Opened 4 years ago

Closed 17 months ago

#16502 closed bug (not reproducible)

FillPolygon to view which is child of BBitmap corruption

Reported by: smallstepforman Owned by: nobody
Priority: normal Milestone: Unscheduled
Component: - General Version: R1/beta2
Keywords: Cc:
Blocked By: Blocking:
Platform: All

Description (last modified by smallstepforman)

I have 2 code paths to draw a BPolygon. The first one draws directly from BView::Draw() (via FillPolygon) and works fine . If I add the same BView to a BBitmap, then there is occassional corruption when drawing the BPolygon. See screenshots. The grey polygon is done directly via BView::Draw::FillPolygon. The coloured polygon is drawn when added to bitmap (it creates a mask used for compositing). Saving the BBitmap data also shows the corruption so it's not the compositing code path.

Attachments (2)

BBitmap_View_Polygon.jpg (245.2 KB ) - added by smallstepforman 4 years ago.
BBitmap::View::FillPolygon
BView_BPolygon.jpg (174.1 KB ) - added by smallstepforman 4 years ago.
BView::FillPolygon

Download all attachments as: .zip

Change History (8)

by smallstepforman, 4 years ago

Attachment: BBitmap_View_Polygon.jpg added

BBitmap::View::FillPolygon

by smallstepforman, 4 years ago

Attachment: BView_BPolygon.jpg added

BView::FillPolygon

comment:1 by smallstepforman, 4 years ago

Description: modified (diff)

The black triangle below 6 is created when drawing via BBitmap::View::FillPolygon. It shouldn't be there (as evidenced when drawing directly to BView (grey screenshot)

Last edited 4 years ago by smallstepforman (previous) (diff)

comment:2 by smallstepforman, 4 years ago

Description: modified (diff)

comment:3 by jackburton, 4 years ago

I see the black triangle in both screenshots, actually

comment:4 by pulkomandy, 4 years ago

It would be a lot easier to investigate if you would share a piece of code reproducing this.

comment:5 by smallstepforman, 4 years ago

Great, another Heisen bug. I've created a test program, and the test program works fine. However, when I combine it with the OpenGL masking code path, it exhibits the problem on some scenarios. Well, considering that the test program always works, I'm sure the issue is somewhere at my end, the Haiku API works as advertised, so it's safe to close this issue. If I discover more information, I'll create a new ticket.

comment:6 by waddlesplash, 17 months ago

Resolution: not reproducible
Status: newclosed
Note: See TracTickets for help on using tickets.