out it won't be
needed in the future ( http://en.wikipedia.org/wiki/You_aren't_gonna_need_it ).

> Just removing
> it does not seem to be a good option. Let me check our developer for this
module
> and see what they say. My thinking is to give an initial value for
> g_GbFontNameMap.

Adding an initializer fixes my problem too, but this change also makes the code
simpler, which seems nice. (But up to you; you're much more familiar with this
than I.)

[Bo Xu, 2014-07-16 03:33:21]

On 2014/07/15 22:17:58, Nico (away) wrote:
> On 2014/07/15 22:10:10, Bo Xu wrote:
> > On 2014/07/15 21:45:15, Nico (away) wrote:
> > > Bo/Jun: Ping.
> > 
> > I looked at this issue. Although the g_GbFontNameMap is not used for now.
The
> > function prototype is reserved to deal with different font maps.
> 
> The guideline we use for this is to not have things just because they might be
> needed one day – that keeps the code simpler, and maybe it turns out it won't
be
> needed in the future ( http://en.wikipedia.org/wiki/You_aren ).
> 
> > Just removing
> > it does not seem to be a good option. Let me check our developer for this
> module
> > and see what they say. My thinking is to give an initial value for
> > g_GbFontNameMap.
> 
> Adding an initializer fixes my problem too, but this change also makes the
code
> simpler, which seems nice. (But up to you; you're much more familiar with this
> than I.)

I agree. Checked with our developer, this is indeed not used. lgtm

[Nico, 2014-07-16 18:07:43]

Committed patchset #1 manually as r27e35a8 (presubmit successful).