[PATCH 2/6] gdi32: Change the compute_metrics() location to get more accurate metrics.

[Byeongsik Jeon]

Marvin wrote:
> Hi,
> 
> While running your changed tests on Windows, I think I found new failures.
> Being a bot and all I'm not very good at pattern recognition, so I might be
> wrong, but could you please double-check?
> 
> Full results can be found at:
> https://testbot.winehq.org/JobDetails.pl?Key=47422
> 
> Your paranoid android.
> 
> 
> === debian9 (32 bit report) ===
> 
> gdi32:
> font.c:1339: Test succeeded inside todo block: got 72, expected 72 (B)
> font.c:1389: Test succeeded inside todo block: abcA(0) and gmptGlyphOrigin.x(0) values are different at width 1
> font.c:1393: Test succeeded inside todo block: abcB(2) and gmBlackBoxX(2) values are different at width 1

I'll change the order of patches to eliminate 'todo_wine'.

However, these patches are abandoned because the problem in coordinate
system transformation has not been solved as I thought. The pixel added
in the subpixel rendering is a problem.

I think these patches are meaningful enough, so I'll post them again.