[PATCH] gdiplus: Avoid filtering of special characters.

[Haoyang Chen]

Symbols such as ←  should also be displayed.

Signed-off-by: Haoyang Chen 
---
 dlls/gdiplus/graphics.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/dlls/gdiplus/graphics.c b/dlls/gdiplus/graphics.c
index df4a89e8294..4a428c49869 100644
--- a/dlls/gdiplus/graphics.c
+++ b/dlls/gdiplus/graphics.c
@@ -5229,9 +5229,6 @@ GpStatus gdip_format_string(HDC hdc,
     hotkeyprefix_count = 0;

for(i = 0, j = 0; i < length; i++){
-        /* FIXME: This makes the indexes passed to callback inaccurate. */
-        if(!iswprint(string[i]) && (string[i] != '\n'))
-            continue;

/* FIXME: tabs should be handled using tabstops from stringformat */
         if (string[i] == '\t')

[Esme Povirk (they/them)]

I'm surprised that iswprint returns false for ←, maybe it's not the
right way to check for nonprintable characters?

[Haoyang Chen]

在 2021/5/12 下午11:59, Esme Povirk (they/them) 写道:
> I'm surprised that iswprint returns false for ←, maybe it's not the
> right way to check for nonprintable characters?

Hi,

I don't think it should be checked. On windows 7, when the 
GdipDrawString() function encounters

a "special character" that cannot be displayed, it will be a blank and 
will not skip it. On wine, if this

happens, if you don't skip it, at worst it will show a default character.

Thanks

[Nikolay Sivov]

On 5/13/21 5:47 AM, Haoyang Chen wrote:
> 在 2021/5/12 下午11:59, Esme Povirk (they/them) 写道:
> > I'm surprised that iswprint returns false for ←, maybe it's not the
> > right way to check for nonprintable characters?
>
> Hi,
>
> I don't think it should be checked. On windows 7, when the
> GdipDrawString() function encounters
>
> a "special character" that cannot be displayed, it will be a blank and
> will not skip it. On wine, if this
>
> happens, if you don't skip it, at worst it will show a default character.
>
> Thanks

What we have now already looks strange, regarding line breaks, or
trailing spaces checks, or using GetTextExtentExPointW() on text parts
within rendered paragraph. It should not be checking arbitrary
characters at that level, control characters are supposed to get forced
zero advances, or have them zeroed by the font. Ideally all that should
go through proper line breaking and shaping steps, maybe with some
preprocessing for key prefixes.

[Esme Povirk (they/them)]

Alright, let's try it and if something breaks we can add a more specific fix.

Signed-off-by: Esme Povirk