[PATCH 1/2] gdi32: Add support for GCP_LIGATE to GetCharacterPlacement.

[Dmitry Timoshkov]

Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru>
---
 dlls/gdi32/font.c | 24 +++++++++++++++++++++++-
 1 file changed, 23 insertions(+), 1 deletion(-)

diff --git a/dlls/gdi32/font.c b/dlls/gdi32/font.c
index 8788426a2c..cc366eb4d7 100644
--- a/dlls/gdi32/font.c
+++ b/dlls/gdi32/font.c
@@ -3309,6 +3309,7 @@ GetCharacterPlacementW(
     DWORD ret=0;
     SIZE size;
     UINT i, nSet;
+    WCHAR *norm = NULL;
 
     TRACE("%s, %d, %d, 0x%08x\n",
           debugstr_wn(lpString, uCount), uCount, nMaxExtent, dwFlags);
@@ -3325,13 +3326,32 @@ GetCharacterPlacementW(
           lpResults->lpDx, lpResults->lpCaretPos, lpResults->lpClass,
           lpResults->lpGlyphs, lpResults->nGlyphs, lpResults->nMaxFit);
 
-    if(dwFlags&(~GCP_REORDER))
+    if (dwFlags & ~(GCP_REORDER | GCP_LIGATE))
         FIXME("flags 0x%08x ignored\n", dwFlags);
     if(lpResults->lpClass)
         FIXME("classes not implemented\n");
     if (lpResults->lpCaretPos && (dwFlags & GCP_REORDER))
         FIXME("Caret positions for complex scripts not implemented\n");
 
+    if (dwFlags & GCP_LIGATE)
+    {
+        nSet = NormalizeString(NormalizationC, lpString, uCount, NULL, 0);
+        if (!nSet) return 0;
+        norm = heap_alloc(nSet * sizeof(WCHAR));
+        if (!norm)
+        {
+            SetLastError(ERROR_NOT_ENOUGH_MEMORY);
+            return 0;
+        }
+        if (!NormalizeString(NormalizationC, lpString, uCount, norm, nSet))
+        {
+            heap_free(norm);
+            return 0;
+        }
+        lpString = norm;
+        uCount = nSet;
+    }
+
     nSet = (UINT)uCount;
     if(nSet > lpResults->nGlyphs)
         nSet = lpResults->nGlyphs;
@@ -3385,6 +3405,8 @@ GetCharacterPlacementW(
     if (GetTextExtentPoint32W(hdc, lpString, uCount, &size))
         ret = MAKELONG(size.cx, size.cy);
 
+    heap_free(norm);
+
     return ret;
 }
 
-- 
2.26.2

[Huw Davies]

Hi Dmitry,

On Thu, Jun 18, 2020 at 05:46:36PM +0800, Dmitry Timoshkov wrote:
> Signed-off-by: Dmitry Timoshkov <dmitry(a)baikal.ru>
> ---
>  dlls/gdi32/font.c | 24 +++++++++++++++++++++++-
>  1 file changed, 23 insertions(+), 1 deletion(-)
>
> diff --git a/dlls/gdi32/font.c b/dlls/gdi32/font.c
> index 8788426a2c..cc366eb4d7 100644
> --- a/dlls/gdi32/font.c
> +++ b/dlls/gdi32/font.c
> @@ -3309,6 +3309,7 @@ GetCharacterPlacementW(
>      DWORD ret=0;
>      SIZE size;
>      UINT i, nSet;
> +    WCHAR *norm = NULL;
>
> TRACE("%s, %d, %d, 0x%08x\n",
>            debugstr_wn(lpString, uCount), uCount, nMaxExtent, dwFlags);
> @@ -3325,13 +3326,32 @@ GetCharacterPlacementW(
>            lpResults->lpDx, lpResults->lpCaretPos, lpResults->lpClass,
>            lpResults->lpGlyphs, lpResults->nGlyphs, lpResults->nMaxFit);
>
> -    if(dwFlags&(~GCP_REORDER))
> +    if (dwFlags & ~(GCP_REORDER | GCP_LIGATE))
>          FIXME("flags 0x%08x ignored\n", dwFlags);
>      if(lpResults->lpClass)
>          FIXME("classes not implemented\n");
>      if (lpResults->lpCaretPos && (dwFlags & GCP_REORDER))
>          FIXME("Caret positions for complex scripts not implemented\n");
>
> +    if (dwFlags & GCP_LIGATE)
> +    {
> +        nSet = NormalizeString(NormalizationC, lpString, uCount, NULL, 0);
> +        if (!nSet) return 0;
> +        norm = heap_alloc(nSet * sizeof(WCHAR));
> +        if (!norm)
> +        {
> +            SetLastError(ERROR_NOT_ENOUGH_MEMORY);
> +            return 0;
> +        }
> +        if (!NormalizeString(NormalizationC, lpString, uCount, norm, nSet))
> +        {
> +            heap_free(norm);
> +            return 0;
> +        }
> +        lpString = norm;
> +        uCount = nSet;
> +    }
> +

This won't work very well if the ligature glyph doesn't exist in the font.

I suspect what you're supposed to do is to use font's gsub table to
find glyphs to combine rather than use Unicode normalization.

Huw.

[Dmitry Timoshkov]

Huw Davies <huw(a)codeweavers.com> wrote:
> This won't work very well if the ligature glyph doesn't exist in the font.
>
> I suspect what you're supposed to do is to use font's gsub table to
> find glyphs to combine rather than use Unicode normalization.

Thanks for the review. What do you think if I skip this patch and resend
2/2 alone? Are there any suggestions regarding that one?

-- 
Dmitry.

[Huw Davies]

On 18 Jun 2020, at 15:00, Dmitry Timoshkov <dmitry(a)baikal.ru> wrote:
> Huw Davies <huw(a)codeweavers.com> wrote:
> > This won't work very well if the ligature glyph doesn't exist in the font.
> >
> > I suspect what you're supposed to do is to use font's gsub table to
> > find glyphs to combine rather than use Unicode normalization.
>
> Thanks for the review. What do you think if I skip this patch and resend
> 2/2 alone? Are there any suggestions regarding that one?

Hi Dmitry,

Yes, please do that.  On first look it seems fine.

Thanks,
Huw.