[PATCH v2 0/2] MR7647: Add single point support for GdipGetPathWorldBounds

[Bartosz Kosiorek (＠gang65)]

--
  v2: gdiplus: Add single point support for GdipGetPathWorldBounds
      gdiplus/test: add tests for GdipGetPathWorldBounds with single point

https://gitlab.winehq.org/wine/wine/-/merge_requests/7647

[PATCH v2 1/2] gdiplus/test: add tests for GdipGetPathWorldBounds with single point

[Bartosz Kosiorek]

From: Bartosz Kosiorek

---
 dlls/gdiplus/tests/graphicspath.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/dlls/gdiplus/tests/graphicspath.c b/dlls/gdiplus/tests/graphicspath.c
index 81b4b40eed4..4640d6bbee6 100644
--- a/dlls/gdiplus/tests/graphicspath.c
+++ b/dlls/gdiplus/tests/graphicspath.c
@@ -586,6 +586,7 @@ static void test_worldbounds(void)
     expectf(266.8, bounds.Width);
     expectf(289.6, bounds.Height);

+    /* Bounds from path without any points */
     GdipCreatePath(FillModeAlternate, &path);
     status = GdipGetPathWorldBounds(path, &bounds, matrix, pen);
     expect(Ok, status);
@@ -596,6 +597,19 @@ static void test_worldbounds(void)
     expectf(0.0, bounds.Width);
     expectf(0.0, bounds.Height);

+    /* Bounds from path with single point */
+    GdipCreatePath(FillModeAlternate, &path);
+    GdipAddPathLine2(path, &(line2_points[0]), 1);
+    status = GdipGetPathWorldBounds(path, &bounds, matrix, pen);
+    expect(Ok, status);
+    GdipDeletePath(path);
+
+    todo_wine expectf(10.40, bounds.X);
+    todo_wine expectf(10.20, bounds.Y);
+    todo_wine expectf(0.0, bounds.Width);
+    todo_wine expectf(0.0, bounds.Height);
+
+    /* Bounds from path with two points */
     GdipCreatePath(FillModeAlternate, &path);
     GdipAddPathLine2(path, &(line2_points[0]), 2);
     status = GdipGetPathWorldBounds(path, &bounds, matrix, pen);

[PATCH v2 2/2] gdiplus: Add single point support for GdipGetPathWorldBounds

[Bartosz Kosiorek]

From: Bartosz Kosiorek

---
 dlls/gdiplus/graphicspath.c       | 7 ++++++-
 dlls/gdiplus/tests/graphicspath.c | 8 ++++----
 2 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/dlls/gdiplus/graphicspath.c b/dlls/gdiplus/graphicspath.c
index cbf8e74340b..0b2af580349 100644
--- a/dlls/gdiplus/graphicspath.c
+++ b/dlls/gdiplus/graphicspath.c
@@ -1624,8 +1624,13 @@ GpStatus WINGDIPAPI GdipGetPathWorldBounds(GpPath* path, GpRectF* bounds,

/* If path is empty just return. */
     count = path->pathdata.Count;
-    if(count == 0){
+    if(count < 2){
         bounds->X = bounds->Y = bounds->Width = bounds->Height = 0.0;
+        if(matrix && (count == 1))
+        {
+            bounds->X = matrix->matrix[4];
+            bounds->Y = matrix->matrix[5];
+        }
         return Ok;
     }

diff --git a/dlls/gdiplus/tests/graphicspath.c b/dlls/gdiplus/tests/graphicspath.c
index 4640d6bbee6..7377875b296 100644
--- a/dlls/gdiplus/tests/graphicspath.c
+++ b/dlls/gdiplus/tests/graphicspath.c
@@ -604,10 +604,10 @@ static void test_worldbounds(void)
     expect(Ok, status);
     GdipDeletePath(path);

-    todo_wine expectf(10.40, bounds.X);
-    todo_wine expectf(10.20, bounds.Y);
-    todo_wine expectf(0.0, bounds.Width);
-    todo_wine expectf(0.0, bounds.Height);
+    expectf(10.40, bounds.X);
+    expectf(10.20, bounds.Y);
+    expectf(0.0, bounds.Width);
+    expectf(0.0, bounds.Height);

/* Bounds from path with two points */
     GdipCreatePath(FillModeAlternate, &path);

[PATCH v2 0/2] MR7647: gdiplus: Add single point support for GdipGetPathWorldBounds

[Esme Povirk (＠madewokherd)]

That's a weird behavior. Is there an application that needs this?

[PATCH v2 0/2] MR7647: gdiplus: Add single point support for GdipGetPathWorldBounds

[Bartosz Kosiorek (＠gang65)]

On Mon Mar 24 09:30:55 2025 +0000, Esme Povirk wrote:
> That's a weird behavior. Is there an application that needs this?

No. In most cases matrix is not existing, or matrix translation is set to 0.0

I thing it is expected behaviour that with single path point (or zero path points) it returns 0.0 for width and height.

[PATCH v2 0/2] MR7647: gdiplus: Add single point support for GdipGetPathWorldBounds

[Esme Povirk (＠madewokherd)]

On Mon Mar 24 15:59:53 2025 +0000, Bartosz Kosiorek wrote:
> No. In most cases matrix is not existing, or matrix translation is set
> to 0.0
> I think it is expected behaviour that with single path point (or zero
> path points) it returns 0.0 for width and height.

Yeah, I think returning early in case of 1 point makes sense. I just wouldn't bother putting the matrix translation into the rect, until we either know something needs it or we have to implement other behaviors that make sense of it.